# The State of Gaming Advertising in 2026: Who sells brand advertising in games, what the standards cover, and how much of the pitch the public record supports
*Gaming advertising is several markets sold under one name. This review found no public US figure for brand spend in games, MRC's record on 29 September 2026 listed no accredited intrinsic in-game metric, and every published field result found for a live brand campaign comes from a party with a stake in it. The paper audits that public proof, not whether games advertising works.*
Evgeny Popov · No Fluff Advisory · Research paper · October 2026 · Research cutoff 1 October 2026
## Abstract
*What brand advertising in games is in 2026, who sells it, and how much of the sellers' pitch the public record can support. Sources were read from 28 September to 1 October 2026, when the evidence was frozen, with the United States as the main market. It audits the public proof, not whether games advertising works.*
> **For buyers.**
> Treat games as three buys: placements inside other companies' games, ads a platform or publisher sells in its own games, and media around games. For each, ask for titles with live dates matched to the publishers' app-ads.txt, for the verifier's own record of the format and environment you are buying, and for a randomised or holdout test with a named funder. Expect no MRC accreditation for intrinsic placements, check the platform's age rules before targeting, and do not add up two networks' reach over the same titles. [What it means](#implications) sets out each buy.
> **In short.**
> Gaming advertising is several markets sold under one name: networks that place ads inside other companies' games, platforms and publishers that sell ads in their own games, and media around games such as streams and communities. The only public US dollar series found counts ads placed in games, mostly bought, EMARKETER says, by game developers; this review found no figure for brand spend. Intrinsic in-game measurement has a finished standard, but the Media Rating Council's (MRC's) record shows no accreditation against it; other ad formats in game apps fall under the general mobile standards. Children's privacy rules and platform rules restrict in-game ad products. Every published field result found for a live brand campaign comes from a party with a stake in it.
### The question
The author's gaming playbook argued that brands underspend in games because they cannot fit them into their planning, safety and measurement systems, not because the audience is small. This paper tests that argument against public evidence and lets it fail; as registered, the test does not cover which constraint causes lower spend. It asks what a brand can buy in and around games today, what of it can be checked, and where the evidence is weak.
### What it found
1. **Three buying tasks.** Networks selling placements inside other companies' games sell mostly on their own documentation. Platforms and publishers selling their own games control the inventory, the rules and, in Roblox's managed campaigns, the reporting that settles the invoice [1]. Much game-adjacent media is bought like other digital media, Twitch through Amazon's demand-side platform (DSP), for example [2].
2. **The numbers.** EMARKETER is the only publisher found with public US dollar figures; its March 2024 forecast put US game ad revenues at $8.59 billion in 2024, 90.4% of it on mobile [3, 4, 5]. As developers buy most of the ads, the series is probably mostly user acquisition, an inference, since no source splits the dollars by objective [6, 7, 8].
3. **Supply.** Six networks' sellers.json files (public lists of their sellers) name 1,619 publisher domains that collapse to 1,381 once duplicates are removed. A seller account shows an authorised connection, not a live title, so the lists cannot be added up as reach [9, 10, 11, 12, 13, 14].
4. **Assurance.** MRC's record, read on 29 September 2026, shows no accredited intrinsic in-game metric and no MRC-confirmed audit for any product scored in the in-game cohorts, and third-party verification rests mostly on announcements (H2, supported) [15, 16]. Standard display and video ads in game apps plausibly fall within DoubleVerify's and IAS's accredited mobile in-app measurement, whose letters name no game [17, 18, 19].
5. **Rules.** Roblox and Microsoft bar personalised ads for users under 18, Discord keeps its age-assurance signals out of targeting, and compliance with the amended Children's Online Privacy Protection (COPPA) Rule was due for most regulated entities by 22 April 2026 (H4, supported) [20, 21, 22, 23].
6. **Console.** Apart from Microsoft's console integrations, whose location its documents do not establish, the Xbox and PlayStation placements found by 28 September 2026 sit outside gameplay or are in test; Electronic Arts sells placements inside its own sports games (H3, unresolved) [24, 25].
7. **Outcomes.** This review found no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign funded by a party without a financial interest [26, 27, 28].
8. **Structure.** At least three specialist in-game ad companies were sold or put into insolvency since 2021, and no record attributes an exit to the format; comparable revenue series are not disclosed (H5, unresolved) [29, 30, 31, 32].
9. **Buyer constraints.** Two surveys put brand safety first among the obstacles, but neither is established as US and neither lead is shown to exceed sampling error (H1, unresolved) [33, 34, 35].
10. **The matrix.** Twitch is the one Most documented product in game-adjacent media, and Anzu, Adverty, Frameplay and Gadsme are Well documented among in-game networks; the platform cohort is too small for bands. [The capability matrix](#matrix) has the scores, and [What it means](#implications) turns them into buyer questions.
### How it was done
AI research agents gathered the evidence, and separate AI verifiers checked all 807 ledger claims against 982 registered sources with orders to refute them. Hypotheses and scoring rules were fixed before research began. The author ran three outside reviews with another AI model and three internal rounds with AI agents, checking every factual finding at its source before changing anything. One scorer scored 23 products; the Play Quotient (PQ) measures documented capability, not performance. This is AI-assisted research, not peer reviewed, externally audited or independently validated. [Method, limits and interests](#method) has the detail.
> **Interests.**
> The author has worked at Samba (formerly Samba TV) since May 2025. At Verve Group he ran a P&L (profit-and-loss account) spanning mobile-gaming inventory. He co-leads the Signals & Measurement working group for AdCP (the Ad Context Protocol) at AgenticAdvertising.org. No vendor funded, reviewed or saw the paper before publication. Every vendor is scored under the same published rules.
## 1. What counts as gaming advertising
*Gaming advertising is several markets sold under one name. This paper records where each ad renders and what it is bought to do as separate fields, scores brand products only, and compares each product only with others in its cohort. The author's GAP model (in-game, around-the-game and away-from-the-game) maps onto those fields in part, and no source screened sizes those lines apart.*
> **In short.**
> Placement and objective are recorded separately because the sources blur them. EMARKETER says most in-game ads promote installs of other mobile games, yet none of the statements screened splits the dollars by objective [6, 7, 8]. User-acquisition and paid-discovery products are listed without scores. Brand products fall into 3 cohorts, scored apart. A supply claim needs named live titles and an audience with a metric and period, and the seller files screened show neither [9, 10, 11, 12, 13, 14]. The GAP model's away-from-the-game line includes app installs, an objective, so this review infers that its lines overlap [36, 8]. No source screened sizes in-game, around-the-game and away-from-the-game advertising as mutually exclusive, reconcilable categories, so GAP totals stay unresolved [37, 4, 38].
This paper classifies every ad by two fields. Placement is where the ad renders: in-game (intrinsic placements, in-game video, rewarded and playable formats, in-game audio), game-adjacent media (streaming, creators, esports, communities, gaming sites, companion apps), or other media reaching players. Objective is what the buyer pays for: brand advertising, or user acquisition and paid discovery. Only brand products are scored. The US comes first, and global vendors are included.
### The sources blur where an ad runs with why it was bought
EMARKETER said in 2023 that most in-game ads promote installs of other mobile games, and in 2024 that game developers buy most of the ads served in games [6, 7, 8]. Neither statement is a dollar split. Reading EMARKETER's US game ad revenue series as probably mostly user acquisition is the author's inference [6, 7, 8]. "Gaming ad spend" also has two opposite meanings. It can mean spend placed in games. Sensor Tower, a data vendor, leaves its term, game ad spend, undefined, but from context it means spend by game advertisers promoting their games; on mobile, most of that flows to app networks such as AppLovin and AdMob [39].
One partnership shows the split. As announced in May 2025, adWMG was to give its Xbox and PlayStation clients, advertisers of console games, access to Gadsme's mobile in-game inventory to acquire players [40, 41]. The placement is in-game on mobile. The objective is user acquisition. Nothing in it puts an ad inside a console title.
Placement follows the surface that draws the ad. Roblox's 24-hour homepage takeover and its tiles on Home and search sit outside gameplay and are not counted as in-game [42, 43]. This review infers that Overwolf's app ads are drawn by a companion app over or beside the game, and keeps them in game-adjacent media as a recorded judgment [44, 45, 46]. The device does not decide placement either. Twitch Premium Video runs in the Twitch app on consoles, and a Discord Play Quest can reward console play while the Quest renders in Discord [47, 48, 49].
### Intrinsic is one kind of in-game ad
An intrinsic in-game ad is part of the game world, such as a billboard or a branded object. The Intrinsic In-Game (IIG) Advertising Measurement Guidelines 2.0 are a final joint release of the IAB (Interactive Advertising Bureau), IAB Tech Lab and the MRC (Media Rating Council), released in August 2022 [17, 50, 51, 52]. They cover placements served by an ad server with an ad session. Hard-coded branded elements are not directly in scope, and interstitials, web banners and in-stream or outstream video fall under the general desktop, mobile web and in-app guidelines, though the guidelines say their in-game-specific guidance should also be considered for those placements [17].
In-game is wider than intrinsic. Anzu bundles intrinsic formats with interstitials and rewarded video, so part of its declared supply path carries impressions that are not intrinsic placements [9, 53, 54]. Gameloft's COMBO! sells in-app units such as display, rewarded and interactive video, and playable ads [55]. Both are in-game placements here, and this paper holds only intrinsic placements to the IIG guidelines.
### User acquisition and paid discovery are filtered out first
Before any product joins a cohort, user-acquisition and paid-discovery products are listed and left unscored. Paid discovery here means a game's owner paying a platform to promote that game to players. Roblox's Plays-objective campaigns run on Home and search, outside experiences [56]; the register lists these as Roblox Sponsored Experiences [57]. Medal Quests, which Medal describes as partnerships with game developers, are listed on the same rule [58], and so is Gadsme's Cross Promo, priced per install [59]. App-install networks are listed too; AppLovin calls Axon Ads Manager its user acquisition solution [60, 61].
The filter applies to each named product, and a scored product can still carry a mixed objective, which is recorded. Discord Quests are scored as a brand product, yet 14 of the 21 success stories on Discord's ads page on 29 September 2026 were games, several reporting acquisition or sales outcomes [62, 48, 63].
### Brand cohorts, each compared only with itself
Brand products are classified by what the buyer purchases. Cohort A, in-game ad networks, sells brand placements rendered inside games across independent third-party titles. Cohort B1, platform and publisher in-game ads, sells brand products on in-game inventory the platform or publisher controls. Cohort B2, game-adjacent media, sells brand products on gaming sites, companion apps, streams, communities and esports media. SSPs (supply-side platforms), DSPs (demand-side platforms) and publisher software are listed, and event sponsorship alone is described. A reseller of a platform's product takes that product's cohort, with the shared supply flagged.
The paper scores 23 products: 9 in cohort A, 4 in B1 and 10 in B2. The product register holds 59 rows, 28 of them listed and 8 historical.
Scores compare only within a cohort because the rules, fixed before scoring, differ by cohort. Coverage does not apply to game-adjacent media, whose ads do not render in a game. Supply means named live titles for a network, the platform's own ad-eligible inventory for B1, and properties or channels with an audience metric for B2. Each cohort's Play Quotient (PQ) is computed against its own field mean. The same supply score rests on different evidence in cohort A and in cohort B2. Chapter 11 holds the scores, which measure documented capability and say nothing about campaign performance.
### A supply claim has to name what is live
A supply claim counts only with its fields filled: named titles, live placements, US availability, ad-eligible audience with a metric and period, selling rights and exclusivity. A sellers.json file, a network's public list of its sellers, shows a supply connection, not scale. In files downloaded on 28 September 2026, six cohort A candidates declared from 30 to 1,117 publisher seller accounts each; the files show no titles, ad-eligible audience or active integrations [9, 10, 11, 12, 13, 14]. They overlap: 1,619 per-vendor unique publisher domains collapse to 1,381 unique domains [9, 11, 12, 10, 13, 14]. Vendor reach is never summed here without deduplication. Entries also outlive their businesses. IION's file, last modified on 28 September 2026, still lists Bidstack as an intermediary seller, although Bidstack Ltd has been in creditors' voluntary liquidation since 18 June 2026 [11, 32, 64].
An installed SDK (software development kit) is a footprint. Kidoz's 2025 annual report on Form 20-F says its SDK is integrated into thousands of mobile apps [65, 66, 67], yet in the pages screened the only game Kidoz names as on its network is Minion Rush, in undated alt text [68, 69, 70, 71, 65, 72]. A live integration needs a named title carrying placements inside the scoring period.
Announcements count once delivery is shown. IION and Frameplay said on 16 June 2025, in the future tense, that IION's 4,000+ titles (vendor-reported) "will be accessible through the Frameplay Exchange in the US" [73, 74]. Frameplay's sellers.json lists IION as an intermediary, and of ten publishers' app-ads.txt files sampled on 28 September 2026, each a publisher's public list of the sellers it authorises, four authorise that route, but this review found no source showing US delivery, volume or a go-live date [10, 75, 76, 77, 78].
Coverage is recorded as tuples of device (mobile, PC, console), distribution (app store, web, cloud, platform launcher) and content environment (premium title, casual title, or UGC platform, for user-generated content). Each distinct tuple counts once, and only when a named live title or documented live integration evidences it. Adverty's SDK supports WebGL (browser graphics) and Windows, but this review found no named live title documenting a web or PC tuple [79, 80, 81, 82].
### The GAP model against the evidence
The author's gaming playbook proposes the GAP model: in-game, around-the-game and away-from-the-game, planned as separate buyer problems and sold as one media line [36]. It maps roughly onto this paper's placement field: in-game to in-game, around-the-game to game-adjacent media, and away-from-the-game in part to other media reaching players. Away-from-the-game lists app install and re-engagement, which is an objective. Since EMARKETER says most in-game ads promote installs, this review infers that an install ad inside a game fits two GAP lines; separate fields remove the overlap [36, 8].
Venatus sells in these terms, and this review infers that its categories do not line up with this paper's placement field. It packages its offer as In the Game, Around the Game and Away from the Game; its home page calls its playable formats around-the-game, although this paper counts playable formats as in-game, and its away-from-the-game formats embed Twitch live streams in website takeovers, which this paper counts as game-adjacent media [83, 84]. A vendor using the labels gives the model no independent support.
The published series do not line up with GAP either. EMARKETER keeps a US total game ad spending series that counts in-game ads on mobile, console, PC and handheld devices and excludes esports, gaming video content and creator sponsored content [85, 86], and separate US series for mobile in-game ads, esports and gaming video content, with values for subscribers only [37, 87, 88]. Together they cover in-game advertising and parts of around-the-game, and the overlap between the esports and video series is unsettled. PwC's published video games and esports segment combines in-game and esports advertising, WPP Media publishes one undefined gaming line, and no publisher screened sizes an away-from-the-game category [37, 4, 38]. That answers QG, the registered question on the model: GAP totals are unresolved, and this paper never adds them. WPP's forecast carries gaming as an ad-revenue line and is not a published planning media mix [89, 90]. IAB Australia used the same three names in July 2023 with different contents, which [How big it is, and by whose definition](#market) sets out [34, 91].
The playbook's line that gaming draws under 5% of worldwide media investment is dropped. Its likely source is dentsu's 2025 Gaming Trends Report, repeated without attribution by EMARKETER, and no accessible text gives a denominator, year or definition; the trace is inferred from wording and dates [36, 92, 93, 94].
### The argument this paper allows to fail
The playbook argues that gaming is a media-mix reallocation problem: buyers hold back because gaming does not fit their planning, safety and measurement systems, whatever the audience's size [36]. This paper tests that argument and allows it to fail. Hypothesis H1, registered before extraction, is supported only if at least two surveys from different publishers, each with a stated base and period, rank measurement, suitability or planning fit above audience reach or scale, and documented ad-eligible inventory or measurement coverage is either stated with a metric smaller than the player audience for the same geography and period or excludes named device or environment combinations. Under that rule the gap between gaming time and ad spend is context only, and EMARKETER's attribution of it to measurement or fragmented buying cites no study [5, 92]. [How big it is, and by whose definition](#market) codes the result, and [What it means](#implications) summarises it.
### What the companion papers own
In-game audio is scored in the companion audio paper, and mixed packages are split. Gadsme's audio format and Cross Promo run through the same contracts and supply files as its managed network; those components go to the audio paper and the listed Cross Promo row [95, 59, 96, 97]. Attention metric definitions belong to the companion attention paper; here attention is only a measurement function a vendor documents. The published mobile app growth paper owns user-acquisition economics, AppLovin and Unity, and retention benchmarks. Its exclusions cover acquisition economics, so any brand product those companies sell is still classified here.
## 2. How big it is, and by whose definition
*This review found no public US figure for brand advertising in games. The one public US dollar series found counts ads placed in games, mostly on mobile, and EMARKETER says most of those ads are bought by game developers promoting other games. The buyer surveys found do not show that measurement and safety hold brands back more than reach.*
> **In short.**
> EMARKETER is the only publisher found with public US dollar figures for game advertising [3, 4, 38, 98]. Its March 2024 forecast put US game ad revenues at $8.59 billion in 2024, 90.4% of it on mobile [5]. Its later public statements give no single year in which the series passes $10 billion: "nearly" by 2029 in one piece, "surpass" by 2029 in another, and $10.2 billion in 2028 in figures Digiday reported in July 2026 [99, 92, 86]. EMARKETER says developers buy most of the ads served in games, so this paper reads the series as probably mostly user acquisition; that reading is an inference, because no source splits the dollars by objective [6, 7, 8]. Player counts are not ad-eligible audiences [100, 3]. Two buyer surveys from different publishers put brand safety first among the obstacles, but neither is established as US, neither lead is shown to exceed sampling error, and the same publisher says the belief that the same audiences can be reached more effectively elsewhere is growing [33, 34, 101]. H1 stays unresolved.
### One US series, and what it counts
EMARKETER's forecast database keeps a US total game ad spending series, which counts in-game ads on mobile, console, PC and handheld devices and excludes esports, gaming video content and creator sponsored content [85, 86]. It also keeps three further US series that touch gaming: mobile game ad spending (ads placed inside games on mobile devices only), esports ad spending (advertising and sponsorship on digital platforms) and gaming video content ad spending (video ads on recorded and live game content, such as YouTube or Twitch) [37, 87, 88]. The values are for subscribers only. What the public can read are the headline numbers in EMARKETER's articles and guides.
Its March 2024 forecast put US game ad revenues at $8.59 billion in 2024, of which $7.77 billion, or 90.4%, was mobile, and put mobile game ad revenues at 2.8% of all US digital ad spending [5]. The same forecast put the year the series passes $10 billion at 2026 [5, 6].
The later public statements do not agree with each other. An EMARKETER article of 26 December 2025 says US game ad revenues "will reach nearly $10 billion in the US by 2029" and puts in-game advertising at 2.3% of US digital ad spending in 2026, falling to 2% by 2029 [99]. EMARKETER's FAQ of 10 January 2026 says they "will surpass $10 billion by 2029" and gives only the 2026 share [92]. Both pieces link to the total game ad spending series, so the two statements conflict, and the vintage behind each is not stated [99, 92]. Digiday reported on 3 July 2026 EMARKETER figures of $9.21 billion in 2026, $9.7 billion in 2027 and $10.2 billion in 2028 for total US game ad spend, excluding esports, gaming video content and creator sponsored content [86]. This paper cites each statement in its own words and derives no crossing year from them.
*Figure fig-estimates. **The US total moves with the vintage, and no source found splits brand from user-acquisition spend** Published estimates of game ad revenue or spend, each shown with its publisher, vintage and year. The EMARKETER series count ads placed in games, mostly on mobile; EMARKETER says developers buy most of them, so reading the series as mostly user acquisition is this paper's inference. WPP Media's figure is global and undefined, so it is not comparable with the US series. Units: $ billion (nominal). Scope: US unless labelled global; forecasts. Sources [5], [108], [86], [38], [99], [92]. Caveat: Each point is a forecast in its own vintage and definition. They are not a time series and must not be summed or averaged. Data: data/figure-data.json#estimates_us*
### Mostly user acquisition, by inference
EMARKETER has said that most in-game ads promote installs of other mobile games (May 2023), that game developers buy most of the ads served in games (May and June 2024) and that most in-game ad spend goes to mobile (June 2024) [8, 7, 6]. The first two describe shares of ads, the third a share of spend by platform. None splits the dollars by buyer or between brand and user-acquisition objectives. This paper therefore reads the EMARKETER series as mixed objective and probably mostly user acquisition. That reading is the author's inference, and the paper never presents it as a measured share.
The word "gaming" makes the numbers harder to compare. Sensor Tower's State of Gaming 2026 blog says the majority of mobile game ad spend flows to app networks such as AppLovin and AdMob, and that PC and console spend is more concentrated in social and video channels, with YouTube taking the largest share [39]. Read in context, that is spend by game companies promoting their games, which the blog does not define [39]. Spend placed inside games and spend by game companies overlap, because an install ad shown in a game is both, but they are different quantities and are never added. User-acquisition economics are covered in the author's published mobile app growth paper; this paper covers brand advertising.
### Other estimates, and why none can be combined
The other publishers screened offer a growth rate, a share or a global total, each with its own definition. WPP Media put global gaming ad revenue at $8.5 billion in 2025, up 29.5% and 0.7% of all global ad revenue, and Marketing Brew reported the same forecast at $10.7 billion for 2026; WPP's public pages do not define gaming [38, 102]. Its June 2026 midyear text says only that gaming continues to grow, and its public chart carries no calendar years [89, 90]. That a global total is smaller than EMARKETER's US-only 2024 forecast suggests, by this review's inference, that the two publishers count different things; WPP publishes no definition to check it against.
The IAB/PwC Internet Advertising Revenue Report for 2025, a report on US advertising revenue, says "video games and esports" advertising revenue grew 22.0% in 2025, after 8.7% in 2024, and publishes no dollar value; the chart labels neither the geography nor the edition of PwC's Outlook it comes from, and the report does not say whether the 2025 figure is a reported actual or a model estimate [4]. PwC's Outlook release puts total global video games revenue at $224 billion in 2024, a whole-industry total that includes consumer spending, and gives a share for "advertising in video games" (32.8% in 2024, rising to 38.5% in 2029) without stating what the percentage is a share of [103, 104]. Circana's $60.7 billion for 2025, in the Entertainment Software Association's (ESA's) report, is US consumer spending on games, not advertising [3].
The estimates table lists each figure with its publisher, vintage, metric, geography and what it includes, excludes or leaves unstated. The "combinable" column is no for every row.
**Market-size reconciliation**
| Estimate | Published | Metric | Value | Year | Geography | Includes | Excludes | Combinable with | Src |
|---|---|---|---|---|---|---|---|---|---|
| EMARKETER (FAQ) | 2026-01-10 | US game ad revenues | will surpass $10 billion | 2029 | US | in-game advertising (FAQ framing: mobile, console, PC) | not stated | no | [92] |
| EMARKETER (FAQ) | 2026-01-10 | in-game advertising share of overall US digital ad spending | 2.3% | 2026 | US | in-game advertising | not stated | no | [92] |
| EMARKETER (article) | 2025-12-26 | US game ad revenues | nearly $10 billion | 2029 | US | game ad revenues (definition not in visible excerpt) | not stated | no | [99] |
| EMARKETER (article) | 2025-12-26 | in-game advertising share of overall US digital ad spending | 2.3% | 2026 | US | in-game advertising | not stated | no | [99] |
| EMARKETER (article) | 2025-12-26 | in-game advertising share of overall US digital ad spending | 2% | 2029 | US | in-game advertising | not stated | no | [99] |
| EMARKETER (guide; March 2024 forecast) | 2024-03 | US game ad revenues | $8.59 billion | 2024 | US | all game ad revenues, mobile and non-mobile | not stated | no | [5] |
| EMARKETER (guide; March 2024 forecast) | 2024-03 | US mobile game ad revenues | $7.77 billion | 2024 | US | mobile game ad revenues | not stated | no | [5] |
| EMARKETER (guide; March 2024 forecast) | 2024-03 | mobile share of US game ad revenues | 90.4% | 2024 | US | mobile game ad revenues over total game ad revenues | | no | [5] |
| EMARKETER (guide; March 2024 forecast) | 2024-03 | mobile game ad revenues share of total US digital ad spending | 2.8% | 2024 | US | mobile game ad revenues | | no | [5] |
| EMARKETER (guide) | 2024 | mobile gaming ad revenues share of total US mobile ad spend | 3.8% | 2024 | US | mobile gaming ad revenues | | no | [5] |
| EMARKETER (guide) | 2024 | time spent in mobile games share of total US mobile time (context, not spend) | 10.7% | 2024 | US | time spent | | no | [5] |
| EMARKETER (guide; March 2024 forecast) | 2024-03 | year US game ad revenues pass $10 billion | passes $10 billion | 2026 | US | all game ad revenues | not stated | no | [5] |
| EMARKETER (article, June 2024; March 2024 forecast) | 2024-03 | US game ad revenues | $8.59 billion (+6.7% vs 2023) | 2024 | US | game ad revenues | not stated | no | [6] |
| eMarketer as cited by IAB | vintage not stated (cited 2024-03-26) | US ad spend in games | $8.53 billion | 2024 | US | not stated | not stated | no | [108] |
| eMarketer as cited by IAB | vintage not stated (cited 2024-03-26) | US ad spend in games | $11.49 billion | 2027 | US | not stated | not stated | no | [108] |
| EMARKETER (report) | 2019-10-17 | US in-game advertising, all devices | will exceed $3 billion | 2019 | US | in-game advertising on all devices; mobile the majority | not stated | no | [544] |
| EMARKETER (FAQ; October 2025 forecast) | 2025-10 | digital gamers, share of US population (audience, not spend) | 59.2% | 2026 | US | internet users of any age playing online or internet-updated games on any device at least monthly (definition in GAM-A-S11) | offline-only play | no | [92] |
| EMARKETER (guide; August 2023 forecast) | 2023-08 | digital gamers, share of US population (audience, not spend) | 54.6% | 2024 | US | digital gamers (definition as above) | | no | [5] |
| ESA Essential Facts 2026 (YouGov survey) | 2026-06 | Americans aged 5-90 who play video games (audience, not spend) | 212.3 million | 2026 | US | people who play at least 1 hour a week on smartphone, tablet, PC, console or VR | ages under 5 and over 90; play under 1 hour a week | no | [3] |
| ESA Essential Facts 2026 (YouGov survey) | 2026-06 | share of US population playing 1+ hours a week (audience, not spend) | 67% | 2026 | US | as above | | no | [3] |
| ESA Essential Facts 2025 | 2025 | Americans who play video games (audience, not spend) | more than 205 million | 2025 | US | not restated on the page | | no | [545] |
| Newzoo (author-confirmed; not opened by agent) | 2026-08-25 | players worldwide (audience, not spend) | 3.7 billion (+4.4%) | 2026 | global | Newzoo player definition (not read) | | no | [105] |
| dentsu 2025 Gaming Trends Report (via Variety) | 2025-07-01 | gamers worldwide (audience, not spend) | over 3.4 billion | not stated | global | not stated | | no | [93] |
| dentsu 2025 Gaming Trends Report (via Variety) | 2025-07-01 | gaming share of global media investment | less than 5% | not stated | global | not stated | not stated | no | [93] |
| WPP Media This Year Next Year (end-of-year) | 2025-12-10 | global gaming ad revenue | $8.5 billion (+29.5%) | 2025 | global | not published | not published | no | [38] |
| WPP Media This Year Next Year (end-of-year) | 2025-12-10 | gaming share of total global ad revenue | 0.7% | 2025 | global | not published | not published | no | [38] |
| WPP Media (as reported by Marketing Brew) | 2025-12-09 | global gaming ad spend | $10.7 billion | 2026 | global | not published | not published | no | [102] |
| PwC Global E&M Outlook 2025-29 | 2025-07-24 | global video games revenue, all sources (not advertising only) | $224 billion | 2024 | global | consumer spending and advertising | hardware | no | [103] |
| PwC Global E&M Outlook 2025-29 | 2025-07-24 | global video games revenue, all sources (not advertising only) | nearly $300 billion | 2029 | global | consumer spending and advertising | hardware | no | [103] |
| PwC Global E&M Outlook 2025-29 | 2025-07-24 | 'advertising in video games' (base not stated in release) | 32.8% | 2024 | global | not stated | not stated | no | [103] |
| PwC Global E&M Outlook 2025-29 | 2025-07-24 | 'advertising in video games' (base not stated in release) | 38.5% | 2029 | global | not stated | not stated | no | [103] |
| IAB/PwC FY2025 report (PwC Outlook data) | 2026-04 | video games and esports advertising revenue, year-on-year growth | +22.0% | 2025 | US | advertising via video games; advertiser spend on esports; digital and non-digital | not stated | no | [4] |
| IAB/PwC FY2025 report (PwC Outlook data) | 2026-04 | video games and esports advertising revenue, year-on-year growth | +8.7% | 2024 | US | as above | not stated | no | [4] |
| Circana (via ESA Essential Facts 2026) | 2026-06 | US consumer spending on video games (not advertising) | $60.7 billion | 2025 | US | content $52.3B, hardware $5.4B, accessories $2.95B | advertising | no | [3] |
| EMARKETER via Digiday | not stated (reported 2026-07-03) | US total game ad spend | $9.21 billion | 2026 | US | ads placed in games on all devices (per report) | esports, gaming video content, creator sponsored content | no | [86] |
| EMARKETER via Digiday | not stated (reported 2026-07-03) | US total game ad spend | $9.7 billion | 2027 | US | ads placed in games on all devices (per report) | esports, gaming video content, creator sponsored content | no | [86] |
| EMARKETER via Digiday | not stated (reported 2026-07-03) | US total game ad spend | $10.2 billion | 2028 | US | ads placed in games on all devices (per report) | esports, gaming video content, creator sponsored content | no | [86] |
### Players are not an ad-eligible audience
Each player count screened uses a different definition, and none measures who can be reached by a brand ad [100, 3, 93]. EMARKETER forecasts that 59.2% of the US population will be digital gamers in 2026, counting internet users of any age who play online or internet-updated games at least once a month [92, 100]. ESA's 2026 Essential Facts reports 212.3 million Americans aged 5 to 90 who play, and 67% of the population playing at least one hour a week, from a YouGov survey of 13,545 people fielded 11 to 25 February 2026 [3]. Newzoo forecasts 3.7 billion players worldwide in 2026, up 4.4% [105]. Newzoo's site blocked this review's tools; the author confirmed the figure and the article date, 25 August 2026, in a browser.
The gap between time spent and money spent is context, not evidence. EMARKETER's 2024 guide sets mobile games' 10.7% of US mobile time against 3.8% of mobile ad spend in 2024, and its January 2026 FAQ sets more than an hour of daily gaming against in-game advertising's 2.3% of digital ad spend in 2026 and says the gap "suggests marketers lack confidence in measurement or struggle with fragmented buying processes" without citing a study [5, 92]. The author's playbook line that gaming gets "under 5%" of worldwide media investment matches dentsu's 2025 Gaming Trends Report as Variety reported it; no accessible text gives that figure a denominator, a year or a definition, so the paper does not use it as a share [36, 93, 94].
### What buyers say holds them back
Hypothesis H1 tests whether buyers cite measurement, suitability or planning fit as constraints more often than audience reach, and whether documented ad-eligible inventory or measurement coverage is narrower than the player audience; as registered, it does not test which constraint causes lower spend. Its survey leg needs two surveys from different publishers, each with a stated base and fieldwork period, that rank measurement, suitability or planning fit above audience reach or scale.
Two surveys qualify on the wording. Advertiser Perceptions' 2024 Gaming Advertising Report asked 200 marketer and agency contacts in July 2024 for the biggest obstacles: brand safety 29%, reaching the same consumers more effectively elsewhere 26%, difficulty of measurement 26%, lack of scale or reach 26%, and organisational resistance 25% [33, 35]. IAB Australia's Game Advertising State of the Nation 2023 asked 100 agency and brand decision makers in June and July 2023 for their three biggest barriers: brand safety 47%, measuring campaign success 36%, brand suitability 36%, evidence of effectiveness 34%, and how games work with other media 32%; no reach barrier appears among the five shown [34, 91].
The support is weak. Neither sample is established as US or limited to brand advertising. On a base of 200, brand safety's three-point lead sits inside the margin of a single share of about six points either way, and whether gaps between answers to the same multiple-choice question are real depends on overlaps neither publisher reports [33, 34]. Counting IAB Australia as ranking safety above reach rests on the inference that its 2023 wave offered a reach option, as its 2022 wave did [34, 106]. And Advertiser Perceptions' own summary says brand-safety concerns are fading while the belief that the same audience can be reached more effectively elsewhere is growing [101, 33]. In the same survey, the top two benefits cited are reach: hard-to-reach audiences 43% and incremental reach 38% [33].
The US evidence points both ways. IAB's March 2024 study with Advertiser Perceptions, sponsored by Admazing and Venatus (both scored in this paper), surveyed 300 US buyers already investing in games and asked no barrier question; 86% of them agreed games advertising is brand safe, second only to digital audio, and 85% said they were confident measuring it, while 79% agreed it is easy to buy, the same as CTV, above online video and below display, social media and digital audio [107, 108]. Among the 51% who treat games as part of other channels rather than a channel of their own, the first of four unnumbered reasons listed is "not yet enough scale", but the report's illustrative quote for that group ties scale to the advertiser's own spend rather than to audience size, so it is weak evidence of a reach constraint [107]. In the sources screened by 28 September 2026 and rechecked on 29 September 2026, this review found no US survey that publishes a base, a fieldwork period and shares for the constraints on brand advertising in games; Advertiser Perceptions' 2025 gaming results, IAS's 2026 Industry Pulse and IAB's August 2026 guide "The Business Case for Gaming" could not be read in full [109, 110, 111].
H1's second leg holds: the documented measurement coverage for in-game brand ads excludes named device and environment combinations. MRC's accredited services use desktop, mobile web, mobile in-app and connected-TV environments, with no gaming or console environment [15, 16]. But H1 needs both legs. Its survey leg is met only on point estimates, and only if IAB Australia's 2023 wave offered a reach option, which is inferred from its 2022 wave; without that inference one usable survey remains, which the registered rule codes as unresolved. Neither sample is shown to be US, and Advertiser Perceptions gives contrary signals. Coded under its registered rule, H1 is unresolved [33, 34, 35]. The public survey record neither shows that infrastructure holds brands back more than audience nor shows the reverse, and no study design in the record identifies which constraint causes lower spend.
### The GAP model meets the published categories
The author's playbook proposed selling gaming as one media line across in-game, around-the-game and away-from-the-game placements [36]. No source screened sizes those three as mutually exclusive, reconcilable categories [37, 4, 38]. EMARKETER comes closest: its total US game ad spending series counts in-game ads on mobile, console, PC and handheld devices and excludes esports, gaming video content and creator sponsored content, and it keeps separate esports and gaming video content series whose overlap it does not settle; most values are for subscribers only [85, 86, 87, 88]. The playbook's own categories overlap, because its away-from-the-game line includes app installs, an objective, while EMARKETER said in 2023 that most in-game ads promote installs of other mobile games [36, 8].
IAB Australia's July 2023 report already sorted game advertising into "in the game", "around the game" and "away from the game", the same three names, with different contents: its around-the-game covers rewarded video, interstitials and overlaid audio ads during the gaming session, and its away-from-the-game covers streaming, esports and influencers [34, 91]. This paper keeps placement and objective as separate fields and adds no GAP totals.
## 3. Who sells inventory inside games
*Specialist networks sell brand placements inside games that other companies publish. Their public files show which publishers have authorised them more clearly than what they sell, where, or to how many players. The same large publishers authorise several networks at once, and some networks are authorised to resell one another.*
> **In short.**
> The cohort A networks are Anzu, Frameplay, IION, Gadsme, Adverty, Gameloft for brands (COMBO!), Super League, Kidoz and Admazing. This review moved AdInMo out of the cohort after its operator applied to be struck off and its ad-serving hostnames did not resolve, and Trivver is listed because this review found no live title for it [112, 113, 114]. Their audience figures are vendor statements, and most lack a metric, a period or both [115, 116, 117]. Six networks' sellers.json files, read on 28 September 2026, hold 1,619 publisher domains between them; deduplicated, they come to 1,381, and miniclip.com appears in five [9, 10, 11, 12, 13, 14]. Frameplay and IION list other cohort A networks as intermediary sellers [10, 11]. IION's route to US buyers through the Frameplay Exchange, announced in the future tense in June 2025, now appears as a reseller line in some publishers' app-ads.txt files, with no go-live record found [73, 10, 75, 77]. Such lines show authorisation, not active buyable supply. Gadsme calls all its publisher relationships direct and exclusive, yet its own file lists three intermediary sellers [118, 12]. Few demand-side platforms (DSPs), the systems buyers bid through, confirm a network integration on their own sites, and no cohort A network documents any state of an agentic buying protocol in the sources screened [119, 120, 121]. No cohort A vendor gives a market share with a denominator in the sources screened, so concentration is unresolved [55, 74, 122].
### The networks and what each names as live
Cohort A holds the networks that sell brand placements rendered inside games across independent third-party titles, the boundary set in What counts as gaming advertising. Anzu, Frameplay, Gadsme and Adverty sell intrinsic display and video drawn into the game's own scenes [123, 124, 59, 125]. COMBO! sells rewarded, interstitial, interactive and playable units in mobile games from Gameloft and partner publishers [55, 126]. IION sells in-game formats beside placements around and away from the game, which Around the game covers [127, 115]. Super League sells managed programmes in Roblox, Minecraft and Fortnite and in mobile and PC titles [128]. Kidoz reaches young audiences through its software development kit (SDK), code a developer builds into an app, which it says is integrated into thousands of mobile apps [65]. Admazing sells brand campaigns in third-party mobile games [129, 130].
The table gives each network's scale claim, the publisher accounts in its sellers.json and the titles that it, a publisher or an independent observation dates inside the scoring period (2 October 2025 to 2 October 2026), the only titles the coding rule for research question 2 (Q2) counts as live.
| Network | Scale as the vendor states it | Publisher accounts in its sellers.json | Titles dated inside the scoring period | Sources |
|---|---|---|---|---|
| Anzu | "Hundreds of live cross-platform games"; 250 million monthly active users (blog, 9 September 2025) | 1,117 | UFL (partnership and inventory announced 13 May 2026, no delivered campaign reported), Gangsta Island: Crime City, Asphalt 8: Airborne and Dragon Mania Legends (case study of 26 November 2025, go-live date not stated), Fashion Battle, Drill & Collect | [123, 131, 9, 132, 133, 134, 135] |
| Frameplay | "a network of 5,000+ mobile games" (9 June 2026) | 30 | none named | [136, 10, 137] |
| IION | "60,000+ games"; "4,000+ gaming titles" (June 2025); "3.6 billion" as "Unique Audience" | 268 | none named as carrying IION placements | [115, 73, 116, 11, 138] |
| Gadsme | "more than 1,600 mobile games" (October 2025) | 223 | Football Manager 26 (PC and Mac): exclusive announced by its publisher on 29 September 2025, just before the period; in-period delivery rests on partner content co-produced by Gadsme (December 2025) | [139, 12, 140, 141, 142] |
| Adverty | "600+ games" (unaudited pages of its 2025 annual report) | 121 (119 later the same day) | Parking Master 3D (14 January 2026) | [143, 13, 82] |
| Gameloft for brands (COMBO!) | 1.3 billion monthly active players (March 2026 data) | 31 | none; its dated campaigns predate the period | [117, 14, 55] |
| Super League | Misfits portfolio of "more than 100 million monthly active users" | no file | Hide or OOF! and My Avatar!, which its 10-Q of 14 August 2026 describes as generating "differentiated inventory for playable advertising"; no campaign in either is dated (it took a 15.0% stake with some brand rights in Hide or OOF! in January 2026 and an ownership interest in My Avatar! in April 2026) | [144, 145, 30, 146] |
| Kidoz | 40,000+ kid-safe games; 500M+ monthly reach | about 270 (29 September 2026) | none; Minion Rush is named without a date | [68, 66] |
| Admazing | 7.3 billion+ impressions served, no period | no file | none named | [130, 147] |
The columns count different things. A seller account records an authorised supply connection; it says nothing about titles, ad-eligible audience or active monetised integrations, and Adverty's count changes from day to day [9, 13]. Anzu's file also reaches beyond games: a keyword screen by this review found at least 118 of its unique publisher domains to be television, streaming, radio, news or weather properties, such as roku.com and cnn.com [9].
### AdInMo and Trivver sit outside the cohort
AdInMo was operating when the scoring period began: PocketGamer.biz reported on 24 October 2025 that it had launched its monetisation technology as a software-as-a-service offering for mobile developers, with free and paid tiers [148]. On 6 September 2026 the sole remaining director of AdInMo Ltd signed a strike-off application declaring that no circumstance in sections 1004 and 1005 of the Companies Act 2006 applied, and section 1004 bars an application by a company that traded in the previous three months [112, 149]. The application was gazetted on 15 September and suspended on 17 September 2026 after an objection, and on 29 September the register showed the company as active with a proposal to strike off [150, 151]. Every Internet Archive capture of its website from 11 July to 27 September 2026 shows a server error, and on 29 September the site still returned one and its ad-serving hostnames did not resolve [152, 113]. This review therefore moved AdInMo to the historical rows, unscored, an inference from those records [112, 153, 113]. Frameplay's and IION's sellers.json files still list it as an intermediary; this review infers that the entries are stale rather than evidence of current trading [10, 11].
Trivver is listed. MRC (the Media Rating Council, which audits and accredits measurement services) shows Trivver's audit as "in process as of 4/2026" across desktop, mobile and gaming consoles, the only gaming entry on its list of services under review [16, 154]. Trivver calls a PlayCanvas SDK for browser games, released on 11 June 2026, its "first commercial deployment", although its use-case page says it already integrates with Unity, Unreal and PlayCanvas [155, 156]. This review found no named title, brand campaign or supply file for Trivver in the scoring period, in the sources screened by 29 September 2026 [114, 157]. The register therefore lists it, and Measurement treats its audit [155, 16].
### Audience figures lack a metric, a period or both
The audience numbers are the vendors' own, and few say what is counted over what period. IION's advertisers page says it is "Home to a premium audience of 3.6 billion" and repeats "3.6 billion" as "Unique Audience" in a block about gaming in general, with no period [116, 115, 73]. The figure matches common estimates of the world's players, so it reads as a player-population figure rather than a measured reach of IION's own supply; that reading is this review's inference [116]. COMBO!'s 1.3 billion monthly active players, from Gameloft data of March 2026, was already its headline at launch in September 2022, and its four regional figures, under four metric names, sum to 1,004 million [117, 158].
Other counts disagree with one another. Kidoz's site claim sits beside 500M+ global monthly unique users in a fact sheet covering its children's and all-ages platforms together, over 350 million a month in its SDK documentation and "hundreds of million" in its annual report [68, 70, 71, 65]. Between December 2024 and May 2026 Gadsme cited over 1,500 games, more than 1,600 mobile titles and more than 100 game developers, and its undated advertisers page claims 200+ game studios, none with a period [159, 139, 118, 160]. Frameplay's "5,000+ mobile games" cannot be reconciled with its own supply records and plausibly includes resold supply, since IION alone was announced as bringing 4,000+ titles (inference) [136, 73, 10]. Super League's Misfits figure has no reference month, measurement source or deduplication across games [144, 161]. Because the same publishers sit behind several networks, these reaches must never be summed [11, 10, 162].
US availability is documented unevenly. Anzu has run US campaigns, and NBCUniversal was named its US and UK sales partner in March 2022 [163, 164]. Kidoz booked $9,041,962 of its $18,433,172 revenue in 2025 in North America, by impressions served [65]. Adverty reports US sales activity, COMBO! fields North America sales staff, Super League targets US audiences and Gadsme opened a New York office in December 2024 [79, 165, 30, 159]. IION's one documented US route is the Frameplay path announced in June 2025 [73, 166]. Frameplay, Adverty, Gadsme, Kidoz and Super League publish no US audience or coverage figure in the sources screened [74, 79, 139, 65, 30]. Admazing's home page cites "191 million US mobile gamers", which this review reads as a market-wide count rather than Admazing's own reach (inference), and COMBO!'s 57 million monthly users covers all of North America [129, 130, 117].
### Seller files record authorisation, and fewer integrations are confirmed
> **How to read a seller file.**
> A sellers.json file, in the format of IAB Tech Lab (the Interactive Advertising Bureau's technical standards body), lists the accounts a network sells for and types each as PUBLISHER, INTERMEDIARY or BOTH; all six cohort A files read on 28 September 2026 declared version 1.0, as did Kidoz's, read a day later [9, 10, 11, 12, 13, 14, 66]. A publisher's app-ads.txt file lists the ad systems it authorises to sell its app inventory, each under a seller ID marked DIRECT or RESELLER [53, 76]. When a DIRECT line matches a PUBLISHER entry, both sides record that the network may sell that publisher's inventory. The match does not say which titles, placements or countries the network sells [14, 53, 76].
Some relationships match from both sides. Anzu's do for seven sampled publishers, among them Miniclip, Zynga, Rovio, Gameloft and Helium9 Games [9, 75, 77, 54, 78, 96]. COMBO!'s do for King, Zynga, Miniclip, Rovio and Outfit7 [14, 53, 77, 75, 54, 76]. IION's in-game supply is documented mainly through Voodoo, whose app-ads.txt authorises the ID that IION assigns to Voodoo SAS; IION's case study on Voodoo reports 70 to 80 million monthly impressions after Voodoo moved to a direct integration across rewarded video, interstitial, banner and native formats, a delivery count for the publisher's monetisation rather than for placements drawn into the game's scenes [138, 167, 11]. Lihuhu, publisher of Adverty's one named live title, authorises the Adverty ID assigned to it [162, 13].
Other accounts do not match. At least five of Frameplay's 30 publisher accounts match a DIRECT line in the publisher's file, while seven sampled accounts have no frameplay.gg line at all [10, 75, 96]. Miniclip and Helium9 Games authorise Adverty seller IDs that Adverty's sellers.json does not contain [75, 96, 13]. IION's file assigns to another publisher, Big Cake, an ID that Lihuhu authorises as a DIRECT IION seller [11, 162]. Admazing serves no sellers.json at the domains checked, so the seller of record behind its exchange's reseller account cannot be resolved [147, 75, 54, 77].
Adding the files together double counts. Across the six sellers.json files read on 28 September 2026, 1,619 per-network unique publisher domains collapse to 1,381 unique domains, or 1,380 with Adverty's later file [9, 10, 11, 12, 13, 14]. Of these, 176 appear in two or more files, 51 in three or more and 10 in four or more; miniclip.com appears in the files of Anzu, IION, Gadsme, Gameloft and Frameplay [9, 11, 12, 14, 10]. A simple sum therefore carries 238 surplus domain appearances [9, 10, 11, 12, 13, 14]. Kidoz's file, read a day later, is outside this count [66].
*Figure fig-sellers. **Seller accounts are not titles, and the networks list many of the same publishers** Publisher accounts declared in each cohort A network's sellers.json, with the unique publisher domains behind them. A seller account shows an authorised supply connection, not a live title, a placement or scale. Anzu's file also includes connected-TV, streaming and news properties. Units: publisher seller accounts. Scope: the six cohort A networks whose sellers.json files were read on 28 September 2026. Sources [9], [11], [12], [13], [14], [10]. Caveat: Files change often; counts are for the reads on 28 September 2026. Kidoz's file, read on 29 September, is outside this count; Super League and Admazing serve no file at the domains checked. Data: data/figure-data.json#sellers*
### Networks are authorised to resell one another
Frameplay's sellers.json lists IION, Adverty and AdInMo as INTERMEDIARY sellers, and IION's lists Anzu, AdInMo, Adverty and Bidstack [10, 11]. These entries authorise resale without showing which inventory is resold, so a buyer of one network may be buying another's supply; that step is an inference [10, 11]. They also go stale: IION's file, modified on 28 September 2026, still lists Bidstack, whose operating company entered creditors' voluntary liquidation on 18 June 2026 [11, 32].
The Frameplay Exchange opened on 9 October 2024 for programmatic and direct buying aimed at buyers in the US and other markets, carrying Adverty's and AdInMo's inventory beside Frameplay's [74, 168]. On 16 June 2025 Frameplay and IION announced, in the future tense, that IION's 4,000+ titles, a vendor figure, "will be accessible through the Frameplay Exchange in the US" [73]. Frameplay's sellers.json, last modified on 19 August 2025, assigns seller ID 9251981480 to IION as an INTERMEDIARY, and the app-ads.txt files of Miniclip, Outfit7, Zynga and Gameloft, four of ten publisher files sampled, authorise "frameplay.gg, 9251981480, RESELLER" [10, 75, 76, 77, 78]. No source found shows US delivery, volume or a go-live date, and Frameplay's resources feed has no post after 4 June 2025 [73, 169].
Large publishers authorise several networks at once. Zynga's app-ads.txt, last modified on 23 September 2026, authorises Gadsme, Anzu, Gameloft, IION and PubMatic as DIRECT sellers and Frameplay, Adverty and Kidoz as resellers [77, 14]. King's file, whose owner domain is Activision Blizzard Media, carries 153 DIRECT and 2,452 RESELLER lines, among them a Gameloft DIRECT line and Anzu and Gadsme reseller entries [53]. King, Zynga, Miniclip and Rovio also authorise Admazing's exchange as a reseller [53, 77, 75, 54]. Gameloft sits on both sides, authorised by five publishers and itself listed as a publisher by Anzu, Gadsme and Frameplay; that some King inventory is offered through both COMBO! and Microsoft's XBOX Media Solutions is this review's inference, unverified at title or placement level [14, 53, 78]. Super League, which serves no seller file at the domains checked, says it sometimes includes inventory from in-game ad firms such as Anzu, Bidstack, AdInMo and Frameplay in its programmes, without saying whose it has used [128, 30, 145]. The same paths carry in-game audio, such as Audiomob in Anzu's and IION's files, which the companion audio paper covers [9, 11].
### Exclusivity claims against the vendors' own files
Gadsme's exclusivity claims are unqualified, and its own documents narrow them. Its pages say "All our publisher relationships are direct and exclusive" and "No reseller stacking" [118, 170]. Its sellers.json lists Digital Turbine, LoopMe and Unity as INTERMEDIARY sellers, and King, Miniclip, Rovio, Gameloft and Zynga authorise gadsme.com as a RESELLER under the Digital Turbine ID [12, 53, 75]. Sixty of its publisher domains also appear in Anzu's file [12, 9]. Its Publisher General Terms, updated on 19 May 2026, restrict exclusivity to 3D and audio ads, apply it per player or game session, and let a publisher test a competing 3D network while Gadsme runs for at least 50% of players [59]. One title-level exclusive is confirmed by its publisher: Sports Interactive announced on 29 September 2025 a three-year exclusive partnership bringing Gadsme's SDK into the PC and Mac edition of Football Manager 26 [140].
Anzu calls its inventory "exclusive", but no publisher- or placement-level exclusivity is documented in the sources screened, and its Marketplace Terms grant publishers a non-exclusive licence to its technology [122, 171]. Super League's exclusive with Misfits, signed on 1 May 2026, covers brand partnerships in Misfits' Roblox games that Misfits has not already discussed; it excludes games Misfits does not solely control, needs Misfits' approval before any brand is approached, and runs for an initial year [172].
### Programmatic rails
Networks name DSPs more often than DSPs name networks. Google's exchange list for Display & Video 360 (DV360), its DSP, updated on 10 June 2026, names 89 exchanges, none of them Anzu, Frameplay, Gadsme, Adverty, AdInMo, IION or Gameloft, though it names intermediaries from their files such as Unity, InMobi and Smaato [119]. Frameplay's SDK guide names DV360 as demand, but the route by which DV360 buys Frameplay inventory is undocumented in the sources screened: GumGum, a named DV360 exchange that an undated testimonial on Frameplay's advertiser page quotes as offering Frameplay inventory, lists no Frameplay entry in its own sellers.json [173, 119, 137, 174].
DSP-side documentation was found for Anzu and Gadsme. Anzu's PRWeb release names The Trade Desk, StackAdapt, Basis, Smadex and Adikteev as onboard for its click-enabled formats; StackAdapt's own page (modified 27 August 2026) says it is partnered with Anzu for non-clickable blended ads, and a Smadex partner post confirms a performance partnership [122, 120, 175]. Gadsme sells through private marketplace (PMP) deals, curated deals and direct programmatic partnerships, and the Eskimi DSP documents a Gadsme integration on a page updated in August 2025 with content from 2022 [118, 176].
Elsewhere the record stops at the vendor's side. The latest public list of Frameplay's DSP integrations found dates from December 2021, and this review found no DSP's own documentation of a current one in the sources screened by 29 September 2026 [177, 119]. IION named DV360 and The Trade Desk in posts of February and March 2025, before the scoring period, and neither DSP's own documentation was found [178, 179]. Adverty named PubMatic as its preferred PMP partner in December 2023, Kidoz offers PMP and programmatic guaranteed deals, and neither names a DSP in the sources screened [180, 181]. COMBO! and Super League claim programmatic routes. At IAB PlayFronts in September 2026, Gameloft for brands' Jean-Baptiste Godinot weighed telling advertisers "you can buy me through DV360 or The Trade Desk", which names DSPs without documenting an integration, and no DSP's own documentation or deal type was found for either network [182, 183, 30, 145].
No cohort A network documents support for an agentic buying protocol, or any protocol state this paper records, in the sources screened by 29 September 2026 [9, 10, 11, 121]. AgenticAdvertising.org's public registry of Ad Context Protocol (AdCP) agents listed 34 agents on 28 and 29 September 2026, none operated by Anzu, Frameplay, IION, Gadsme, Adverty, Gameloft, Super League, Kidoz, Admazing or AdInMo, and AdCP's schema can represent in-game products by channel but defers in-game creative formats [121, 184, 185]. Exits, platform growth and the agentic question takes up what the protocols could carry.
### Concentration without a denominator
No cohort A vendor publishes a market share with a stated denominator in the sources screened. Gameloft calls COMBO! "the largest premium in-game ad network", Frameplay's president called its exchange "by far, the biggest pool that you can access" at its launch in October 2024 while declining to share figures, and Anzu calls itself "the leading intrinsic in-game ad solution" [55, 74, 122]. The operators that file public reports publish revenue, in different scopes. Adverty reported net sales of 12.9 million Swedish kronor for the second quarter of 2026, up 79% year on year [79]. Super League reported media and advertising revenue of $1.4 million for the same quarter [30]. Kidoz Inc.'s $18.4 million of 2025 revenue mixes programmatic, partner and direct brand sales, cost-per-install units and an all-ages division [65]. Gadsme publishes no revenue, and Anzu discloses funding statements without a revenue series, in the sources screened [186, 164].
Without a denominator, shares and concentration stay unresolved under the rule for Q2. A ranking by declared supply would mislead, because one publisher can sit behind five networks [9, 10, 11, 12, 14]. The same gap leaves part (b) of H5 untested.
## 4. Platforms and publishers selling ads in their own games
*Roblox, Microsoft, Electronic Arts and Zynga sell brand ads inside games they own or run. They sell mostly through their own teams, and their programmatic routes are documented mainly in announcements. None of them reports the revenue of its in-game brand product in the filings and releases this review screened.*
> **In short.**
> Roblox sells Immersive Ads and Rewarded Video; its programmatic routes appear only in announcements and marketing dated April 2024 to September 2026, and no demand-side platform (DSP) documenting a live integration was found in the sources screened [187, 188, 42, 189, 183]. Its Advertising Standards bar personalised ads for users under 18 [20]. Two Roblox documents still disagreed on 29 September 2026 about when developers start paying Roblox a revenue share on brand integrations: 30 September 2026 or 1 January 2027 [57, 190]. XBOX Media Solutions sells formats in Candy Crush Saga and Microsoft Casual Games, and its gaming pages screened name no programmatic route for them [191, 192]. Electronic Arts (EA) launched EA Advertising on 15 June 2026, and by 15 September 2026 Digiday reported programmatic buying across EA's mobile games [25, 193]. Zynga Ads sells brand formats inside Zynga's own mobile games and announced a PubMatic partnership in July 2026 [194, 195, 196]. Among the cohort's parents only Take-Two reports advertising in its games as a separate revenue line, and that line covers every ad type; the disclosed revenue series differ in metric, currency, period and scope, so part (b) of the market-structure hypothesis cannot be tested [197, 198, 199, 200, 143, 128].
### Owned inventory, and what sits beside it
Cohort B1 holds brand ad products sold on in-game inventory that the platform or publisher controls. It has 4 scored rows at the cutoff: Roblox's brand ads, XBOX Media Solutions' in-game ads, EA Advertising and Zynga Ads. Placements these companies sell outside the game are listed and not scored. They include Roblox's homepage takeover, Homepage and Search tiles and a Homepage Feature still labelled beta on 29 September 2026 [188, 201, 42, 43], and Microsoft's ad-supported cloud streaming test, begun for Xbox Insiders on 23 July 2026, which plays ads before sessions [202]. Zynga Ads+, a third-party network announced on 17 September 2026, is listed and not scored because no participating app was named and no product documentation was found [183, 203].
Roblox's product also reaches buyers through resellers: Venatus in the UK, France and Germany, Playwire in select Asia-Pacific markets, and SuperAwesome as exclusive partner for under-13 inventory [204, 205, 206]. A resale route carries the same inventory, so this paper never adds its reach to Roblox's.
The capability matrix shows quotient bands only for a cohort with at least 6 band-eligible rows, and this cohort has 4. A band measures a row's distance from its own cohort's mean, and in a cohort this small each product is a large part of that mean, so one product's evidence moves the reference for the others. The chapter on the capability matrix and the Play Quotient therefore reports this cohort's quotients and coverage without bands, and nothing here compares these products with the networks in "Who sells inventory inside games".
On 29 September 2026 the Media Rating Council (MRC) showed no accreditation and no MRC-confirmed audit for any product scored in cohorts A and B1; the one gaming audit in MRC's process is Trivver's, a listed product [15, 16].
### Roblox sells two formats through routes announced over two years
Roblox sells two in-experience formats that creators place as ad units and Roblox serves [197, 207]. Immersive Ads are video billboards of up to 30 seconds, static image ads and portal ads that teleport users into an advertiser's experience; Rewarded Video is a full-screen ad of 6 to 30 seconds shown only to users 13 and over [208, 207, 20]. Roblox reported Rewarded Video in over 400 experiences on 6 January 2026 and over 1,200 games on 20 May 2026, said over 1,000 brands had run it, and warned that supply "will take some time to scale" [188, 209, 201]. Its Q1 2026 letter told creators running independent brand integrations to register campaigns in Roblox's Ads Manager and to move to native formats such as Rewarded Video and immersive billboards [210, 57].
The programmatic routes arrive as a sequence of announcements [187, 188, 189]. PubMatic said on 10 April 2024 that Roblox video inventory would "ultimately" reach its buyers through programmatic guaranteed deals, private marketplaces, auction packages and open exchange [189]. On 1 April 2025 Roblox announced Rewarded Video through Google Ad Manager and direct insertion orders "in the coming weeks", with Billboards and Google Authorized Buyers to follow [187]. On 6 January 2026 it announced expansion "this year" with the DSPs Amazon DSP and Liftoff and the supply-side platforms (SSPs) Index Exchange, Magnite and PubMatic; Magnite's post that day said its integration "will make" Rewarded Video available to its buyers [187, 188, 42, 189, 211]. Roblox's brands page, read on 29 September 2026, calls "Magnite, Index Exchange, and Display & Video 360" major DSPs, although the January release classed two of them as SSPs, and limits programmatic targeting to 18-plus inventory [42]. AdExchanger reported that at the Interactive Advertising Bureau's (IAB) PlayFronts event on 17 September 2026 Roblox pointed to integrations with Display & Video 360 and other programmatic platforms [183]. This review found no Amazon DSP or Liftoff page about Roblox in the searches logged [212, 211, 189].
### Roblox's rules bind its own ads and every other ad in its games
Roblox's Advertising Standards, last updated 5 June 2026, state that ads may not be personalised for users under 18, allow rewarded video only for users 13 and over, and ban rewards for under-13 ad engagement [20, 213, 214]. The under-18 personalisation rule reaches most of the age-checked base (73%, derived): of users who had age-checked in the seven days to 30 June 2026, 35% were under 13 and 38% were 13 to 17, and Roblox warns that this group may not represent all users [43, 210, 197].
Roblox announced on 20 March 2026 that from 4 May 2026 age-appropriate formats, billboards included, could reach users under 13 [215]. Its frequently asked questions say those ads "will be contextual and sold directly via Roblox or creators" and will not run through programmatic partners such as Google [215, 216]. Roblox's brand-safety page still said on 29 September 2026 that users under 13 cannot see ads [217].
Under Roblox's managed-campaign terms, effective 30 April 2026, Roblox's reporting is the sole and controlling measurement, advertisers get aggregate reporting only, and an advertiser can dispute invalid-traffic counts, through the invoice dispute process, only when its approved verifier finds a gap above 10% [1].
Ads that developers place outside Roblox's own products fall under the Advertising Integrations Terms, effective 1 May 2026 [57]. They require Roblox's prior approval of every asset, bar programmatic calls to third-party advertising and measurement services, and bar sharing personally identifiable engagement data with advertisers [57]. Roblox's 20 March 2026 announcement also says third parties may not run software development kits (SDKs) that serve, target or measure ads [57, 215]. The Terms keep a registered, Roblox-approved integrations programme for work that does not replace Roblox's formats, so the Q1 letter describes a channelling rule rather than a ban [210, 57].
Roblox's own documents conflict on the developer revenue share [57, 190]. The Terms, created 27 March 2026 and last edited 30 April 2026, say it begins on 30 September 2026, charged per thousand visits while an integration is present, at unstated pricing [57]. The Creator Hub page, live on 29 September 2026, applies it to campaigns starting on or after 1 January 2027, at $1.50 per 1,000 qualified visits in the US for the first 28 days and $0.10 afterwards, with a $400,000 cap per 2027 campaign [57, 190, 215, 218]. In Roblox's public documentation repository the 2027 date first appears in a commit of 17 June 2026, though a Roblox forum post of 20 March 2026 and a Reuters report of 27 March 2026 had already placed the start in 2027 [190, 219, 215, 218]. The two documents still conflicted on 29 September 2026, one day before the Terms' date [57, 190]. This paper records both with their access dates and treats neither as settled.
H4, the paper's hypothesis on children's and platform rules, is supported, because for at least five products or platforms, against a registered threshold of three, a primary rule document restricts targeting, personalisation, third-party calls or measurement: Roblox's rules on personalisation, third-party calls and measurement on its platform, Microsoft's Privacy Statement ruling out personalised ads for Microsoft accounts of users under 18, and rules from Discord, Epic for Fortnite and Google Play [20, 213, 57, 21, 22, 220, 221]. The Federal Trade Commission's amended COPPA Rule, made under the Children's Online Privacy Protection Act, adds separate parental consent for advertising disclosures in child-directed games [23]. The registered rule names Roblox's two revenue-share dates as an example of an unresolved condition; on this review's reading, that conflict concerns developer fees rather than any restriction counted, so it does not block the verdict [57, 190]. The chapter on suitability, user-generated worlds and children's rules covers the law.
### XBOX Media Solutions sells Microsoft's casual games
XBOX Media Solutions sells formats inside Microsoft-owned casual games; the pages screened name no programmatic route for them and do not establish any console placement inside gameplay [191, 192, 24].
It is the former Activision Blizzard Media, renamed in an announcement of 4 February 2026 that describes it as part of Microsoft Advertising [222, 223]. Its pages document in-game Rewarded Video, In-Game Interactive Video, Playables and custom integrations: "'Candy'fied Sponsored Moments" in Candy Crush Saga, a Solitaire Event Takeover in Microsoft Casual Games on PC and mobile, and console "XBOX Experiences" described as native display formats and branded content [191, 224, 24]. The pages call these integrations "100% owned and operated" and do not say which console placements, if any, render inside gameplay [191, 224, 24].
On 12 May 2026 Microsoft announced Rewarded Portals, a Candy Crush Saga format on the level screen, as "currently available in closed beta across select markets" [225, 226]. Candy Crush Saga is the one title named in dated campaigns inside the scoring period (Hershey's, holiday 2025; KitKat, published 18 August 2026); Solitaire appears only on an undated page, and no console title is named [226, 227, 24, 228].
The four gaming pages opened on 28 September 2026 name no DSP, private marketplace, deal type or other programmatic route for the in-game formats [192, 191]. They route buyers to account managers and a contact form [191, 225]. King's inventory is also authorised to or sold through other routes, including Gameloft's COMBO!, Anzu and Gadsme reseller paths, Meta Audience Network (facebook.com DIRECT lines in King's file; Meta's story cites Q4 2020 data) and Microsoft Advertising's own gaming deals, described on 26 September 2024 with their status in the scoring period not established; this paper never adds reach across those routes [53, 229, 230].
The unit's general manager, describes Xbox's approach as "the idea of game-adjacent rather than in-game", and no Microsoft document screened by 29 September 2026 confirms a live in-gameplay console product [231, 232, 233]. The chapter on console, PC and cloud explains why the console hypothesis, H3, stays unresolved.
Microsoft's Privacy Statement (September 2026) says Microsoft does not deliver personalised advertising to children whose account birthdate shows them to be under 18 [21], while King's privacy policy (26 February 2026) says King does not knowingly target interest-based advertising to anyone under 13 [234]. No document read states which rule governs XBOX Media Solutions inventory in King titles [21, 228, 234].
### EA Advertising launched in June 2026, weeks before EA went private
EA Advertising sells units inside EA SPORTS gameplay; it was sold direct at launch, and mobile programmatic buying was reported by September 2026 [25, 193].
Electronic Arts launched it on 15 June 2026, saying advertisers "can now" reach players through native ad units (digital ad boards, scoreboards and brand broadcast overlays) served dynamically in select, unnamed EA SPORTS games from its own ad server and SDK built for the Frostbite engine [25]. Digiday reported on 3 July 2026 that the ad server was live on a handful of titles [25, 86]. Comcast dated dynamic LED ad-boards in EA SPORTS FC 26 Ultimate Team matches in the US on 17 November 2025; no source screened states which platforms carry them [235, 236, 25, 237]. Madden NFL 26, dated only by a fan database, and College Football 26, whose integrations went live in July 2025, before the scoring period, bring the EA console titles with documented brand placements to three [25, 235, 238, 239]. EA documents none of the three as an ad placement buyable in the US, so at most three qualify under H3, and any count from this screen is a floor for the market, because other publishers' console titles were not screened [25].
At launch EA sold console integrations and ad units directly and bespoke, with "expanded buying capabilities" planned [25]. On 3 July 2026 its advertising vice president said EA was not yet working with programmatic partners such as The Trade Desk or Criteo [25, 86]. By 15 September 2026 Digiday reported that EA already offered programmatic buying across its mobile games and was extending more dynamic capabilities to console and PC titles; no DSP or deal type is named [86, 193]. EA's app-ads.txt, the file in which an app declares who may sell its inventory, names Anzu as manager domain and authorises Anzu as a direct seller under IDs that Anzu's sellers.json assigns to Electronic Arts and Playdemic; its one IION DIRECT line uses ID 10329, which IION's sellers.json assigns to another publisher, Big Cake [240, 9, 11, 25].
Electronic Arts has been privately owned by PIF (Saudi Arabia's Public Investment Fund), Silver Lake and Affinity Partners since the acquisition closed (Form 8-K and Form 25-NSE filed 4 August 2026) [241]. A Form 15-12G filed on 14 August 2026 gives notice of the termination of its registration with the Securities and Exchange Commission (SEC) and of the suspension of its duty to file reports, so filings will no longer provide new EA facts [241].
### Zynga Ads sells Zynga's own mobile games
Zynga Ads sells brand formats and integrations inside Zynga's mobile games, and Take-Two, which owns Zynga, is the only parent in the cohort that reports advertising in its games as a separate revenue line [194, 242, 200, 197, 198, 243].
In-period activations name titles: Survivor's 50th season across five Zynga games in March 2026, with commercial terms not stated, and the Hyundai NEXT Cup Tour in Top Eleven in June 2026 [194, 242, 200]. Zynga Ads names seven brand offers, from non-skip video to sponsorships, and claims 85 to 95% video completion and 96% viewability, but its readable pages publish no specifications or counting approach [244, 245].
Zynga has a programmatic sales team [195]. In July 2026 it announced a partnership with PubMatic under which Zynga will integrate PubMatic's OpenWrap SDK and advertisers will be able to buy Zynga inventory through PubMatic's Activate and AgenticOS [196]. No DSP, deal type or protocol state is named [195, 196]. Zynga's app-ads.txt, last modified 23 September 2026, also authorises Gadsme, Anzu, Gameloft, IION and PubMatic as direct sellers, and reach is never summed across these paths [139, 77, 14, 200].
### Disclosed revenue series do not compare
The series that are disclosed differ in metric, currency, period and scope [200, 198, 197, 243, 143, 128, 65].
| Company | What is disclosed | Latest figure | Period | Limit |
|---|---|---|---|---|
| Roblox | Advertising and licensing revenue called "insignificant" | none | Fiscal 2025 and second quarter of 2026 | No figure [197, 246] |
| Microsoft | XBOX revenue, advertising included | $21,790 million | Year to 30 June 2026 | Includes hardware, content, subscriptions, cloud [198] |
| Electronic Arts | Advertising inside live services and other net revenue | none | Fiscal 2026 | No figure; no further filings [243, 241] |
| Take-Two | Advertising net revenue, primarily mobile | $494.3 million | Year to 31 March 2026 | All ad types and offers [200] |
| Amazon | Advertising services, company-wide | $68,635 million | 2025 | Twitch not broken out [199] |
| Adverty | Net sales | SEK 33.7 million | 2025 | In Swedish kronor (SEK) [143] |
| Super League | Media and advertising revenue | $5.6 million | Fiscal 2025 | Includes influencer work [128] |
| Kidoz | Revenue | $18.4 million | 2025 | Mixes performance units and the all-ages Prado division [65] |
Roblox's 10-K for fiscal 2025 (filed 11 February 2026) and its 10-Q for the second quarter of 2026 (filed 30 July 2026) describe revenue from advertising and licensing as "insignificant" and report no advertising line [197, 246]. In its Q4 2025 letter Roblox called advertising revenue "modest" and said its nearly $7 billion of 2025 bookings came "without significant reliance on advertising" [201].
Microsoft's XBOX revenue was $21,790 million for the year to 30 June 2026, against $23,455 million the year before, in a line that combines advertising with hardware, content, Game Pass and cloud gaming [198]. Its new reporting for fiscal 2027 leaves XBOX unchanged, so gaming advertising stays inside XBOX content and services [247]. EA's fiscal 2026 10-K folds advertising into live services and other net revenue [243].
Take-Two reports Advertising net revenue of $494.3 million for the year to 31 March 2026, $466.1 million for fiscal 2025 and $656.1 million for fiscal 2024, or 7.4%, 8.3% and 12.3% of total net revenue (derived); its 10-K places this advertising primarily in mobile titles and includes offers and sponsorships [200]. Take-Two contracts with ad networks, agencies, brokers and directly with advertisers, so the line cannot isolate Zynga Ads' brand sales [194, 242, 200].
Part (b) of H5, the market-structure hypothesis, asks whether platform ad revenue grew faster than specialist revenue over the same periods. It cannot be tested: Roblox gives no figure, Microsoft reports gaming advertising only inside a broader gaming line, Amazon reports only company-wide advertising services, and the disclosed series (Take-Two in dollars for years to March, Adverty net sales in kronor, Super League including influencer work) differ in metric, currency, period and scope [197, 198, 199, 200, 143, 128]. Part (a) meets its threshold: primary records show three specialist companies sold or put into insolvency since 2021 (Bloxbiz, Bidstack and Let's Bounce), and none of the records attributes an exit to the in-game format [29, 31, 32, 30]. H5 is unresolved; chapter 10 classifies each event.
## 5. Around the game: game-adjacent media
*Brands buy the media around games through a few routes: Amazon for Twitch, a sales team for Discord, and direct, private or outsourced programmatic deals for companion apps, gaming sites and esports channels. The accreditor's list names none of the products scored here. Every audience figure is the seller's own, counted its own way.*
> **In short.**
> Twitch is sold by Amazon, self-service through Amazon's demand-side platform (DSP) or as a managed service [2]. During the scoring period Discord sold Quests only through its sales team, and on 29 September 2026 it announced an auction-based Ads Manager in managed access [48, 248]. Overwolf's companion-app ads, gaming sites and esports channels sell direct, through private marketplaces and through outsourced intermediaries [44, 249, 250, 251]. The Media Rating Council (MRC) accredits Amazon DSP for supply sources that do not include Twitch, and its list names no product scored in this cohort [15]. Audience claims run from Medal's 10 million monthly users to IGN's more than 440 million, on different metrics, and cannot be added [252, 253]. YouTube is listed only: the one named gaming product found in Google's own sources, a Shorts lineup announced in April 2024, could not be confirmed live [254, 255].
### Where the ad renders sets the cohort
This paper puts a product in cohort B2 when the ad renders in media around a game: a stream, a community app, a companion app, a clip site, a gaming website or an esports channel. Event sponsorship on its own is described and never scored. Placement and objective are separate fields ([What counts as gaming advertising](#defining)), so a game publisher buying play counts as paid discovery wherever the ad appears. [The capability matrix and the Play Quotient](#matrix) scores 10 products here: Twitch, Discord Quests, Overwolf, IGN, Venatus, Playwire, Enthusiast Gaming, ESL FACEIT Group, LIVAD and Medal.
YouTube is listed only. The one named gaming-specific YouTube buying product found in Google's own sources is a gaming lineup of YouTube Select Shorts, announced on 17 April 2024 for six markets including the US, and its status at the cutoff could not be established [254, 255].
The author's GAP model splits the category into in the game, around the game and away from the game. Venatus sells in those three categories, but it puts in-game audio under the first and embeds Twitch streams in website takeovers, so its categories, by inference, overlap this paper's [83, 256, 84]. No source screened by 28 September 2026 sizes the three as mutually exclusive, reconcilable categories, so GAP totals stay unresolved [37, 4, 38].
### Streams: Twitch through Amazon, overlays through LIVAD
Twitch media is bought from Amazon. Amazon Ads sells video and display on Twitch's homepage, discovery pages and streamer channels, self-service through Amazon DSP or managed with an unpublished minimum spend, and sells creator activations through account executives [2]. Twitch's Terms of Service give Twitch the exclusive right to sell and serve ads on Twitch and bar third-party ad networks [257]. Amazon DSP is the only demand-side platform named in the documents opened: display as open auction from 14 December 2022 in the US, and video through premium preferred deals, generally available to self-service buyers from 25 October 2023 [258, 259]. Amazon's streaming TV ads, which Amazon says can appear on Twitch, are a further route open to US brands [260].
Creator routes are sold separately: Bounty Board sponsorships, Channel Takeovers with streamer-read ads, and a Brand Partnership Studio [261, 262, 2]. One Twitch product renders inside games: Streamables plays an opt-in, non-skippable 30-second livestream in "Twitch Partnered" mobile games, none of them named [263]. Premium Video also runs on consoles, where it is not clickable [47]. By inference, it plays in the Twitch app outside gameplay, which makes it adjacent media on a console device that [Console, PC and cloud](#console) weighs under H3, the paper's console hypothesis [47, 264].
Twitch documents advertiser suitability controls in operational detail. Since 20 June 2023 streamers must apply Content Classification Labels (six at launch, seven in Twitch's current help article and API reference), and a Mature-rated game label is added automatically to streams of games the Entertainment Software Rating Board (ESRB) rates M [265, 266, 267, 268]. Amazon's guide of 22 February 2026 says the labels inform Twitch's suitability tools, among them three inventory tiers (Expanded, Standard, Restrictive) and Contextual Collections that target or exclude content such as Family Friendly games, and that new creators cannot carry ads [266]. A launch note of 6 November 2023 named the tiers Expanded, Moderate (Default) and Restrictive; neither document says what each tier excludes [269, 266].
LIVAD sells brand overlays inside creators' live streams, with gaming as its documented focus and no published split of its streams by category [270, 271, 272]. Trade reporting of 18 September 2026 describes overlays and sidebars as its core business, but no named campaign inside the scoring period was found [273, 274, 272]. Its streamer contract defines a view as the count of viewers during each branded overlay session, billed per view or per thousand [275]. Its ads.txt (a site's public list of authorised sellers) names Adform, Improve Digital and the in-game network Anzu, and Adform's and Anzu's sellers.json files (intermediaries' public lists of their sellers) list LIVAD, while Improve Digital's does not; no buyer-side route is documented [276, 277, 9, 278]. Its contracting company is in Istanbul, and no US campaign, sales contact or buyer terms were found [279, 275].
### Communities and clips: Discord Quests and Medal
Discord sold Quests, rewarded ad units inside the Discord client, only through its sales team during the scoring period, and its advertiser pages still said so on 30 September 2026 [48]. On 29 September 2026 it announced Discord Ads Manager, its own auction-based campaign tool with objective-based bidding, available with managed access to a select group of advertisers and planned as fully self-serve in early 2027, with a new mobile-first Video Quests format sold through it; the earlier Video Quests became Video Quests Reserve, a reservation-based offering [248]. Buyers sign an insertion order under the Interactive Advertising Bureau (IAB) Standard Terms, as amended by Discord's addendum, updated 12 May 2026 [280]. The addendum treats Quests as flat-fee deliverables with no third-party ad server, gives Discord "ultimate authority and control" over the placement and frequency of Quest prompts, and makes impression and reach numbers non-binding estimates [280].
The formats move faster than the advertiser pages, which listed two on 29 September 2026: Play Quests and Video Quests [62]. Discord's releases added Video Quests on Mobile and an Arena Quests alpha (2 October 2025), announced Instant Play Quests as forthcoming (9 March 2026), announced Mission Quests on 22 July 2026 after a beta running by December 2025, and announced Play Quest+ on 20 August 2026, with a first campaign for Electronic Arts' Battlefield 6 planned for September 2026 [62, 281, 282, 283, 284]. No later Discord source found states the Arena Quests alpha's status [281].
Quests mix objectives. Of the 21 success stories on Discord's ads page on 29 September 2026, 14 are games, several reporting acquisition or sales outcomes; 4 are film or TV releases; and 3 are non-entertainment consumer brands (Mountain Dew, Wendy's, Walmart), two of them tied to entertainment properties (World of Warcraft; Netflix's Wednesday) [62, 48, 63]. Play Quests are completed by playing the advertised game for a set time or by a linked in-game action [285, 62]. Bought by a game publisher, a Play Quest sits close to paid discovery. This review keeps Quests as one brand row with the mix recorded, since no split by objective is public [62, 63].
Targeting is audience-based: countries, interests, games already played and broad age ranges from self-declared age, with no buyer-selectable contextual signal documented [62, 286, 22, 287]. Age rules restrict the product itself. Users aged 13 to 17 can complete at most three Quests in 24 hours, and Discord says it uses age-assurance signals only to keep adult ads away from teens and does not use them to target ads [285, 22]. Its Ads Policy bars ads intended solely for children under 13, its minimum age [288]. On 22 September 2026 Discord began placing accounts worldwide into adult (18+) and teen (13 to 17) age groups; the post, as read, says nothing about ads [289]. These rules are among those that support H4, as [Suitability, user-generated worlds and children's rules](#safety) sets out [22].
Medal, a gameplay-clip app and website, entered the register after the second AI-model review. On 29 September 2026 it sold IAB-standard units, a homepage roadblock, sponsored stories, branded watermarks and clip contests in its own app and site, and its ads.txt listed Google, OpenX, Magnite and PubMatic as direct sellers [252, 290]. Its one case study is undated and reports game-promotion outcomes, and Medal Quests are partnerships with game developers, so the Quests are listed as paid discovery [252, 58]. The ad units themselves are scored in cohort B2 [252].
### Companion apps: Overwolf
Overwolf's app ads render inside third-party desktop companion apps that run over the game [44]. Overwolf says they appear within the app's boundaries, during downtime or when the player opens the app, with placement set by the app developer, and it describes them as served in-game on the game screen [44, 291]. This review keeps the product in cohort B2 as a recorded judgment. By inference, Overwolf's software development kit draws the ad inside the companion app's window, so the ad is rendered by the companion app, not by the game [44, 45, 46]. Its custom activations in Minecraft and Fortnite maps are listed separately [292].
Overwolf also sells ads on more than 500 gaming websites through Nitro, formerly NitroPay, whose acquisition it announced on 8 May 2024 [293]. Buyers deal direct or through a Deal ID or private marketplace (PMP) [44]. Its FAQ names The Trade Desk and other DSPs within a list of mostly supply-side platforms, and no confirmation from the DSPs themselves was found [44, 294, 46].
### Gaming sites: IGN, Venatus, Playwire and Enthusiast Gaming
Gaming sites sell through direct teams, private marketplaces and outsourced monetisation firms. IGN advertising is sold by IGN Entertainment, Inc., a Ziff Davis subsidiary, through sales teams in the US, UK and Australia [295, 253, 296, 297]. It publishes specifications for 24 ad units across eight named sites but no counting approach for impressions or views [297, 298]. IGN documents measurement only as support for the buyer's own vendors. It supports monitoring tags but no blocking tags, because gaming content is often falsely flagged as unsafe [299, 298]. Its IMAGINE audience platform, announced on 24 September 2025, sat behind a sign-in on 29 September 2026, so its general availability at the cutoff is unconfirmed [300, 301].
Venatus changed owner in the scoring period. On 19 June 2026 Media Scale Capital Limited, a UK holding company incorporated that month and controlled by the Dutch company To Media Groep, took control of Venatus's holding chain, and the private equity fund Livingbridge 7 Global LP ceased to control it the same day [302, 303, 304, 305]. Venatus sells takeovers, video and banners on a publisher network it says reaches 400 million or more monthly active users [84, 256]. Buyers use a curated PMP or direct sales with guaranteed impressions, and no Venatus page opened names a DSP [249, 306]. Its claim of exclusive inventory on OP.GG does not hold globally or in the US: OP.GG's ads.txt of 28 September 2026 authorises Nitro and Overwolf as direct sellers and names Venatus as manager only for the UK, France, Germany, Australia and New Zealand [84, 307, 308, 309].
Playwire, a publisher monetisation company, sells an advertiser-facing product, "Access 74M passionate gamers across 200+ premium communities", bought on insertion orders under terms of 21 November 2025 [310, 311, 312]. Its 74M+ figure, labelled both "Monthly Active Gamers" and "total US gaming reach", has no measurement provider or dated period [313]. Its PC in-game video and Minecraft video are listed, because its pages do not say where either ad renders [314, 313]. Playwire places the Minecraft video on third-party multiplayer servers whose core audience it describes as aged 13 to 15 [315, 316].
Enthusiast Gaming sold its direct sales business to Vertiqal Studios for C$900,000 in cash plus assumed liabilities, closing on 30 August 2025, and kept U.GG, Icy Veins, The Sims Resource, PocketGamer and other sites [317]. Its filings say the retained inventory is sold programmatically and only "to a much lesser extent" through direct sales [318, 250]. They name Playwire as its partner, yet on 29 September 2026 the ads.txt files of all five retained sites checked were hosted by Freestar, with Freestar as manager [250, 319, 320]. No measurement function or independent verification was documented for the retained inventory in the sources screened [15, 250]. A sale does not show that a format failed; [Exits, platform growth and the agentic question](#structure) covers the category's exits.
### Esports media: ESL FACEIT Group
ESL FACEIT Group (EFG) sells media on its esports broadcasts and channels. Savvy Gaming Group, wholly owned by Saudi Arabia's Public Investment Fund, bought ESL and FACEIT and merged them in a deal announced on 24 January 2022, and the ESL sale closed on 21 April 2022 [321, 322]. EFG's Ad Solutions page, published on 28 July 2026, sells premium media on live and on-demand broadcasts and its own channels, alongside custom integrations such as Fortnite worlds; only the premium media line is the cohort B2 product [323]. The page names no DSP or deal type, but EFG's web inventory is offered programmatically through Publift for faceit.com and through Nitro for mobalytics.gg [251, 324, 325, 326]. Esports Insider reported on 17 March 2025 that EFG had acquired the analytics platform Mobalytics [327]. Much of the audience sits on other platforms: in December 2025 EFG listed 374 Twitch channels and 277 YouTube channels [328]. A FACEIT mission privacy notice, updated 22 July 2026, says it does not provide personalised ads to players under 18 [329, 330]. EFG's case studies report delivery metrics only, and no US sales contact or buyer terms were found [323, 331].
### What is measured, and by whom
MRC's accredited list, read on 29 September 2026, names no product scored in this cohort [15]. Its Amazon DSP entry covers Retail and IMDb, Amazon Publisher Services, Fire TV, third-party exchanges and Alexa connected devices, and no entry names Twitch, Discord, Overwolf, Playwire, Venatus, Enthusiast Gaming or ESL FACEIT Group [15]. MRC's letter of 7 November 2025 accrediting the Amazon DSP integration of Integral Ad Science (IAS) says "FireTV and Twitch supply sources are excluded" [332]. Searches of MRC's list and of the Trustworthy Accountability Group (TAG) registry found no record for LIVAD, IGN or Medal [15, 333, 334].
MRC's accreditations for Google's video tools and for YouTube data are wider and have no gaming scope. They cover Google Ads video impressions and TrueView views, including on connected TV; video viewability on desktop, mobile web and mobile app only; YouTube Brand Safety; and DoubleVerify, IAS and Zefr reporting of YouTube data [15]. No entry is gaming-specific [15].
Third-party verification rests on announcements. DoubleVerify (31 October 2023) and IAS (29 May 2024) each announced server-to-server measurement of viewability and invalid traffic on Twitch [335, 336]. TAG records Twitch as Brand Safety Certified for its global operations and Venatus for its UK operations only [337, 338]. Overwolf's FAQ still lists MOAT among its monitoring tools, although Oracle's end-of-life FAQ set 30 September 2024 as Moat's end date [44, 339]. Discord uses no third-party ad server for Quests, which leaves no documented route for verification tags; it measures through Kantar brand lift, AppsFlyer and a Gamesight integration in beta [280, 340, 281].
[Measurement](#measurement) finds H2 supported for cohorts A and B1: no accredited intrinsic in-game metric, and verification resting mostly on partnership announcements dated 2022 to 2025. H2's rule does not cover this cohort; by inference, the game-adjacent record shows the same pattern. Outcome figures here also come from sellers, such as Discord's Kantar study for DOOM: The Dark Ages, whose headline tiles cite "Discord Internal Data, 2025" [341]. [What the outcome evidence shows](#outcomes) weighs such results; attention claims belong to the companion attention paper.
### Audience figures do not add up
Every audience figure in this cohort is the seller's own, and each counts something different.
| Seller | Figure | What it counts |
|---|---|---|
| IGN | more than 440 million | monthly users of 15 properties, worldwide (Ziff Davis 10-K); no period or US split [253] |
| Venatus | 400 million or more | monthly active users across its publisher network; undated [84] |
| ESL FACEIT Group | 287 million | social fan reach, labelled 2025 [342] |
| Discord | 200 million or more; 90 million or more | monthly active users (2025 releases); daily active users (fourth quarter of 2025); global [281, 48] |
| Overwolf | 113 million and 100 million | monthly active gamers, both shown; no period or deduplication [343, 291] |
| Twitch | 105 million | average monthly site visitors, global, 2023 [2] |
| Enthusiast Gaming | 89 million | unique visitors in the quarter ("gamers served"), second quarter of 2026 [318, 344] |
| Playwire | 74 million or more | "monthly active gamers", also called "total US gaming reach"; no provider or period [313] |
| Medal | 10 million | monthly active users, 45% in the US; listed under "In the past 12 months" on an undated page; base unstated [252] |
| LIVAD | none | "thousands" of creators; no figure with a metric and period [345, 346] |
These figures cannot be summed. They count site visitors, quarterly unique visitors, daily users and social reach, several have no period, and none is stated as an ad-eligible audience [281, 2, 344, 342]. The supply overlaps as well. Playwire's sellers.json lists 22 Enthusiast Gaming entries, 34 Overwolf entries and OP.GG, and shares roughly 66 to 67 root domains with Venatus's file and 82 to 84 with Nitro's [309, 311, 308, 326]. IGN sells pre-roll on its YouTube channels, and EFG's reach sits partly on Twitch and YouTube [298, 328]. Without a deduplicated measurement, audience overlap across these properties stays unresolved.
## 6. Measurement: standards, verification and accreditation
*Intrinsic in-game advertising has a finished measurement standard and no product accredited against it. The Media Rating Council's (MRC's) own record, read on 29 September 2026, shows no accredited intrinsic in-game metric and no MRC-confirmed audit for any in-game product this paper scores. Interstitial, banner and video ads in game apps fall under the general mobile in-app standards, where DoubleVerify and IAS hold accreditations that name no game. What buyers can measure is documented mostly by sellers; who checked it is documented mostly in partnership announcements.*
> **In short.**
> The Intrinsic In-Game (IIG) Measurement Guidelines 2.0 are a final joint release of the Interactive Advertising Bureau (IAB), IAB Tech Lab and MRC, dated August 2022, and set counting and viewability rules for ad-served intrinsic placements only [17, 52]. The IAB Gaming Measurement Framework of 26 June 2025, issued by the IAB alone, is a menu of metrics without thresholds [347, 348]. On 29 September 2026 MRC's accredited digital list had no gaming environment, and the one gaming audit in MRC's process was Trivver's, a product this review lists but does not score [15, 16]. MRC removed Integral Ad Science's (IAS's) in-game service from its process in January 2026 after IAS withdrew [349]. DoubleVerify's and IAS's current MRC letters name no gaming or console environment [18, 19]. Third-party verification rests on announcements dated 2022 to 2025, and several sellers' certification claims are missing from the certifiers' own lists [350, 351, 333, 352]. The paper's hypothesis H2, that verification and accreditation of in-game products are thin, is supported.
### Two documents with different jobs
The IIG Guidelines 2.0 are a joint release of the IAB, IAB Tech Lab and MRC, the US body that accredits measurement services [17, 15]. The first version appeared in September 2009 [17]. Version 2.0 opened for public comment on 14 June 2022, closed on 15 July 2022 and was released in August 2022; MRC's standards list, read on 28 September 2026, still shows August 2022 as the last update [17, 50, 51, 52].
The guidelines cover impression, viewable-impression and audience measurement of in-game placements served by an ad server with an ad session, on desktop, mobile and console, including augmented and virtual reality headsets [17]. Hard-coded branded elements sit outside direct scope because they generally lack an ad session [17]. Interstitials, web banners and in-stream or outstream video fall under the general IAB and MRC guidelines, and interaction, engagement and outcomes are not covered [17]. A claim of IIG compliance therefore speaks mainly to ad-served intrinsic placements; for the excluded formats the guidelines say only that their in-game guidance should also be considered [17].
> **What the IIG guidelines count.**
> An impression counts on the client once the ad has loaded and begun to render [17]. Nothing need be visible, and server-side count-on-decision is not acceptable [17]. A viewable impression needs at least 50% of the creative in view for one continuous second (two for video) and a visible ad area of at least 1.5% of the screen [17]. The 55-degree angle limit on any one axis is a recommendation, and measurers may use other thresholds they can support empirically [17]. Occluded impressions are reported separately, polling runs every 200 milliseconds unless a measurer validates and discloses a less frequent interval, and the power state of the TV or monitor must be measured and considered [17]. Reporting is split by platform and by format [17].
The guidelines strongly encourage independent audits of in-game ad serving but do not require them [17]. They do require measurers to vet and periodically review software development kit (SDK) and application programming interface (API) inputs to measurement [17]. On this review's reading, a seller can claim compliance without breaching the guidelines and without anyone having checked [17, 353].
The IAB Gaming Measurement Framework does another job [347]. The IAB released it alone on 26 June 2025, with no version number, and the ungated text of its guideline page does not mention MRC, accreditation or audit [348, 354]. It lists gaming ad formats by media type (display, video, audio, custom) and sets baseline and additional metrics for each; viewability is a baseline metric, with no counting thresholds [347]. Its display matrices refer the intrinsic column to the IIG guidelines, and its closing page calls them "the primary guideline for gaming platforms and providers", although the guidelines' own scope stops at ad-served intrinsic placements [347, 17]. It says its general metrics, invalid traffic among them, are not always readily available on PC and console [347].
### What buyers can measure
Sellers document their own counting [355, 59]. Frameplay aggregates player view metrics for up to 60 seconds per displayed ad into one impression, weighted by time on screen, screen share, orientation, skew and obstruction [355]. Its 2022 release says viewability follows Frameplay's own proprietary requirements, and no source opened states alignment with the IIG guidelines [355, 356, 357]. Gadsme's terms count a display impression when the ad is displayed and visible and a video impression at video start [59, 95]. For publisher-sold campaigns on its supply-side platform (SSP), display impressions are counted by "the proprietary viewability method designed by Gadsme" [59].
Anzu's developer guide of 4 December 2024 lists the IAB and MRC viewability thresholds and defines "Billable Impressions" as views that meet all visibility criteria [358]. Its measurement guide of 22 August 2024, under a heading about separating impressions from viewable impressions, misstates the standard: it says an ad must be viewed for at least one second at under 55 degrees "to be considered an impression" and that "having the ad load no longer counts" [359, 17]. Counting only viewable impressions may be Anzu's own stricter policy; the guide presents it as the rule [359].
Roblox's image-ad impression uses the same headline numbers as the IIG criteria for a viewable display impression: 50% of pixels, 1 second, 1.5% of the viewport and an angle of up to 55 degrees [208, 17]. Its wording does not say the second is continuous, that the 1.5% refers to the visible part of the ad, or how repeat exposures in a session are counted, so equivalence is unverified [208, 17]. Its autoplay video impression, at 0.5 seconds, is shorter than the guidelines' two continuous seconds [208, 17]. Roblox applies these definitions to its Ads Manager (beta) reports and does not claim alignment with the guidelines [208].
IION lists attention metrics from Amplified Intelligence, brand-lift surveys and viewability guarantees of 90% for display and interstitials and 95% for rewarded video, figures its page calls illustrative, and it documents no counting aligned to the IIG guidelines [127, 360]. Adverty added Open Measurement SDK (OM SDK) viewability reporting on Android on 31 March 2026 and on iOS on 19 June 2026 [80]. Electronic Arts (EA) says its impression measurement is "aligned to IAB standards" without naming the standard [25, 237]. Super League's 10-K lists brand lift studies and measurement and reporting, with no viewability, attention or reach function and no method [128, 30]. XBOX Media Solutions names eight measurement partners without saying what each measures, and footnotes its headline 98% viewability and 96% completion figures to Moat mobile in-app benchmarks from Q1 2023 [361, 362, 224].
Contracts decide whose numbers settle the invoice. Roblox's managed-campaign terms, effective 30 April 2026, make its own reporting "the controlling measurement for all campaigns" and give advertisers aggregate reporting only [1]. An advertiser may raise an invalid-traffic discrepancy through the invoice dispute process when its approved third-party verifier finds one exceeding 10% [1]. Anzu's advertiser terms make its reports conclusive unless the advertiser's weekly independent reports differ by more than 5%, in which case an Anzu-approved third-party vendor settles the amount [363].
Oracle's end-of-life FAQ of 12 August 2024 set 30 September 2024 as the end date for all its advertising products, Moat Analytics and Moat Reach included [339]. On 28 September 2026 Anzu's media-value page still listed MOAT among its measurement partners [364, 339]. Overwolf's advertising FAQ and Playwire's advertiser terms also still name MOAT [44, 312].
### Who checked it: MRC accreditation
MRC's record shows no accredited intrinsic in-game metric [15]. Its Accredited Services: Digital list uses four environment codes (desktop, mobile web, mobile in-app, connected TV) and has no gaming, in-game or console entry [15]. On 29 September 2026 MRC's records showed no accreditation and no MRC-confirmed audit for any of the 9 cohort A and 4 cohort B1 products scored in this paper [15, 16].
The one gaming entry in MRC's process is Trivver: pre-audit reviewed July 2022, audit in process as of April 2026, for impressions, "T-Views" and product actions across desktop, mobile web, mobile in-app and gaming consoles [16]. MRC's entry does not name the standard applied, and an audit in process confers no accreditation [16]. This review lists Trivver's product and does not score it: Trivver calls a browser-game SDK released on 11 June 2026 its "first commercial deployment", and this review found no named live title or brand campaign in the scoring period, in the sources screened [155, 16].
A verifier's in-game service entered MRC's process and left it [349]. MRC's under-review list, captured on 10 October 2025, showed "IAS In-Game" in pre-audit for display and video rendered and viewable impressions in desktop, mobile and gaming console environments [365]. In January 2026 MRC removed it from in-process status because IAS did not intend to engage in a third-party audit and withdrew from the accreditation process [349, 16].
MRC's current letters for DoubleVerify (13 April 2026) and IAS (9 July 2026) accredit measurement in desktop, mobile web, mobile in-app and, for some metrics, connected TV [18, 19]. None of their current letters, including IAS's 7 November 2025 letter for third-party reporting of Amazon's demand-side platform (DSP), names a gaming or console environment or lists the IIG guidelines among the standards the accredited statistics comply with [18, 19, 332].
Both accreditations reach mobile in-app traffic measured through IAB Tech Lab's OM SDK [18, 19]. This review infers that standard display and video ads served through the OM SDK in mobile game apps plausibly fall within accredited measurement, while rendered 3D intrinsic placements, billboards on user-generated platforms and console environments fall outside it [18, 366]. MRC's record names no game, so that reach is unattested [18, 15].
For game-adjacent media, MRC's accreditations touch YouTube only in general terms, with no gaming-specific scope, and no MRC accreditation was found for Twitch, Discord or Venatus; MRC's letter for IAS's Amazon DSP reporting excludes Twitch supply [15, 332]. The chapter Around the game covers those products.
*Figure fig-assurance. **MRC's record lists no accredited intrinsic in-game metric; assurance rests on verifier announcements** Scores from the capability matrix for the cohort A and B1 rows on four dimensions. No product reaches 4 on independent assurance, which needs third-party verification documented in the product's own environments together with a certification on the certifier's record, or with an MRC accreditation for an adjacent environment or an audit in progress. MRC's own record, read on 29 September 2026, shows no accredited intrinsic in-game metric for any of these products; interstitial, banner and video ads in game apps fall under the general mobile in-app standards instead. Third-party verification rests mostly on partnership announcements dated 2022 to 2025. Units: score 0 to 5 (n/e = no evidence). Scope: cohort A and B1 rows admitted to the matrix. Sources [546], [396], [130], [15], [16], [389], [147], [53], [547], [129], [548], [80], [353], [549], [386], [385], [13], [81], [550], [125], [123], [364], [394], [358], [350], [551], [339], [383], [382], [9], [171], [397], [363], [445], [356], [552], [475], [479], [553], [369], [406], [355], [357], [387], [10], [75], [96], [78], [76], [77], [464], [554], [555], [466], [372], [373], [118], [531], [59], [556], [392], [398], [12], [97], [170], [95], [444], [55], [377], [14], [126], [54], [405], [127], [360], [388], [116], [11], [162], [557], [558], [559], [399], [560], [561], [562], [563], [564], [66], [565], [69], [181], [65], [128], [30], [566], [145], [400], [567], [568], [569], [57], [25], [376], [240], [410], [1], [42], [187], [351], [370], [486], [488], [208], [17], [570], [380], [538], [571], [572], [539], [573], [217], [20], [213], [214], [190], [414], [215], [408], [361], [362], [227], [226], [224], [390], [391], [352], [223], [574], [409], [234], [21], [228], [425], [244], [203], [378], [411]. Caveat: Scores measure public documentation against the anchors, capped by evidence grade. They are not measures of product quality. Data: data/capability-matrix.json*
### Verification rests on announcements
Third-party verification of in-game inventory is documented mainly in partnership announcements [350, 351]. IAS's own in-game announcements name viewability and invalid-traffic (IVT) functions only: Anzu on mobile (28 July 2022), extended in an IAS post of 11 July 2023 to 2D and intrinsic 3D placements on mobile, PC and console, Gadsme (15 December 2022), Frameplay (5 October 2023, IVT and impression counts) and Roblox (announced on 1 May 2024 for later that year, with an IAS page dated 12 September 2024) [350, 367, 368, 369, 370, 371]. In the sources opened, none dated inside the scoring period confirms that the Frameplay integration is still live, and no verifier record confirms that Gadsme's is running in 2025 or 2026 [369, 137, 372, 373]. Roblox restates its relationships on a current page: on 28 September 2026 its brands page named IAS and DoubleVerify for independent verification [42].
DoubleVerify formally entered in-game measurement on 27 July 2023, Digiday reported, and a DoubleVerify product director said none of its tools needed a significant update [366]. Its Roblox post of 1 April 2025 splits coverage by format: for in-game billboards it extended fraud and viewability measurement through a custom integration on Roblox's direct-sold inventory; for Rewarded Video, then about to launch, it described its functions in the future tense [351]. In the sources screened by 29 September 2026, DoubleVerify's own records document gaming coverage in two places, Roblox and Microsoft Casual Games (named in a 15 October 2025 release among Microsoft properties), and an IAB event listing for a Zynga Ads session reports a DoubleVerify attention figure for Zynga [351, 374, 375, 203]. No DoubleVerify record was found for Anzu, Frameplay, Gadsme, Adverty, IION, Gameloft for brands, Super League, Kidoz or Admazing [351, 375, 203].
Other sellers state partnerships for which no verifier record was found [25, 55]. EA says it measures viewability and invalid traffic "in partnership with Integral Ad Science", but IAS's published sitemaps, read on 29 September 2026, hold no page about Electronic Arts while they hold in-game pages for Anzu and Gadsme [25, 376]. Gameloft labels its COMBO! offer "IAS compliant" and cites an IAS data study; no IAS record of the relationship was found [55, 377]. Zynga shows DoubleVerify's logo on its ad-solutions page; no verifier or MRC record for Zynga Ads was found, and on 29 September 2026 the Trustworthy Accountability Group (TAG) registry listed Zynga as enrolled in Verified by TAG with no seal [244, 378, 333, 379].
DoubleVerify agreed on 6 August 2026 to be acquired by the parent company of Nielsen Company (US), LLC, with closing expected by Q1 2027 subject to a stockholder vote and regulatory approvals [380, 381]. Roblox named DoubleVerify and Nielsen as new measurement partners on 1 April 2025, so if the deal closes, two of Roblox's named measurement partners would be under common ownership; this is an inference from the records [187, 381].
### Certification claims against certifier records
Several sellers claim certifications that the certifiers' own lists did not show on 28 or 29 September 2026; each absence is bounded to the names queried, and none is a documented lapse [333, 352]. Anzu announced on 12 August 2021 that it had earned TAG's Brand Safety Certified seal, validated by the third-party auditor ABC, TAG's Certified Against Fraud seal and IAB UK Gold Standard 2.0 [382]. Neither the TAG registry nor IAB UK's certified and registered list showed Anzu on 28 September 2026, while a control query for Gadsme returned its record [382, 333, 352, 383].
Anzu's boilerplate of January 2024 called Anzu "compliant with the industry measurement standards outlined by the MRC and IAB", and Adverty's home page says its tracking is "verified by leading industry stakeholders and 100% in line with the latest IAB and MRC measurement guidelines" without naming a verifier; neither appears on MRC's lists [384, 353, 15, 16]. The TAG registry has no Adverty record, and IAB Tech Lab lists it only in an other-industry-compliance tab linked to IAB UK's member directory [385, 386]. Frameplay's about page says it was among the first to be Certified for Transparency by TAG in 2023; the registry returned no Frameplay record [357, 387].
XBOX Media Solutions claims "IAB UK Gold Standard 2.1 Certified", but IAB UK's list does not include XBOX Media Solutions, Activision Blizzard Media, King or Microsoft [223, 352]. IION says it is "officially accredited" for the OM SDK (10 September 2024) and shows TAG seal images, yet IAB Tech Lab's compliant-companies page has no IION entry and the TAG registry returned no company [388, 116, 386, 385]. Gameloft's only third-party assurance on record is IAB Tech Lab's OM SDK compliance for its mobile apps, last certified 16 January 2026; its "IAS compliant" and "IAB gold standard in 2018" labels have no certifier record behind them in the sources opened [386, 389].
TAG does record some gaming sellers: Activision Blizzard Media and King.com Ltd. each hold Brand Safety Certified for the United Kingdom only, by independent validation, and Certified Against Fraud by self-attestation for North America and other regions [390, 391]. Gadsme is enrolled in Verified by TAG with no seal, and DoubleVerify and IAS hold both seals globally [333, 390, 391, 392]. TAG seals certify company-level programmes on fraud, brand safety and transparency; they say nothing about in-game measurement [333]. IAB Tech Lab's programmes page, read on 28 September 2026, lists no gaming-specific compliance programme, and its OM SDK compliance certifies an SDK integration only [393, 386].
### Attention metrics in games
Attention measurement in games is offered by several parties, each on its own statement, and no accreditor record found attests it for an in-game environment [351, 18, 356]. DoubleVerify said on 1 April 2025 that advertisers would be able to measure attention with DV Authentic Attention on Roblox Rewarded Video, then about to launch [351]. Its MRC accreditation for Authentic Attention covers display and video in desktop, mobile web and mobile in-app and does not name Roblox or in-game environments [351, 18]. An IAB event listing for a Zynga Ads session reports a DoubleVerify Attention Index of 204 across more than 250 million Zynga impressions measured from April to June 2026 [203]. Frameplay announced an "Intrinsic Time-in-View" metric on 5 April 2022, which it says Lumen and eye square each validated against their own eye tracking on one campaign [356]. IION offers attention metrics from Amplified Intelligence, and Anzu's release of 27 March 2025 says its partnerships already covered attention [360, 394]. Around the game, Overwolf cites Adelaide's Attention Score on its audience page and lists Adelaide among its "Measurement & Analytics" service providers, and Venatus announced a Lumen attention analysis on 18 March 2026 with no sample, period or geography [294, 343, 395].
The IAB framework lists attention, brand lift and incremental sales as additional metrics that are not guaranteed for every implementation [347]. What attention metrics define, and whether they are valid, is the subject of the companion paper on attention measurement. The results reported with them are weighed in the chapter What the outcome evidence shows.
### H2 coded as supported
H2 holds that intrinsic in-game measurement is standardised but that independent verification and MRC accreditation of in-game ad products are thin. The first part of its registered rule holds: MRC lists the IIG Guidelines 2.0 as last updated in August 2022, a final release of the issuing bodies, and the IAB published its framework on 26 June 2025 [52, 17, 348]. The second part needs MRC's record to show accreditation for no more than one third of the products scored in cohorts A and B1, or verification documented in operational detail for no more than half. MRC's record shows no accredited intrinsic in-game metric and no MRC-confirmed audit for any of them; the one gaming audit in process is Trivver's, a listed product [15, 16]. Third-party verification rests on announcements dated 2022 to 2025 for Anzu, Gadsme, Frameplay and Roblox, while EA Advertising and Admazing state IAS or DoubleVerify partnerships and Zynga Ads shows DoubleVerify's logo, none with a verifier record [350, 368, 369, 351, 25, 396, 244]. The rule would leave H2 unresolved with fewer than six scored products; 9 cohort A and 4 cohort B1 products are scored. No product is accredited, so H2 is supported.
MRC's Multimedia and Services Outside the U.S. pages returned only short pages, so the absence rests mainly on the Digital lists, read on 29 September 2026 and unchanged at the rechecks on 30 September and 1 October 2026 [15, 16]. If accredited mobile in-app measurement reaches rewarded and interstitial formats in mobile games, as inferred above, assurance for those formats is thicker than for intrinsic placements [18, 19].
## 7. Suitability, user-generated worlds and children's rules
*Suitability in games is written mostly by the platforms that own the worlds, on top of age ratings and labels that creators declare. Third-party verifiers document little suitability coverage inside games. Children's privacy rules tightened in 2025 and 2026, and Roblox, Microsoft and Discord restrict functions of their own ad products by age.*
> **In short.**
> Among in-game ad networks (cohort A), Anzu, Gadsme, Adverty and Kidoz document suitability controls in operational detail, IION only in its inventory-partner terms, and Super League and Admazing only in marketing language in the pages opened; AdInMo, since moved out of the cohort, did so in a copy archived in October 2025 [397, 398, 81, 69, 399, 400, 129, 401]. Content ratings are the common signal in the controls documented, and a title's rating does not cover what players build in user-generated content (UGC) worlds [20, 266, 220]. Roblox restricts ad content by rating for each age band and, since 4 May 2026, allows direct-sold, mainly contextual ads to under-13s [20, 215, 216]. The Global Alliance for Responsible Media (GARM) was discontinued in August 2024, yet guidance from IAB Technology Laboratory (IAB Tech Lab), the Interactive Advertising Bureau's technical standards body, still cites its framework [402, 403]. DoubleVerify described suitability measurement on Roblox Rewarded Video in the future tense [351]. The amended Children's Online Privacy Protection (COPPA) Rule took effect on 23 June 2025, with compliance due for most regulated entities by 22 April 2026 [23]. Hypothesis H4 is supported: rule documents from Roblox, Microsoft, Discord, Epic and Google restrict ad targeting, personalisation, third-party calls or measurement [20, 57, 21, 22, 220, 221]. The verdict rests on rule documents and does not show that any product complies with them.
### Suitability controls, product by product
> **How this review codes a control.**
> A control counts when its documentation states what is set, how it is applied and its scope. A marketing adjective or a partner's name does not count. Whether a control works needs an independent test.
Among in-game ad networks (cohort A), Anzu, Gadsme, Adverty and Kidoz document controls in operational detail, Kidoz on its marketing pages [397, 398, 81, 69, 181]. Gameloft for brands and Frameplay document them in outline, and Super League and Admazing show marketing language only in the pages opened [404, 405, 137, 406, 400, 129]. For IION, only its inventory-partner terms document a control in operational detail: partners may not deliver its ads to properties with excluded content, including content directed to children under 13 [127, 399]. AdInMo, which this review moved out of the cohort on 29 September 2026, documented controls in operational detail in a copy archived on 7 October 2025 [401]. Among platform sellers screened (cohort B1), Roblox documents ad-content and publisher rules in operational detail but defines its advertiser-selectable tiers only in a January 2024 post, and XBOX Media Solutions documented controls in operational detail only on a page archived on 21 January 2026 that its live site no longer serves, its live pages using marketing language [20, 407, 408, 409, 224]. Of the cohort's other two sellers, EA Advertising has no brand-suitability control documented in the pages this review read, and Zynga Ads offers only a claim that its owned platform guarantees "a 100% brand safe environment" [410, 244, 411]. Among game-adjacent media screened (cohort B2), Twitch documents advertiser controls in operational detail and Discord an ad-content policy with no placement control [266, 288].
Who applies a control differs by network [397, 398]. Anzu gives advertisers allow and block lists by publisher, game, country and platform, checks direct creatives by hand, and aims for suitability to each game's PEGI rating, the European age rating, on a reasonable-efforts basis [397]. Gadsme reviews each new game's PEGI rating, checks placements before a game goes live, and blocks brand-safety discriminators by default on campaigns it manages [398]. Frameplay's Terms of Use, updated 1 January 2023, make publishers "solely responsible" for each game's content restrictions [406, 137]. XBOX Media Solutions sells "100% brand safe" rewarded video, and the former Activision Blizzard Media page that let buyers avoid age-restricted titles now redirects to Microsoft's gaming hub [224, 409, 191].
Content ratings are the signal these documents share [20, 397]. Roblox restricts ad content by Entertainment Software Rating Board (ESRB) and PEGI level per age band, Twitch labels streams of ESRB M-rated games automatically, and Epic requires ratings from the International Age Rating Coalition (IARC) for every Fortnite island [20, 266, 220]. A title rating describes the game, not what players build inside it; Roblox, Twitch and SuperAwesome add experience-level or stream-level signals [412, 265, 266, 206].
### User-generated worlds write their own rules
Roblox turns film, television and game ratings into ad rules by audience age [20]. Its Advertising Standards, launched on 15 March 2023, were last updated on 5 June 2026 [20]. Self-serve and independent ad integrations may not promote Mature-rated games or R and TV-MA content, and Roblox-managed ads may promote them only to audiences aged 18 and over [20]. Ads to users aged 9 to 12 may not promote M or PEGI 18 games, and ads to users aged 5 to 8 also may not promote Teen or PEGI-12 games or PG-13 and TV-14 content [20]. These age-band rules were added on 5 June 2026 to an under-13 section that took effect on 4 May 2026 [20, 213, 214].
Which experiences carry ads rests largely on what publishers declare [407, 408]. An experience qualifies for immersive ads only if it is public, its owner is ID-verified with two-factor authentication, Roblox has approved its Maturity and Compliance Questionnaire, and it keeps at least 2,000 unique visitors a month [407, 20]. Publishers may not misclassify an experience's guideline rating to raise ad delivery, and in the sources screened this review found no published audit of questionnaire accuracy [20, 407].
The advertiser side is documented in less detail [42, 217]. Roblox's brand pages name "Suitability Tiers, Age and Geo controls" and Advertiser Select, a hand-curated list of experiences, without operating detail [42, 217]. The only tier definitions found come from a January 2024 post, which keyed them to experience age recommendations and said they "may change over time" [408].
Experiences carry Roblox's Content Maturity Labels (Minimal, Mild, Moderate and Restricted), set from creators' questionnaire answers and content descriptors; Restricted experiences are open only to age-checked users aged 18 and over [412, 413]. Roblox announced on 3 September 2025 that IARC ratings (ESRB in the US) would replace the labels "when implemented", but its Creator Hub documentation, updated 26 September 2026, still describes its own four [413, 412, 414]. The labels do not cover user-generated items that players bring into an experience [414].
Since 4 May 2026, Roblox's rules allow ads to under-13s [215]. The FAQ of its 20 March 2026 announcement, in the edited version read on 29 September 2026, says all under-13 ads "will be contextual and sold directly via Roblox or creators", with no programmatic partners such as Google [215]. AdExchanger reported on 4 June 2026 that SuperAwesome had been named exclusive partner for under-13 users, and a Roblox spokesperson said these ads exclude rewarded video and can also be targeted by country, gender and age range [216]. The sources opened do not show when under-13 ads first served, and Roblox's brand-safety page still said on 29 September 2026 that users under 13 are not eligible to see ads [216, 217].
SuperAwesome says its staff play each title by hand and check it against 11 brand-suitability flags, which it does not list in full [206]. How manual play keeps pace with the roughly 15,000 new games it says are uploaded daily is not documented [415, 206].
Epic's Fortnite Developer Rules, last updated 29 April 2026, require IARC ratings for islands, cap content at Fortnite's maximum rating, and refuse content aimed mainly at under-13s or island promotion that targets them [220]. Paid commercial content must be disclosed, individual developers must be at least 18 to create or publish it, and commercial content may not promote categories including tobacco, alcohol, marijuana and firearms [220]. This review found no Epic ad-serving product for brands in the sources screened [220].
Minecraft's undated usage guidelines bar brands from using Minecraft gameplay to promote unrelated products without written approval from Mojang or Microsoft, and require all servers and advertising to suit audiences of all ages [416]. Discord's Ads Policy, updated 9 September 2026, covers all ads, including in Activities, and bans ads intended solely for children under 13 [288]. It documents no advertiser-side placement control [288].
Twitch builds advertiser controls on labels that streamers apply [265, 266]. Since 20 June 2023 streamers must apply whichever Content Classification Labels fit their stream; there were six at launch, and Twitch's current help article and application programming interface (API) reference list seven after a Politics and Sensitive Social Issues label was added in November 2024 [265, 266, 267, 268, 417]. Per Amazon's guide of 22 February 2026, the labels inform three inventory tiers and impression-level content reporting, which the guide, as modified on 29 September 2026, places in Amazon Ads Agent; the guide does not say what each tier excludes [266, 269].
### GARM ended, and verifiers say little about suitability inside games
GARM is gone, but its framework is still referenced [402, 403]. The World Federation of Advertisers (WFA) discontinued GARM on 9 August 2024 [402]. On 29 July 2026, announcing a settlement of the GARM litigation with X Corp., WFA said it will not form or restart GARM or a similar initiative [402, 418].
IAB Tech Lab's Content Taxonomy 3.0 guidance, in a version last changed on 5 August 2025, maps game genres to taxonomy categories and defines suitability risk levels by reference to the September 2020 framework of GARM and the American Association of Advertising Agencies' (4A's) Advertiser Protection Bureau [419, 403]. WFA's notice does not say that framework was withdrawn [419, 403, 402].
Verifier coverage of suitability inside games is thin in the documents opened [351, 371]. DoubleVerify's post of 1 April 2025 extended only invalid-traffic (IVT) and viewability measurement to Roblox's in-game billboards, through a custom integration on direct-sold inventory [351]. For Rewarded Video, then about to launch, it said advertisers "will be able to" measure brand safety and suitability, and no later record opened confirms that function in operation [351]. Integral Ad Science (IAS) announcements for Roblox, Anzu and Gadsme name only viewability and IVT functions, and its Frameplay page names IVT and impression counts [371, 370, 350, 368, 420]. Roblox says IAS pre-bid suitability is available for programmatic Roblox buys through enabled demand-side platforms (DSPs), but no IAS document opened describes suitability classification inside an in-game placement [187, 371]. In DoubleVerify's own records screened on 29 September 2026, including its sitemap, this review found none naming a cohort A network; Gadsme's policy says only that advertisers can wrap their tags with IAS or DoubleVerify [351, 375, 398]. On Twitch, Amazon's guide says the two verifiers are approved for brand safety at domain or app level, which, on this review's reading, rates Twitch as a whole rather than the stream an ad runs in [266].
### Children's rules
Federal children's privacy law tightened in 2025 [23]. The Federal Trade Commission (FTC) announced its amended COPPA Rule on 16 January 2025; it took effect on 23 June 2025, and most regulated entities had until 22 April 2026 to comply [23, 421].
A covered operator needs separate verifiable parental consent before disclosing a child's personal information to third parties unless the disclosure is integral to the service, and the Commission gives third-party advertising as an example of a disclosure that is not integral [23, 421]. The amended Rule lists serving contextual advertising and capping ad frequency among the activities that support internal operations, and the Commission also places ad attribution within them [23]. It exempts from parental consent only the collection of a persistent identifier, with no other personal information, used solely for that support and disclosed in the operator's notice, and it bars using that information for behavioural advertising or profiling [23]. The Commission's statement says the Rule does not prohibit collecting personal information for targeted advertising where the parent gives opt-in consent, which must be off by default and cannot be a condition of the child's access [23]. Read together, which is an inference, contextual in-game ads in child-directed games can run without parental consent under those conditions, while behaviourally targeted ads need prior parental opt-in consent, plus separate consent where the child's data goes to third parties [23].
The FTC's policy statement on age verification of 25 February 2026, adopted 2-0, applies only to mixed-audience and general-audience operators; services primarily directed to children must still treat every user as a child [422, 423]. The FTC will not bring COPPA actions over personal information collected, used or disclosed solely to determine age if six conditions are met: use only for age verification, disclosure only to third parties that give written assurances, prompt deletion, notice in the privacy policy, reasonable security, and reasonable steps to ensure accurate results [422, 423]. The statement never mentions advertising [422]. Its first condition bars any other use of age-verification data, so, on this review's reading, an operator relying on it cannot use that data to target or measure ads [422].
FTC COPPA enforcement in games has named game makers and platform owners [424, 425, 426]. The FTC announced a $275 million civil penalty for Epic Games on 19 December 2022, alleging that Fortnite collected children's personal information without parental notice or consent [424]. Microsoft's $20 million settlement, announced on 5 June 2023, concerned an Xbox sign-up flow that, the complaint said, included a pre-checked box allowing data sharing with advertisers until 2019 [425]. Genshin Impact maker Cognosphere's $20 million penalty, announced on 17 January 2025, was over alleged COPPA violations that included sharing user IDs and device identifiers with analytics firms and advertisers [426]. The FTC's COPPA Safe Harbor page, read on 29 September 2026, lists 20 related case entries, and its only ad-technology respondent is OpenX [427]. In that list this review found no FTC COPPA action against an in-game ad network or game-ad vendor filed from January 2023 to 29 September 2026 [427].
State laws make app stores and operating systems a source of age data, and courts have shaped when they apply [428, 429, 430]. Texas SB 2420, the App Store Accountability Act, was due to take effect on 1 January 2026 and was preliminarily enjoined in December 2025 [428]. After the Fifth Circuit stayed that injunction, the Supreme Court declined on 6 July 2026 to reinstate it, so the law can be enforced while the appeal continues [428, 431, 221]. Utah's HB 498 moved the app-store and developer duties of its 2025 act from 6 May 2026 to 6 May 2027 [432, 433]. Louisiana's 2025 act, HB 570, never took effect: Act 185 of 2026, signed on 15 May 2026, provides that it shall not become effective and re-enacts app-store rules from 1 July 2027 [429, 434]. From 1 January 2027, California's Digital Age Assurance Act, as amended on 10 September 2026, deems a developer that receives an age-bracket signal to know the user's age range unless it holds clear and convincing internal information of a different age, which then governs, and bars sharing the signal with a third party for a purpose the Act does not require [430, 435]. This review reads that as bringing COPPA's actual-knowledge duties to under-13 signals in general-audience games, an inference the Act does not state [430].
Maryland's Online Data Privacy Act took effect on 1 October 2025 and bars a controller from processing a consumer's personal data for targeted advertising, or selling it, if the controller knew or should have known the consumer is under 18 [436].
Google wrote an advertising ban into its age signals; Apple limited their purpose without naming advertising [221, 437]. Google's Play Age Signals terms, last modified October 2025, forbid use of the signals for advertising, marketing, user profiling or analytics [221]. Apple's Declared Age Range API returns the Texas law's age brackets for new Texas accounts from 4 June 2026, and Apple's developer agreement, updated 18 August 2026, allows requests only for age-appropriate content or features or to comply with laws that require age-range information [438, 439, 437, 440].
Platforms write rules on top of the law [20, 57]. Roblox bars personalised ads for users under 18, bars developers from rewarding under-13s for engaging with ads, and bans whole categories, including all food and beverage, when the intended audience includes users under 13 [20, 213, 214]. Its Advertising Integrations Terms, effective 1 May 2026, bar ads that developers place in experiences from making programmatic calls to third-party advertising and measurement services [57]. Its managed-campaign terms give advertisers only aggregate reporting and make Roblox's reporting the controlling measurement [1]. Two Roblox documents give different start dates for a developer revenue share on integrations, 30 September 2026 and 1 January 2027, and no document read says which governs [57, 190].
Microsoft's Privacy Statement says it does not deliver personalised advertising to children whose Microsoft account birthdate shows them to be under 18, while King, whose games carry XBOX Media Solutions ads, says it does not knowingly target interest-based ads to anyone under 13 [21, 234]. No document read says which rule governs XBOX Media Solutions inventory in King titles [21, 234]. Discord says signals from its age-assurance process are never used to target ads, although advertisers can target broad age ranges from signals such as self-declared age [285, 22]. None of the Twitch documents opened states an age-based rule for ad delivery, targeting or personalisation [257, 441, 442].
*Figure fig-rules. **Children's rules tightened through 2025 and 2026, and platforms wrote their own on top** Dates from the Federal Register, the FTC, state statutes and court orders, and Roblox's own documents. Roblox's Advertising Integrations Terms and its Creator Hub give different start dates for the developer revenue share; both are shown and neither is settled. Units: dated events. Scope: US federal and state rules and Roblox platform rules. Sources [23], [422], [436], [434], [433], [57], [190]. Caveat: Documentation is not legal compliance. Utah's duties and the Creator Hub date fall after the cutoff. Data: data/figure-data.json#rules*
**Rules and governance instruments**
| Jurisdiction | Instrument | What it requires | Effective | Scope | Law or practice | Src |
|---|---|---|---|---|---|---|
| United States (federal) | COPPA Rule as amended (16 CFR Part 312; 90 FR 16918) | Separate verifiable parental consent before disclosing children's data to third parties unless integral (advertising disclosures are not integral); written data retention policy, no indefinite retention; neutral age screen for mixed-audience services; notice naming third parties; contextual ads and frequency capping remain within the internal-operations exception | Effective 2025-06-23; compliance 2026-04-22 | Operators of services directed to children under 13, mixed-audience services for users identified as under 13, and operators with actual knowledge | Federal regulation | [23] |
| United States (federal) | COPPA safe-harbor duties (16 CFR 312.11) | Public member lists by 2025-07-21, updated every six months; annual report by 2025-10-22; proposed guideline modifications by 2025-10-22; technological-capabilities report by 2028-04-22 and every three years | 2025-07-21; 2025-10-22; 2028-04-22 | FTC-approved safe-harbor programmes | Federal regulation | [23] |
| United States (federal) | FTC Enforcement Policy Statement Promoting the Adoption of Age-Verification Technology | No COPPA enforcement for data collected, used or disclosed solely to determine age, if six conditions are met (purpose limit, vetted third parties with written assurances, prompt deletion, notice, security, reasonable accuracy) and COPPA is otherwise complied with; creates no substantive rights | Issued 2026-02-25; until final rule amendments or withdrawal | General-audience and mixed-audience operators; not services primarily directed to children | Regulator enforcement policy statement (non-binding) | [422] |
| Texas | SB 2420 App Store Accountability Act | App-store age assurance and parental consent for minors' downloads, purchases and significant changes; age categories under 13, 13-15, 16-17, over 18 | Scheduled 2026-01-01; enjoined December 2025; injunction stayed by the Fifth Circuit June 2026; Supreme Court declined to reinstate the injunction 2026-07-06; merits appeal pending | App stores and developers; users in Texas | State statute | [428, 438, 431] |
| Utah | App Store Accountability Act (SB 142, 2025; amended by HB 498, 2026) | Developers verify age category and parental consent through the app store; may use age category data only for age restrictions, legal compliance and safety features; may not share age category data | Developer and app-store duties from 2027-05-06 (HB 498); SB 142 had set 2026-05-06 | App store providers and developers; users located in Utah | State statute | [433, 432] |
| Louisiana | HB 570 (2025 Regular Session), R.S. 51:1771-1775 | As enacted in 2025: developers verify age category and obtain verifiable parental consent for minors via the app store; may not share age category data; Attorney General enforcement, fines up to $10,000 per violation. The re-enacted rules from 2027-07-01 bar developers from selling age category data | Never took effect: Act 185 of 2026, signed 2026-05-15, provides that the 2025 act shall not become effective and re-enacts app-store rules from 2027-07-01 | Covered application stores and developers; users in Louisiana | State statute | [429, 434] |
| California | Digital Age Assurance Act (AB 1043, Ch. 675, Stats. 2025) | OS providers and app stores send age-bracket signals; a developer receiving a signal has actual knowledge of the age range; may not share the signal with third parties for other purposes | Operative 2027-01-01 | Operating system providers, covered application stores and developers | State statute | [430] |
| Maryland | Maryland Online Data Privacy Act (Ch. 455 of 2024) | No processing for targeted advertising and no sale of personal data where the controller knew or should have known the consumer is under 18; contextual, first-party and measurement-only processing excluded from targeted advertising | Effective 2025-10-01; § 14-4612 not applied to processing before 2026-04-01 | Controllers meeting the Act's thresholds (not checked here) | State statute | [436] |
| California | Age-Appropriate Design Code Act | Default high-privacy settings and age estimation for services likely to be accessed by under-18s; certain data-use and dark-patterns provisions enjoined | Partly enjoined; Ninth Circuit narrowed the injunction on 2026-03-12; remanded | Businesses with online services likely to be accessed by children under 18 | State statute (in litigation) | [575] |
| European Union | Digital Services Act, Article 28(2) | No ads based on profiling with personal data when the platform is aware with reasonable certainty that the recipient is a minor | Applies from 2024-02-17 | Providers of online platforms | EU regulation | [576] |
| United Kingdom | ICO Age Appropriate Design Code, standard 12 (Profiling) | Profiling off by default for children unless there is a compelling reason; a privacy setting for behavioural advertising | In force (date not checked in this review) | Online services likely to be accessed by children under 18 | Statutory code of practice | [577] |
| Platform: Roblox (global) | Roblox Advertising Standards | No personalised ads for users under 18; restricted ad categories where the intended audience includes under-13s; no reward mechanisms for under-13 ad engagement; rewarded video only for 13+ via Roblox's product | Last updated 2026-06-05; U13 section effective 2026-05-04 | Advertisers and publishers on Roblox | Platform policy | [20] |
| Platform: Roblox (global) | Roblox Advertising Integrations Terms | Developer-placed ads may not make programmatic calls to third-party advertising and measurement services; no personally identifiable engagement data to advertisers; revenue share per thousand visits from 2026-09-30 | Effective 2026-05-01 | Developers placing ads directly in experiences | Platform terms | [57] |
| Platform: Roblox (global) | Roblox Creator Hub, Ad integrations | Revenue share on qualified visits for campaigns starting on or after 2027-01-01 (US $1.50 per thousand for 28 days, then $0.10); campaigns under review served only to users 18 and over | Published text read 2026-09-28; conflicts with the Integrations Terms | Developers running ad integration campaigns | Platform documentation | [190] |
| Platform: Roblox (global) | Roblox Advertising Terms (Managed Campaigns) | No personal data exchanged with advertisers; aggregate reporting only; Roblox reporting is the controlling measurement | Effective 2026-04-30 | Advertisers buying managed campaigns | Platform contract terms | [1] |
| Platform: Microsoft (global) | Microsoft Privacy Statement | No personalised advertising to children whose Microsoft account birthdate shows under 18 | Last updated September 2026 | Microsoft products and accounts | Company policy | [21] |
| Platform: King / Microsoft (global) | King Privacy Policy | No knowing targeting of interest-based ads to anyone under 13 (or higher where local law requires); opt-in to targeted ads by default where local law allows | Last updated 2026-02-26 | King games and services | Company policy | [234] |
| Platform: Discord (global) | Discord Quests FAQ; age-group help article | Age-assurance signals used only for safety and not to target ads; teens 13-17 limited to three Quests per day; users can opt out of Quest personalisation | Articles edited 2026-08-31 and 2026-09-24 | Discord users and Quest advertisers | Platform policy | [22, 285] |
| Platform: Epic Games Fortnite (global) | Fortnite Developer Rules | No gathering of personal information in islands; no targeting of under-13s with island advertising; paid commercial content disclosed; only developers 18+ may publish paid commercial content | Last updated 2026-04-29 | Fortnite island developers and brand collaborations | Platform rules | [220] |
| Platform: Google Play (global) | Play Age Signals API terms of service | Age signals may be used only for age-appropriate content and experiences in compliance with laws; not for advertising, marketing, user profiling or analytics | Terms last modified October 2025; Texas signals for accounts created after 2026-05-28; Brazil from 2026-03-17 | Apps distributed through Google Play using the API | App-store rule | [221] |
| Platform: Apple App Store (global) | Declared Age Range API notices | Age categories shared on developer request for new accounts in Utah (from 2026-05-06 per Apple) and Texas (from 2026-06-04, after a pause from 2025-12-23 while the injunction stood); Apple's February 2026 notice also named Louisiana from 2026-07-01, before Louisiana's Act 185 of 2026 stopped that state's 2025 law from taking effect | 2026-05-06 (Utah, per Apple); 2026-06-04 (Texas) | Apps distributed through the App Store | App-store rule | [578, 431, 439, 434] |
### Age data, litigation and the H4 verdict
Roblox reported that 57% of its global daily active users (DAUs) had completed an age check by the end of Q2 2026, and more than 70% in the US [43]. Of users who had age-checked globally, 35% were under 13, 38% were 13 to 17 and 27% were 18 or over in the seven days ended 30 June 2026 [43, 246]. Roblox says this population may not represent all users, and its Form 10-Q quarterly report says its user metrics have not been validated by an independent third party [43, 246]. This review infers no user's age from these shares, which are neither impressions nor ad-eligible reach [43].
On that global split, 73% of age-checked users were under 18, and Roblox's standards bar personalised ads for all of them [43, 20]. Contextual Roblox ads can still reach users aged 13 to 17 and, subject to the under-13 category and format restrictions, users under 13 [43, 20].
Roblox's Form 10-K for 2025 says its age-check methodology has misclassified users' ages [197]. AdExchanger reported on 4 June 2026 that Roblox faced at least 130 child-safety lawsuits in federal court plus suits by several states; the federal cases are centralised as a multidistrict litigation (MDL No. 3166) [216, 443]. Roblox's Q2 2026 10-Q lists settlement agreements with states including Alabama, Mississippi, Nevada, South Dakota and West Virginia, and $91 million accrued in the first half of 2026 for settlement agreements and settlement discussions [216, 246]. The sources screened document no advertiser pulling spend from Roblox over these matters, and none of them is a finding of liability [216].
H4 is supported on its registered rule, which needs at least three products or platforms. This review found five for which a primary rule document restricts a targeting, personalisation, third-party-call or measurement function the product offers elsewhere: Roblox (personalisation under 18, third-party calls in developer-placed ads, reporting beyond aggregates), Microsoft (personalised ads for accounts under 18), Discord (age-assurance signals in targeting), Fortnite (island advertising aimed at under-13s) and Google Play (advertising use of age signals) [20, 213, 57, 21, 22, 220, 221, 23]. COPPA's separate-consent rule adds a legal restriction for child-directed games, and the registered rule names the conflict between Roblox's two revenue-share documents as an example of an unresolved condition [23, 57, 190]. On this review's reading, that conflict concerns developer fees rather than any restriction counted here, so it does not block the verdict [57, 190]. The verdict does not show that products comply or that restrictions are enforced, and most of the strongest restrictions are platform practice that a platform can change on its own [23, 20].
Children's-data documentation among the in-game networks is thin [444, 445]. Gadsme's privacy policy of 24 February 2026 lists a "COPPA flag" among the data its services collect without saying how flagged traffic is handled [444]. Anzu's December 2025 policy offers publishers a Kids Safe-Mode that treats users as children and limits collection, although the publisher may share additional information at its discretion [445]. Documentation is not legal compliance, and its absence is not evidence of non-compliance. Chapter 11 scores these documents under its dimension for suitability, age and data controls.
## 8. Console, PC and cloud
*On consoles and cloud services, most ad placements that platforms document sit on the dashboard, in the store, in a TV app or in the queue before play. The documented exception is a publisher, Electronic Arts, selling boards and overlays inside its own sports games. Microsoft markets console integrations its own pages cannot place, so the console hypothesis stays unresolved.*
> **In short.**
> Xbox sells dashboard and store placements, and its cloud streaming test, begun on 23 July 2026, plays ads before sessions begin [24, 202]. Live TV on PS5 (PlayStation 5), launched in the US on 31 August 2026, is TV inventory in the console's Media section [446]. GeForce NOW has shown free members ads before play since March 2024, and a 2025 explainer says Amazon Luna shows ads while connecting [447, 448]. Reported plans, patent applications and hiring are not products [231, 449, 450, 451]. This review found no Nintendo ad product in the sources screened by 28 September 2026 [452]. Steam does not support advertising models in games it distributes, and Epic requires paid placements to be disclosed [453, 454]. EA names brand integrations in two console titles and in its Madden NFL franchise, and documents none as an ad placement buyable in the US [25]. This review codes H3, the console hypothesis, as unresolved because Microsoft markets console integrations "into gameplay" and names no title [24].
### Most console placements sit outside gameplay
Console and cloud platforms document ad placements on the dashboard, in the store, in TV apps and before streamed sessions; in the sources screened, the one documented case of brand placements inside the gameplay of console titles is a publisher selling its own games [24, 446, 447, 25]. H3, the console hypothesis this review registered on 28 September 2026, tracks each placement on Xbox, PlayStation, Nintendo and cloud-gaming services separately, with a dated source for each stage, and first asks whether the ad renders inside gameplay. A console device does not make a placement in-game. Twitch Premium Video runs in the Twitch app on consoles, and Discord Play Quests reward play on console while the Quest renders in Discord [47, 48, 49]. Discord also announced on 9 March 2026 that Instant Play Quests, game trials streamed from the cloud and played directly on Discord, "will be made available" for games on NVIDIA GeForce NOW or Amazon GameLift Streams [282]. [Around the game](#adjacent) covers all three.
*Figure fig-console. **On console, most ad placements sit outside gameplay; the in-game exception is a publisher selling its own games** Dated placements from platform and publisher documents and trade reporting. Apart from Microsoft's console integrations, whose location its own documents do not establish, the Xbox and PlayStation placements found by the cutoff are outside gameplay (dashboard, store, TV apps, pre-session ads) or in test; EA sells placements rendered inside its own sports games. Microsoft's console custom integrations could not be placed inside or outside gameplay from its own documents, so H3 stays unresolved. Units: dated events. Scope: console, cloud and EA titles, 2022 to the cutoff. Sources [231], [202], [446], [447], [448], [235], [25]. Caveat: A console device does not make a placement in-game. Patent applications and reported plans are not products. Data: data/figure-data.json#console*
### Xbox sells around play, plus one product its pages do not place
Xbox's documented console placements render outside gameplay or leave their location unstated [24, 455, 456, 457]. On 30 June 2025 Microsoft announced an updated Xbox Activities offering, "available now as part of Xbox Landing Experiences", and its retired Invest page describes an invite-only private marketplace for Xbox Home and Store banners [455, 457]. CNBC reported that in 2024 some players complained about a McDonald's ad on a screen for selecting games [458].
XBOX Media Solutions, Microsoft's gaming ad unit, markets the one console product whose location its own pages do not establish [24]. Its custom-integrations page says it integrates brands "into gameplay across mobile, PC, and console", in solutions that range "from immersive in-game moments to XBOX dashboard & store placements" [24]. Its console product, XBOX Experiences, is built on "native display formats and branded content" [24]. The page names no console title, launch date, specification, US availability or programmatic route, and does not say whether XBOX Experiences render inside gameplay [24]. A sponsored article produced by The Drum Studios in January 2026, built on an interview with the unit's general manager, Greg Carroll, describes Xbox's approach through "the idea of game-adjacent rather than in-game" [233].
Xbox's ad-supported cloud streaming is a test: on 23 July 2026 Xbox began a limited-time test, open to Xbox Insiders, in which players stream games they own free with ads [202]. Ads play before sessions begin, sessions are limited to one hour, and Microsoft states that it does not interrupt the flow of gameplay [202]. Variety reported the same day a two-minute block of ads from market-specific brands before the player joins the cloud queue [459]. The five games Variety named, from Forza Horizon 6 to War Thunder, are titles eligible to stream, not titles carrying ads [459]. Xbox's 3 September 2026 announcement of cloud-gaming hour limits from November 2026 does not mention the ad test, and this review found no dated general-availability plan for ad-supported streaming in the sources screened by 28 September 2026 [460].
A 2022 press report and a 2026 Microsoft patent application point at ads inside play, and neither is a product [231, 449]. In April 2022 TechCrunch and XDA, relaying Business Insider's unnamed sources, reported a program for brands to advertise in free-to-play Xbox games, for example as rendered billboards in racing games, through a private marketplace, set to launch by the third quarter of 2022 [231, 461]. Microsoft said it had nothing further to share, and this review found no product-specific evidence of a launch in the sources screened by 28 September 2026 [231, 461]. Microsoft Technology Licensing's patent application US 2026/0260258 A1, filed on 23 April 2026 as a continuation of an abandoned 2023 application and published on 3 September 2026, describes in its abstract a system that suspends a player's interaction with a game on a promoted-content trigger, presents promoted audiovisual content and grants a credit that delays the next suspension [449].
### Networks claim console reach and name no console title
Anzu, an in-game ad network, calls itself "the sole officially licensed in-game ad provider for Xbox" on an undated page and announced its participation in the Xbox Tools and Middleware Program in November 2019 [232, 462]. This review found no Microsoft record confirming the licence or its exclusivity; Anzu's own listing on Microsoft Marketplace repeats the claim in copy the publisher supplies [463]. It also found no console title named as carrying Anzu placements in the twelve months to the cutoff [232, 462, 164].
Frameplay's developer page claims console availability, and its Unity changelog lists "PS5 Ad Loading Support (Experimental)" in November 2023; no live console title was found in the sources screened by 29 September 2026 [464, 465, 466]. Gadsme markets "mobile, PC and console" placements, but no source opened names a console title, and Gadsme's own press page describes the adWMG deal cited for it on Xbox and PlayStation as mobile inventory for console-focused advertisers ([What counts as gaming advertising](#defining)) [170, 40, 467]. Roblox's client runs on Xbox and PlayStation, but no Roblox document opened states on which devices its immersive ads are served [197, 207]. DoubleVerify's post of 1 April 2025, before the twelve-month window, says its Roblox coverage spans mobile, desktop and console, without naming a console title or experience [351].
### PlayStation sells TV on the console, and patents and hires for more
The one PlayStation ad placement this review found launched by the cutoff is TV inventory [446, 451]. Live TV on PS5 launched for players in the US on 31 August 2026 as a free ad-supported service from Sony Pictures Television, with more than 100 live channels in the console's Media section, outside gameplay [446]. Variety reported, in an interview with Sony Pictures Entertainment's Peter Wood, that Sony will work with Publica and PubMatic, two companies it describes as focused on connected-TV (CTV) advertising [468]. In the sources screened by 28 September 2026, no Sony, Publica or PubMatic release was found, and deal types are not described, so reading the buying route as a CTV route is inference [468]. The same report cites more than 95 million PS5 units sold worldwide as of 30 June 2026, which counts consoles and does not measure viewers eligible for ads [468].
Sony's other signals are a patent application and a job posting [450, 451]. Sony Interactive Entertainment's application US 2026/0138036 A1, filed on 15 November 2024 and published on 21 May 2026, describes in its abstract a console that determines that an appropriate time for alternate content is about to begin during gameplay, estimates how long it will last and shows alternate content on part of the screen for a shorter duration [450]. Its figure shows a "waiting in queue" screen, and the abstract does not mention advertising [450]. Digiday reported on 5 August 2026 that Sony Interactive Entertainment Media Solutions posted a global director role in mid-July 2026 to scale client-partnership sales in programmatic, FAST (free ad-supported streaming TV) and CTV and "emerging ad solutions" [451]. This review found no PlayStation in-game ad product in the sources screened by 28 September 2026 [451].
### Nintendo: one reported rule, no product found
This review found no Nintendo advertising product and no public Nintendo policy text on in-game advertising in the sources screened by 28 September 2026 [452]. The one record found is a developer's account: AUTOMATON reported on 20 May 2025 that Nintendo Switch store review flagged an indie game's parody commercials as in-game advertising "against the platform's rules", and that the developer had to remove them [452]. Nintendo's developer terms are gated, so the rule's scope is unknown [452].
### Cloud services put ads in the queue
NVIDIA's GeForce NOW help page says free members have seen up to two minutes of video ads before each free session, while in queue, since 5 March 2024, and that ads do not interrupt gameplay [447]. Members under 16 cannot opt in to personalised ads [447]. On 28 September 2026 the page said NVIDIA is moving to a new advertising provider, so free members may see fewer ads or none; the page does not disclose the provider, advertisers, buying route or volume [447].
In the sources screened by 28 September 2026, this review found no Amazon document on Luna ads [448]. An Android Authority explainer dated 6 May 2025 says Luna shows ads while connecting to a game and not while the game loads or during gameplay [448].
### Steam and Epic set rules for ads inside games
Valve's Steamworks documentation states that Steam contains no paid advertising and does not support advertising models in games distributed on Steam [453]. It bars requiring players to watch ads to play and rewarding them for engaging with ads, and it allows real brands as product placement when they are not disruptive [453]. Whether programmatically served intrinsic placements in a Steam build count as product placement or as an advertising model is not stated, and the page documents no enforcement [453]. The Epic Games Store requires developers to disclose paid advertisements and promotional content, including product placement, logos and slogans for real-world brands, and labels a declared offer as promotional content [454]. This review found no Epic Games Store ad product for advertisers in the sources screened [454].
On 13 May 2026 Anzu announced a partnership with Strikerz Inc. to sell LED perimeter boards, pitch-side assets, branded scoreboards and referee kits in the free-to-play football game UFL on PC only, although UFL launched on console in December 2024 [469]. Its 7 million players worldwide is a cumulative claim, and no ad-eligible audience is given [469]. UFL's PC version is reportedly distributed through the game's own launcher; on 29 September 2026 its Steam page said the game was not yet available there [470, 471, 132]. This review infers that Steam's policy does not govern Anzu's UFL inventory [453, 471]. Gadsme's documented PC integration is in Football Manager 26, which is sold on Steam [170, 141, 472]. No document read says how Steam's policy classifies that placement [453].
### EA sells placements inside its own console games
Electronic Arts (EA) launched EA Advertising on 15 June 2026, with native ad units such as digital ad boards, scoreboards and brand broadcast overlays dynamically served in select, unnamed EA SPORTS games from its own ad server and SDK (software development kit) built for the Frostbite engine [25]. [Platforms and publishers selling ads in their own games](#platforms) covers the product and its buying routes.
EA's release names brand integrations in two console titles, EA SPORTS FC 26 and EA SPORTS College Football 26, and in its Madden NFL franchise [25]. In EA SPORTS FC 26, Comcast's release of 17 November 2025 dates dynamic LED ad-boards for Xfinity and Peacock in Ultimate Team matches in the US [25, 235]. FC 26 runs on PlayStation 5, Xbox Series X|S, PC, Nintendo Switch and Switch 2, but no source opened states which platforms carry the ad-boards [235, 236]. In EA SPORTS College Football 26, Mountain Dew's "DEW University" went live on 15 July 2025, and Lowe's release of 11 April 2025 said Lowe's would feature on the Stadium Pulse broadcast overlay [25, 239, 473]. MUT.GG, a fan database, places a Lowe's Ultimate Team program, with releases on 27 September and 18 October 2025, in Madden NFL 26; EA's release names only the franchise [25, 238]. The release reports engagement with these integrations as games played, matches played and challenges completed, and no impressions [25].
### Why H3 stays unresolved
This review codes H3 as unresolved under its registered rule [25, 202, 24, 446, 447, 469]. Support needs fewer than ten named console or cloud titles documented, by the network, the platform or the game's publisher, as carrying live in-game brand placements buyable in the US in the twelve months to the cutoff, no console platform documenting a generally available in-game brand ad product, and other placements outside gameplay or in test. The rule codes H3 as unresolved when a placement's location or status cannot be established from the platform's, network's or publisher's own documents.
The title count is far below ten. At most three EA console titles qualify [25, 235, 238, 239]. FC 26 has a dated, US-specific counterparty record inside the window [235]. Madden NFL 26 is dated only by a fan database [238]. College Football 26's integrations went live in July 2025, before the twelve-month window opened, so their presence inside it is inferred [239]. EA documents none of the three as an ad placement buyable in the US; US availability is read from the US advertisers named and from Comcast's "in the US" [25, 235]. Any count from this screen is a floor for the market, because this review did not screen other publishers' console titles and networks may run console campaigns they do not publicise [25].
The platform clause is where the coding stops. In the sources screened, no console platform documents a named, generally available in-game brand ad product with a buying route [24, 446, 452]. Microsoft, a console platform, does market brand integration "into gameplay" on console, sold through its gaming experts and a contact form, without naming a title [24, 191]. Its own documents do not establish whether those integrations render inside gameplay or what stage they are at, which is the unresolved condition as registered [24]. The 2022 plan stays unresolved for the same reason: the rule asks for product-specific evidence, and none was found [231, 461].
Two disclosures would move the coding. If Microsoft named a console title and placement for XBOX Experiences, the platform clause could be tested. If EA named its "select EA SPORTS games", the title count could rise, though it would still rest on one publisher [25]. The contradiction clause's product test names console platforms, not publishers, so EA's launched product could trigger it only through a count of ten or more titles.
## 9. What the outcome evidence shows
*The published outcome evidence for brand ads in and around games Every published field result found for a brand campaign in or around games comes from a party with a stake in it: a seller of the inventory, a measurement firm selling the method or an agency that trades on its norms; the academic studies are few and mostly laboratory work. The agency-led dentsu study across three gaming formats found high attention and overall brand recall below the agency's own norms, with livestream ads the exception on recall. This review found no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign whose funding is documented as coming from a party without a financial interest, in the sources screened by 29 September 2026.*
> **In short.**
> The Play Attention report by dentsu and Lumen, published in January 2024, found gaming ads more viewable and more attended than dentsu's norms, and brand recall of 21%, lower than those norms [27, 384]. Sellers' own studies compare exposed players with a survey control group or with a measurement firm's database, and Frameplay supplied about half of the campaigns (nearly 90 of 173) in the in-game benchmark it reports beating (inference) [474, 475, 28]. Roblox, Discord and Microsoft publish lifts from their own studies or internal data, often without magnitudes, units or method [28, 341, 476, 227]. The field comparison with the largest survey base found, an Ipsos analysis drawing on 37,000 interviews in 16 European countries, rests on self-selected exposure [477]. The only study with documented independent funding is a laboratory experiment, in which 26.6% of 79 players reported noticing any in-game ad unaided [26]. This review found no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign whose funding is documented as coming from a party without a financial interest, in the sources screened by 29 September 2026 [26, 27, 28].
### How the studies were coded
Research question 8 asks what outcome evidence exists for brand advertising in and around games, and who paid for it. Each study is coded by design, sample, comparator, sponsor and funding. A study counts as independent only when its funding is documented as coming from a party without a financial interest; unstated funding is recorded as unknown. Attention metrics are defined in the companion paper on attention measurement, and [Measurement](#measurement) covers who offers attention measurement in games.
The screen combined web searches for brand lift, attention, sales and incrementality studies, sellers' newsrooms and case-study pages, and title searches of scholarly records from 2019 to the cutoff, followed by a verifier re-screen. Field experiments with live in-game campaigns and academic studies of children and Roblox advertising were screened by title search only, after the web-search budget ran out, so the absence finding is weaker on those topics. Form-gated reports from Gameloft, Roblox and Anzu with Lumen were not opened.
This review found no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign whose funding is documented as coming from a party without a financial interest, in the sources screened by 29 September 2026 [26, 27, 28]. Every field result for a specific live brand campaign was published by the seller of the measured inventory, by a measurement firm selling the method, or by an agency that trades on its norms [27, 28, 477]. The table at the end of this chapter holds 34 study records, one per study, with every publication of a study attached to its record; the figure counts the same records. Each record says what its design can support (a causal result for the tested stimulus, an association, a description, or nothing assessable) separately from who paid for it [475, 478, 479, 477]. Ipsos's GameTrack analysis, discussed below, is not in the table [477].
*Figure fig-studies. **Most outcome studies of in-game and game-adjacent ads were run or paid for by interested parties, and none is an independent field test** Studies screened by the cutoff that report brand lift, recall, attention, sales or incrementality for in-game or game-adjacent brand ads, classed by who ran or paid for them, with their most common designs. Seller studies often compare the seller with its own benchmark. Units: studies. Scope: studies found by the review, published by the cutoff. Sources [482], [483], [474], [579], [478], [475], [479], [28], [484], [486], [488], [341], [476], [489], [490], [580], [492], [494], [495], [395], [496], [118], [356], [553], [477], [27], [498], [499], [500], [505], [503], [501], [502], [26]. Caveat: A bounded count of the sources screened, not a census. Funding is recorded as stated; "not stated" is not independence. Data: data/figure-data.json#studies*
### The dentsu study: more attention, less recall
The Play Attention report, published by dentsu in January 2024 with the research firm Lumen, measured livestream ads charted as Twitch, Activision Blizzard mobile rewarded video and intrinsic in-game ads with Anzu [27, 384]. Gaming scored 99% viewability, 6,736 attentive seconds per 1,000 impressions (APM) and a $3.38 attention cost per 1,000 (aCPM), all better than dentsu's norms [27, 384]. Brand recall was 21%, which dentsu calls highly variable across gaming environments and lower than its norms [27]. The comparator is dentsu's cross-media normative database, and the report gives no sample size, fieldwork dates, uncertainty or funding [27].
The formats diverged. Rewarded video had 10,043 APM against 6,100 for dentsu's online video norm [27]. Intrinsic ads with Anzu had 3,442 APM, above online display at 1,416 but below social at 3,975 and online video [27]. APM depends on dwell time and viewability, so a full-screen, opt-in rewarded video starts with an advantage over a billboard in a scene (inference) [27]. Livestream ads, shown in a simulated livestream environment and charted as Twitch, produced 57% brand recall against a dentsu norm of 38%, and a 17% average brand-choice uplift against dentsu norms of 7.25%, with uplift measured from a control-group baseline [27]. The report itself does not describe Twitch's role, but Lumen's release of it carries a Twitch Ads company description beside Anzu's, and Microsoft's October 2024 post credits the study to dentsu in partnership with Lumen, Twitch Ads, Activision Blizzard Media and Anzu; not all ad categories shown were gaming-specific [27, 384, 480].
Anzu's blog is the only source found for the sample: 200 gamers aged 18 or over in a Las Vegas lab and 250 online controls [481]. It gives Anzu's attention figure as 3,422 where dentsu and Lumen give 3,442 [481, 27, 384].
Microsoft's April 2026 gaming ad blog says, citing research with dentsu, that "100% of gaming ads are fully viewed", although the report gives 100% on-screen only for rewarded video [456, 27]. The same passage says immersion predicts sales with 83% accuracy; that figure is not in dentsu's report, and no source for it was found [456, 27]. Microsoft's October 2024 post credits its immersion findings to a seller-commissioned study run by Alter Agents in March 2024, with no public method [480].
### Seller studies against seller benchmarks
Specialist in-game networks report results against a measurement partner's benchmark, and the benchmarks shift. In October 2022 Anzu reported in-game ads viewed for 2.9 seconds against a Lumen benchmark of 1.5 seconds; in June 2023 it reported 3.1 seconds against a digital ad norm of 2.9 seconds [482, 483]. Neither release gives geography or a sample for these results, or says whether the benchmarks share ad sets or definitions [482, 483]. The same releases report 36% average spontaneous recall in 2022 and 49% average prompted recall in 2023; these are different measures, and both are levels among exposed players [482, 483].
An Anzu and Happydemics study of 110 intrinsic campaigns, released on 26 March 2024, reports ad recall 13 points above the digital media average and purchase consideration 5 points above the online media average [474]. The comparator is Happydemics' database of over 600 advertisers, with no control group, and the release does not say who commissioned the study [474].
Frameplay's release of 27 October 2025 reports a Happydemics in-game benchmark of 173 campaigns, nearly 90 of them Frameplay's [475]. In-game ads had the highest ad recall of any digital format at 32%, a level among respondents with no control reported, and Frameplay scored 24% on consideration, 2 points above the in-game benchmark [475]. With nearly 90 of the 173 campaigns, Frameplay is compared with a benchmark about half made up of its own campaigns, unless Happydemics excluded them, which the release does not say (inference) [475]. Gadsme's advertiser page states an average 35% ad recall and +25 points on brand image and consideration from brand lift surveys on its inventory, with no provider, sample or period [118].
### Platforms that measure their own ads
Roblox, Discord and Microsoft sell the inventory and publish the studies of it. Roblox's own brand lift study, run with Latitude from 20 to 27 March 2024, surveyed 2,100 US monthly users: 300 in a control group and 1,800 exposed users aged 13 to 49 [28]. It reports statistically significant increases in awareness, opinion, consideration and recommendation averaged across three brands, without magnitudes, and stated intentions up by 3 to 10 points [28]. How exposure was assigned is not stated, and the sample includes 13- to 17-year-olds, which [Suitability, user-generated worlds and children's rules](#safety) takes up [28].
Digiday reported on 12 May 2025 that a MediaScience study for Roblox measured 140 US users aged 13 to 34 with eye tracking and skin-conductance sensors, and that Roblox figures showed unaided recall 211% higher than for social ads [484]. For recall against streaming TV, Digiday reports 95% and Tubefilter 38%; no Roblox-issued document was found, and funding is not stated [484, 485]. Roblox's Immersive Is Effective study with Ipsos, fielded from 6 to 27 May 2026, found that 52% of 200 Roblox users said they pay more attention to ads on Roblox than they usually do, 13 points above a separately recruited streaming base [486]. That is self-reported attention, and the release does not say who paid [486, 487]. EDO, which named Roblox its first gaming measurement partner on 18 June 2026, reported measurable lifts from a pilot but published no effect sizes [488].
In AdExchanger's report on the Interactive Advertising Bureau's (IAB's) PlayFronts event of September 2026, a Roblox executive pointed to an 18% lift in sales per cardholder and a 6% lift in conversions for a Sam's Club rewarded video campaign, attributed to Cint, with no published design [183]. Agency marketers had told Digiday in May 2025 that Roblox would need metrics such as direct sales conversions to convince brands more firmly [484].
Discord's Quest case studies cite Kantar and Discord's own data. For Bethesda's DOOM: The Dark Ages, a 7-day Video Quest in the US and Canada, the page reports +11.3 points in ad awareness and +3.4 points in play intent at 90% confidence, with the headline figures sourced to Discord internal data; the advertiser is a game publisher [341]. Its film cases for The Black Phone 2 and Obsession do not say whether their Kantar lifts are points or relative changes [476, 489]. Mountain Dew's World of Warcraft Quest reports +16 points in brand favorability from Discord internal data with no named partner or method, and Wendy's reports only a 1.30% lift in Discord ad awareness [490, 491].
Microsoft's XBOX Media Solutions case studies for KitKat and Hershey's in Candy Crush Saga report ad recall lifts of 29 and 55 percentage points, KitKat's measured in the UK, France, Germany, India and Australia and Hershey's in markets not stated, naming no measurer, sample, control design or period [227, 226]. KitKat's campaign also ran retail and product activations, so its lift cannot be credited to in-game exposure alone [227]. Gameloft for brands, in paid content on The Drum on 1 June 2026, reported 18% more positive attention for an in-game interactive video ad than for a YouTube pre-roll, giving no sample, period or funding details [492, 493].
### Attention figures for video in and around games
Sellers of mobile and gaming-site video publish attention multiples against short benchmarks. Digital Turbine and Amplified Intelligence reported in April 2023 that, across seven brands in June 2022, viewers attended to Digital Turbine's in-game mobile video ads for a combined 22 seconds, against 1.6 seconds for mobile web video and 2.5 seconds for social video [494]. The release calls the 22 seconds "combined" and credits the ads' end cards with "an attention boost" without saying what the total adds together; if it counts end-card time, the units are not like-for-like with the web and social video figures (inference) [494]. IION's Garnier case, measured with Amplified Intelligence across 674,343 impressions in 2024, reports 7.38 seconds of attention per video impression against a 1.10-second benchmark, and gives two different display figures, 1.48 and 2.4 seconds [495]. A Venatus release of 18 March 2026 says Lumen found Venatus video audiences viewing ads for 8 seconds on average against a 2-second benchmark, with no sample, period, geography, benchmark source or funding [395]. Whether such measures predict brand outcomes is examined in the companion attention paper.
### Sales and incrementality claims
The purchase outcomes found for in-game brand activations are vendor claims without a published method. Super League's case studies report a 6.3% lift in purchase rate for a Lunchables virtual world designed to keep children coming back, and a highest-ever mobile sales day for Chipotle's Roblox maze [496]. Anzu said on 27 March 2025 that a December 2024 beta with Claritas for a major US advertiser found an incremental lift in conversions, giving no figure or method [394]. Its April 2025 release claims that intrinsic ads beat advertisers' cost-per-acquisition (CPA) goals by 21% on average and gives no campaign count, objective mix, period or method; user-acquisition economics belong to the published mobile app growth paper [497].
The field comparison with the largest survey base found is Ipsos's analysis of its GameTrack survey, published on 7 August 2026: 37,000 interviews in 16 European countries in the first quarter of 2026, run on behalf of the trade body Video Games Europe; Ipsos does not publish the sizes of the exposed and control samples for the nine campaigns [477]. Across nine campaigns, players of titles with brand integrations showed an average uplift of 31% on seeing the brand as one for people like them, against 6% for brands that did not advertise in the games; the campaigns include Electronic Arts' EA SPORTS FC integrations for Beats, Pepsi and Visa [477]. The non-advertising brands serve as a placebo check, but exposure is self-selected by who played which game, the uplifts are relative, and Ipsos markets the method [477]. This review found no public randomised or holdout sales-lift or incrementality study with published design and results for in-game brand advertising, in the sources screened by the cutoff [496, 488].
### What academic research finds
Academic work on in-game ads is small, mostly run in laboratories, and less favourable on memory than the seller studies (inference). In a peer-reviewed eye-tracking study with no external funding, published in December 2025, 79 adults played a mobile tennis game with embedded static brand ads for ten minutes; only 21 of them, 26.6%, reported noticing any advertisement unaided [26]. Whether an ad sat in a task-relevant zone predicted recall better than fixation time did, in one game with a convenience sample of young adults in Turkey [26]. Seller recall levels of 32% to 49% come from live campaigns whose recall questions and timing are not published; Anzu's 36% is spontaneous recall, the same kind of unaided measure, and its 49% is prompted, so the figures differ in setting, exposure and wording and sit in tension with that result without contradicting it (inference) [475, 118, 482, 483, 26].
A 2021 meta-analysis of 38 data sets found more positive attitudes toward advergames themselves than toward other advertising, but brand information in them seemed less likely to be remembered, they were less often recognised as advertising, and younger consumers were more susceptible [498]. Advergames are games built by brands, so the result bears on branded worlds on user-generated platforms by analogy only (inference) [498]. Experiments by Ingendahl and colleagues find that brand placements gain in attitude only when met during positive in-game experiences, and the 2022 paper calls the overall effect of placements on brand attitudes small and variable between studies [499, 500]. Experiments published in 2026 measure attitudes and intentions only, such as tap intention for rewarded banners among 600 Japanese smartphone users [501, 502].
The only peer-reviewed field-data study found, by Huang and Morozov in Marketing Science in February 2025, followed 60,000 Twitch streamers for eight months [503, 504]. It estimates an elasticity of 0.027 of a game's concurrent players to its live-stream viewers, an effect the authors call moderate and short-lived, and suggests that only about one sixth of games profit from sponsored streams [504]. Its subject is game promotion through creators, and its funding is unknown, as is the funding of most academic papers read only as abstracts [503, 498, 505].
### What the record supports
The record supports statements about viewability, attention and survey-measured recall in campaigns that sellers chose to publish. Its weakness is design, not volume: the comparators are a measurement firm's database, a benchmark partly built from the seller's own campaigns, a survey control with unstated assignment, or a self-selected exposed group [474, 475, 28, 477]. Many published lifts give no unit, sample or uncertainty, so a buyer cannot tell a point change from a relative one [476, 489].
The evidence that would change this finding is a randomised or holdout test of an in-game brand campaign, with published design and results, funded by a party without a financial interest. Ipsos's placebo check and Roblox's control group move toward that design, but neither documents randomised exposure or disinterested funding [477, 28]. [The capability matrix and the Play Quotient](#matrix) scores what each product documents about outcome measurement; those scores record documented capability and say nothing about how the ads perform.
**Study evidence table**
| Study | Year | Design | What it can support | Sample | Setting | Format / channel | Outcome measure | Headline result | Uncertainty | Sponsor | Generalisability | Src |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| dentsu / Lumen: Play Attention (Twitch, Activision Blizzard, Anzu) | 2024 | lab eye-tracking + survey; online control cell | Association (control cell, assignment not documented) | 200 gamers + 250 online controls (per Anzu) | lab, Las Vegas; real games on mobile and PC; simulated livestream | livestream video; rewarded video; intrinsic | viewability, APM, aCPM, brand recall, brand choice | gaming recall 21% (below norms); Twitch recall 57% vs 38% norm; Twitch brand choice uplift 17% vs 7.25% norm; rewarded APM 10,043 vs online video 6,100; intrinsic APM 3,442 vs display 1,416 and social 3,975 | not reported | dentsu (agency); sellers collaborated | US adults 18+; lab exposure; norms are cross-media database | [27] |
| Anzu / Lumen campaign attention tests | 2022 | eye-tracking panel on live campaigns; Lumen benchmark | Descriptive (against a norm or benchmark) | 5,000 respondents (quoted) | Anzu mobile and PC campaigns | intrinsic | % viewed, view time, spontaneous recall, purchase intent | 83% viewed vs 62%; 2.9 s vs 1.5 s; spontaneous recall 36%; purchase intent 61% ('uplift of 9%', comparator unstated) | not reported | Anzu (vendor) with Lumen | global; vendor campaigns | [482] |
| Anzu / Lumen State of In-Game Advertising | 2023 | eye-tracking; Lumen digital norm | Descriptive (against a norm or benchmark) | not stated | two-year study | intrinsic | viewability, view time, prompted recall | 98% vs 78% norm; 3.1 s vs 2.9 s; prompted recall 49% (high 97%) | not reported | Anzu (vendor) with Lumen | global; report gated | [483] |
| Anzu / Happydemics full-funnel benchmark | 2024 | brand lift surveys vs normative database | Descriptive (against a norm or benchmark) | 110 campaigns | intrinsic campaigns, multiple countries | intrinsic | ad recall, brand identification, purchase consideration | +13 pts recall, +11 pts identification, +5 pts consideration vs digital/online media averages | not reported | Anzu (vendor) with Happydemics | global; database comparator | [474] |
| Anzu creative best practices | 2024 | observational: blind creative scoring matched to lift results | Association (observational) | 55 campaigns; 15K+ respondents | Anzu campaigns | intrinsic video and display | brand recommendation, purchase intent | guideline-following video +21% recommendation, +5% intent; display +8%, +9% | not reported | Anzu (vendor) | global; correlational | [579] |
| Frameplay / Happydemics meta-analysis | 2024 | brand lift surveys vs channel benchmarks | Descriptive (against a norm or benchmark) | 29 studies | Frameplay network | intrinsic | attribution, likeability, consideration | attribution 55% vs 40% all media; likeability 52% vs 49%; consideration +21 pts vs +19 | not reported | Frameplay (vendor) with Happydemics | global; vendor campaigns | [478, 478] |
| Frameplay / Happydemics in-game benchmark | 2025 | brand lift surveys vs normative database | Descriptive (against a norm or benchmark) | 173 in-game campaigns (Frameplay nearly 90) | Happydemics database since 2021 | in-game (mostly intrinsic) | ad recall, attribution, likeability, consideration | in-game ad recall 32% (highest format); attribution 52% (+9 vs media average); Frameplay 54% attribution and likeability, 24% consideration | not reported | Frameplay (vendor) with Happydemics | global; benchmark about half seller's own campaigns | [475, 475] |
| Comscore / Frameplay snack-brand survey | 2021 | exposed vs control brand lift survey | Association (exposed vs control, assignment not documented) | not stated | one campaign | intrinsic | ad recall, sponsor fit, purchase intent | +31.8 pts mobile ad recall; +26.2 pts purchase intent | not reported | Comscore and Frameplay (both interested) | single campaign | [479, 479] |
| Roblox / Latitude audience impact and brand lift | 2024 | exposed vs control survey | Association (exposed vs control, assignment not documented) | 2,100 (1,800 exposed, 300 control) | US monthly Roblox users 13-49 | Roblox immersive video ads | brand and ad awareness, opinion, consideration, recommendation; intended actions | significant increases in all key metrics (no magnitudes); visit brand's Roblox experience +10 pts; search +3 pts | significance stated; level not given | Roblox (seller) | US; includes teens; three brands | [28] |
| Roblox / MediaScience biometric study | 2025 | lab eye tracking + electrodermal | Descriptive (lab measures; comparison design not published) | 140 | US Roblox users 13-34; branded experiences vs feed ads and streaming TV | branded Roblox experiences | attention, unaided recall | attention 100x social, 35x streaming; unaided recall +211% vs social; vs streaming +95% (Digiday) or +38% (Tubefilter) | not reported | Roblox (seller) with MediaScience | US; lab; format not like-for-like | [484] |
| Roblox / Ipsos Immersive Is Effective | 2026 | self-report survey of three platform bases | Descriptive (self-report) | 600 (200 per base) | Roblox, streaming and social users 13-34 | Roblox ads (unspecified) | self-reported attention, memorability | 52% pay more attention to ads on Roblox (+13 pts vs streaming, +11 vs social); 63% say product info memorable | not reported | Roblox (seller) with Ipsos | geography not stated; no exposure measurement | [486] |
| EDO / Roblox outcomes pilot | 2026 | engagement-rate norms; event-window branded-search lift vs adjacent weeks | Association (before-and-after search lift) | not stated | Roblox campaigns and branded events | rewarded video, billboard video, branded events | branded search engagement | 'clear, measurable lifts' (no figures published) | not reported | EDO (measurement vendor) and Roblox | US; proxy outcome | [488] |
| Discord / Kantar: DOOM: The Dark Ages | 2025 | Kantar brand lift survey | Association (exposed vs control, assignment not documented) | not stated | 7-day Video Quest, US and Canada | Video Quest (rewarded video) | ad awareness, play intent, top-of-mind | +11.3 pts ad awareness; +3.4 pts play intent; +5.7 pts TOMA (frequent FPS players) | play intent at 90% confidence | Discord (seller); Kantar | game advertiser; US and Canada | [341] |
| Discord / Kantar: The Black Phone 2 | 2025 | Kantar brand lift survey | Association (exposed vs control, assignment not documented) | not stated | multi-Quest film campaign | Video Quests + Mission Quest | awareness, favorability, intent | ad awareness +30.7%; favorability +4.8%; action intent +2.7%; TOMA +1.9% (units unstated) | 90% significance stated | Discord (seller); Kantar | film advertiser | [476] |
| Discord / Kantar: Obsession | 2026 | Kantar brand lift survey | Association (exposed vs control, assignment not documented) | not stated | three-phase film campaign | Video Quests + Mission Quest | awareness, action intent | +12.9% awareness; +5.2% action intent (units unstated) | not reported | Discord (seller); Kantar | film advertiser | [489] |
| Discord: World of Warcraft x Mountain Dew | 2025 | survey of exposed users (method not stated) | Not assessable (method not published) | not stated | Quest: stream WoW to a friend | Quest | brand favorability, perception | +16 pts favorability; +6 pts positive perception | not reported | Discord (seller), internal data | non-endemic FMCG brand | [490] |
| Activision Blizzard Media rewarded video (unnamed research) | 2025 | not stated | Not assessable (method not published) | not stated | mobile games | rewarded video | ad recall | +17% ad recall lift | not reported | Microsoft (seller) | no method | [580] |
| Gameloft for brands: Attention & Emotion in Play | 2026 | lab eye tracking + facial coding across three environments | Descriptive (lab measures across environments) | not stated | in-game interactive video vs YouTube pre-roll vs web video | in-game interactive video | positive attention, emotional engagement | +18% positive attention and +24% engagement vs YouTube | not reported | Gameloft (seller); protocol by Mediamento and ActFuture | report gated | [492] |
| Digital Turbine / Amplified Intelligence | 2023 | attention panel + virtual-store brand choice (STAS) | Association (attention and brand choice, assignment not documented) | seven brands (respondents not stated) | mobile games | in-game mobile video with end card | active attention, STAS | 22 s combined active attention vs 1.6 s mobile web video, 2.5 s social video; STAS 160 vs 100 | not reported | Digital Turbine (seller) with Amplified | not like-for-like durations | [494] |
| iion / Amplified Intelligence: Garnier | 2024 | attention model on campaign impressions | Descriptive (modelled attention) | 674,343 impressions | 'Around the Game' placements on mobile and tablet | display and video between levels | active attention per impression | video 7.38 s vs 1.10 s benchmark; display 1.48 s vs 0.40 s (also stated as 2.4 s) | not reported | iion (seller) with Amplified | single campaign; internal inconsistency | [495] |
| Venatus / Lumen 'In the Zone' | 2026 | eye-tracking attention model | Descriptive (modelled attention) | not stated | Venatus gaming, sports and entertainment sites | video (web) | % viewed, view time, APM | about 2x likely to view; 8 s vs 2 s; up to 5x APM | not reported | Venatus (seller) with Lumen | gaming websites, not in-game | [395] |
| Super League: Lunchables Playables world | undated | case study; method not stated | Not assessable (method not published) | not stated | virtual world for children | branded world | purchase rate | +6.3% lift in purchase rate | not reported | Super League (seller) | no method; child audience | [496] |
| Gadsme inventory brand lift surveys | undated | brand lift surveys (count and provider not stated) | Not assessable (method not published) | not stated | Gadsme inventory | intrinsic | ad recall, brand image, consideration | 35% average ad recall; +25 pts brand image and consideration | not reported | Gadsme (seller) | no method | [118] |
| Büyükakgül, Yüce, Katırcı: in-app ads in a mobile sports game | 2025 | lab eye tracking (Tobii Pro Glasses 2), GLMM | Association (lab observation) | 79 | adults, mean age 25.8, Turkey; 10 min play | embedded static ads in mobile tennis game | noticing, unaided and aided recall | 26.6% noticed any ad unaided; task-relevant placement strongest predictor of recall (p < 0.001) | p-values reported | no external funding (university) | one game, convenience sample, non-US | [26] |
| van Berlo, van Reijmersdal, Eisend: advergame meta-analysis | 2021 | meta-analysis | Pooled estimate across studies (meta-analysis) | 38 data sets | published studies | advergames | attitude, memory, persuasion, choice, ad recognition | more positive attitudes; lower brand memory than other ads; less recognised as advertising; younger more susceptible | effect sizes not read (abstract only) | funding unknown | advergames, not in-game ads | [498] |
| Ingendahl, Vogel, Maedche, Wänke: brand placements and in-game experiences | 2022 | two lab experiments | Causal, for the lab stimulus (randomised experiment) | not in abstract | 3D first-person game | brand placement | brand attitude | only placements met during positive experiences benefit; overall placement effect small and variable across studies | not read | funding unknown | lab | [499] |
| Ingendahl, Brückner, Vogel: attribute associations | 2024 | 2 x 2 within-subjects lab experiment | Causal, for the lab stimulus (randomised experiment) | 76 | third-person adventure game | brand placement | brand attitude, attribute associations | attitudes more positive with positive in-game experiences; attribute associations mediate | not read | funding unknown | lab | [500] |
| Min, Kim, Ko: esports in-game situations and ad animation | 2025 | 2 x 2 between-subjects lab eye tracking | Causal, for the lab stimulus (randomised experiment) | not in abstract | recorded esports broadcast | in-broadcast ads | fixation, memory, brand attitude, intention | non-battle moments raise fixations; fixations raise memory; implicit memory raises attitude and intention | not read | funding unknown | esports viewing, lab | [505] |
| Huang, Morozov: promotional effects of Twitch live streams | 2025 | observational panel: high-frequency data on 60,000 Twitch streamers sampled every 10 minutes for eight months, with within-day variation in top influencers' broadcast hours | Quasi-experimental (observational panel) | 60,000 streamers | Twitch influencer live streams | sponsored game streams | game usage and purchases | elasticity of 0.027 of a game's concurrent players to its live-stream viewers, which the authors call moderate and short-lived; about one sixth of games profit from sponsored streams | not stated in the abstract | funding unknown | game promotion, not non-endemic brand | [503, 504] |
| Zagni, Rizzo, Filieri: perceived benefit and in-game ads | 2026 | 2 x 2 between-subjects experiment + qualitative | Causal, for the lab stimulus (randomised experiment) | not in abstract | not stated | in-game ads (congruity x interactivity) | perceived benefit, attitude toward IGA | interactivity and congruity raise perceived benefits and attitudes | not read | funding unknown | attitudes only | [501] |
| Hatakeyama: rewarded in-game banner tap intention | 2026 | 2 x 3 between-subjects online experiment | Causal, for the lab stimulus (randomised experiment) | 600 | Japanese smartphone users with RPG involvement | rewarded in-game banner | tap intention | in-game currency raised tap intention vs no reward; not vs e-money points; category fit no effect | significance stated | funding unknown | Japan; intention only | [502] |
| Lumen and eye square validation of Intrinsic Time-in-View | 2022 | eye-tracking comparison | Descriptive (method comparison) | not stated | one McCormick campaign in Basketball Battle | intrinsic | viewable time, attention | 1.4x attention vs Lumen mobile display norm | not reported | Frameplay (release); study funding unknown | one campaign, one title | [356] |
| Kochava attribution case | undated (before 2024-10) | attribution counts, no control | Descriptive (attribution counts, no control) | 1 campaign | power-sports brand; supply-side partner apps | intrinsic | web engagements | 60,000+ attributed web engagements from 2.7M impressions | not reported | Kochava (partner) | single campaign | [553] |
| Ipsos GameTrack: brand integrations in games | 2026 | exposed vs unexposed survey across nine in-game campaigns, with brands that did not advertise in the games as a placebo check; exposure self-selected by who played which game | Association (exposed vs unexposed, self-selected; placebo brands) | 37,000 interviews, 16 European countries | GameTrack consumer survey, first quarter of 2026 | brand integrations in games, including EA SPORTS FC | "a brand for people like me" | average uplift of 31% across nine campaigns, against 6% for brands that did not advertise in the games; EA SPORTS FC integrations: Beats +97%, Pepsi +26%, Visa +18% | not stated; uplifts are relative | Ipsos (measurement firm) for Video Games Europe (trade body) | Europe, not US; players only; self-selected exposure | [477] |
## 10. Exits, platform growth and the agentic question
*Since 2021, specialist in-game ad companies have been bought, merged, put into insolvency or wound down, while platforms and publishers launched or widened their own ad sales. No record reviewed says the in-game format caused an exit, and no comparable revenue series shows which side grew faster. Artificial intelligence (AI) features are few and mostly in test, and in the sources screened no in-game seller documents use of an agentic buying protocol.*
> **In short.**
> Primary records show at least three specialist companies sold or put into insolvency since 1 January 2021: Super League bought Bloxbiz and Let's Bounce, and Bidstack Group Plc entered administration with a pre-packaged sale in March 2024, and its former operating company, Bidstack Ltd, entered creditors' voluntary liquidation in June 2026; Super League's purchase of the Misfits Ads business line from an esports company and Admix's merger with LandVault, announced in 2022, are recorded but not counted [29, 30, 31, 32, 506, 507]. Bidstack's administrators cite liquidity and record that a 242-party sale process drew no third-party offer amid doubts about future revenue and sector outlook; no record reviewed attributes an exit to the in-game format [31, 507]. Roblox announced programmatic routes from April 2025, Electronic Arts (EA) launched EA Advertising in June 2026 and Zynga announced Zynga Ads+ in September 2026 [187, 188, 25, 183]. H5, the market-structure hypothesis, is unresolved: part (a) holds, and part (b) cannot be tested because comparable revenue series are not disclosed [197, 198, 199, 200, 143, 128]. The Ad Context Protocol (AdCP) can represent a gaming channel at schema level, but in the sources screened no in-game seller shows a connector, test, production use or transaction [184, 185, 121]. Documented AI features for in-game products are few: Roblox's are in test or announced, and Super League reports a partner's AI creative service, announced in January 2026, with no documented controls [209, 508, 30].
### Each event, classified from its record
The primary records since 2021 show acquisitions, a merger and pivot, an administration with a pre-packaged sale, a dissolution, creditors' voluntary liquidations, an undated court liquidation, a strike-off application, and one product retirement announced and another in progress [29, 30, 507, 31, 509, 32, 510, 150, 339, 511].
| Company or product | Lane | Classification | Date | Record |
|---|---|---|---|---|
| Bloxbiz | Specialist | Acquisition (asset purchase by Super League) | 4 October 2021 | Form 8-K [29] |
| ESL Gaming | Around the game | Acquisition (by Savvy Gaming Group) | 21 April 2022 | Seller's release [322] |
| Admix | Specialist | Merger with LandVault, and pivot | 21 June 2022 | Party release [507] |
| Bidstack Group Plc | Specialist | Administration with pre-packaged sale | 22 March 2024 | Administrators' proposals, registry [31, 512] |
| NitroPay | Around the game | Acquisition (by Overwolf) | 8 May 2024 | Buyer's release [293] |
| Oracle Moat | Measurement | Retirement announced, effective 30 September 2024 | FAQ of 12 August 2024 | Oracle notice [339] |
| Enthusiast Gaming | Around the game | Sale of its direct-sales business line | 30 August 2025 | Seller's release [317] |
| Let's Bounce | Specialist | Acquisition (asset purchase by Super League) | 5 January 2026 | Form 10-Q [30] |
| Misfits Ads | Specialist | Acquisition (asset purchase by Super League) | 1 May 2026 | Form 8-K [506] |
| Bidstack Ltd and its parent | Specialist | Creditors' voluntary liquidation | 18 June 2026 | Registry [32, 64] |
| Venatus | Around the game | Acquisition (change of control) | 19 June 2026 | Registry [302, 303] |
| AdInMo Ltd | Specialist | Strike-off application, suspended | 7 and 17 September 2026 | Registry [112, 151] |
| Super Biz (formerly Bloxbiz) | Specialist | Retirement (in progress) | Undated; read 28 September 2026 | Vendor page [511] |
| Sayollo Media Ltd | Specialist | Court liquidation | Undated in the record | Israeli registry [510] |
Super League bought Bloxbiz, Let's Bounce and the Misfits Ads business line [29, 30, 506]. For Bloxbiz it paid $6.0 million at closing plus up to $11.5 million contingent on revenue milestones through fiscal 2023 [29]. Bloxbiz became Super Biz, and the Super Biz site now says the platform is being deprecated, with sign-ups closed and work continuing with a small group of studios [128, 511]. For Misfits Ads, a business line of Esports Now, LLC, it paid $1.5 million in cash plus shares and warrants on 1 May 2026, with a fixed $300,000 cash payment due on the first anniversary and an earn-out of up to $1.2 million in cash and 105,571 shares [506]. On 5 January 2026 it had bought Let's Bounce, Inc., a marketing technology company for brand engagement inside gaming and user-generated content environments, in an asset acquisition priced at $200,000, with up to $325,000 more payable on net revenue milestones for 2026 [30]. The buyer may itself change hands: under agreements signed on 18 August 2026, Metaplanet would hold about 95.7% of Super League's common stock, subject to a stockholder vote on 16 October 2026, after the cutoff [513, 514].
Admix, an in-play product-placement company, announced a merger with the metaverse builder LandVault in a release dated 21 June 2022 [507]. The merged entity would carry the LandVault name, Admix's founder and chief executive would lead it and LandVault's founders would leave, and the release gives the growth of Admix's metaverse revenue as context, so this paper classifies the event as a merger, rebrand and pivot away from in-game placements, not a sale [507]. Infinite Reality announced a $450 million all-stock acquisition of Landvault on 9 July 2024; this review did not establish that it completed [515]. An "Admix" ad network now runs at admixplay.com under W Media International Ltd, and no record opened links that operator to the original company [516, 517].
Israel's Registrar of Companies records Sayollo Media Ltd as "in liquidation by court", with 2021 as its last annual report and no order date [510]. In December 2021 it had launched gComm, an in-game marketplace, which suggests the liquidation came later (inference) [518]. The record gives no cause and no order date, and H5 needs a primary record dating each event after 1 January 2021, so Sayollo is left out of the H5 count [510]. AdInMo Ltd applied for voluntary strike-off on 7 September 2026, and the action was suspended on 17 September after an objection [112, 150, 151]. That is neither an insolvency nor a completed dissolution; [Who sells inventory inside games](#networks) explains why AdInMo moved to the historical rows [150, 153]. The product register holds 8 historical rows, none of them scored.
*Figure fig-events. **Specialists were sold, merged or wound up while platforms and publishers built their own ad sales** Events classified from primary records (filings, registries, party releases); grey dots mark insolvency events (administration, liquidation). An exit does not show that a format failed: Bidstack's administrators cite liquidity, and no record reviewed attributes an exit to the in-game format. Units: dated events. Scope: in-game specialists, platforms, game-adjacent sellers and measurement, 2021 to the cutoff. Sources [29], [507], [31], [30], [506], [32], [522], [187], [25], [293], [317], [303], [402], [339]. Caveat: Comparable revenue series are not disclosed, so the growth half of H5 stays unresolved. Data: data/figure-data.json#events*
### Bidstack: what the administrators recorded
Bidstack's record is the fullest of the exits, and its administrators name liquidity as the cause [31]. Bidstack Group Plc, a listed holding company, entered administration on 22 March 2024 [512]. The same day its administrators sold the business and assets, including its shareholdings in subsidiaries, for £804,848 to Virtual Sport Technology Limited, a vehicle owned by the plc's then management team [31, 512, 519]. A marketing process that contacted 242 parties had produced no third-party offer [31].
The administrators attribute the administration to liquidity constraints [31]. Further drawdowns on an October 2023 convertible loan note facility from Irdeto could not be made, because its conditions, City Code requirements among them, could not be met [31]. Group turnover, across all Bidstack activities, had fallen to £3.1 million in 2023 from £5.3 million in 2022, on unaudited management accounts [31]. The parties approached in the sale cited strategic fit, timelines, uncertainty over future revenues and the outlook for the media, advertising and gaming sectors [31]. Those are doubts about one company's revenue and a sector's prospects; the administrators do not say the in-game format failed [31].
The plc moved from administration to dissolution and was dissolved on 26 June 2025 [509, 512]. Bidstack Ltd and its parent, Virtual Sport Technology Limited, entered creditors' voluntary liquidation on 18 June 2026 [32, 64]. IION's sellers.json, last modified on 28 September 2026, still lists Bidstack as an intermediary seller, although Bidstack Ltd has been in creditors' voluntary liquidation since 18 June 2026 [11, 32].
### Around the game, owners changed and the products kept selling
Game-adjacent sellers changed owners, and the records show the products continuing [323, 293, 520, 521]. Savvy Gaming Group, wholly owned by Saudi Arabia's Public Investment Fund (PIF), bought ESL and FACEIT and merged them, and the ESL sale closed on 21 April 2022 [321, 322]. Overwolf announced on 8 May 2024 that it had bought NitroPay, renamed Nitro and run as a division [293]. Enthusiast Gaming sold its direct-sales business line to Vertiqal Studios for C$900,000 in cash plus assumed liabilities, closing on 30 August 2025, and kept reporting quarterly results through the second quarter of 2026 [317, 520]. On 19 June 2026 Media Scale Capital Limited, controlled by the Dutch company To Media Groep, took control of Venatus's holding chain from the private equity fund Livingbridge 7 Global LP and two individuals; Venatus's operating company stayed active on the UK register [302, 303, 305, 521].
Measurement was set to lose a service: Oracle's end-of-life FAQ of 12 August 2024 states that Oracle would end all its advertising products, Moat Analytics and Moat Reach included, on 30 September 2024, and this review opened no record of the retirement itself [339]. [Measurement](#measurement) records the pages that still name Moat [364].
### Platforms launched or widened their own ad sales
Platforms and publishers widened their own ad sales over the same years, and the growth shows in products and routes, not in disclosed revenue [187, 188, 25, 197]. Microsoft completed its acquisition of Activision Blizzard on 13 October 2023, and the former Activision Blizzard Media now sells as XBOX Media Solutions, a rename announced on 4 February 2026 [522, 222, 223].
Roblox announced Rewarded Video through Google Ad Manager and direct insertion orders on 1 April 2025 [187]. On 6 January 2026 it announced expansion with the demand-side platforms (DSPs) Amazon DSP and Liftoff and the supply-side platforms Index Exchange, Magnite and PubMatic; no DSP's own documentation of a live integration was found in the sources screened [187, 188, 42]. Its filings call advertising and licensing revenue "insignificant" and say it plans to "invest in and expand" its advertising business [197, 246].
EA launched EA Advertising on 15 June 2026, saying advertisers "can now" reach players through native ad units served dynamically in select, unnamed EA SPORTS games [25]. At launch EA described direct, bespoke work with brands, with expanded buying planned, and on 3 July 2026 an EA advertising executive told Digiday that EA was not yet working with partners such as The Trade Desk or Criteo; on 15 September 2026 Digiday reported that EA already offers programmatic buying across its mobile games and is extending more dynamic capabilities to its console and PC titles, with no DSP or deal type named [25, 86, 193]. Zynga, a Take-Two label, presented Zynga Ads+ on 17 September 2026 as a third-party publisher network of about 100 casual mobile games; no participating app is named, so this paper lists it as unconfirmed [183, 203, 200]. A platform-owned seller is moving into third-party supply, the specialists' business; that reading is this review's inference [183].
### H5: part (a) holds, part (b) cannot be tested
H5 needs two parts to hold: (a) at least three specialist in-game ad companies entered insolvency, liquidation or dissolution, retired their product, or were sold since 1 January 2021, each shown by a primary record; and (b) comparable disclosed series showing platform ad revenue growing faster than specialist revenue.
Part (a) holds on the registered wording: primary records show three specialist companies sold or put into insolvency, Bloxbiz and Let's Bounce bought by Super League and Bidstack's plc in administration and its operating company in liquidation, and Super League's purchase of the Misfits Ads business line from an esports company is recorded and not counted [29, 30, 31, 32, 506]. Admix's merger, announced in June 2022, under which Admix's own chief executive would lead the combined company under the LandVault name, is a merger, rebrand and pivot, which the rule does not list, so it is recorded and not counted [507]. Super Biz, Sayollo and AdInMo are also recorded beside them and not counted [511, 510, 150]. Three of these events, Bloxbiz, Let's Bounce and Misfits Ads, were purchases by one buyer, Super League, which reads as consolidation as much as departure; that reading is an inference [29, 30, 506]. Anzu, Gadsme, IION, Adverty and Frameplay remain active on registry or vendor evidence, Kidoz and Admazing were live in the period, and Zynga presented Zynga Ads+ in September 2026, so exits alone do not show the category shrinking [523, 524, 525, 143, 10, 136, 65, 526, 183]. Adverty, a listed specialist, reported net sales of 33.7 million Swedish kronor for 2025 against 31.8 million for 2024 [143, 79].
Part (b) cannot be tested: Roblox gives no advertising figure, Microsoft reports gaming advertising only inside an XBOX line that also holds hardware, content, subscriptions and cloud gaming, and Amazon reports only company-wide advertising services [197, 198, 199]. The disclosed series, Take-Two's advertising in dollars, Adverty's net sales in kronor and Super League's media revenue including influencer work, differ in metric, currency, period and scope [200, 143, 128]. [Platforms and publishers selling ads in their own games](#platforms) sets them side by side. H5 stays unresolved overall, and part (a) stands as a count of classified events [29, 30, 507, 31, 506].
### AI features are few, and most are in test
Documented AI capabilities for advertising inside games are thin at the cutoff: an asset generator for a paid-discovery product, tests and announcements, and claims without documented controls [527, 209, 508, 528]. Roblox's Ads Manager offers "AI generate", which creates three thumbnail variations from a text prompt for Sponsored games on the Home page and in search, a product this paper lists and does not score; no AI generation or AI-assisted placement is documented for Immersive Ads or Rewarded Video in the Roblox documentation read [527, 529, 530]. On 20 May 2026 Roblox was testing a drag-and-drop Studio plugin for Rewarded Video [209]. MediaPost reported on 15 September 2026 a Studio plug-in with rewarded-video templates "expected to launch in Q4 2026", and said a Roblox presentation described the upcoming placement options as no-code and using AI; no primary Roblox document for the timing or the AI was found [209, 508].
Super League's 10-K for 2025 states that it does not currently incorporate AI or machine learning into its products [128]. Its second-quarter 10-Q reports its partner AdArcade's "AI-powered Creative-as-a-Service" that converts brands' video ads into playables, announced in January 2026, with no controls documented in the filing [128, 30]. Anzu's dynamic channels change the creative after an impression and a cool-off period, and Gadsme's SDK, the software development kit publishers build in, selects ads from several demand sources at run time; no AI-assisted placement was found in the Gadsme pages opened [358, 59, 531]. Around the game, Amazon documents user review and creative moderation for a generative-AI workflow that makes Twitch-sized assets [532]. Under the coding rule for research question 11, an announcement counts as announced, and AI-assisted placement counts only with documented controls; [The capability matrix and the Play Quotient](#matrix) applies that rule.
### Agentic buying: a gaming channel in the schema, no seller on it
At the cutoff a gaming channel exists in the AdCP schema, and in the sources screened no in-game seller documents a connector, test, production use or transaction under it or under any other agentic protocol [185, 121, 9]. This paper records five protocol states for each seller: schema representation, implemented connector, demonstrated interoperability test, documented production use and demonstrated transaction. Each state needs its own source; membership of a protocol body, working-group roles and announcements earn no credit.
**AI and agentic capability map**
| Platform / product | Capability | Status | Date | Independently observable? | What the record supports | Src |
|---|---|---|---|---|---|---|
| Roblox | AI-generated ad thumbnails in Ads Manager ('AI generate', three variations from a text prompt) | live | documentation as of 2026-09-28 | Creator Hub documentation | Applies to Sponsored games and search ads (paid discovery, listed not scored); not documented for Immersive Ads or Rewarded Video; no AI-specific controls documented | [527] |
| Roblox | Drag-and-drop Rewarded Video plugin in Studio (no-code placement) | unconfirmed | 2026-05-20 | Developer Forum announcement | Q1 2026 letter: releasing 'in the coming months'; AI element not stated by Roblox | [209] |
| Roblox | 'Join with Reward' rewarded placement on the experience details page before entry | unconfirmed | 2026-05-20 | Developer Forum announcement | Renders before gameplay, outside the experience | [209] |
| Roblox | Studio plug-in with 2D and 3D templates, pre-gameplay rewarded surface, multi-reward rotation; described as no-code and using AI | unconfirmed | 2026-09-15 | trade press report read in summary | No primary Roblox source for the date or the AI element found | [508] |
| Roblox | Homepage Feature: brand video turned into an immersive 3D environment without custom development | beta | 2026-01-06 | vendor release; 8-K exhibit | Template-based asset tooling; AI not stated; homepage is outside gameplay | [188] |
| Microsoft (XBOX Media Solutions) | AI or agentic features for in-game formats | unconfirmed | 2026-09-28 | four gaming pages searched | Bounded to pages opened | [191] |
| Microsoft Advertising | MCP Server for AI assistants (campaigns, ad groups, ads, keywords, performance metrics) | live | accessed 2026-09-28 | vendor page | Not documented for XBOX Media Solutions in-game inventory | [192] |
| Super League | AI-powered Creative-as-a-Service (AdArcade Native Playables) converting video ads to playables in rewarded mobile video inventory | announced | 2026-01 | 10-Q narrative | Partner technology; 10-K of 31 Mar 2026 says the company does not incorporate AI/ML into its products; no controls documented | [30] |
| Super League | AI and data-driven play-persona engine | unconfirmed | accessed 2026-09-28 | website | Planning, not placement; no controls documented | [528] |
| AdCP (AgenticAdvertising.org) | Protocol release | live | 2026-09-23; 2026-09-27 | GitHub releases | v3.1.24 channels enum includes 'gaming' ('In-game advertising across platforms', present since v3.0.0); canonical creative formats defer in-game ('Out of scope until cross-engine standards land') | [184, 185] |
| AdCP (AgenticAdvertising.org) | Rule-12 states for in-game sellers in the product register (schema representation; connector; interoperability test; production use; transaction) | unconfirmed | 2026-09-28 | public agent registry (34 agents) and v3.1.24 docs | Public tier only; members-only agents not visible | [121, 185] |
| IAB Tech Lab | ARTF (Agentic Real Time Framework) | live | 2026-07-20 | GitHub release; standard page | Bidstream container framework; not game-specific | [533] |
| IAB Tech Lab | AAMP 2.0 buyer and seller agent SDKs; seller agent v2.4.2 | live | 2026-04-23; 2026-08-18 | blog post; GitHub release | Connectors to Google Ad Manager, FreeWheel, PubMatic, Index Exchange, Magnite; no game seller named | [536] |
| IAB Tech Lab | Agentic Mobile specification with REWARDED placement type | beta | accessed 2026-09-28 | specification text | Schema representation of rewarded in-app placements only; no intrinsic in-game object | [537] |
AdCP's latest stable release is v3.1.24 (23 September 2026) [184]. Its channels list includes "gaming" ("In-game advertising across platforms", present since v3.0.0), which its taxonomy defines to include in-game display, rewarded video and playable ads, while its canonical creative formats defer in-game ("Out of scope until cross-engine standards land") [185]. Its public agent registry listed 34 agents on 28 and 29 September 2026, none operated by Roblox, Microsoft, EA, Zynga, Anzu, Frameplay, Gadsme, Adverty, Gameloft, IION, Super League, Kidoz, Admazing or AdInMo [121]. Two of the 34 entries are listed under the name of No Fluff Advisory, this paper's publisher [121]. No in-game seller in this paper's register therefore shows an implemented AdCP connector, interoperability test, production use or transaction in the sources screened; the registry's public tier omits members-only agents [184, 185, 121].
In the sources screened, the cohort A networks document no agentic protocol at all; the only supply-chain standard observed in use is the sellers.json format, version 1.0, published by six of them [9, 10, 11, 12, 13, 14]. Microsoft Advertising's Model Context Protocol (MCP) Server describes access to campaigns, ad groups, ads and keywords without mentioning gaming inventory [192, 121]. Zynga and PubMatic announced in July 2026 that advertisers will be able to buy Zynga inventory through PubMatic's AgenticOS; the announcement names no protocol state [196].
The agentic specifications screened from the Interactive Advertising Bureau's Tech Lab (IAB Tech Lab) have no game-specific object [533, 534, 535, 536, 537]. IAB Tech Lab's seller agent, released in its AAMP 2.0 kits, connects to Google Ad Manager, FreeWheel, PubMatic, Index Exchange and Magnite, with no game seller named [535, 536]. Its Agentic Mobile specification, a draft for public comment, has a REWARDED placement type for mobile in-app inventory and no intrinsic in-game object [537]. Some intermediaries that Roblox's app-ads.txt authorises list sales agents in AdCP's registry or connect to IAB Tech Lab's seller agent, but no source screened shows game inventory offered to buyers' agents through any of them [538, 539, 121, 536].
A buyer seeking in-game inventory through an agent would find a gaming category in AdCP's schema, no seller from this paper's register among its agents, and one announced route, Zynga through PubMatic's AgenticOS, with no protocol state documented [185, 121, 536, 196].
## 11. The capability matrix and the Play Quotient
*The matrix scores 23 products on what their public record documents, within three cohorts that are never compared with each other. The field is thin in similar places in every cohort: creative tooling and format specifications score low throughout, coverage is the lowest-scoring dimension among in-game networks, and independent assurance has the most cells with no evidence to score.*
> **In short.**
> The Play Quotient (PQ) summarises documented capability within a cohort; it is not a measure of performance. In-game ad networks (cohort A) average a field PQ of 46: 4 rows are Well documented, 4 Partly documented and 1 Least documented. Platform and publisher products (cohort B1) average 46, but with 4 rows the cohort is too small to carry bands. Game-adjacent media (cohort B2) average 47, with 1 Most documented row, 2 Well documented and 7 Partly documented. Across the matrix 249 of the 276 cells carry a score, 17 are n/e and 10 are n/a. Vendor documentation caps a cell at 4 and vendor marketing at 3, and 52% of the scored cells rest on vendor marketing, so the scores measure how much a seller has put in writing and how independently it can be checked. No in-game product reaches 4 on independent assurance, and MRC's own record shows no accredited intrinsic in-game metric for any of them [15, 16].
### How to read a cell
Each cell scores one product on one dimension from 0 to 5 against written anchors, then applies the evidence cap: vendor documentation (grade C) caps a cell at 4, vendor marketing (grade D) at 3, so the type of evidence is folded into the number. A 5 needs an official document, a filing or independent observation. A cell with no basis to score is n/e and never counts as zero; n/a comes only from the cohort's mask, which in game-adjacent media removes the coverage dimension. Only cells a verifier re-checked count. Open any cell in the matrix below for its rationale, availability, evidence grade, date and sources.
The anchors describe a brand buyer who wants to buy games like other digital media: by context rather than personal data, through named programmatic routes, with creative that can change in flight and measurement someone else can check. Three choices follow and were fixed before scoring: targeting that uses only personal data scores no higher than targeting without a named signal; the top creative score needs AI-assisted tooling with documented controls; and a product sold only direct scores 0 on programmatic access. A managed sponsorship, a seller-built branded world or a contextual specialist can be the right buy and still score low there, so a buyer with such a brief should read the matching dimension columns, not the composite; reweighting the groups cannot undo a preference built into an anchor. [What it means](#implications) turns the cells into buyer questions.
Availability is a label on each cell. A cell is live when a source dated inside the scoring period, or a vendor page this review found still served, shows the capability. It is unconfirmed when the only evidence is older and survives in a case study, a slide, a reseller's copy or an archive. Under the registered rules an unconfirmed cell still counts, so the quotients include 9 such cells. Treating them as n/e changes 0 bands in cohort A and 1 in cohort B2: LIVAD livestream video ads, whose record rests mostly on material from before the period, falls to a thin record.
Rows are compared only within their cohort. A network selling placements inside other companies' games, a platform selling its own inventory and a streaming service answer different buying tasks, and the anchors for supply and measurement differ between them. Positions inside a band move under reasonable weights, so the paper reports bands, not ranks.
**The capability matrix** (0–5 evidence-bounded scores; n/e = insufficient evidence; n/a = outside the cohort’s mask; unverified cells are marked and counted as n/e; availability, evidence grade and verification status per cell are in data/capability-matrix-long.csv; one section per cohort; banded products ordered by PQ, the Play Quotient, within their cohort, and unbanded rows listed A to Z)
*Cohort A: in-game ad networks*
| Vendor | Product | Group | PQ (band) | Formats | Targeting | Measurement | Outcomes | Creative | Supply | Coverage | Programmatic | US | Assurance | Transparency | Suitability | Counted |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Anzu | Anzu in-game ad network | network | 57 (Well documented) | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 2 | 3 | 4 | 12/12 |
| Adverty in-game ad network | Adverty in-game ad network | network | 52 (Well documented) | 2 | 4 | 3 | 2 | 3 | 3 | 1 | 3 | 3 | 1 | 4 | 3 | 12/12 |
| Frameplay | Frameplay intrinsic in-game network | network | 51 (Well documented) | 2 | 3 | 3 | 3 | 3 | 2 | 1 | 3 | 3 | 3 (unconfirmed) | 3 | 2 | 12/12 |
| Gadsme | Gadsme Managed Ad Network | network | 49 (Well documented) | 1 | 3 | 3 | 2 | 3 | 2 | 2 | 3 | 3 | 2 | 3 | 3 | 12/12 |
| Kidoz kid-safe ad network | Kidoz kid-safe ad network | network | 47 (Partly documented) | 1 | 3 | 1 | 3 | 1 | 3 | 1 | 3 | 3 | 3 | 3 | 3 | 12/12 |
| Gameloft for brands | Gameloft for brands (COMBO!) | network | 45 (Partly documented) | 1 | 2 | 3 | 2 | 1 | 3 | n/e | 1 | 3 | 3 | 3 | 2 | 11/12 |
| Admazing mobile gaming ads for brands | Admazing mobile gaming ads for brands | network | 41 (Partly documented, provisional: Least documented with n/e = 1, Well documented with n/e = 4) | 1 | 2 | 3 | 2 | 1 | 2 | n/e | 1 | 3 | n/e | n/e | 2 | 9/12 |
| IION in-game brand network | IION in-game brand network | network | 40 (Partly documented) | 1 | 3 | 3 | 2 | 1 | 2 | 1 | 2 | 1 | 1 | 3 | 2 | 12/12 |
| Super League | Super League in-game advertising | network | 38 (Least documented, provisional: Least documented with n/e = 1, Partly documented with n/e = 4) | 1 | 1 | 3 | 3 | 1 | 2 | n/e | 1 | 3 | n/e | 2 | 1 | 10/12 |
*Cohort B1: platform and publisher in-game ads*
| Vendor | Product | Group | PQ (band) | Formats | Targeting | Measurement | Outcomes | Creative | Supply | Coverage | Programmatic | US | Assurance | Transparency | Suitability | Counted |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EA Advertising | EA Advertising | platform | 44 (cohort too small) | 1 | 2 | 3 | 3 | 1 | 3 | 2 | 1 | 3 | 1 | 3 (unconfirmed) | 3 | 12/12 |
| Roblox | Roblox brand ads (Immersive Ads and Rewarded Video) | platform | 56 (cohort too small) | 4 | 3 | 3 | 3 | 3 | 3 | 1 | 3 | 3 | 2 | 3 | 4 | 12/12 |
| XBOX Media Solutions (Microsoft) | XBOX Media Solutions in-game ads | platform | 43 (cohort too small) | 1 | 2 (unconfirmed) | 3 | 2 | 1 | 3 | 2 | n/e | 2 | 3 | 2 | 2 | 11/12 |
| Zynga Ads | Zynga Ads | platform | 41 (cohort too small) | 1 | 2 | 3 | 2 | 1 | 3 | 1 | 2 | 3 | 1 | 3 | 2 | 12/12 |
*Cohort B2: game-adjacent media*
| Vendor | Product | Group | PQ (band) | Formats | Targeting | Measurement | Outcomes | Creative | Supply | Coverage | Programmatic | US | Assurance | Transparency | Suitability | Counted |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Twitch (Amazon) | Twitch advertising | adjacent | 64 (Most documented) | 3 | 4 | 4 | 3 | 4 | 3 | n/a | 3 | 3 | 3 | 4 | 4 | 11/11 |
| Overwolf ads | Overwolf ads (companion apps and gaming websites) | adjacent | 54 (Well documented) | 2 | 3 | 3 | 2 | 3 | 3 | n/a | 3 | 3 | 2 | 3 | 4 | 11/11 |
| IGN advertising | IGN advertising | adjacent | 50 (Well documented, provisional: Partly documented with n/e = 1, Well documented with n/e = 4) | 3 | 2 | 2 | 1 | 3 | 3 | n/a | 3 | 4 | n/e | 3 | 2 | 10/11 |
| Venatus gaming media | Venatus gaming media | adjacent | 47 (Partly documented) | 1 | 2 | 3 | 2 | 1 | 3 | n/a | 3 | 3 | 3 | 3 | 2 | 11/11 |
| LIVAD livestream video ads | LIVAD livestream video ads | adjacent | 47 (Partly documented, provisional: Partly documented with n/e = 1, Well documented with n/e = 4) | 3 (unconfirmed) | 2 (unconfirmed) | 3 (unconfirmed) | 3 (unconfirmed) | 1 | 3 | n/a | 2 | 1 (unconfirmed) | n/e | 3 (unconfirmed) | 2 | 10/11 |
| Discord Quests | Discord Quests | adjacent | 46 (Partly documented) | 2 | 1 | 3 | 3 | 1 | 3 | n/a | 0 | 3 | n/e | 3 | 3 | 10/11 |
| Playwire gaming media | Playwire gaming media | adjacent | 43 (Partly documented) | 2 | 2 | 2 | 1 | 1 | 3 | n/a | n/e | 3 | 1 | 3 | 2 | 10/11 |
| Medal advertising | Medal advertising (ad units and sponsored placements) | adjacent | 42 (Partly documented) | 1 | 1 | 1 | 2 | 1 | 3 | n/a | 2 | 2 | n/e | 3 | 3 | 10/11 |
| Enthusiast Gaming media | Enthusiast Gaming media | adjacent | 41 (Partly documented, provisional: Least documented with n/e = 1, Well documented with n/e = 4) | 2 | 1 | n/e | n/e | n/e | 3 | n/a | 2 | 2 | n/e | 3 | 1 | 7/11 |
| ESL FACEIT Group esports media | ESL FACEIT Group esports media | adjacent | 41 (Partly documented, provisional: Least documented with n/e = 1, Partly documented with n/e = 4) | 2 | 2 | 1 | 2 | 1 | 3 | n/a | 2 | 2 | n/e | 2 | 2 | 10/11 |
### Cohort A: in-game ad networks
The 4 Well documented networks are Anzu, Adverty, Frameplay and Gadsme. Each documents at least two contextual targeting signals, at least one measurement function and a sellers.json file, and each has a counted score on every dimension; Anzu, Adverty and Frameplay also document specifications for at least one format. Kidoz, Gameloft for brands (COMBO!), IION and Admazing are Partly documented. Super League is Least documented.
What separates the bands is documentation, not size. Super League files with the SEC, but its filings describe its product only in the company's own words, which grades those cells C: they name formats without specifications, claim programmatic access without naming a DSP or deal type, and document no suitability control in operational detail; no assurance record was found [128, 30, 15]. Admazing sells brand campaigns in third-party mobile games but names no titles, gives its inventory count without a period, and serves no sellers.json, ads.txt or app-ads.txt at the domains checked, so its supply cell is 2 and its transparency cell has no basis to score [130, 129, 147]. Its assurance cell is n/e: it names IAS and DoubleVerify as delivery-verification partners, and a partner's name alone is not evidence under the anchors [396, 15].
Across the cohort, the weakest dimensions are coverage, format execution and creative tooling: most networks name their formats without publishing specifications, a counting approach and creative guidelines together. Coverage is scored only where a named live title or documented integration shows a device, distribution and environment combination; Anzu reaches three such tuples, and the count depends on an inferred mobile device for its Gameloft titles and on classing Asphalt 8: Airborne as premium, both flagged in the cell [134, 135, 132]. Independent assurance tops out at 3: third-party verification rests on announcements dated 2022 to 2025, and several certification claims are missing from the certifiers' own lists [350, 351, 333]. Chapter 3 has the supply evidence behind these rows; chapter 6 the measurement.
### Cohort B1: platforms and publishers
Roblox documents the most in the cohort on formats and on suitability, age and data controls, its two cells scored 4. It documents unit-level specifications (unit sizes and video lengths, not advertiser asset specifications), a counting approach and creative guidelines for every in-game brand format it sells, and its filed audience and age mix, combined with its published age rules, would put its ad-eligible base above the supply threshold only by this review's derivation, an inference that counts accounts rather than people, so its supply cell stays at 3 [208, 207, 246, 43]. Its coverage cell is 1, because the PC and console tuples rest on a partner's announcement from before the scoring period, and its assurance cell is 2, because verification is documented for only some formats and environments [211, 351].
XBOX Media Solutions, EA Advertising and Zynga Ads have thinner records. None names DSP integrations with deal types for its in-game inventory; the best documented route is Zynga's exchange partnership with PubMatic [195, 378]. In the sources screened, none publishes specifications for its brand formats. The cohort has 4 rows, below the 6 the rules require before any band is shown, so its quotients are shown without bands. That is a limit of the market's public record: in the sources screened, few platforms and publishers document brand ads on their own games in public.
### Cohort B2: game-adjacent media
Twitch is the one Most documented row, and it holds that band under all 4 reweightings tested. Its record is documented in vendor detail on most dimensions: formats with specifications for at least one unit, three or more measurement functions, Amazon DSP as a named buying route, advertiser suitability tiers built on mandatory content labels, and a TAG Brand Safety certification on TAG's own record [540, 264, 2, 266, 337]. It stops short of 5 on measurement because its evidence is vendor documentation, which caps the cell at 4, and MRC's letter on IAS's Amazon DSP integration does not extend to Twitch supply [332].
Overwolf and IGN are Well documented: Overwolf on documented programmatic access, transparency and suitability controls, IGN on documented US availability, formats and creative tooling. 7 rows are Partly documented. Discord Quests scores 0 on programmatic access, because Discord's advertiser pages state that its ads can only be purchased through its sales team; the Ads Manager it announced on 29 September 2026 is an auction on Discord's own platform, in managed access, with no DSP or deal type named [48, 280, 248]. Medal, added after the second outside review, documents formats only by name and targeting only through personal data, and is Partly documented [252, 541]. Independent assurance is the thinnest dimension here: for most rows this review found no assurance record, so the cell is n/e rather than a low score.
### How stable the results are
*Figure fig-quotient. **Read the band, not the rank: one product clears the most documented line, and four rows are not banded.** The Play Quotient for the products with enough evidence to band, one section per cohort. The Play Quotient is built like the author’s Contextual and Native Quotients, except that its field mean is the cohort’s mean Raw score rather than pooled group scores; that does not validate it. Depth, Reach, Standing group means are combined 45/25/30 and shrunk toward the cohort’s field mean (the mean Raw score of its admitted rows) with a prior strength of 4; bands come from the unrounded distance to that field score, and only verified or qualified cells count. A band needs at least 6 counted cells with one in every applicable group, in a cohort with at least 6 such rows. Across the reweightings, positions move by up to 2 places within a cohort, and 15 of the 19 banded products keep their band under every weighting. PQ measures publicly documented capability within a cohort, not product performance. Units: PQ points on a 0–100 scale, pulled toward the cohort field mean; shown rounded, plotted unrounded. Scope: the banded products in the capability matrix, scored on 29 September 2026, with later score changes logged in the corrections ledger. Caveat: One full scoring pass, with blind AI re-scoring of the rows nearest a band line. The index inherits every limit of the dimension scores and is sensitive to the weights; the bands, not the positions, are the finding. Data: data/pq.csv*
The tests are reported per cohort in the matrix and in the downloads. Under 4 reweightings of Depth, Reach and Standing, positions move by at most 1 place in cohort A and 2 in cohort B2. Three rows in cohort A change band under at least one reweighting, IION between Partly and Least documented, Kidoz between Partly and Well documented and Super League between Least and Partly documented, and one in cohort B2, Venatus between Partly and Well documented.
Missing evidence is not neutral. When each row's n/e cells are filled with 1 and then with 4, with every other row frozen, 2 rows in cohort A and 4 in cohort B2 change band: Admazing and Super League in cohort A, and Enthusiast Gaming, ESL FACEIT Group, IGN and LIVAD in cohort B2. Their bands are provisional, and the matrix marks them ±. Deleting one counted cell at a time moves a quotient by at most 3 points up and 3 down in cohort A; the largest single effect is Adverty in-game ad network's independent assurance score of 1, whose removal lifts its quotient from 52 to 55. A low score that is documented costs more than a missing one, which is why the paper publishes coverage for every row and flags the band-sensitive ones. Setting the prior strength to 0 instead of 4, which removes the shrinkage, changes 3 bands in cohort A and 2 in cohort B2; at 2 the changes are 0 and 0, and at 8 they are 1 and 0.
These tests vary the weights, the prior strength and the missing cells. They do not vary evidence grades or the scorer's reading of an anchor, so stability under reweighting is not the same as confidence in the evidence: 52% of the scored cells rest on vendor marketing, and one scorer read every anchor.
A blind second scoring pass tested the scorer. Separate AI scorers re-scored every applicable cell of the 10 banded rows nearest a band line, those within 2 points of a line or whose band moves under a reweighting or prior strength, from the rules, each row's sources and the verified claims, without the first scores, grades or reasons. They matched the first pass on 98 of 115 cells (85%), came within one point on 106 of the 106 cells both passes scored, a quadratic-weighted kappa of 0.9, and differed on whether 3 cells could be scored at all. The 5 material disagreements were adjudicated against the deciding sources and none showed an error in the first pass, so no score changed; Kidoz and Venatus, which the second pass would lift from Partly to Well documented, are marked †. Both passes are AI scorers reading the same rules and the same verified claims, some of which were written with the anchors in mind, so the test shows reproducibility, not independent human judgement. In all, 8 of the 19 banded rows carry ± or †; data/second-pass.csv holds both passes cell by cell.
### The deep profiles
The profiles cover seven scored products and one verification vendor, DoubleVerify, whose gaming product is described but not scored. Each profile gives what is sold, who buys it, how it earns, its supply dependencies, its capabilities by dimension with availability and evidence grade, its performance evidence with sponsorship, and a SWOT analysis (strengths, weaknesses, opportunities and threats) in which every line carries a source and a basis. Opportunities and threats are the author's reading of dated facts.
### Anzu (active)
*Anzu Virtual Reality Ltd. (Israeli private company no. 515647808); its advertiser terms also name a subsidiary, Anzu Inc., whose jurisdiction this review did not establish*
**SWOT (2026-09-29).** Anzu is a privately held Israeli in-game ad network with the widest declared seller file in its cohort and documented buying routes, but thin independent assurance and no disclosed revenue or dated audience series.
- **Strengths.** On 29 September 2026 its sellers.json listed about 1,117 publisher seller accounts, none confidential; these are declared accounts, not titles, audience or active integrations. (vendor documentation) [9] It sells intrinsic display and video across mobile, PC and console, added web games on 19 November 2025 and click-enabled formats on 9 February 2026. (vendor claim) [123, 133, 122] Buyers can use IO, open auction or Deal ID private auction; StackAdapt (August 2026) and Smadex (February 2026) confirm the partnership from the DSP side. (vendor documentation) [397, 120, 175] Its brand-safety policy documents allow and block lists, category blocking, manual creative checks within 24 hours and takedown within one Israeli business day. (vendor documentation) [397]
- **Weaknesses.** On 29 September 2026 Anzu was on neither MRC digital list, and the TAG registry and IAB UK Gold Standard lists returned no Anzu record. (record) [15, 16, 383, 352] The only verifier-published verification record is IAS mobile viewability and invalid-traffic measurement from July 2022; no verifier document covers web or click-enabled formats. (vendor documentation) [350, 551] Its sellers.json is not game-only: a keyword screen found at least 118 of about 958 publisher domains to be TV, streaming, radio, news or weather properties. (inference) [9] No revenue or dated impression series is disclosed, and no public source documents title-level reporting to buyers as standard; advertiser terms leave billing reports to each IO. (vendor documentation) [164, 131, 363]
- **Opportunities.** Click-enabled formats sold through named performance DSPs (PRWeb release, 9 February 2026) give Anzu a route to performance budgets beyond brand campaigns. (inference) [122, 175] The Playgama web route (19 November 2025) and the UFL PC partnership (13 May 2026) widen supply beyond mobile, though UFL reported no delivered campaign. (inference) [133, 132] Trivver is the only gaming service under MRC review (audit in process April 2026), so an Anzu audit would close its assurance gap before peers pull ahead. (inference) [16, 15]
- **Threats.** July 2026 Marketplace Terms bar primarily child-directed apps, yet sellers.json lists children's video properties as publishers; an undocumented mode invites regulatory and buyer scrutiny. (inference) [171, 9] Its media-value page still names Moat, retired 30 September 2024, and its brand-safety policy dates from April 2023, which buyers auditing assurance may discount. (inference) [364, 339, 397] Super League's FY2025 10-K (filed 31 March 2026) names Anzu as a competitor while sometimes packaging such firms' inventory, so resellers can also compete for budgets. (inference) [128]
- **What is sold.** Brand placements rendered inside third-party games: intrinsic ('blended') display and video on in-game surfaces such as billboards and stadium boards, plus custom ads, sold to advertisers by insertion order and programmatically through open and private auctions (GAM-ANZU-F04, GAM-ANZU-F05). Since 9 February 2026 Anzu also sells click-enabled intrinsic formats to performance DSPs and bundles intrinsic units with interstitials and rewarded video (GAM-ANZU-F04); that performance and user-acquisition use is recorded here but its economics belong to the mobile app growth paper. Since September 2025 it also sells third-party audience-segment targeting (GAM-ANZU-F17).
- **Who buys it.** Brands and agencies buying direct (named US examples: SUPERPRETZEL via 160over90 in Philadelphia and Pittsburgh, QuikTrip), programmatic buyers through DSPs (Anzu names The Trade Desk, StackAdapt, Basis, Smadex and Adikteev), and resellers: NBCUniversal was its third-party sales partner for the US and UK from March 2022 (current status not established), and Super League says it sometimes includes such networks' inventory in its brand programmes (GAM-ANZU-F06, GAM-ANZU-F25, GAM-ANZU-F33).
- **How it earns.** Advertisers pay Anzu per IO, in advance unless the IO says otherwise, at rates set in the IO, with Anzu's own reports conclusive unless an advertiser's weekly independent reports differ by more than 5% (GAM-ANZU-F15). As an SSP it sells publisher inventory through real-time bidding under its Marketplace Terms and pays publishers; the revenue share is not public. It recruits Unity mobile developers with a bonus of $100 per $500 earned in the first three months, capped at $10,000 (GAM-ANZU-F36). No revenue, margin or impression series is disclosed; funding totals are vendor statements (GAM-ANZU-F03).
- **Supply and data dependencies.** Supply depends on developers integrating Anzu's SDK and pre-approving placements; publishers keep discretion over whether ads display (GAM-ANZU-F05, GAM-ANZU-F15). Part of the declared supply path runs through 58 intermediaries including Unity, Digital Turbine, Liftoff, InMobi, Ogury and Smaato, and through web distribution partner Playgama's Bridge SDK (GAM-ANZU-F09, GAM-ANZU-F14). Audience targeting depends on Eyeota, Predactiv and Semcasting; measurement on IAS (announced 2022 and 2023), Comscore, Human, Claritas and research partners Lumen and Happydemics (GAM-ANZU-F17, GAM-ANZU-F18, GAM-ANZU-F19). Console access rests on a vendor-claimed Xbox middleware licence with no Microsoft record found (GAM-ANZU-F13). End-user data is hosted on servers in the US (GAM-ANZU-F23).
- **Differentiation.** The largest declared supply connection among cohort A candidates: 1,117 PUBLISHER accounts in sellers.json, against 268 for IION and 223 for Gadsme (stream B, verified). This is not scale in titles or audience (GAM-ANZU-F07; GAM-B-F01). One of the few cohort A networks with a published privacy-mode design that separates contextual, behavioural and child-directed ('Kids Safe-Mode') data processing (GAM-ANZU-F23). A documented intrinsic web-game route through Playgama since November 2025, which Anzu calls the first for this format (a vendor claim) (GAM-ANZU-F14). It sells brand (non-clickable) and performance (click-enabled) demand on the same placements through named DSPs (GAM-ANZU-F04, GAM-ANZU-F06).
- **Risks.** Supply claims and the sellers.json overstate game-only scale: at least 118 publisher domains are TV, streaming, news or weather properties, and 58 intermediaries may carry non-intrinsic supply (GAM-ANZU-F08, GAM-ANZU-F09). Shared supply with IION, Gadsme, Adverty, COMBO! and resellers means buyers can meet the same inventory through several routes, so reach cannot be added (GAM-ANZU-F27, GAM-ANZU-F33). Stale assurance signals: MOAT is still listed as a partner after its retirement, TAG seals announced in 2021 are absent from the registry, and IAS is not named on the current measurement page (GAM-ANZU-F18, GAM-ANZU-F19, GAM-ANZU-F21). Age-related tension (inference, H4): the June 2026 blog pitches family-friendly titles to reach kids, while the privacy policy disclaims knowing collection from under-16s and the SSP terms require properties not directed to minors (GAM-ANZU-F24). The move into click-enabled performance formats may shift inventory toward user-acquisition budgets; this is inference from the vendor's own framing of two budget pools (GAM-ANZU-F04, GAM-ANZU-F31).
- **Best fit.** Brand buyers who want intrinsic display or video in named mobile titles, with allow and block lists by game and PMP or IO buying (GAM-ANZU-F05, GAM-ANZU-F22). Programmatic buyers already on StackAdapt or Smadex, where the DSP confirms the integration (GAM-ANZU-F06). Buyers who need a documented contextual or child-directed data mode on the publisher side (GAM-ANZU-F23).
- **Poor fit.** Buyers who need MRC-accredited in-game measurement or a current TAG seal (GAM-ANZU-F20, GAM-ANZU-F21). Plans that need named console titles or verified Xbox inventory in 2025-2026 (GAM-ANZU-F13). Buyers who need a deduplicated US ad-eligible audience figure before committing budget (GAM-ANZU-F11). Campaigns that must avoid audiences under 18 without their own title-level checks, given the family-audience marketing and the advertiser-side responsibility for audience eligibility (GAM-ANZU-F24).
- **Limits.** No ad-eligible audience figure with metric and period, and no US coverage figure, was found; supply scale rests on six vendor-named titles and an unquantified 'hundreds of live cross-platform games' (GAM-ANZU-F10, GAM-ANZU-F11). No console title carrying Anzu placements is named in the scoring period, and the Xbox licence claim has no Microsoft record in the sources screened (GAM-ANZU-F13). No MRC accreditation or audit in process, and no current TAG seal on the registry (GAM-ANZU-F20, GAM-ANZU-F21). Third-party verification rests on 2022-2023 IAS announcements covering viewability and IVT only; no verifier coverage was found for web games or for click-enabled and custom formats (GAM-ANZU-F19). The SDK documentation portal is login-gated, so format specifications, counting logic and age-mode behaviour were read from partner and blog sources instead (GAM-ANZU-F35). Title-level reporting as standard for buyers is not documented (GAM-ANZU-F26). All outcome evidence is vendor-published or commissioned; none is independent (GAM-ANZU-F28, GAM-ANZU-F29, GAM-ANZU-F30).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | D | Formats offered: blended display, blended video and custom ads (developer page), each clickable or non-clickable, with click-enabled variants sold to performance DSPs since February 2026 (GAM-ANZU-F04). Specifications: Playgama's integration guide for Anzu's SDK sets placement aspect ratios (1.55 landscape, 0.67 portrait), image and video allowance and clickable, click-and-hold, double-click, UI and sprite placement types (GAM-ANZU-F14). Counting approach: Anzu's placement guide says billable impressions are recorded only when viewability criteria are met (at least 1.5% of screen, 50% visible, 1 second for display and 2 for video, angle under 55 degrees) (GAM-ANZU-F15). Creative guidelines: video of 6 seconds or less, sound off, logo and CTA in the first frames, headlines of seven words or less (GAM-ANZU-F16). Custom ads (e.g. UFL referee kits, scoreboards) have no published specifications, and no counting rule for clicks on click-enabled units was found. [614, 122, 615, 358, 616, 132] |
| Targeting and context | live | D | Listed signals documented: (1) named title or title list: allow and block lists by publisher, app or game, country and platform for open- and private-auction campaigns, and a pre-approved allow list for IO and PMP buys (brand-safety policy); (2) genre: Anzu's CPO says in-game targeting was 'largely limited to contextual targeting based on genre or title' until September 2025, and its targeting guide describes game-genre targeting; (3) contextual segments without personal data: the guide describes keyword and site-list audiences translated into game bundle IDs, and the advertiser page claims 'Advanced contextual targeting with no reliance on personal data' (GAM-ANZU-F17). Content or age rating is not buyer-selectable: Anzu itself uses 'reasonable efforts' to suit ads to each game's PEGI rating (GAM-ANZU-F22). In-game or session context is not documented. Personal-data targeting (62,400 third-party segments; age, gender and hashed emails in Behavioral mode) is recorded but earns nothing here. [397, 617, 618, 123, 445] |
| Measurement functionality | live | D | Listed functions described: viewability (Anzu's own 'patented viewability tech' and IAS viewability, announced 2022 and 2023), brand lift (Anzu 'invests in brand lift studies'; Comscore named), attention and incremental reach (named as existing partnerships in the Claritas release), and outcome measurement (Claritas attribution and incremental lift, announced 27 March 2025, beta with one US advertiser in December 2024, no method or result published) (GAM-ANZU-F18, GAM-ANZU-F19). Aligned counting: Anzu calls its formats 'IAB-compliant' and uses the IIG 2.0 viewable-impression thresholds for billing, but its own guides misstate the guidelines' served-impression rule (GAM-ANZU-F15; stream finding GAM-C-F19, verified). The media-value page still names MOAT, which Oracle retired on 30 September 2024 (GAM-ANZU-F18). [123, 364, 394, 358, 350, 551, 339] |
| Optimisation and outcomes | live | D | Outcome objectives: since February 2026 performance DSPs can run 'UA, retargeting, and lower-funnel campaigns' on click-enabled intrinsic units (release only). In-flight optimisation is described only as marketing ('more stable delivery, greater control'). Results with a stated method, all vendor-published or commissioned: 55 campaigns scored blind against brand lift studies with more than 15,000 respondents (June 2024); Happydemics' 110 campaigns compared with its database of over 600 advertisers (March 2024); dentsu's lab study of 200 gamers with a 250-person online control cell (January 2024) (GAM-ANZU-F28, GAM-ANZU-F29, GAM-ANZU-F30). Performance claims (21% lower CPA than advertisers' goals; CTR up 13% for video and 5% for display 'based on Anzu proprietary data') are user-acquisition economics for the mobile paper and are not used here (GAM-ANZU-F31). [122, 474, 579, 481] |
| Creative and insertion tools | live | C | Dynamic insertion is documented: Playgama's integration guide for Anzu's SDK has an 'Is Dynamic' setting ('Enable if the ad content should change dynamically'), and Anzu's placement guide describes dynamic channels that change the creative automatically after an impression and a cool-off period (GAM-ANZU-F14, GAM-ANZU-F16). Asset tooling is described only in marketing ('proprietary rendering functions ... dynamically match any advertisement to its environment'); creative adaptation appears as a managed service in case studies. No AI-assisted placement or generation with controls was found. [615, 358, 123, 616] |
| Supply | live | D | Six named live titles in the scoring period, all named by Anzu: UFL (PC), Gangsta Island: Crime City (web), Asphalt 8: Airborne and Dragon Mania Legends (Gameloft), Fashion Battle and Drill & Collect (APPS) (GAM-ANZU-F10). The aggregate claim ('Hundreds of live cross-platform games') has no number, metric or period, and no ad-eligible audience figure was found (GAM-ANZU-F11). The sellers.json count of 1,117 publisher accounts is a supply connection, not titles or scale, and includes non-game properties (GAM-ANZU-F07, GAM-ANZU-F08). Exclusivity is claimed but not documented at publisher or placement level (GAM-ANZU-F27). Overlapping supply is flagged (see shared_supply_flags). [132, 133, 134, 135, 123, 9] |
| Coverage | live | D | Three evidenced tuples in the scoring period: mobile / app store / premium title (Asphalt 8: Airborne); mobile / app store / casual title (Fashion Battle, Drill & Collect, Dragon Mania Legends); PC / platform launcher (inferred) / premium title (UFL). Not counted: web / casual (Gangsta Island: Crime City) because no device is stated; console, because no console title is named in the period; UGC, because the Roblox claim dates from 2022 (GAM-ANZU-F10, GAM-ANZU-F12, GAM-ANZU-F13, GAM-ANZU-F14). [134, 135, 132, 471, 133, 619] |
| Programmatic access | live | D | Named DSP integrations: Anzu names The Trade Desk, StackAdapt, Basis, Smadex and Adikteev as 'already onboard' (9 February 2026). Two are confirmed by the DSP: StackAdapt's own in-game page (published 6 January 2025, modified 27 August 2026) and a Smadex-authored partner post on Business of Apps (5 February 2026) (GAM-ANZU-F06). Deal types: open auction (OMP), private auction with a Deal ID (PMP) and insertion orders (brand-safety policy); Anzu also sells through its own SSP under Marketplace Terms (GAM-ANZU-F05). The '40+ programmatic partners including all major DSPs' claim is unverified; stream B found no cohort A network among DV360's named exchanges (GAM-B-F15, verified). Protocol states are recorded separately and earn no score. [122, 120, 175, 397, 171, 123] |
| US availability and coverage | live | D | Available to US buyers: a US campaign for SUPERPRETZEL in Philadelphia and Pittsburgh across Trackmania, Asphalt Unite, Cooking Fever and Slapshot Rebound (case study published 25 March 2025); a US geo-targeted campaign for QuikTrip (2023); a February 2026 release datelined New York; NBCUniversal named US and UK sales partner in March 2022; an October 2025 US gamer audience report with Semcasting; end-user data hosted on US servers; a CCPA notice (GAM-ANZU-F25, GAM-ANZU-F23). US coverage is not quantified for the scoring period: the four US campaign titles predate it, and the only US figure is one campaign's delivery reach (3 million unique players across 80 games), not coverage (GAM-ANZU-F11). The contact page lists no US office. [163, 617, 122, 164, 620, 445, 621, 123] |
| Independent assurance | live | D | (a) MRC: Anzu is on neither MRC's accredited digital services list nor its digital services-under-review list, read on MRC's own site on 29 September 2026; the only gaming entry under review is Trivver (GAM-ANZU-F20). (b) Third-party verification: IAS announced IVT and viewability measurement on Anzu's mobile inventory (28 July 2022); Anzu announced the joint solution across mobile, PC and console, display and video (11 July 2023). Neither names a suitability function. No verifier document covers the web environment (from November 2025) or the click-enabled and custom formats, and Anzu's current media-value page names Comscore, MOAT (retired), Tag and Human, not IAS (GAM-ANZU-F19). (c) Certification: Anzu announced TAG Brand Safety Certified and Certified Against Fraud seals in August 2021, but TAG's registry returned no Anzu record on 29 September 2026 (control query for Gadsme returned its record). Its sellers.json carries a TAG-ID, which is a registration identifier, not a seal (GAM-ANZU-F21). Compliance claims ('IAB-compliant', 'GDPR, CCPA, and COPPA compliant', ISO 27001) are vendor statements with no certifier record opened. [15, 16, 350, 551, 364, 383, 382, 9] |
| Transparency | live | B | Supply-path disclosure is published and was observed directly: sellers.json v1.0 with 1,176 entries, a seller_type for every entry, no confidential entries and a TAG-ID (GAM-ANZU-F07). The Marketplace Terms require publishers to adopt ads.txt v1.1 and list Anzu as a certified seller (GAM-ANZU-F26). Stream B matched DIRECT IDs from both sides for seven publishers (GAM-B-F11, verified). Title-level transparency is partial: open-auction bid requests carry the game and country, and advertisers receive live screenshots, but advertiser reports come 'in such form, content, and frequency ... as to be mutually communicated', and title-level reporting as standard is not documented (GAM-ANZU-F26). The sellers.json mixes non-game properties with game publishers (GAM-ANZU-F08). [9, 171, 397, 363] |
| Suitability, age and data controls | live | C | Brand-suitability controls, operational detail: the April 2023 policy, still linked live, sets advertiser allow and block lists by publisher, app or game, country and platform for OMP and PMP; a pre-approved allow list for IO and PMP; a category block list; a manual human check of all direct creatives, targeted within 24 hours; real-time creative scanning for OMP and PMP; PEGI-rating suitability on a 'reasonable efforts' basis; and takedown within one Israeli business day (GAM-ANZU-F22). Age controls, outline: the publisher can choose a 'Kids Safe-Mode' category in which end users 'shall be treated as children' and data are limited to a truncated IP, country or state location, preference strings and interaction. Anzu says it does not knowingly collect information from children under 16. Its advertiser terms put responsibility on the advertiser to ensure games are not intended for an ineligible audience, and its SSP terms require publisher properties not to be directed to minors. How Safe-Mode restricts ad categories, and any per-user age gating, are not documented (GAM-ANZU-F23, GAM-ANZU-F24). Data practices, operational detail: the December 2025 privacy policy lists data collected per mode (Contextual, Behavioral, Kids Safe-Mode), purposes, legal bases, IP truncation, US-hosted servers and retention criteria (GAM-ANZU-F23). No regulator or court finding against Anzu was found in the searches logged. [397, 445, 363, 171] |
Performance evidence: In dentsu's Play Attention lab study, Anzu intrinsic ads recorded 3,442 attentive seconds per 1,000 impressions against 1,416 for online display, with 99% on-screen visibility (Lumen page); Anzu's own blog gives 3,422. Gaming recall across environments was 21%, which dentsu calls highly variable and lower than its norms. (commissioned measurement, sponsor dentsu (agency) with Lumen; Anzu, Twitch and Activision Blizzard collaborated as measured sellers; funding split not stated) [384, 481]; Anzu's account of the dentsu study reports a significant correlation between average viewed time and prompted brand recall (R² = 94%). (vendor assertion, sponsor Anzu (reporting a dentsu study)) [481]; Across 110 intrinsic in-game brand campaigns, Happydemics reports purchase consideration 5 percentage points above the online media average, ad recall 13 points and brand identification 11 points above the digital average. (commissioned measurement, sponsor Anzu released the study with Happydemics, a survey vendor; funding not stated) [474]; Across 55 Anzu campaigns scored blind for adherence to creative guidelines and matched to brand lift studies with over 15,000 respondents, guideline-following video saw a 21% lift in brand recommendation and a 5% lift in purchase intent, and display 8% and 9%. (vendor assertion, sponsor Anzu) [579]; A December 2024 beta with a major US advertiser found that intrinsic in-game advertising 'generated an incremental lift in conversions', measured by Claritas. (vendor assertion, sponsor Anzu and Claritas (partners)) [394]; SUPERPRETZEL's US campaign reported +29 points in brand interest, +22 in brand image, +23 in purchase intent and an 81% video completion rate. (vendor assertion, sponsor Anzu (case study)) [163]; Performance claims: intrinsic ads delivered a 21% lower CPA than advertisers' goals on average, and clickable intrinsic CTRs exceeded online benchmarks by 13% (video) and 5% (display), 'based on Anzu proprietary data'. (vendor assertion, sponsor Anzu) [122]; Supply-side results: Gameloft reports a 10% ads-ARPDAU uplift, 5.8x intrinsic ad revenue growth and 3x impressions per DAU; APPS reports eCPM up 47% and ARPDAU up 50% on Fashion Battle after clicks were enabled (four months before vs eleven months after). (vendor assertion, sponsor Anzu (developer case study and blog quoting the developers)) [134, 135]
### DoubleVerify (active; acquisition pending)
*DoubleVerify Holdings, Inc. (Delaware corporation, SEC CIK 0001819928); its operating subsidiary DoubleVerify, Inc. is the addressee of MRC's accreditation letter*
**SWOT (2026-09-29).** DoubleVerify is a listed verification vendor with a current MRC accreditation and TAG certifications, but its gaming coverage is documented in three places (Roblox, Microsoft Casual Games, Zynga), none of its accreditations names gaming, and Nielsen's pending acquisition leaves ownership unsettled.
- **Strengths.** MRC's letter of 13 April 2026 continues accreditation of DoubleVerify's Pinnacle platform, covering impressions on desktop, mobile web, mobile app and CTV, viewability, Authentic Attention and pre-bid. (record) [18] TAG's registry lists DoubleVerify as Certified Against Fraud and Brand Safety Certified through an independent third-party auditor, globally, and Certified For Transparency in North America and Europe. (record) [581] DoubleVerify says it has extended fraud/IVT and viewability measurement to Roblox in-game billboard image and video ads through a custom integration on Roblox's direct-sold inventory. (vendor claim) [351] DoubleVerify reported FY2025 revenue of $748.3 million, 90% from advertisers, charged as a Measured Transaction Fee per thousand impressions. (record) [582]
- **Weaknesses.** None of MRC's accreditation letter, MRC's accredited list or DoubleVerify's 10-K, 10-Q and merger proxy mentions gaming; MRC's standards exclude intrinsic in-game ads. (record) [18, 15, 582, 583, 584] DoubleVerify's site lists no gaming channel or gaming product page; its Roblox Rewarded Video functions were framed prospectively in the 1 April 2025 post. (vendor documentation) [375, 351] Roblox's Managed Campaign terms make Roblox's reporting the controlling measurement; an advertiser may raise a verifier's IVT discrepancy above 10% through invoice disputes. (record) [1] Q2 2026 revenue grew 3% to $193.8 million, and Activation revenue fell 1% on lower programmatic revenue. (record) [583]
- **Opportunities.** Zynga launched Zynga Ads+ on 17 September 2026 with inventory scored by DoubleVerify, reading as a route into mobile game supply beyond Roblox. (inference) [183, 203] Its 15 October 2025 Microsoft release names Microsoft Casual Games; placement and devices are unstated, so a scoped gaming product is still unclaimed. (inference) [374] Its MRC mobile in-app accreditation, including OM SDK, plausibly covers standard display and video in mobile game apps, though no record attests it. (inference) [18, 15]
- **Threats.** Nielsen's parent agreed 6 August 2026 to acquire DoubleVerify for $13.60 a share; close is expected by Q1 2027; no HSR outcome was public by 29 September. (record) [380, 584] Roblox named both DoubleVerify and Nielsen as measurement partners on 1 April 2025; if the deal closes, two of its measurement partners share an owner. (inference) [187, 381] No DoubleVerify record was found for any intrinsic in-game network screened, such as Anzu, Frameplay or Gadsme, where rendered 3D formats sit outside its accreditation. (inference) [375, 398]
- **What is sold.** Software that measures and verifies digital advertising: post-bid measurement of fraud and invalid traffic, viewability, brand safety and suitability, geography and attention ('Measurement'), the same signals applied before purchase through demand-side platforms and ad platforms ('Activation'), and verification and analytics sold to publishers and platforms ('Supply-side') (GAM-DOUBLEVERIFY-F02). In games DoubleVerify sells no inventory, and no DoubleVerify gaming product page was found (GAM-DOUBLEVERIFY-F25). Its documented gaming activity consists of: its standard measurement applied to in-game ads from 27 July 2023, which it said needed no significant tool update (F07); a custom integration for IVT and viewability measurement of Roblox in-game billboard image and video ads on Roblox's direct-sold inventory, plus tag-based Rewarded Video coverage announced in the future tense (1 April 2025) (F09); approval on Twitch for invalid traffic, viewability and domain- or app-level brand safety (F16, game-adjacent); and a measurement-partner role for Zynga, whose Zynga Ads+ network was announced on 17 September 2026 (F17).
- **Who buys it.** Advertisers and agencies generate most revenue (90% in 2025; over 2,500 customers, 131 of them at $1 million or more of annual revenue); publishers and platforms buy supply-side products (GAM-DOUBLEVERIFY-F02). In gaming, buyers are advertisers placing DV tags on Roblox campaigns (F09, F10) and advertisers who wrap tags on networks that allow it, such as Gadsme (F19); a publisher, Zynga, is reported as using DoubleVerify as its measurement partner for Zynga Ads+ (F17). Whether Zynga pays DoubleVerify as a supply-side customer is not stated in the sources opened.
- **How it earns.** Advertisers pay a Measured Transaction Fee per thousand impressions measured; most prices are fixed. Programmatic activation is bought through DSPs, which collect the fee and remit it; DoubleVerify books the gross amount and records the DSP's share as cost of revenue. Supply-side customers pay under monthly or annual contracts with minimum guarantees and tiered pricing (GAM-DOUBLEVERIFY-F02). Revenue from gaming environments is not disclosed: the FY2025 10-K, the Q2 2026 10-Q and the merger proxy contain no occurrence of 'game', 'gaming' or 'console' (F04).
- **Supply and data dependencies.** DoubleVerify's reach in any environment depends on the integration the platform or publisher permits. On Roblox, billboard measurement runs through a custom integration on direct-sold inventory (F09); Roblox's managed-campaign terms make Roblox's own reporting 'the controlling measurement for all campaigns' and let an advertiser raise an invalid-traffic discrepancy from its 'approved third-party verification partner' only above 10% (F11). In mobile apps its accredited viewability measurement includes traffic measured through IAB Tech Lab's OM SDK (F12). Pre-bid products reach buyers through DSP integrations (Amazon, Google, The Trade Desk, Yahoo, Nexxen, Infillion, Adobe, StackAdapt named in the 10-K) (F14). Data collected by its solutions include IP address, user agent, app attributes (including the app's age rating) and ad exposure and engagement events, with personal information kept no longer than 45 days (F21). Gaming coverage outside Roblox, Twitch and Zynga rests on partner mentions (F18, F19, F20).
- **Differentiation.** A documented custom integration on a gaming platform's in-world billboards (Roblox, direct-sold inventory), beyond tag-based coverage (F09). MRC accreditation of Authentic Attention metrics in desktop, mobile web and mobile in-app, the attention product it offers on Roblox Rewarded Video and that Zynga reports (F12, F09, F17); the accreditation does not extend to game environments. Named by Roblox for independent verification, approved by Amazon on Twitch, and named by Zynga as measurement partner (F10, F16, F17). TAG Certified Against Fraud and Brand Safety Certified by independent validation for global operations (F15).
- **Risks.** Ownership change: the pending Nielsen acquisition would place DoubleVerify under the same owner as Nielsen, which Roblox also names as a measurement partner and which DoubleVerify's own 10-K lists among its primary competitors; whether buyers treat the combined company as independent of Roblox's other measurement partners is an open question (F05, F06, F28). Deal risk: closing depends on a stockholder vote and antitrust approvals in the US, Australia, Cyprus, Germany and Morocco, with an outside date of 6 May 2027, extendable to 6 August 2027 (F05). Platform control: Roblox's terms make Roblox's reporting controlling and limit verifier IVT disputes to discrepancies above 10%, so DoubleVerify's figures on Roblox are advisory for billing purposes (F11). Scope drift in marketing: DoubleVerify's Roblox post claims coverage of 'all of Roblox's advertising inventory globally' across mobile, desktop and console, while its format-level detail and Roblox's documents support a narrower scope (F09, F10). Gaming apps as a fraud vector: DoubleVerify's own research describes gaming apps used to fake impressions, which bears on buyers of standard in-app inventory in games (F23).
- **Best fit.** Advertisers buying Roblox direct-sold billboard campaigns who want an independent IVT and viewability read alongside Roblox's controlling reporting (F09, F11). Advertisers buying standard display and video in mobile game apps through DSPs, where DoubleVerify's accredited mobile in-app measurement and app-level pre-bid controls apply to the app environment (inference; the record does not name games) (F12, F14, F27). Game-adjacent buys on Twitch where IVT, viewability and domain/app-level brand safety are sufficient (F16).
- **Poor fit.** Buyers who need MRC-accredited measurement of intrinsic in-game placements: no verifier's accreditation, including DoubleVerify's, covers them (F12, F13). Console in-game campaigns: no documented DoubleVerify console integration beyond its unconfirmed Roblox device claim (F09). Buyers who need content-level suitability inside UGC experiences or livestreams: DoubleVerify documents only app/domain-level safety on Twitch and a future-tense Rewarded Video announcement on Roblox (F09, F16). Age-based controls: DoubleVerify documents none (F21).
- **Limits.** DoubleVerify's MRC accreditation covers desktop, mobile web, mobile in-app and CTV, with no gaming, console or intrinsic environment and no Intrinsic In-Game guideline among the standards applied (GAM-DOUBLEVERIFY-F12, F13). The only gaming integration DoubleVerify documents as built is Roblox's billboard IVT and viewability measurement on direct-sold inventory; the Roblox Rewarded Video suitability, attention and pre-bid functions were announced in the future tense and no later go-live record was found (F09, F10). No DoubleVerify record was found for intrinsic in-game networks (Anzu, Frameplay, AdInMo, Adverty, IION, Gameloft for brands, Super League) in the searches logged; Gadsme's tag-wrapping statement and XBOX Media Solutions' and Overwolf's partner lists are the counterparties' own words, without scope from DoubleVerify (F18, F19, F20, F26). No age controls are documented by DoubleVerify; age eligibility in the environments it verifies is set by the platforms (F21, F10). No public method describes how DoubleVerify counts an impression or viewable impression for a rendered 3D in-world billboard (F12, F26). Filings do not disclose gaming revenue or volume, and DoubleVerify's device split of measured transactions (mobile, desktop, CTV) has no console category (F04).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | D | DoubleVerify sells no ad format; it measures formats others sell. The gaming formats it names are Roblox in-game billboard image and video ads and Rewarded Video (F09). Its counting approach is documented by the accreditor only for display and video impressions and viewability in desktop, mobile web, mobile in-app (inclusive of OM SDK traffic) and CTV, under the IAB and MRC display, video and mobile in-app guidelines; the MRC letter does not list the Intrinsic In-Game Advertising Measurement Guidelines 2.0 and contains no reference to games (F12, F13). No public DoubleVerify document describes how an impression or viewable impression is counted for a rendered in-world billboard. Creative guidelines are outside a verifier's offer. [351, 18, 15] |
| Targeting and context | live | D | Inside in-game placements DoubleVerify documents no buyer-selectable targeting by genre, title, session context or content rating; for Roblox Rewarded Video it announced fraud, viewability and brand-suitability pre-bid solutions in the future tense (F09). For standard programmatic inventory, MRC's accredited pre-bid list includes App Store Categories (mobile app, CTV), Site and App Inclusion/Exclusion (desktop web, mobile web, mobile app, CTV), IAB contextual categories and client custom categories, activated through the buyer's DSP (F14). Its solutions collect each app's store, developer, star rating and age rating to apply customer profile settings (F21). Whether game apps sit inside these app-level controls is inference: no DoubleVerify document names games. [351, 18, 622] |
| Measurement functionality | live | D | Roblox billboards: fraud/IVT and viewability through a custom integration on direct-sold inventory (vendor statement, present perfect). Roblox Rewarded Video: fraud, brand safety and suitability, viewability, geo-relevance and DV Authentic Attention through DV tags, announced in the future tense (F09). Zynga: a DoubleVerify Attention Index is reported for more than 250 million impressions measured April to June 2026 (F17). Twitch (game-adjacent): IVT, viewability and domain/app-level brand safety, per Amazon (F16). Accreditation: MRC accredits DoubleVerify's viewability and Authentic Attention metrics only in desktop, mobile web and mobile in-app (impressions also CTV), with no gaming environment and no IIG alignment (F12, F13). Attention definitions belong to the attention paper. [351, 203, 266, 18, 15] |
| Optimisation and outcomes | unconfirmed | D | DoubleVerify's filings describe pre-bid avoidance through DSPs and Scibids AI, which builds bidding models toward client KPIs from DSP impression-level data (F14, F02); none is documented for a gaming environment. On Roblox, pre-bid fraud, viewability and brand-suitability solutions for Rewarded Video were announced in the future tense (F09). No DoubleVerify outcome study for in-game advertising was found; Gadsme's page attributes to 'IAS and DoubleVerify' a claim that in-game campaigns outperform traditional advertising, with no method or DoubleVerify source (F19). [582, 351, 118] |
| Creative and insertion tools | unconfirmed | E | No record: DoubleVerify does not describe dynamic insertion, asset tooling or AI-assisted placement for games, and creative tooling is outside what a verifier sells. No source documents absence in the anchor's sense. |
| Supply | live | A | DoubleVerify holds no inventory and no selling rights; it earns fees on ads it measures (F02). Environments where its measurement is documented or reported in games: Roblox (UGC platform; vendor and platform statements), Zynga's own mobile games (attention figure reported for April to June 2026), and Twitch (game-adjacent). Audience or title counts in those sources belong to the sellers and are not DoubleVerify reach (F09, F10, F16, F17). [582, 351, 203] |
| Coverage | live | D | Distinct tuples with DoubleVerify measurement documented or reported: mobile x app store x UGC platform (Roblox; vendor and platform statements) and mobile x app store x casual title (Zynga's own games; inference from a reported attention figure). PC x platform launcher x UGC platform and console x app store x UGC platform rest only on DoubleVerify's statement that its Roblox coverage spans desktop and console, which no Roblox document opened confirms (see coverage_tuples). No intrinsic in-game network, premium console title or cloud distribution is documented (F26). [351, 42, 623, 203] |
| Programmatic access | live | A | DoubleVerify's pre-bid products are activated through DSPs; its 10-K names Amazon, Google, The Trade Desk, Yahoo, Nexxen, Infillion, Adobe and StackAdapt as DSP integration partners, and MRC's letter says advertisers activate segments through their DSP's interface (F14). None of these integrations is documented for a gaming environment; for Roblox, DoubleVerify's Rewarded Video pre-bid was announced in the future tense and no DSP is named for it, while Roblox's own footnote on programmatic pre-bid names only IAS (F09, F10). Deal types are not applicable to a verifier. Protocol states: OM SDK documented for mobile in-app measurement; no AdCP agent registered (protocol_states). [582, 18, 351, 187, 121] |
| US availability and coverage | live | D | DoubleVerify is US-headquartered, its customer base is 'predominately U.S.-based' and its accreditation is by MRC, a US body (F02, F12). Gaming coverage in the US is not quantified: DoubleVerify's Roblox coverage is described as global without a US figure (F09); Zynga Ads+ was announced in New York (F17); its Pangle measurement covers the US but the release does not mention games (F24). [582, 351, 183, 624] |
| Independent assurance | live | A | (a) MRC, read from MRC's own record: continuing accreditation (letter of 13 April 2026) of DoubleVerify's Pinnacle platform for display and video impressions (desktop, mobile web, mobile in-app, CTV, with SIVT filtration), viewability (desktop, mobile web, mobile in-app, inclusive of OM SDK traffic; CTV for certified apps), property-level ad verification, Authentic Attention (desktop, mobile web, mobile in-app) and pre-bid signals; initial accreditation of further Authentic Attention reporting and TikTok viewability. No game, gaming, console or intrinsic environment appears in the letter or on MRC's accredited list, the standards applied do not include the Intrinsic In-Game Guidelines 2.0, and MRC's services-under-review list shows no DoubleVerify service (the only gaming entry is Trivver) (F12, F13). (b) Third-party verification of DoubleVerify's own environments does not apply to a verifier. (c) TAG registry: Certified Against Fraud and Brand Safety Certified by independent validation for global operations, and Certified for Transparency by TAG validation for North America and Europe; company-level seals, not in-game measurement (F15). For scored rows: DoubleVerify's accreditation does not extend to in-game environments, so a row's credit for DoubleVerify verification rests on vendor or platform announcements (F09, F10, F27). [18, 15, 16, 581] |
| Transparency | live | C | MRC accredits DoubleVerify's property-level ad verification, with domain and mobile-app-level classification, in desktop, mobile web, mobile in-app and CTV (F12). Its privacy notice says standard reporting is anonymised and aggregated and that impression- or URL-level reporting is available on request, generally de-identified (F21). For games, no document shows title-, experience- or placement-level reporting from DoubleVerify; on Roblox, the managed-campaign terms give advertisers aggregate reporting only and make Roblox's reporting controlling (F11). DoubleVerify is not a seller, so sellers.json and app-ads.txt disclosure do not apply; it holds TAG's Certified for Transparency seal for North America and Europe (F15). [18, 622, 1, 581] |
| Suitability, age and data controls | live | C | Brand-suitability controls: documented in operational detail for web, app and CTV environments (over 100 avoidance categories; pre-bid brand safety and suitability and property-level verification accredited by MRC in desktop, mobile web, mobile in-app and CTV) (F12, F14, F22); for games only in outline, as a future-tense announcement of suitability measurement and pre-bid controls for Roblox Rewarded Video, with none stated for Roblox billboards, and domain/app-level (not stream-level) brand safety on Twitch (F09, F16). No DoubleVerify classification of in-game content, UGC experiences or game titles is documented. Age controls: none documented by DoubleVerify; it does not process user age, collects an app's age rating as a context attribute, and age eligibility on Roblox is set by Roblox (F21, F10). Data practices: documented in operational detail in the Solutions Privacy Notice (effective 9 October 2025): categories collected (IP address, user agent, app attributes including age rating, viewability and engagement events), purposes, precise geolocation never collected, personal information kept no longer than 45 days, fraud tables kept indefinitely, solutions not directed at children and no knowing collection of children's personal information (F21). Documentation is not legal compliance. [622, 625, 18, 351, 266] |
Performance evidence: Across more than 250 million impressions measured April to June 2026, Zynga delivered an average Attention Index of 204, which DoubleVerify classifies as a very high performer (scores above about 133). (vendor assertion, sponsor Zynga (seller of the inventory), in an IAB event listing; the metric is DoubleVerify's proprietary index) [203]; Gadsme states: 'According to IAS and DoubleVerify, in-game campaigns consistently outperform traditional advertising'. (vendor assertion, sponsor Gadsme (network naming the verifiers); no IAS or DoubleVerify source cited) [118]; The DV Fraud Lab says it uncovered the 'SkyWalk' scheme, in which iOS gaming apps load hidden gaming websites to generate fake impressions, and nearly 9,000 fraudulent mobile apps in 2025. (vendor assertion, sponsor DoubleVerify) [626]
### Frameplay (active (operating; no exit, insolvency or retirement record found))
*Frameplay Corporation (per its Terms of Use and Technology Privacy Policy and the IAB Europe Global Vendor List; state registry record not opened)*
**SWOT (2026-09-29).** Frameplay is a US intrinsic in-game network and exchange whose own supply is partly confirmed in publisher files, whose resold supply overlaps other networks, and whose measurement is unaccredited.
- **Strengths.** The Frameplay Exchange opened on 9 October 2024 for programmatic and direct buys aimed at US buyers, reselling Adverty and AdInMo inventory as listed intermediaries. (record) [74, 10] Publisher app-ads.txt files read on 29 September 2026 show matching DIRECT authorisations for at least five of Frameplay's 30 publisher accounts, including Miniclip and Helium9 Games. (record) [75, 96, 585, 586, 587] The SDK supports Unity, Unreal and Cocos2dx and documents a single-sided billboard with seven image ratios, four of them video-capable. (vendor documentation) [466, 124, 465] Frameplay and IAS both document IAS invalid-traffic measurement and impression counts for Frameplay campaigns; neither names viewability or brand-suitability verification. (vendor claim) [369, 420]
- **Weaknesses.** Frameplay is on neither MRC list as of 29 September 2026, and counts impressions with its own method; no source states alignment with IAB in-game guidelines. (record) [15, 16, 355] Its about page claims TAG Certified for Transparency, but a TAG registry search returned no Frameplay entry. (record) [357, 387] The IAB Europe vendor list records Frameplay's TCF registration as deleted on 24 March 2025, while its privacy policy still cites vendor ID 970. (record) [555, 406] Frameplay supplied nearly 90 of the 173 campaigns in the October 2025 Happydemics benchmark, so beating that benchmark is largely a comparison with itself. (inference) [475]
- **Opportunities.** The June 2025 iion announcement said iion's 4,000+ titles will be accessible through the Exchange in the US; a go-live would widen resold supply. (inference) [73] Trivver is the only gaming service under MRC review at 29 September 2026, so an audit building on the IAS integration would set Frameplay apart. (inference) [16, 369] The 9 June 2026 NexTide partnership, a Scott's Lawn Care livestream campaign reaching 3.3 million views, extends selling into game-adjacent streaming. (inference) [136]
- **Threats.** Much of the Exchange's supply belongs to Adverty, AdInMo and iion, which buyers can reach directly, and the 5,000+ games figure cannot be reconciled with supply records. (inference) [10, 136, 73] Frameplay's own releases thin after mid-2025 (newest Unity SDK 16 October 2024, newest resources post 4 June 2025), which buyers may read as reduced investment. (inference) [465, 169, 588] Without a live TCF registration since March 2025, consent-based buying of EEA supply depends on developer CMPs that may not list Frameplay. (inference) [555, 554]
- **What is sold.** Intrinsic in-game brand placements: billboard-style "Ad Spaces" rendered inside the 3D environments of third-party games that integrate Frameplay's SDK, carrying static image or video creative in 7 image and 4 video aspect ratios [GAM-FRAMEPLAY-F09]. They are sold direct and through the Frameplay Exchange, which is open to programmatic and direct buyers [GAM-FRAMEPLAY-F03]. The June 2026 boilerplate also lists rewarded experiences, sponsorships and live streaming [GAM-FRAMEPLAY-F07]. The livestream offer with NexTide renders on streamers' channels, not inside games, so it belongs to game-adjacent media (cohort B2) and is excluded from this row [GAM-FRAMEPLAY-F30].
- **Who buys it.** Brands and agencies, primarily US media buyers according to the Exchange's launch reporting [GAM-FRAMEPLAY-F03, GAM-FRAMEPLAY-F28]. Named buyers in the sources are all before the scoring period or unnamed: dentsu for McCormick's Frank's RedHot (2022) [GAM-FRAMEPLAY-F19], an unnamed snack brand (2021) [GAM-FRAMEPLAY-F24], an international power-sports company (Kochava case) [GAM-FRAMEPLAY-F20], and unnamed brands in the case studies [GAM-FRAMEPLAY-F07]. The only named brand inside the period is Scott's Lawn Care, in a livestream campaign outside this row [GAM-FRAMEPLAY-F30]. GumGum appears as a reseller in an undated testimonial, with no documented integration [GAM-FRAMEPLAY-F26].
- **How it earns.** Advertisers pay per impression. Frameplay's developer guide calls this the CPI (cost per impression) and ties the CPI to impression quality and to targeting data such as player age and sex [GAM-FRAMEPLAY-F10, GAM-FRAMEPLAY-F12]. Game developers who host Ad Spaces earn a share of that revenue (developer revenue guide and developer page) [GAM-FRAMEPLAY-S10, GAM-FRAMEPLAY-S03]. The revenue split, pricing and company revenue are not public. The company is private, and no public filing or financial statement was found [GAM-FRAMEPLAY-F33].
- **Supply and data dependencies.** Supply depends on three things. First, developers integrating the SDK (Unity, Unreal, Cocos2dx), with production ads loading only after account and game approval [GAM-FRAMEPLAY-F27]. Second, publisher authorisations in app-ads.txt [GAM-FRAMEPLAY-F06]. Third, other networks' inventory resold through the Exchange: AdInMo, Adverty and iion are intermediaries in Frameplay's sellers.json. AdInMo's operating company has a suspended voluntary strike-off recorded in workstream U (GAM-U-F22) [GAM-FRAMEPLAY-F05, GAM-FRAMEPLAY-F31]. Measurement depends on partners: IAS for IVT and impression counts, Happydemics and earlier Comscore for brand lift, and Kochava for attribution [GAM-FRAMEPLAY-F18, GAM-FRAMEPLAY-F20, GAM-FRAMEPLAY-F21, GAM-FRAMEPLAY-F24]. Data depends on developers obtaining consent (a CMP in the EEA, ATT on iOS), on the device advertising ID, on developer-passed age and sex, and on third-party data providers [GAM-FRAMEPLAY-F12, GAM-FRAMEPLAY-F15].
- **Differentiation.** Runs an exchange that aggregates competing intrinsic networks' inventory (Adverty and AdInMo from October 2024; iion announced 2025), rather than selling only its own SDK supply [GAM-FRAMEPLAY-F03, GAM-FRAMEPLAY-F04]. Documents an impression-quality model in its developer docs: a 60-second aggregation weighted by on-screen share, orientation, skew and obstruction. It has marketed a proprietary time-in-view metric since 2022 [GAM-FRAMEPLAY-F10, GAM-FRAMEPLAY-F19]. US-headquartered and positioned toward US media buyers, while most cohort A peers are based outside the US [GAM-FRAMEPLAY-F01, GAM-FRAMEPLAY-F28]. Has a large volume of brand lift studies with Happydemics, nearly 90 since 2021, though these are vendor-released normative comparisons [GAM-FRAMEPLAY-F21].
- **Risks.** Public activity has thinned since mid-2025: no resources post after 4 June 2025, no Unity SDK release after 16 October 2024, sellers.json unchanged since August 2025, and the TCF registration deleted in March 2025. These are monitoring signals, not evidence of an exit [GAM-FRAMEPLAY-F32, GAM-FRAMEPLAY-F14]. Dependence on partner networks for exchange supply, including AdInMo, whose operating company has a suspended strike-off (GAM-U-F22), and iion, whose US go-live is undocumented [GAM-FRAMEPLAY-F04, GAM-FRAMEPLAY-F05]. Scale claims that cannot be reconciled with supply records (5,000+ games against 30 publisher accounts and two confirmed direct authorisations) invite buyer scepticism [GAM-FRAMEPLAY-F07, GAM-FRAMEPLAY-F08, GAM-FRAMEPLAY-F06]. Out-of-date compliance and assurance statements: a TCF participation claim after the GVL deletion, and a TAG certification claim with no registry record [GAM-FRAMEPLAY-F14, GAM-FRAMEPLAY-F17]. Signs of a widening offer beyond in-game (livestream with NexTide) could dilute the intrinsic network focus. This is an inference from one release [GAM-FRAMEPLAY-F30].
- **Best fit.** US brand campaigns that want intrinsic billboard exposure in casual mobile games through one buying point, and that accept network-level rather than title-level planning [GAM-FRAMEPLAY-F03, GAM-FRAMEPLAY-F06, GAM-FRAMEPLAY-F07]. Upper-funnel objectives measured by vendor-arranged brand lift studies (Happydemics) or attribution feeds, where buyers accept normative benchmarks [GAM-FRAMEPLAY-F20, GAM-FRAMEPLAY-F21]. Buyers who already buy Adverty or AdInMo inventory and want it in the same deal as Frameplay's own supply, provided reach is deduplicated [GAM-FRAMEPLAY-F03, GAM-FRAMEPLAY-F31].
- **Poor fit.** Buyers who require MRC-accredited or IIG-aligned measurement, or third-party viewability and suitability verification inside placements [GAM-FRAMEPLAY-F10, GAM-FRAMEPLAY-F16, GAM-FRAMEPLAY-F18]. Plans that need named titles, US audience figures or exclusivity [GAM-FRAMEPLAY-F07]. Console or premium PC campaigns: no live console title was found [GAM-FRAMEPLAY-F27, GAM-FRAMEPLAY-F03]. Campaigns directed at children or needing documented age-based delivery limits: the controls are in outline only, and the services are not directed to children [GAM-FRAMEPLAY-F11, GAM-FRAMEPLAY-F13]. EEA personalised campaigns relying on TCF consent, given the GVL deletion [GAM-FRAMEPLAY-F14].
- **Limits.** No named live title and no ad-eligible audience figure with metric and period for the scoring period; supply is described only by an undefined "5,000+ mobile games" and "billions of impressions" [GAM-FRAMEPLAY-F07]. Impression counting and viewability are Frameplay's own definitions; no source states alignment with the IIG Guidelines 2.0, and no MRC accreditation or audit exists [GAM-FRAMEPLAY-F10, GAM-FRAMEPLAY-F16]. Third-party verification is documented only by a 2023 IAS announcement covering IVT and impression counts; no viewability or suitability verification is named [GAM-FRAMEPLAY-F18]. Brand-suitability and age controls are documented only in outline, in 2023 Terms describing a self-serve dashboard; how they apply to Exchange or DSP buys is not stated [GAM-FRAMEPLAY-F11, GAM-FRAMEPLAY-F29]. Programmatic rails are thinly documented for the period: one DSP (DV360) is named in developer docs but absent from DV360's exchange list; the multi-DSP list dates from 2021 [GAM-FRAMEPLAY-F25, GAM-FRAMEPLAY-F26]. A large, unquantified share of the reachable inventory may belong to other networks (Adverty, AdInMo, iion) [GAM-FRAMEPLAY-F04, GAM-FRAMEPLAY-F08, GAM-FRAMEPLAY-F31].
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | C | Specifications: the SDK documentation defines the intrinsic Ad Space as a single-sided virtual billboard with 7 image aspect ratios (8:1, 6:1, 16:9, 4:3, 9:16, 2:3, 1:1), 4 of which can also carry video, and static or dynamic content types (dynamic Ad Spaces load a new ad every 60 seconds) [GAM-FRAMEPLAY-F09]. Counting approach: the developer revenue guide says view metrics are captured for up to 60 seconds per displayed ad and aggregated into one impression, weighted by on-screen time, screen share, ad-space share, orientation, skew and obstruction; viewability follows Frameplay's own proprietary requirements, and no alignment with the IIG Guidelines 2.0 is stated [GAM-FRAMEPLAY-F10]. Creative guidelines for brand creative: none found; the best-practice pages address developers' placement choices, not advertisers' assets. The other formats in the June 2026 boilerplate (rewarded experiences, sponsorships, live streaming) have no specifications in the sources opened [GAM-FRAMEPLAY-F07]. Breadth attribute: static image and video intrinsic billboards. [466, 173, 124, 355, 356, 136] |
| Targeting and context | live | C | Frameplay's Terms of Use (updated 1 January 2023, live at access) describe an advertiser dashboard where buyers name the genre and gaming platforms to target and choose audience by location, language, gender, age range and content restrictions; publishers register each game's title, genre, rating and content restrictions [GAM-FRAMEPLAY-F11]. Listed contextual signals documented: (1) genre (also the genre list on the advertiser page); (2) content or age rating, through buyer-set content restrictions matched to publisher-set game rating and restrictions. Not documented for buyers: named title or title lists, in-game or session context, contextual segments without personal data. Personal-data targeting is documented and earns no points: developers are asked to pass player age and sex, the SDK uses the advertising ID on opt-in, and the Technology Privacy Policy allows third-party data to target ads in games [GAM-FRAMEPLAY-F12]. Whether Exchange and DSP buyers get the dashboard's controls is not stated. [406, 137, 355, 466] |
| Measurement functionality | live | D | Network-cohort anchors applied. Listed functions documented: (1) viewability and attention through Frameplay's own Intrinsic Time-in-View metric (launched April 2022; the home page claims a 33-second network average without date or method) [GAM-FRAMEPLAY-F19]; (2) brand lift through Happydemics studies (nearly 90 Frameplay studies, 40 in 2024) and earlier Comscore surveys [GAM-FRAMEPLAY-F21, GAM-FRAMEPLAY-F24]; (3) outcome attribution through Kochava's Atlas Performance, with the impression feed documented, on an undated partner page that predates October 2024 [GAM-FRAMEPLAY-F20]. IAS provides IVT measurement and impression counts, which are not listed functions [GAM-FRAMEPLAY-F18]. The SDK collects camera-based session viewability data [GAM-FRAMEPLAY-F13]. Aligned counting: not documented; Frameplay's impression is its own up-to-60-second aggregate and its viewability test is proprietary; its role in drafting the IIG guidelines is not a statement that the product counts to them [GAM-FRAMEPLAY-F10]. [356, 552, 475, 479, 553, 369, 406, 355, 357] |
| Optimisation and outcomes | live | D | In-flight optimisation: Kochava's case study says attribution signals let Frameplay optimise toward the highest-performing placements and creatives (partner page, undated, before October 2024) [GAM-FRAMEPLAY-F20]. Result with a stated method: the October 2025 release reports a Happydemics benchmark of brand lift surveys (173 in-game campaigns since 2021; comparator: Happydemics format benchmarks), without sample sizes, a control design or uncertainty, and with about half of the benchmark being Frameplay's own campaigns [GAM-FRAMEPLAY-F21, GAM-FRAMEPLAY-F22]. Buyer-settable outcome objectives: not documented beyond the Terms' reference to "defining your goals" when creating an ad [GAM-FRAMEPLAY-F11]. [553, 475, 406] |
| Creative and insertion tools | live | C | Dynamic insertion is documented: when an Ad Space activates, an ad is requested from Frameplay's servers and downloaded at runtime, and dynamic Ad Spaces load a new ad every 60 seconds, so creative changes without a game update [GAM-FRAMEPLAY-F09]. Asset tooling (templates, builders, 2D or 3D conversion): not documented. AI-assisted placement or generation with controls: not documented. Developers set placeholders and approve brands, which are publisher controls, not creative tools [GAM-FRAMEPLAY-F29]. [124, 173, 464] |
| Supply | live | D | Network-cohort anchors applied. Named live titles in the scoring period (2 October 2025 to 2 October 2026): none found. The advertiser and case-study pages name no titles, and the latest title Frameplay names in the sources opened is Basketball Battle (April 2022) [GAM-FRAMEPLAY-F07, GAM-FRAMEPLAY-F19]. Aggregate claims: "a network of 5,000+ mobile games" in Frameplay's boilerplate within a NexTide release of 9 June 2026 (metric undefined, no period, not stated as active) and "Billions of impressions available" (undated; impressions are not audience) [GAM-FRAMEPLAY-F07]. Supply records: sellers.json lists 30 publisher accounts on 27 domains and 3 intermediaries (AdInMo, Adverty, iion) [GAM-FRAMEPLAY-F05]; only Miniclip and Helium9 Games show matching DIRECT authorisations among six sampled Frameplay publisher accounts [GAM-FRAMEPLAY-F06]. Ad-eligible audience with metric and period: not published. Selling rights and exclusivity: not documented in the period. Shared supply: Adverty, AdInMo and (by announcement) iion inventory reached through the Exchange is those networks' supply, and the 5,000+ figure plausibly includes it (inference) [GAM-FRAMEPLAY-F04, GAM-FRAMEPLAY-F08, GAM-FRAMEPLAY-F31]. [136, 137, 627, 10, 75, 96, 73] |
| Coverage | live | B | One tuple evidenced: mobile x app store x casual title. Evidence: publisher-declared DIRECT authorisations of frameplay.gg by Miniclip and Helium9 Games matching Frameplay's sellers.json, observed 29 September 2026, plus the claim of 5,000+ mobile games (June 2026) [GAM-FRAMEPLAY-F06, GAM-FRAMEPLAY-F07]. The casual-title environment is inferred from the publishers' portfolios, not stated by a source. Not counted: PC (Digiday reported in October 2024 that most Exchange inventory was in free-to-play mobile or PC games; outside the period and not a named integration) [GAM-FRAMEPLAY-F03]; console (developer-page platform label and an experimental PS5 build from November 2023; no live title) [GAM-FRAMEPLAY-F27]. See coverage_tuples. [10, 75, 96, 136, 74, 464, 465] |
| Programmatic access | live | C | Programmatic access exists through the Frameplay Exchange, opened 9 October 2024 for programmatic purchasing and direct buys (Digiday) [GAM-FRAMEPLAY-F03]. Named DSP in vendor documentation: Google DV360, named in the SDK guide (updated 18 July 2024) as demand for video-capable ad spaces [GAM-FRAMEPLAY-F26]; DV360's own exchange enum (updated 10 June 2026) does not list Frameplay, so the buying route is undocumented [GAM-FRAMEPLAY-F26]. The only multi-partner list is from December 2021: Magnite, Adelphic, Yahoo, InMobi and The Trade Desk (three DSPs), with private marketplaces and direct deals; nothing opened confirms these are current [GAM-FRAMEPLAY-F25]. GumGum, named in an undated testimonial, carries no Frameplay entry in its sellers.json [GAM-FRAMEPLAY-F26]. Deal types documented in the period: none beyond "programmatic" and "direct" (2024). Protocol states are recorded in protocol_states; no OpenRTB version is documented. [74, 173, 119, 177, 137, 174] |
| US availability and coverage | live | B | Available to US buyers: Frameplay is headquartered in San Francisco; Digiday (October 2024) describes the Exchange as a primary point of exchange for media buyers in the United States; the June 2025 iion release places the combined supply in the Frameplay Exchange in the US; Adverty's US test ran through Frameplay [GAM-FRAMEPLAY-F01, GAM-FRAMEPLAY-F28, GAM-FRAMEPLAY-F31]. US coverage quantified: no. No US-available titles are named and no US audience figure with metric and period exists [GAM-FRAMEPLAY-F07]. The only US brand campaign dated inside the scoring period (Scott's Lawn Care, June 2026) is livestream advertising outside this row [GAM-FRAMEPLAY-F30]. International breadth (APAC via iion; worldwide offices claimed) earns no points. [357, 74, 73, 628, 136] |
| Independent assurance | unconfirmed | D | (a) MRC: Frameplay is on neither MRC's accredited digital list nor its digital services-under-review list (MRC's own record, read 29 September 2026); MRC's environment key has no gaming environment [GAM-FRAMEPLAY-F16]. (b) Third-party verification: Frameplay announced on 5 October 2023 that IAS provides IVT measurement and independent impression counts for its intrinsic campaigns "across all environments" (unnamed); no viewability or suitability verification is named, and no source inside the scoring period confirms the integration is live [GAM-FRAMEPLAY-F18]. (c) Certification: the about page claims TAG Certified for Transparency (2023), but TAG's registry returned no Frameplay record on 29 September 2026; no documented denial or lapse was found [GAM-FRAMEPLAY-F17]. Guideline role: a claimed contribution to the IIG guidelines is not assurance [GAM-FRAMEPLAY-F10]. [15, 16, 369, 357, 387] |
| Transparency | live | B | Supply-path disclosure is published: Frameplay's sellers.json (version 1.0, TAG-ID, 30 PUBLISHER and 3 INTERMEDIARY entries, last modified 19 August 2025) separates publishers from intermediaries, observed directly on 29 September 2026 [GAM-FRAMEPLAY-F05]. Quality: of six sampled Frameplay publisher accounts, only two have matching DIRECT authorisations, and the file still lists AdInMo [GAM-FRAMEPLAY-F05, GAM-FRAMEPLAY-F06]. Title- or property-level reporting to buyers is not documented as standard. Developers get game-level dashboard and API reporting, and the IAS release promises "daily reporting" without saying at what level [GAM-FRAMEPLAY-F18, GAM-FRAMEPLAY-F29]. [10, 75, 96, 78, 76, 77, 464, 369] |
| Suitability, age and data controls | live | C | Brand suitability, in outline: the Terms let advertisers set content restrictions, and publishers set each game's rating and content restrictions, with the publisher solely responsible [GAM-FRAMEPLAY-F11]; the developer page gives developers control of ad formats and brand approvals, and the SDK guide collects each game's ESRB rating and "White Listed Advertisement References" [GAM-FRAMEPLAY-F29]. How the restrictions filter delivery, and their scope, are not stated. No pre-bid or post-bid suitability verification is documented: the IAS partnership names IVT only [GAM-FRAMEPLAY-F18]. Age controls, in outline: games and ads carry a COPPA-compliance flag and buyers can pick an age range [GAM-FRAMEPLAY-F11]; the Technology Privacy Policy says the services are not directed to children and Frameplay does not knowingly collect children's data [GAM-FRAMEPLAY-F13]. How COPPA-flagged games are served, and whether developer-passed player age limits personalisation, are not stated [GAM-FRAMEPLAY-F12]. Data practices, in operational detail: the Technology Privacy Policy lists the device, session and user data the SDK collects, its purposes, recipients and a purpose-bound retention principle [GAM-FRAMEPLAY-F13]; the SDK guide lists the data collected, consent requirements (a CMP in the EEA, ATT, AD_ID) and Apple privacy-manifest declarations [GAM-FRAMEPLAY-F15]. Tension: the policy still claims TCF participation as vendor 970, which IAB Europe's register records as deleted on 24 March 2025 [GAM-FRAMEPLAY-F14]. [406, 464, 554, 369, 555, 466] |
Performance evidence: Happydemics in-game benchmark (173 campaigns since 2021, nearly 90 Frameplay): in-game ad recall 32%, attribution 52%; Frameplay 54% attribution and likeability, 24% consideration (2 points above the in-game benchmark) (vendor assertion, sponsor Frameplay (release issuer); Happydemics sells the studies) [475]; Happydemics meta-analysis of 29 Frameplay studies: attribution 55% vs online video 39%; likeability 52% vs 50%; consideration +21 vs +17 points (vendor assertion, sponsor Frameplay (release issuer)) [478]; Comscore single-campaign brand lift: +31.8 points mobile ad recall, +26.2 points purchase intent (unnamed snack brand) (vendor assertion, sponsor Comscore (release issuer); both parties commercially interested) [479]; Intrinsic Time-in-View validated against eye tracking by Lumen and eye square; Lumen reported 1.4x the attention of its mobile display norm (vendor assertion, sponsor Frameplay (release issuer); study funding unknown) [356]; Kochava attribution case: 2.7 million impressions and over 60,000 attributed web engagements for a power-sports brand; optimisation toward top placements and creatives (vendor assertion, sponsor Kochava (partner selling the attribution product)) [553]; Home page network statistics: 33-second average Intrinsic Time-in-View ("33x the minimum standard"), 1:18 viewable time per session, 69% would consider or were likely to purchase (vendor assertion, sponsor Frameplay) [552]; NexTide livestream campaign for Scott's Lawn Care: 3.3 million views; engagement rates said to be 4x standard display in the first 32 days (vendor assertion, sponsor NexTide Media (release issuer)) [136]
### Gadsme (active)
*GADSME (Societe par actions simplifiee), SIREN 884197799, per the French company register; Gadsme's Publisher General Terms also name 'Gadsme FRANCE SAS' and 'Gadsme Ltd' as contracting entities*
**SWOT (2026-09-29).** Gadsme is a small private Paris in-game ad network with one publisher-confirmed premium PC exclusive and checkable supply files, but no independent accreditation and marketing that its own contracts qualify.
- **Strengths.** Sports Interactive announced a three-year exclusive on 29 September 2025, and its partners page still names Gadsme its exclusive in-game ads partner, starting with Football Manager 26. (vendor documentation) [140, 589] On 29 September 2026 its sellers.json listed 226 entries, including a Sports Interactive Games publisher account, and its app-ads.txt held 898 lines across 166 ad systems. (vendor documentation) [12, 97] It holds US patent 12,211,141 B2, issued 28 January 2025, for calculating object visibility in a 3D scene; the record documents a method, not its use in counting. (record) [590] Its February 2026 brand-safety policy requires a PEGI and environment review for each new game, checks placements before live and commits to takedown within 48 hours. (vendor documentation) [398]
- **Weaknesses.** On 29 September 2026 Gadsme was on neither MRC digital list, and the TAG registry showed it enrolled in Verified by TAG with no seal. (record) [15, 16, 392] Marketing says '100% exclusive' and 'No multi-SSP exposure', but its terms limit exclusivity to 3D and audio ads, and its CEO named Magnite, Xandr and Equativ routes. (vendor documentation) [95, 59, 591] Its advertisers page reports 35% ad recall, IVT below 0.5% and 98% viewability without naming a provider, sample, period or method. (vendor claim) [118, 592] Developer documentation sits behind a sign-in, no public creative specifications were found, and the register put 2023 headcount at 6 to 9 salaried employees. (record) [593, 594, 186]
- **Opportunities.** ARROVA, a Hakuhodo DY ONE subsidiary, began selling Gadsme in Japan on 13 May 2026; this reads as a route into a new market through an agency-group seller. (inference) [160] Football Manager 27 is announced for November 2026 and the Sports Interactive exclusive runs three years from September 2025, which reads as a second premium PC season. (inference) [595, 140] Its SDK was extended in October 2024 to Unreal, Godot, Cocos2d and custom C++ including console, yet no console title is named; console supply remains open. (inference) [596, 40]
- **Threats.** Sixty of its 222 publisher domains also appear in Anzu's sellers.json on 29 September 2026, so buyers can reach part of the same supply through a competitor. (inference) [12, 9] King's and Zynga's app-ads.txt authorise Gadsme only as a reseller under a Digital Turbine seller ID, which reads as supply dependence on an intermediary. (inference) [53, 77, 12] Google's DV360 API v4 exchange list of 10 June 2026 has no Gadsme value, which limits direct selection by DV360 buyers. (inference) [119]
- **What is sold.** The Gadsme Managed Ad Network: Gadsme buys ad space inside third-party games that run its SDK and fills it with ads from its buyer partners, sold direct and programmatically through private marketplace deals, curated deals and direct programmatic partnerships (GAM-GADSME-F03, GAM-GADSME-F17). Its contract catalogue lists display and video ads in menus, cutscenes and 2D and 3D game scenes, plus audio ads; click-enabled units are marketed (GAM-GADSME-F04). Audio ads belong to the audio paper, and the SSP and Cross Promo services are separate listed rows (GAM-GADSME-F27).
- **Who buys it.** Brands and agencies, direct or through DSPs and private marketplaces (GAM-GADSME-F17). Buyers named in opened records: Easy Cash through L'Agence Bradford (France, Q4 2025, GAM-GADSME-F10) and Make-A-Wish International (December 2025, GAM-GADSME-F11). Resellers: ARROVA, a Hakuhodo DY ONE subsidiary, in Japan from 13 May 2026 (GAM-GADSME-F15), and adWMG for console-focused advertisers (GAM-GADSME-F14). Brand logos on Gadsme's pages are not treated as evidence of buying. Game studios are the supply-side counterparties.
- **How it earns.** Under the Publisher General Terms (updated 19 May 2026), Gadsme purchases the publisher's ad space, bills and collects from buyers, and pays the publisher a fee set in each term sheet out of 'Transaction Revenues' (amounts received less agency commissions, DSP and exchange fees, invalid-traffic deductions, serving and measurement fees, refunds and taxes); revenues are stated in USD (GAM-GADSME-F03). Cross Promo is priced per install or as a fixed fee and is not scored. No revenue figure is published (GAM-GADSME-F29).
- **Supply and data dependencies.** Supply depends on game studios integrating the Gadsme SDK. Exclusivity is contractual, per player or game session, and covers only 3D and audio ads; a publisher may A/B test a competing 3D network provided Gadsme runs for at least 50% of players (GAM-GADSME-F06). Gadsme's sellers.json declares three intermediaries (Digital Turbine, LoopMe, Unity), and its app-ads.txt, which Zynga and Helium9 Games reference through inventorypartnerdomain, authorises 166 ad-system domains, mostly as RESELLER (GAM-GADSME-F07 to F09). The SDK collects device advertising IDs, IP address, location and ad-view history for targeting, segmenting and frequency capping (GAM-GADSME-F24).
- **Differentiation.** A publisher-confirmed, three-year exclusive on a premium PC franchise (Football Manager 26, PC/Mac), which the other cohort A networks screened do not document for a named PC title in the scoring period (GAM-GADSME-F10; comparison is bounded to streams B and G). A published contract (Publisher General Terms) that defines the managed network, its counting rules, its revenue waterfall and the scope of exclusivity, which is more specific than most network marketing pages (GAM-GADSME-F03, GAM-GADSME-F04, GAM-GADSME-F06). Click-enabled in-game units and an SDK for non-Unity engines, reported by AdExchanger with a named publisher (GAM-GADSME-F26, GAM-GADSME-F31).
- **Risks.** Exclusivity messaging ('100% exclusive', 'No reseller stacking', 'No multi-SSP exposure') conflicts with Gadsme's own supply-path files and contract terms; a buyer relying on the claim for supply-path decisions may be misled (GAM-GADSME-F06 to F09). Thin independent assurance: no MRC record, no TAG seal, and an IAS integration last documented in 2022 without PC or console scope (GAM-GADSME-F18 to F20). Children's audiences: casual mobile titles are part of the supply, but no age control on delivery is documented and the handling of the COPPA flag is not stated (GAM-GADSME-F24). Concentration of the PC story in one franchise: the FM26 exclusive runs three years from September 2025; renewal is not public (GAM-GADSME-F10).
- **Best fit.** Brands wanting in-match stadium or cutscene billboards in Football Manager 26 on PC, bought through Gadsme (inference from GAM-GADSME-F10). European programmatic buyers using PMPs or curated deals for intrinsic mobile placements, where IAS or DoubleVerify tag wrapping is acceptable as buyer-side verification (inference from GAM-GADSME-F05, GAM-GADSME-F17, GAM-GADSME-F23).
- **Poor fit.** US buyers who need quantified US reach, named US titles or MRC-accredited in-game measurement before committing budget (GAM-GADSME-F16, GAM-GADSME-F19). Buyers requiring console in-game placements with named titles (GAM-GADSME-F14). Buyers whose supply-path policy requires a single direct path without reseller authorisations (GAM-GADSME-F08, GAM-GADSME-F09).
- **Limits.** Buyer-facing creative specifications and developer documentation are not public: docs.gadsme.com redirects to a sign-in page, and the Terms of Use treat documentation as confidential (GAM-GADSME-F05). Format and creative cells are scored on contract terms and marketing pages only. Only one named live title inside the scoring period (Football Manager 26), and its live status rests on vendor-involved content, not independent observation (GAM-GADSME-F10). No ad-eligible audience figure with metric and period is published; the figures found are title counts, campaign delivery and a campaign reach forecast (GAM-GADSME-F11, GAM-GADSME-F12). No MRC accreditation or audit, and no TAG seal (GAM-GADSME-F19, GAM-GADSME-F20); the only verifier record is a 2022 IAS release for mobile and web (GAM-GADSME-F18). The sellers.json and app-ads.txt files cover all Gadsme services (managed network, SSP, audio); the managed network's own supply path cannot be separated from them (GAM-GADSME-F07, GAM-GADSME-F09). Console coverage is claimed but not evidenced by any named console title (GAM-GADSME-F14).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | C | Formats named: display, video, in-menu placements and audio integrations (programmatic page); the contract catalogue is 'non-disruptive display and video ads to be positioned in menu, cut scenes, 2D and 3D scenes of games and audio ads'. Counting approach documented in the Publisher General Terms: display counted when displayed and visible, video at video start, audio at audio start. Partial behaviour specifications: video sound off at launch, formats auto-refresh. No public creative specifications (sizes, durations, file weights for brand formats) or creative guidelines were found: developer documentation redirects to a sign-in page and the Terms of Use treat documentation as confidential. The only file-weight limits published apply to Cross Promo creatives, which are not scored. Audio is scored in the audio paper. [59, 95, 594, 593] |
| Targeting and context | live | C | Listed signals documented: (1) in-game context: brand-safety 'discriminators' made of 'persona behaviors and/or in game scenes', negatively targeted by default on managed campaigns, with 'Specific discriminators' added on client request; (2) title or title list: advertiser exclusion lists are honoured on all campaigns, and the advertisers page offers 'Curated games' and 'Themed game' routes (labels only). Not counted: the contextual categories in the Publisher Terms (e.g. 'news, lifestyle, business, sports, or women') are Gadsme's own classification for optimising monetisation, not documented as buyer-selectable; PEGI is checked at game onboarding, not offered as a buyer setting. Personal-data targeting (advertising IDs used for 'segmenting users') is recorded and earns no points. [398, 118, 59, 444] |
| Measurement functionality | live | D | Viewability and IVT through IAS Signal were announced by IAS on 15 December 2022 for Gadsme inventory ('include mobile and web browsing campaigns'); the IAS page was modified on 15 May 2025 and remains live, but no 2025-2026 record confirms the integration is running. Native reporting: 'Track impressions, exposure time and campaign performance in real time' (advertisers page) and 'Viewability and exposure metrics tied to gameplay' (publishers page, publisher-facing); the terms' SSP appendix counts display impressions by 'the proprietary viewability method designed by Gadsme', which is not described. Brand lift is claimed as survey results on Gadsme inventory without provider or method. No claim of counting aligned to the Intrinsic In-Game Guidelines 2.0 was found in the pages opened; the documented counting (display when displayed and visible, video at start) is not IIG counting. Attention is not documented as a defined metric (exposure time is recorded, and definitions belong to the attention paper). [372, 373, 118, 531, 59, 556] |
| Optimisation and outcomes | live | D | Marketing states 'Campaigns are activated and optimized in real time to maximize performance' and 'Click-enabled formats support full-funnel objectives with transparent reporting and real-time optimisation'; the advertisers page offers 'Control over pricing, pacing, placements and frequency'. Results are reported without method: Easy Cash reach of about 560,000 people and 0.42% CTR (France, Q3-Q4 2025, in content produced with Gadsme and the agency) and brand-lift averages of 35% ad recall and +25 points on brand image and consideration. No outcome objective setting or in-flight optimisation is documented in vendor documentation. [118, 170, 142] |
| Creative and insertion tools | live | C | Dynamic insertion is documented: the SDK 'Tag' calls and displays ads, 'Gadsme Format shall auto refresh' (Publisher Terms), and 'Ad Content Pooling' selects content from several demand sources 'in real time' (publishers page). Asset handling: the advertisers page says 'Standard IAS creatives are rendered directly within gameplay environments' (wording as published) and the publishers page says placements follow 'lighting, perspective, and environment constraints'; no buyer-facing templates, builders or 2D-to-3D conversion tool is documented. No AI-assisted placement or generation was found in the pages opened. [59, 531, 118, 596] |
| Supply | live | D | Named live titles inside the scoring period: one, Football Manager 26 (PC/Mac), with publisher-confirmed three-year exclusivity (29 September 2025) and a vendor-involved report of a live programmatic activation (11 December 2025). Aggregate counts without metric period: 'more than 1,600 mobile titles' (October 2025, reported from Gadsme), 'thousands of games' and '200+ game studios' (advertisers page), 'more than 100' developers (ARROVA boilerplate, May 2026). Publisher deals with Zynga and Konami name no titles. No ad-eligible audience figure with metric and period was found. Selling rights are documented at contract level (session-level exclusivity for 3D and audio ads) and for FM26 by its publisher. 223 PUBLISHER seller accounts are supply connections, not titles. [140, 142, 139, 118, 160, 12, 59] |
| Coverage | live | D | Two evidenced tuples: PC x platform launcher x premium title (Football Manager 26 on Steam) and mobile x app store x casual title (December 2025 campaigns across mobile games from Miniclip, Lion Studios, Voodoo and Ubisoft; titles not named, distribution inferred). Not evidenced: console (marketing claims mobile, PC and console; adWMG claims Xbox and PlayStation titles without naming any, and Gadsme's own press list describes that deal as mobile inventory for console-focused advertisers); web (IAS 2022 mentions web browsing campaigns, outside the scoring period); mobile premium titles (Zynga and Konami titles unnamed). [140, 141, 142, 472, 40, 467] |
| Programmatic access | live | C | Deal types named in vendor documentation: 'Private Marketplace deals (PMPs)', 'Curated deals' and 'Direct programmatic partnerships' (programmatic page); 'Deal-ID activation' (publishers page). One DSP integration is confirmed by the DSP's own page: Eskimi (page updated 6 August 2025, original content from 2022), and Gadsme's app-ads.txt carries eskimi.com lines. 'Gadsme own DSP' and 'Directly connected to the largest programmatic DSPs' are claims without named DSPs. Google's DV360 API v4 Exchange enum (last updated 10 June 2026) names no Gadsme exchange. Protocol states are recorded separately; no AdCP state was found. [95, 531, 176, 97, 119, 118] |
| US availability and coverage | live | D | Gadsme opened a US office in New York with a VP US (reported 4 December 2024 and restated in October 2025). No named US campaign, US title list or US audience figure was found in the sources opened. The Privacy Policy includes California (CCPA) rights, which concern data subjects, not US sales. International breadth (France, Japan via ARROVA) is recorded as an attribute only. [159, 139, 444] |
| Independent assurance | live | D | (a) MRC: Gadsme is on neither MRC's Accredited Services: Digital list nor its Services Under Review: Digital list (29 September 2026); the only gaming entry under review is Trivver. (b) Third-party verification: IAS announced global viewability and IVT verification of Gadsme inventory on 15 December 2022, covering 'mobile and web browsing campaigns'; no IAS or Gadsme record covers PC or console inventory such as FM26, and no 2025-2026 confirmation that it is running was found. Advertisers may wrap tags with IAS or DoubleVerify (brand-safety policy, February 2026), which is buyer-side tagging, not a verifier's record of in-game coverage. (c) Certification: Gadsme is enrolled in Verified by TAG with no seal listed. Gadsme's claims of COPPA compliance and of audits 'with independent market consulting verification firms' are vendor claims with no named certifier. [15, 16, 372, 373, 392, 398, 118] |
| Transparency | live | C | Supply-path disclosure is published: sellers.json (226 entries with seller types, TAG-ID and DUNS) and app-ads.txt (898 lines). The disclosure is incomplete and in tension with marketing: three INTERMEDIARY sellers and 837 RESELLER authorisation lines sit beside 'No reseller stacking' and 'No multi-SSP exposure'; no SEGA or Sports Interactive account appears in sellers.json although FM26 is Gadsme's flagship PC title; one Zynga DIRECT ID is assigned to an unrelated developer. Title- or placement-level reporting to buyers is not documented as standard: the brand-safety page says 'All routes are transparent and reported in real time through the Gadsme Console', and 'Placement-level performance insights' is a publisher-side feature. [12, 97, 77, 398, 531, 170, 95] |
| Suitability, age and data controls | live | C | Brand suitability, operational detail: Brand Safety & Takedown Policy (updated February 2026) states a review of PEGI, game type and environments for every new game, a check of all placements before a game goes live, default negative targeting of brand-safety discriminators on managed campaigns (8 stated, four listed), advertiser exclusion lists honoured on all campaigns, and takedown within 48 hours of written notice. Age controls, not documented: no age-gating or age-based limit on delivery, targeting or personalisation was found; the Privacy Policy lists a 'COPPA flag' among app data collected but does not say how flagged traffic is handled, and the publishers page states 'Ads are all GDPR, LDPR, CCPA, and COPPA compliant' (a compliance claim, not a control). Data practices, operational detail: the Privacy Policy (24 February 2026) states what is collected (IP, IDFA, Google Advertising ID, location, ad-view history, app data), uses (targeting, segmenting, frequency capping), retention (user data up to 12 months, IP in a security log for 4 months, cookie up to 365 days), opt-out routes and CCPA rights, and that Gadsme does not knowingly collect personal information from children under 13. Documentation is not legal compliance. [398, 444, 531] |
Performance evidence: 'Brand lift surveys conducted on Gadsme inventory shows an average 35% in ad recall, and uplift of +25 pts on both brand image and consideration'. (vendor assertion, sponsor unknown (vendor page names no survey provider or funder)) [118]; Easy Cash campaign (France, Q3-Q4 2025) including a Football Manager 2026 PC activation: reach of about 560,000 people and average CTR of 0.42%, 'top' of Gadsme's campaigns in France for Q3 and Q4 on performance. (vendor assertion, sponsor Content produced in collaboration with L'Agence Bradford and Gadsme (Minted's own label)) [142]; Make-A-Wish campaign in December 2025 'delivering over 5M impressions every single day'; page metadata says it 'will reach more than 300 million mobile gamers'. (vendor assertion, sponsor unknown (announcement by the campaign parties, published by Mobidictum)) [472]; 'Only less than 0.5% of IVT is measured across Gadsme portfolio'. (vendor assertion, sponsor Gadsme) [118]; 'According to IAS and DoubleVerify, in-game campaigns consistently outperform traditional advertising'. (vendor assertion, sponsor Gadsme (attribution to IAS and DoubleVerify without a cited study)) [118]; 'Publishers typically achieve 8% - 20% ARPDAU uplift without impacting any other monetization setup or player experience'. (vendor assertion, sponsor Gadsme) [531]
### Roblox (active)
*Roblox Corporation (advertiser contracts for managed campaigns are with Roblox Advertising Services, LLC)*
**SWOT (2026-09-29).** Roblox sells in-game brand ads on a platform with 123 million average daily active users in Q2 2026 and documents its formats and age rules, but it discloses no advertising revenue, holds no MRC or TAG record and treats its own reporting as the controlling measurement.
- **Strengths.** Roblox reported 123 million average DAUs in Q2 2026, about 82% outside the US and Canada; DAUs count accounts, not people, and are not an ad-eligible audience. (record) [246] Its Creator Hub documentation defines a counting rule for each format, from autoplay billboard video (0.5 seconds looked at) to Rewarded Video (first frame counted). (vendor documentation) [208, 207] Its Advertising Standards (updated 5 June 2026) bar personalised ads to under-18s, limit rewarded video to users 13 and over, and ban categories for under-13 audiences. (vendor documentation) [20] Roblox says over 1,000 brands had run Rewarded Video by 5 February 2026 and over 1,200 games had used it by 20 May 2026. (vendor claim) [201, 209]
- **Weaknesses.** Roblox does not disclose advertising revenue; its filings call it insignificant and its Q4 2025 letter says it currently remains modest. (record) [197, 246, 201] On 29 September 2026 neither MRC digital list had a Roblox entry, and a TAG registry search returned no Roblox record. (record) [15, 16, 570] Its managed-campaign terms make Roblox's own reporting the controlling measurement, give advertisers aggregate reports only, and allow IVT disputes only above a 10% discrepancy. (vendor documentation) [1] On 29 September 2026 its brand-safety page still said under-13s cannot see ads, and its Terms and Creator Hub gave different integration revenue-share start dates. (vendor documentation) [217, 57, 190]
- **Opportunities.** Magnite (6 January 2026) and PubMatic (April 2024) describe programmatic access to Roblox inventory in future tense; if delivered, buyers gain routes beyond managed campaigns. (inference) [211, 189] A Join with Reward test (20 May 2026) and Studio plug-in expected in Q4 2026 could widen Rewarded Video supply, which Roblox says will take time to scale. (inference) [209, 508, 188] Since 4 May 2026 under-13 users can see contextual billboard ads sold directly or through SuperAwesome, which opens an age band closed to most programmatic sellers. (inference) [215, 216, 206] Roblox added EDO and Ipsos as measurement partners in June 2026; a published EDO result could add outcome evidence, independent only with disclosed funding and design. (inference) [486, 488]
- **Threats.** Its Q2 2026 10-Q records youth-safety settlements with five states and says its age checks have misclassified users, which bears on every age-gated ad rule. (inference) [246] Nielsen agreed on 6 August 2026 to acquire DoubleVerify, a named Roblox verification partner; the deal was still pending on 29 September 2026. (inference) [380, 597] Magnite and Index Exchange sellers.json files list Roblox as reached through Google's exchange bidding, a declared supply connection through Google; no delivery through that route was observed. (inference) [539, 573]
- **What is sold.** Two in-experience brand products that Roblox serves into ad units creators place in their games (GAM-ROBLOX-F06, F07): Immersive Ads, meaning billboards (video up to 30 seconds, click-to-play or autoplay, and static image ads) and portal ads that teleport users into an advertiser's experience; and Rewarded Video, a full-screen 6 to 30 second opt-in video that grants an in-game reward and is shown only to users 13 and over (GAM-ROBLOX-F11). Managed campaigns are sold by insertion order under Roblox Advertising Services, LLC terms (GAM-ROBLOX-F17); programmatic access is marketed through Google (Ad Manager and Display & Video 360) and named SSPs and DSPs (GAM-ROBLOX-F14, F15). Roblox also sells placements outside gameplay (Homepage Hero, Homepage Feature in beta, Home and Search tiles) and paid discovery (Sponsored Experiences); these are outside this scored product (GAM-ROBLOX-F27). Developer-placed Advertising Integrations are a separate, creator-sold route that Roblox regulates and plans to charge for (GAM-ROBLOX-F21, F22).
- **Who buys it.** Brands and agencies, through Roblox's managed service or programmatically (GAM-ROBLOX-F14). Roblox says over 1,000 brands have run Rewarded Video (vendor claim, GAM-ROBLOX-F10). US examples in the record include a Sam's Club rewarded video campaign described by a Roblox executive at IAB PlayFronts in September 2026 and H&R Block's 2025 paid-media campaign (GAM-ROBLOX-F29). Campaigns aimed at under-13 audiences are sold by direct deal only, with SuperAwesome as exclusive partner (GAM-ROBLOX-F12). Resellers sell the product in named markets outside the US: Venatus (UK, France, Germany) and Playwire (select APAC markets) (GAM-ROBLOX-F23).
- **How it earns.** Roblox sells the inventory and shares ad revenue with the publishing creators in Robux: per impression for image and autoplay video ads, per 15-second view for click-to-play video, per teleport for portals, and per impression (EPM) for Rewarded Video (GAM-ROBLOX-F07). Roblox does not disclose advertising revenue; its filings call advertising and licensing revenue 'insignificant' and its Q4 2025 letter calls advertising revenue 'modest' (GAM-ROBLOX-F02). Separately, Roblox plans to charge developers a revenue share on Advertising Integrations they sell directly, with conflicting start dates and bases in its own documents (GAM-ROBLOX-F22).
- **Supply and data dependencies.** Supply depends on creators inserting ad units into experiences that pass Roblox's publisher gates (public game, ID-verified owner with 2FA, approved Maturity and Compliance Questionnaire, at least 2,000 unique monthly visitors) and on users being eligible by age, region, device and campaign settings (GAM-ROBLOX-F08, F11). Age eligibility rests on Roblox's own age-check data, which covered 57% of global DAUs at the end of Q2 2026 (GAM-ROBLOX-F05). Programmatic demand reaches Roblox through Google (Ad Manager and Open Bidding) and SSP connections declared in app-ads.txt and SSP sellers.json files (GAM-ROBLOX-F16). Advertisers receive aggregate reporting only, no personal data is exchanged under the managed-campaign terms, and Roblox's own reporting is the controlling measurement (GAM-ROBLOX-F17, F33). Third-party ad SDKs and programmatic calls to third-party ad and measurement services are barred for developer-placed ads (GAM-ROBLOX-F21). Measurement beyond Roblox's own reporting depends on named partners (DoubleVerify, IAS, Nielsen, Kantar, Cint, iSpot.tv, EDO) whose integration scope is documented only in announcements (GAM-ROBLOX-F17, F18).
- **Differentiation.** Platform-controlled inventory inside a UGC platform with a filed audience of 123 million average DAUs in Q2 2026 (GAM-ROBLOX-F04). Detailed published age rules for ads, rather than a compliance claim: no personalisation under 18, Rewarded Video 13+, age-band category bans for under-13s, programmatic 18+ only (GAM-ROBLOX-F11, F12). Native ad units served by the platform, with dynamic per-user insertion and published counting definitions per format (GAM-ROBLOX-F07, F32). Roblox restricts competing third-party ad routes inside its platform: developer-placed ads may not replace billboards and portals, rewarded ads must use Roblox's product, and third-party ad and measurement calls or SDKs are barred (GAM-ROBLOX-F21).
- **Risks.** Child-safety litigation and state actions: Roblox had settlement agreements with Alabama, Mississippi, Nevada, South Dakota and West Virginia as of 30 July 2026 and is negotiating with other states (GAM-ROBLOX-F30). Roblox's own documents conflict on the start date and basis of the Advertising Integrations revenue share, a governance and contract-certainty risk for creators and the brands that buy integrations (GAM-ROBLOX-F22). Roblox's public pages conflict on who can see ads and who can publish them (brand-safety page: publishers 18+, users under 13 not eligible; Creator Hub and Developer Forum: publishers 13+, under-13 billboard ads from 4 May 2026) (GAM-ROBLOX-F08, F12). Suitability tiers are keyed to creator-declared maturity data and do not cover items players bring into experiences; no independent test of any Roblox control was found (GAM-ROBLOX-F20, F13). Independence of measurement: Nielsen agreed on 6 August 2026 to acquire DoubleVerify (closing expected by Q1 2027), and both are named Roblox measurement partners (GAM-ROBLOX-F18). Roblox says bookings growth slowed in Q2 2026 and guides Q3 2026 bookings down 14% to 18% year over year; the effect on ad supply is not stated (inference; GAM-ROBLOX-S03).
- **Best fit.** Brand campaigns aimed at users aged 13 and over that can run on opt-in rewarded video or in-world billboards, bought direct or through Google's DV360 (GAM-ROBLOX-F07, F14). Advertisers who need strict published age rules and no personal-data exchange, and who accept aggregate reporting (GAM-ROBLOX-F11, F17, F33). Contextual campaigns to under-13 audiences that can be bought by direct deal through Roblox or SuperAwesome, within the category bans (GAM-ROBLOX-F12).
- **Poor fit.** Buyers who need MRC-accredited in-game metrics or independent verification across every format and device (GAM-ROBLOX-F18, F19). Buyers who need title-level reporting, user-level data or their own third-party ad tags inside experiences (GAM-ROBLOX-F17, F21). Programmatic buyers targeting under-18 audiences, or US buyers expecting a quantified US ad-eligible audience (GAM-ROBLOX-F12, F14, F29). Console- or PC-specific plans, since delivery on those devices is not documented by Roblox (GAM-ROBLOX-F06).
- **Limits.** Advertising revenue is not disclosed, so the size of the business behind the product cannot be measured from public sources (GAM-ROBLOX-F02). No ad-eligible audience is published for the scoring period; the last figure is from Q4 2024 on self-reported ages, and DAUs count accounts, not people (GAM-ROBLOX-F04, F09). Age data cover only age-checked users (57% of DAUs at the end of Q2 2026) and Roblox warns they may not be representative (GAM-ROBLOX-F05). Device coverage of the ads is documented only for mobile in-app; PC and console delivery rests on a partner's 2025 announcement (GAM-ROBLOX-F06, F18). Advertisers get aggregate reporting only, and Roblox's reporting is the controlling measurement; title-level reporting is not documented (GAM-ROBLOX-F17). Third-party verification is documented for part of the product and only in announcements; no MRC accreditation or audit exists on MRC's record (GAM-ROBLOX-F18, F19). Programmatic targeting is limited to 18+ inventory, and under-13 inventory is direct-sold only (GAM-ROBLOX-F12, F14).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | C | Roblox's Creator Hub documents specifications and a counting approach for every in-experience brand format in this product: billboard video up to 30 seconds (click-to-play counted as a '15 second view'; autoplay counted as a video impression when a user looks at the ad for at least 0.5 s, it fills at least 1.5% of the viewport, the angle is up to 55 degrees and at least 50% of pixels are visible); static image ads (impression at 1 s with the same thresholds); portal ads (counted per successful teleport); Rewarded Video (full-screen, 6 to 30 s, 'The first frame of a video ad shown to a user counts as an impression'); ad-unit size limits of 8 x 4.5 to 32 x 18 studs. Creative guidelines for advertisers sit in the Advertising Standards (disclosure, prohibited and age-restricted content, Rewarded Video disclosure and opt-in rules). Advertiser-side asset specifications (file types, resolutions) were not found on Roblox pages opened; a reseller page states VAST without VPAID and 6 to 15 s preferred. [208, 207, 20, 205] |
| Targeting and context | live | D | Listed contextual signals named by Roblox: content or age rating (Suitability Tiers on the brands page; the January 2024 tier definitions keyed to experience age recommendations, whose current mapping is unconfirmed); named experience list (Advertiser Select, a 'hand curated list of brand safe experiences', no size or criteria given); genre targeting (May 2024 release); 'Contextual targeting for all eligible users (13+)' (brands page, no signal named). Under-13 campaigns are targeted 'mainly contextually based on the Roblox experience in which the ad is shown' plus generalised demographics (Roblox spokesperson via AdExchanger), with SuperAwesome's 80+ contextual categories (partner). Personal-data targeting: gender, device and country criteria (Creator Hub) and interest and demographic personalisation for users 18+ only (Ads Preferences); programmatic targeting only on 18+ inventory. These earn no points here. [42, 217, 629, 408, 216, 206, 208, 213] |
| Measurement functionality | live | D | Roblox's own reporting is 'the sole means of tracking and reporting delivery and is the controlling measurement for all campaigns' (managed-campaign terms). Roblox markets first-party viewability ('70%+ viewability packages' for Billboard Video; '95% viewability' for Rewarded Video). Other functions are through named partners, documented only in announcements: viewability and IVT (DoubleVerify custom integration for billboards on direct-sold inventory; IAS for Immersive Ads on mobile, 2024), brand lift (Cint, Kantar), deduplicated reach and frequency (Nielsen ONE Ads), audience verification and incrementality (iSpot.tv), engagement and search-lift benchmarking (EDO). Counting: Roblox's image-ad impression thresholds (50% pixels, 1 s, 1.5% of screen, 55 degrees) numerically match the IIG Guidelines 2.0 criteria for a viewable display impression, but Roblox does not state alignment, and its autoplay video impression (0.5 s) is shorter than the IIG's 2 continuous seconds for a viewable video impression. [1, 42, 187, 351, 370, 486, 488, 208, 17] |
| Optimisation and outcomes | live | D | Buying objectives are delivery or engagement events set by bid type (15-second view, video or image impression, teleport, Rewarded Video impression); no in-flight optimisation toward an outcome is documented for Immersive Ads or Rewarded Video in the pages opened. Reported results with a stated method: Roblox's own brand lift study with Latitude (20 to 27 March 2024; 2,100 US monthly users, 1,800 exposed and 300 control, ages 13 to 49; significant increases averaged across brands, magnitudes given only for intended actions, such as +10 points for visiting the brand's Roblox experience), paid for by Roblox. Roblox's 2026 Ipsos survey is self-reported attention without exposure measurement; EDO's pilot published no figures. [208, 207, 629, 486, 488] |
| Creative and insertion tools | live | C | Dynamic insertion is documented: ad units are served per user ('two users might simultaneously see different display ads or teleport to different advertiser games from the exact same ad unit'), with a fallback image for ineligible users, and the 10-K says creators insert ad units 'for Roblox to programmatically serve ad content from advertisers'. Asset tooling for the scored products appears only in marketing (a 'Rewarded Video Experience' that turns 'standard assets into immersive experiences'); the Homepage Feature template sits outside gameplay. A drag-and-drop Rewarded Video plugin for creators was in test in May 2026, and Q4 2026 tools described as 'no-code' and using AI were reported from RDC 2026, both unlaunched. The only documented AI generation (Ads Manager thumbnails) is for Sponsored games, a listed paid-discovery product. [208, 197, 630, 209, 508] |
| Supply | live | C | Filed platform audience: 123 million average DAUs in Q2 2026 (accounts, not unique people). Roblox publishes no ad-eligible audience for the scoring period; its last figure is 'over 61% ... (more than 50 million)' DAUs over 13 and eligible to view ads in Q4 2024. Current eligibility rules are vendor-documented: Rewarded Video to users 13 and over only; billboard, homepage and integration ads to all age bands from 4 May 2026 under category restrictions; programmatic targeting on 18+ inventory. Filed age data: 57% of global DAUs age-checked by the end of Q2 2026, of whom 35% under 13, 38% aged 13 to 17 and 27% 18 or over. Derived floors: about 45.6 million daily age-checked accounts aged 13+ and about 18.9 million aged 18+. Rewarded Video is integrated in over 1,200 games (May 2026); named experiences carrying it include Grow a Garden, Dress to Impress, Brookhaven and Plants vs. Brainrots. [246, 43, 187, 20, 215, 42, 188, 209] |
| Coverage | live | D | One tuple is evidenced inside the scoring period: mobile x app store x UGC platform (Magnite, January 2026: Rewarded Video 'in mobile in-app environments'; IAS 2024 scope 'mobile web and in-app'). The Roblox client runs on iOS, Android, Windows, Mac, Xbox, PlayStation, Chromebook and select VR hardware (10-K), and DoubleVerify said in April 2025 that its Roblox coverage spans mobile, desktop and console, but no Roblox document opened states that Immersive Ads or Rewarded Video serve on PC or console, so those tuples stay unconfirmed. [211, 370, 197, 351] |
| Programmatic access | live | D | Named DSPs: Display & Video 360 (listed on the brands page as a 'major DSP'; Roblox 'touted its integrations with DV360' at PlayFronts on 17 September 2026), Amazon DSP and Liftoff (announced 6 January 2026 for expansion 'this year'; no DSP's own documentation found). Named SSPs and exchanges: Google Ad Manager and Authorized Buyers (Google's own blog, April 2025), Magnite (own blog, January 2026), PubMatic (own release, April 2024), Index Exchange. The brands page wrongly calls Magnite and Index Exchange DSPs. Deal types: PubMatic's April 2024 release says Roblox video inventory 'will ultimately be made available' via programmatic guaranteed, PMPs, auction packages and open exchange (future tense); Roblox's own pages name no deal types; direct IO is documented. Supply path observed: PubMatic lists Roblox as a PUBLISHER; Magnite and Index list Roblox inventory via Google Exchange Bidding/Open Bidding as intermediary. Programmatic targeting is limited to 18+ inventory and under-13 inventory is never programmatic. [42, 183, 188, 187, 212, 211, 189, 572, 539, 573, 216] |
| US availability and coverage | live | D | Available to US buyers: managed campaigns contract with Roblox Advertising Services, LLC; US advertisers' campaigns are documented (Sam's Club rewarded video campaign described by Roblox at PlayFronts, September 2026; H&R Block's 2025 paid-media campaign; Universal's 'Jurassic World Rebirth' at about 2 billion impressions, per Roblox). US coverage is not quantified for the ad product: filings give only the US and Canada region (about 18% of 123 million DAUs in Q2 2026, derived as about 22 million) and US age-check penetration above 70%; no US ad-eligible audience or list of five US-available titles with ads was found. [1, 183, 188, 246, 43] |
| Independent assurance | live | D | (a) MRC: no Roblox entry on MRC's accredited digital services list or its services-under-review list on 29 September 2026 (the only gaming entry under review is Trivver). (b) Third-party verification in Roblox environments is documented only in partner announcements and for part of the product: DoubleVerify's custom integration covers IVT and viewability for in-game billboard image and video ads on direct-sold inventory, while its Rewarded Video functions (including suitability and attention) were announced in the future tense in April 2025; IAS announced 3D viewability and IVT for Immersive Ads on mobile web and in-app in May 2024, with no dated general-availability notice found. The managed-campaign terms refer to an advertiser's 'approved third-party verification partner' for IVT disputes above 10%. (c) TAG registry: no Roblox record. Pending: Nielsen agreed on 6 August 2026 to acquire DoubleVerify, so two of Roblox's named measurement partners may come under common ownership. [15, 16, 351, 370, 1, 570, 380] |
| Transparency | live | C | Supply-path disclosure is published and was observed directly: Roblox's app-ads.txt (15 lines: DIRECT for Google x3, PubMatic x2, Magnite, Index Exchange, Smart, Sharethrough, InMobi and Meta; RESELLER for LoopMe, Smaato, PubNative and Verve) and ads.txt (8 DIRECT lines); SSP sellers.json entries label Roblox as PUBLISHER (PubMatic) or as reached via Google Exchange Bidding (Magnite, Index Exchange). Named resellers are announced (Venatus, Playwire, SuperAwesome). Title- or experience-level reporting to buyers is not documented: the managed-campaign terms promise 'standard reporting' and 'aggregate reporting and metrics about campaign performance (e.g., number of impressions)'; Advertiser Select is a curated experience list, not a report. [538, 571, 572, 539, 573, 1, 217] |
| Suitability, age and data controls | live | C | Brand-suitability controls (outline for the scoring period): the brands pages name a Brand Suitability Suite (Suitability Tiers, Advertiser Select, age and geo controls, IAS and DoubleVerify) without operating detail; the only operational definition of the tiers is a January 2024 forum post keyed to experience age recommendations that Roblox's current maturity labels no longer use, and the post warned the tiers 'may change'. Publisher-side gates are documented (approved maturity questionnaire, ban on misclassifying the experience guideline rating to increase ad delivery), but maturity labels do not cover user-generated items players bring into an experience. Age controls (operational detail): no personalised ads under 18; Rewarded Video only to users 13 and over; category bans for audiences including under-13s, with extra restrictions for ages 5 to 8 and 9 to 12; no reward mechanisms for under-13 ad engagement; programmatic only on 18+ inventory; ad-integration assets under review served only to users 18 and over; Kids and Select accounts limit which experiences younger users reach. Data practices (operational detail): personalised ads and data selling or sharing only for users 18+, who can switch them off; 'We do not sell or share the personal information of known users younger than eighteen'; no personal data exchanged with advertisers and aggregate-only reporting under the managed terms; no personally identifiable engagement data to advertisers and no third-party ad or measurement calls or SDKs in developer-placed ads; retention periods not stated in the pages opened. [20, 213, 214, 1, 57, 190, 414, 42, 217, 215, 408] |
Performance evidence: Rewarded Video delivers 'completion rates of over 90% and viewability rates of 95%' for over 1,000 brands; the brands page repeats '90% completion rates and 95% viewability'. (vendor assertion, sponsor Roblox) [201, 42]; Billboard Video is sold with '70%+ viewability packages available'. (vendor assertion, sponsor Roblox) [42]; Roblox video ads produced statistically significant increases in brand awareness, ad awareness, opinion, consideration and recommendation averaged across brands, and higher intended actions (+10 points for visiting the brand's Roblox experience; +3 points for looking for more information). (commissioned measurement, sponsor Roblox (study conducted via Latitude)) [629]; 52% of surveyed 13- to 34-year-old Roblox users say they spend more time paying attention to ads on Roblox than they usually do, 13 points above streaming users and 11 above social users. (commissioned measurement, sponsor Roblox (with Ipsos)) [486]; EDO's Roblox pilot 'surfaced clear, measurable lifts in brand engagement' for branded events, measured as branded-search engagement against adjacent-week baselines, and benchmarked rewarded, billboard and homepage campaigns against EDO's convergent TV norms. (vendor assertion, sponsor EDO and Roblox) [488, 486]
### Super League (active; listed; pending change of control (not closed at the cutoff))
*Super League Enterprise, Inc. (formerly Nth Games, Inc. and Super League Gaming, Inc.)*
**SWOT (2026-09-29).** Super League is a small Nasdaq-listed seller of managed brand programs inside Roblox, Minecraft and Fortnite whose supply, measurement and ownership are all in flux at the cutoff.
- **Strengths.** SEC filings document its products, pricing, revenue recognition and deals, so buyers can check terms against primary records rather than marketing. (record) [128, 30] It runs fully managed programs for US brands and agencies across Roblox, Minecraft and Fortnite, priced by CPM, other cost-per-measure or flat fee. (record) [128, 30] In 2026 it added supply rights: 15% of Hide or OOF! with exclusive brand rights, a stake in My Avatar!, and the Misfits Ads business assets. (record) [30, 506] About 60% of revenue comes through agencies and over half from returning clients, per the FY2025 10-K. (record) [128]
- **Weaknesses.** FY2025 revenue fell 30% to $11.3 million with a $20.7 million net loss; two customers made up 22% of revenue. (record) [128] Much supply runs through platforms it does not own, such as Roblox Immersive Ads; its own Super Biz platform is being deprecated, sign-ups closed. (record) [30, 128, 511] No viewability, attention or IIG-aligned counting is documented; published outcomes such as '13x ROAS' are case-study figures without method. (vendor documentation) [128, 496, 528] It reports programmatic sales, but no DSP or exchange is named and no major SSP sellers.json listed it on 29 September 2026. (record) [145, 572, 598]
- **Opportunities.** The Misfits brand-partnership agreement of 1 May 2026 gives first access to Misfits' Roblox games, which could deepen owned supply if Misfits approves deals. (inference) [172] Q2 2026 on-platform revenue rose $431,000 and programmatic $199,000; with new US sellers hired, the ad line could stabilise. (inference) [30, 598] The June 2026 Play.Works CTV role and the Poki web-games partnership extend selling beyond UGC platforms, outside the scored product. (inference) [599, 567]
- **Threats.** Roblox's Advertising Integrations Terms (effective 1 May 2026) require asset approval and bar programmatic third-party ad calls, narrowing what Super League can sell there. (inference) [57, 600] The Q2 2026 call cites evolving Roblox policies affecting brand activations; the 10-K names platform operators like Roblox as potential direct competitors. (inference) [598, 128] Metaplanet would own about 95.7% after a Q4 2026 closing, with a bitcoin treasury expected to drive value, leaving advertising a secondary segment. (inference) [513, 514, 601]
- **What is sold.** Fully managed advertising and branded-content programs aimed at people who play video games, sold to brands and agencies (GAM-SUPERLEAGUE-F03). The scored product is the media part: brand placements inside third-party Roblox, Minecraft and Fortnite experiences. The filings name the on-platform media assets as static billboards, video billboards, portals, 3D characters and Pop Ups, sold on a CPM, other cost-per-measure or flat-fee basis over one week to two or three months (GAM-SUPERLEAGUE-F04). The company also sells playable and rewarded video ads in mobile inventory (partly through AdArcade), in-game ads in mobile and PC titles (partly using other networks' inventory), CTV advergaming as Play.Works' US sales partner, web-game integrations with Poki, and influencer campaigns (GAM-SUPERLEAGUE-F05, GAM-SUPERLEAGUE-F23). Custom builds, custom integrations and sponsorships inside Roblox, Fortnite and Minecraft are its larger 'publishing and content studio' line; they are recorded descriptively and not scored (GAM-SUPERLEAGUE-F13). Since May 2026 it also holds preferred, qualified rights to present brand partnerships in Misfits' Roblox games (GAM-SUPERLEAGUE-F08) and ownership or economic interests in two Roblox experiences, Hide or OOF! and My Avatar! (GAM-SUPERLEAGUE-F10).
- **Who buys it.** Brands and agencies. About 60% of revenue comes through agency relationships and 40% from direct brand engagements; over 50% comes from returning clients; two customers accounted for 22% of FY2025 revenue (GAM-SUPERLEAGUE-F28). The filings describe the programs as targeting US audiences (GAM-SUPERLEAGUE-F22). Named FY2025 custom-experience clients include Lions Gate, Panda Express, Chipotle's CMG Strategy Co., Google and Logitech (GAM-SUPERLEAGUE-F13); case studies name Panda Express, Chipotle, Maybelline and Lunchables (GAM-SUPERLEAGUE-F21).
- **How it earns.** Two revenue lines. The first is media and advertising: on-platform media, off-platform media (including Native Playables and gamified video), programmatic media sales and influencer marketing. It was $5.6 million in FY2025 and $3.1 million in H1 2026, and it is not an in-game ad series. The second is publishing and content studio (custom game development and immersive experiences), $5.4 million in FY2025 (GAM-SUPERLEAGUE-F13, GAM-SUPERLEAGUE-F14). Revenue is reported gross as principal except where a reseller partner is used (GAM-SUPERLEAGUE-F05). On-platform media is recognised impression by impression against a contracted number of impressions (GAM-SUPERLEAGUE-F04). The company also earns a share of revenue from game properties, for example 15.0% of Hide or OOF!'s gross revenue and a 12.0% fee on certain direct brand transactions (GAM-SUPERLEAGUE-F10). It reported a net loss of $20.7 million for FY2025 and $8.4 million for H1 2026 (GAM-SUPERLEAGUE-F13, GAM-SUPERLEAGUE-F14).
- **Supply and data dependencies.** Most of the supply is owned or controlled by others. (1) Roblox: media is delivered through the Roblox Immersive Ads platform (a resale of Roblox's own product, cohort B1) and through the Super Biz platform, which is being deprecated (GAM-SUPERLEAGUE-F05, GAM-SUPERLEAGUE-F11). Placements a developer puts directly into a Roblox experience fall under Roblox's Advertising Integrations Terms, effective 1 May 2026: registration, Roblox review of assets, no programmatic calls to third-party ad and measurement services, and billboard-style ads routed to Roblox's immersive ads product (GAM-SUPERLEAGUE-F19). (2) Misfits: its preferred rights exclude games Misfits does not solely control, need Misfits' prior approval before a brand is approached, and leave Misfits with sole discretion over terms and pricing (GAM-SUPERLEAGUE-F08). (3) Other in-game networks: Super League sometimes includes inventory from Anzu, Bidstack, AdInMo and Frameplay in its programs (GAM-SUPERLEAGUE-F05). (4) Partners outside the scored placements: AdArcade (rewarded mobile video), Play.Works (CTV games) and Poki (web games) (GAM-SUPERLEAGUE-F05, GAM-SUPERLEAGUE-F23). Data inputs: 'proprietary gameplay and campaign data as well as third-party inputs', psychographic data from Solsten (a $200,000 SAFE investment), and LiveRamp identity matching on its own websites (GAM-SUPERLEAGUE-F18, GAM-SUPERLEAGUE-F20, GAM-SUPERLEAGUE-F25).
- **Differentiation.** It is the only cohort A seller whose supply rights, deal terms and revenue lines can be read in SEC filings, including the full executed Misfits Brand Partnership Agreement (GAM-SUPERLEAGUE-F06, GAM-SUPERLEAGUE-F08). It combines placements in UGC platforms (Roblox, Fortnite, Minecraft) with custom builds and influencer work in one managed program, across platforms rather than tied to one network (GAM-SUPERLEAGUE-F03). It holds equity or economic interests in Roblox experiences it sells into, a 'Strategic Properties' model that differs from pure network resale (GAM-SUPERLEAGUE-F10).
- **Risks.** Pending change of control: about 95.7% Metaplanet ownership, a rename to Superplanet, Inc. and a bitcoin-treasury strategy, if stockholders approve on 16 October 2026 (after the cutoff). The DEF 14A says the existing business is expected to continue under current management (GAM-SUPERLEAGUE-F02). Platform rule risk: Roblox's Advertising Integrations Terms (effective 1 May 2026) require registration and Roblox approval of assets, bar programmatic calls to third-party ad and measurement services, route billboard-style ads to Roblox's immersive ads product, and add a revenue share per thousand visits from 30 September 2026 (per the Terms; the brief records a conflicting date on another Roblox page, not re-opened here). This limits how Super League can sell and measure its own placements in Roblox experiences (inference) (GAM-SUPERLEAGUE-F19). Platform competition: the 10-K says platform operators, including Roblox, may compete directly by offering their own native advertising products (GAM-SUPERLEAGUE-F03). Losses and scale: FY2025 revenue fell 30% to $11.3 million with a $20.7 million net loss, and H1 2026 showed a net loss of $8.4 million. Two customers made up 22% of FY2025 revenue (GAM-SUPERLEAGUE-F13, GAM-SUPERLEAGUE-F14, GAM-SUPERLEAGUE-F28). Children: the company sells youth-oriented media (the Youth and Family Marketplace, Lunchables) but publishes no operational age or data controls for its ad product. Its privacy policy dates from September 2023 (GAM-SUPERLEAGUE-F18). Contingent supply: the one-year Misfits agreement, Misfits' discretion and the deprecation of Super Biz make the continuity of named supply uncertain (GAM-SUPERLEAGUE-F08, GAM-SUPERLEAGUE-F11).
- **Best fit.** US brands and agencies that want a managed, cross-platform program in Roblox, Fortnite and Minecraft that combines placements with custom builds and creators, and that accept vendor-reported results (GAM-SUPERLEAGUE-F03, GAM-SUPERLEAGUE-F22). Buyers seeking brand partnerships in specific Roblox experiences where Super League holds rights (Hide or OOF!, My Avatar!, Misfits Games subject to Misfits' approval) (GAM-SUPERLEAGUE-F08, GAM-SUPERLEAGUE-F10).
- **Poor fit.** Programmatic buyers needing named DSP paths, deal IDs, sellers.json or supply-path verification (GAM-SUPERLEAGUE-F15). Buyers requiring MRC-accredited or independently verified in-game measurement, or IIG-aligned counting (GAM-SUPERLEAGUE-F16, GAM-SUPERLEAGUE-F17). Buyers needing documented, operational suitability and age controls for UGC or child audiences (GAM-SUPERLEAGUE-F18). Buyers who need title-level reach or ad-eligible audience figures to plan against (GAM-SUPERLEAGUE-F09, GAM-SUPERLEAGUE-F10).
- **Limits.** Named supply is thin and mostly contingent. Only two Roblox experiences are named with rights, and neither is shown carrying a dated campaign in the scoring period. The Misfits rights depend on Misfits' discretion over terms, pricing and each approach to a brand, exclude games Misfits does not solely control, and run for one year from 1 May 2026 (GAM-SUPERLEAGUE-F08, GAM-SUPERLEAGUE-F10). No ad-eligible audience figure is published. Every figure is portfolio MAU, visits, DAU or partner impressions, and none is ad-eligible (GAM-SUPERLEAGUE-F09, GAM-SUPERLEAGUE-F10). No format specifications, creative guidelines, named DSPs, deal types, sellers.json or verification partners were found in the sources opened (GAM-SUPERLEAGUE-F04, GAM-SUPERLEAGUE-F15, GAM-SUPERLEAGUE-F17). The revenue lines are not in-game ad series. Media and advertising mixes on-platform media, off-platform media, programmatic sales and influencer work, and the level of on-platform revenue is not disclosed (GAM-SUPERLEAGUE-F13, GAM-SUPERLEAGUE-F14). The product is a managed service: creative, targeting and measurement are delivered by Super League's team, not through buyer tools (GAM-SUPERLEAGUE-F03, GAM-SUPERLEAGUE-F04).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | C | The filings name the on-platform media assets: 'static billboards, video billboards, portals, 3D characters, Pop Ups and other media products'. They also name playable ads, rewarded video and in-game advertising as formats. The revenue note describes a counting basis: the brand benefits when the company 'delivers the impression on the platform, and the media product is viewed or interacted with by the on-platform user', against a contracted number of impressions. No format specification (sizes, durations, file types, placement rules) and no creative guidelines were found in the filings, releases or site pages opened. superleague.com has no product or spec pages. [128, 30, 528] |
| Targeting and context | live | C | The filings say the company works with brands to 'target ad units and target demographics'. They list 'Audience segmentation and targeting based on observed engagement patterns and inferred consumer motivations'. The home page sells 'play personas' from an 'AI & data-driven engine' powered by Solsten's psychological research, including 'PLATFORM TARGETING'. None of the listed contextual signals is documented as buyer-selectable: genre, named title or title list, in-game or session context, content or age rating, or no-personal-data contextual segments. Demographic and psychographic persona targeting is personal-data audience targeting. [128, 528] |
| Measurement functionality | live | A | The 10-K lists 'Brand lift and measurement solutions designed to assess campaign effectiveness' among its offerings, and 'Measurement and reporting, including brand lift analysis and performance tracking' among its services. No method, provider or sample rules are published. No viewability, attention, reach-and-frequency or sales-measurement function is documented, and the company claims no counting aligned to the Intrinsic In-Game Guidelines. The Bounce acquisition is described as giving 'a roadmap to more automated campaign measurement', which is future tense. No verifier integration in Super League's own placements was found. [128, 30] |
| Optimisation and outcomes | live | C | The 10-K documents objectives the buyer sets ('awareness, engagement, and customer acquisition') and in-flight optimisation in outline ('Campaign optimization through ongoing analysis of user interaction and performance metrics'; insights applied to 'in-flight campaign optimization'). Reported results have no stated method: '+6.3% lift in purchase rate' (Lunchables), 'Highest-ever mobile sales day' (Chipotle), '+31% purchase intent lift vs baseline' and '13x ROAS' (home page), and '50% potential ROAS uplift using our intelligence' (home page counter). None gives a design, sample, comparator, period or measurement party. [128, 528, 496] |
| Creative and insertion tools | live | C | Creative is produced by Super League as a fully managed service: 'The Company customizes the media advertising campaign and media products with applicable branding, images and design and place the media on the various digital platforms.' The filing's 'reusable frameworks, including interactive formats, content templates, and measurement tools' are internal studio assets, not documented buyer-facing tooling. No dynamic insertion by Super League is documented at the cutoff. Super Biz, which provided in-experience insertion, is being deprecated. The Misfits agreement says Super League will not be 'permitted to, expected to, or be obligated to, make actual changes to Misfits Games'. Roblox's Advertising Integrations Terms require Roblox to review and approve assets and route billboard-style ads to Roblox's immersive ads product. AdArcade's AI Creative-as-a-Service is partner technology for rewarded mobile video, outside this product's placements, with no documented controls. The FY2025 10-K states 'We currently do not incorporate AI/ML into our products.' [128, 172, 57, 631, 511] |
| Supply | live | A | Two Roblox experiences are named with rights in filings. Hide or OOF! carries 15.0% equity, a revenue share and 'exclusive rights with respect to certain brand partnership opportunities and related placement economics'. My Avatar! is an ownership interest described as generating 'differentiated inventory for playable advertising'. A title-level count with metric and period appears in the 10-Q: Hide or OOF! 'has been visited more than 570 million times in 16 months' (visits, sourced by the filer to Rotrends). Misfits' portfolio carries preferred, qualified rights with no game named in the agreement, and the vendor figure of 'more than 100 million monthly active users' has no period and is not ad-eligible. No ad-eligible audience figure was found. Other supply is resold or shared: Roblox Immersive Ads, and other networks' inventory (Anzu, Bidstack, AdInMo, Frameplay). The 2024 claim of 3,500 experiences with 140 million monthly active users belongs to the Super Biz network, which predates the scoring period and is being deprecated. [30, 172, 146, 161, 128] |
| Coverage | live | E | The filings document live on-platform revenue in Roblox, Fortnite and Minecraft in Q2 2026. That gives the environment (UGC platform) and the distribution (platform launcher), but no source opened states the device (mobile, PC, console) on which Super League's placements render, and no named live title is tied to a device. The 10-K's general line 'Advertising placements embedded within gameplay environments across mobile, immersive and PC games' is not tied to any title or combination. Web (Poki) is a custom integration; mobile rewarded video (AdArcade) and CTV (Play.Works) are partner inventory outside the product definition. [30, 128, 567, 631] |
| Programmatic access | live | C | The Q2 2026 10-Q reports a '$199,000 increase in programmatic media sales revenue'. The Misfits releases say the deal 'adds programmatic revenue'. The Q2 release says the Youth and Family Marketplace is sold 'through both programmatic buying and Super League’s managed-services team'. No DSP, exchange, SSP or deal type (PMP, PG, open auction, curated) is named in any source opened, and no sellers.json, ads.txt or app-ads.txt is published at superleague.com or superbiz.gg (HTTP 404, 29 September 2026). Roblox's Advertising Integrations Terms bar developer-placed integrations from making 'programmatic calls to third-party advertising and measurement services'. [30, 161, 568, 632, 145, 57] |
| US availability and coverage | live | C | The filings state that the company 'generates revenue from brands and agencies by executing programs targeting U.S. audiences'. The Q2 2026 release documents 'new revenue leadership and experienced sales talent across key U.S. advertising markets'. US brand campaigns are named (Chipotle, Panda Express). The company is US-listed and based in Santa Monica. No US coverage figure for the product was found: no five US-available titles named as carrying placements, and no US audience with metric and period. The '200 million consumers in the U.S.' in the Misfits release is gamer population, not coverage. [128, 30, 568, 496, 161] |
| Independent assurance | unconfirmed | E | No assurance record was found. On 29 September 2026, MRC's list of accredited digital services has no Super League, Misfits or gaming entry; its environment key covers only desktop, mobile web, mobile in-app and CTV. MRC's digital services-under-review list shows only Trivver as a gaming audit in process. The TAG registry returned no company for 'Super League' or 'Misfits'. No third-party verification (IVT, viewability, suitability) in Super League's own placements was documented in the sources opened, and no vendor claim of compliance with the IIG guidelines was found either. Any verification attached to Roblox's own ad products belongs to the Roblox row and is not inherited. [15, 16, 566] |
| Transparency | live | A | No title- or property-level reporting to buyers is documented; case studies report delivery and engagement figures without a report format. Super League publishes no sellers.json, ads.txt or app-ads.txt at superleague.com or superbiz.gg (all HTTP 404 on 29 September 2026). The filings do give a partial, company-level supply-path disclosure. They name the delivery routes ('Super Biz Roblox platform, the Roblox Immersive Ads platform, other platforms'). They say the company acts as principal 'except in situations where the Company utilizes a reseller partner'. And they say it sometimes includes inventory from Anzu, Bidstack, AdInMo and Frameplay in its programs. [128, 30, 145] |
| Suitability, age and data controls | live | C | (a) Brand suitability: marketing only. The Common Sense Networks initiative (26 March 2024) promises reaching young audiences 'in safe and suitable ways'. The Poki web offer is described as 'curated, brand-safe environments' (17 September 2026). No exclusion lists, rating tiers, UGC exclusions or pre- or post-bid controls are documented. Misfits' approval rights over brand partnerships are publisher controls, not buyer suitability controls. (b) Age controls: outline at most. The Youth and Family Marketplace offers 'kid-safe media across gaming channels' with no mechanism stated (Q2 2026 release). The FY2025 10-K COPPA risk factor says it 'retained a COPPA expert as a consultant' and takes precautions for 'our platform'. The age rules Roblox applies to its own products (under-18 non-personalisation) are Roblox's documentation, not Super League's. (c) Data practices: outline. The privacy policy (last updated 20 September 2023) covers the corporate websites and properties, and still lists minehut.com, sold in February 2024. It states categories collected, sharing of hashed email with LiveRamp for advertising, retention 'as long as necessary', and that 'SLE does not knowingly collect information for any child under the age of 13 without parental consent'. It does not cover data in in-game ad delivery. The 10-K says the company aggregates 'proprietary gameplay and campaign data as well as third-party inputs'. Roblox's terms separately bar developers from sharing 'personally identifiable user engagement information with advertisers'. [400, 567, 568, 128, 569, 57] |
Performance evidence: Lunchables Playables virtual world: '+6.3% lift in purchase rate', with 10 million hours of engagement, 7.4 billion points collected and 7.3 million virtual pets redeemed. The world was 'designed to keep kids coming back'. (vendor assertion, sponsor Super League (seller)) [496]; Chipotle 'Boorito Maze' on Roblox: '3.5M+ unique players engaged', 'Highest-ever mobile sales day' and 'Record app downloads and digital transactions in a single day'. (vendor assertion, sponsor Super League (seller)) [496]; Home page: '+31% PURCHASE INTENT Lift vs Baseline', '13x ROAS' and a '50% POTENTIAL ROAS UPLIFT USING OUR INTELLIGENCE' counter. (vendor assertion, sponsor Super League (seller)) [528]; Maybelline Sunkisser Blush takeover of 'a popular Roblox roleplay world': '25M+ virtual item try-ons', '38.5M brand impressions' and '9 minutes average time spent engaging'. (vendor assertion, sponsor Super League (seller)) [496]; CeraVe Pharmacy inside Poki's Sprint League: players spent 'an average of five minutes engaging with the brand, 231 times longer than a typical piece of social content'. (vendor assertion, sponsor Super League (seller)) [567]; Play.Works CTV network: 'more than 150 million monthly ad impressions with a 95% view-through-rate (VTR)', 'more than 19 minutes of session time per user' and 'more than 100 million households'. (vendor assertion, sponsor Super League and Play.Works (interested parties)) [599]
### Twitch (Amazon) (active)
*Twitch Interactive, Inc. (registered in Delaware, file no. 4221511, per Twitch's Terms of Service). Twitch advertising is sold through Amazon Ads (Amazon DSP self-service or managed service); the Amazon contracting entity for Twitch ad buys was not established in the sources opened.*
**SWOT (2026-09-29).** Twitch sells documented video, display and creator formats through Amazon Ads with first-party suitability controls and a TAG seal, but no MRC accreditation names Twitch and its audience figures are internal and dated.
- **Strengths.** Twitch media has a documented Amazon DSP route: display open auction since December 2022, premium preferred deals since October 2023, and desktop vertical video via programmatic guaranteed. (vendor documentation) [258, 259, 602] Amazon documents three inventory tiers, contextual collections such as Family Friendly games, mandatory Content Classification Labels and impression-level content reporting as standard for Twitch buyers. (vendor documentation) [266, 265] TAG's registry records Twitch with the Brand Safety Certified seal, validated by a third-party auditor as a Direct Seller and applying to global operations (read 29 September 2026). (record) [337] Premium Video runs pre-roll and mid-roll on desktop, mobile, console and streaming TV, and server-side stitching reaches console, CTV and ad-block users on desktop. (vendor documentation) [47, 264]
- **Weaknesses.** No MRC record names Twitch as accredited: the November 2025 IAS letter excludes Twitch supply, and the May 2026 Amazon DSP letter does not mention Twitch. (record) [332, 603, 15] None of the eleven Twitch spec pages defines when an impression, view or completion is counted; they list tracking events and creative rules only. (vendor documentation) [47, 604] Amazon's Twitch audience figures are internal, global and dated 2022-2023; the only in-period US reach figure pools Twitch with Prime Video, Fire TV and exchanges. (vendor claim) [2, 260] Some DIRECT lines in Twitch's ads.txt resolve to Amazon-held seller accounts, and APS sellers.json has no entry for the APS seller ID Twitch lists. (record) [605, 606, 573, 572]
- **Opportunities.** Sponsorships opened to Creator Sponsorship Certified Affiliates on 20 August 2026, with Wehype surfacing offers, which widens the creator pool game publishers can buy. (inference) [607] Chat Sentiment Analysis and Campaign Assist, announced by 9 June 2026, could give creator campaigns in-stream reporting; general availability is not stated. (inference) [608] Amazon says Twitch campaigns can be measured through Amazon Marketing Cloud including store sales (June 2026), which lets game advertisers read sales outcomes, not only reach. (inference) [608]
- **Threats.** A 2025 Marketing Science article finds Twitch influencer streams have a small positive effect on game usage and purchases, suggesting most games do not profit from sponsoring them. (inference) [503] Twitch added teen Parental Controls in July 2026 but no document states an age-based ad rule, and its ad-policy URL now redirects to Amazon's global policies. (inference) [442, 441, 609] Turkey's data regulator fined Twitch 2 million lira in November 2024 over a breach, and Twitch's privacy notice allows combining Twitch data with Amazon customer data. (inference) [610, 611]
- **What is sold.** Two families of brand advertising on Twitch, sold by Amazon Ads (GAM-TWITCH-F03). (1) Twitch media: video and display ads on the Twitch homepage, discovery pages and streamer channels, bought self-service in Amazon DSP or through managed service. Named units are First Impression Takeover (one advertiser per day per geo, daily rate), Headliner, Premium Video (pre-roll and mid-roll in live broadcasts across desktop, mobile, tablet, console and streaming TV; 6 to 60 seconds; unskippable), Vertical Video Ads, Stream Display Ads (beside or below the live player), Mid-Feed Ads, Super Leaderboard, Medium Rectangle and Homepage Carousel (GAM-TWITCH-F04). Twitch video inventory is also reachable inside Amazon's Streaming TV ads product, which bundles Prime Video, Twitch, Fire TV Channels and third-party publishers (GAM-TWITCH-F07). (2) Creator activations: Creator Sponsorships, Bounty Board (creators accept paid bounties to play a game or watch a trailer on stream), Channel Takeover (channel skins, 30-60 second streamer-read ads, follower promotions, sponsored subscriptions) and custom work by the Brand Partnership Studio (GAM-TWITCH-F05). Streamables, an opt-in 30-second livestream unit shown inside 'Twitch Partnered' mobile games, is sold from the same catalogue but renders inside third-party games, so it is flagged as a possible separate product row rather than scored here (GAM-TWITCH-F06). Ads render on a game-adjacent livestreaming service, never inside gameplay; on consoles they run in the Twitch app and are not clickable (GAM-TWITCH-F28).
- **Who buys it.** Brands and agencies through Amazon Ads: Amazon says 'Advertisers can buy Twitch ads whether or not they sell products in the Amazon store' and names auto, tech, apparel and food and beverage advertisers as buyers (GAM-TWITCH-F03). Programmatic buyers use Amazon DSP (self-service or managed; premium preferred deals are for self-service advertisers); smaller advertisers can reach Twitch through Streaming TV ads with no minimum spend (GAM-TWITCH-F07). Game publishers are a named buyer group for Bounty Board ('promote gameplay features and trailers'), which is paid discovery for games and sits outside the brand-advertising objective (GAM-TWITCH-F05). Case studies opened name KITKAT, Kraft Heinz Canada, Reebok, Premier Inn and Lacoste; none is a US case with a stated method (GAM-TWITCH-S01, GAM-TWITCH-S14, GAM-TWITCH-S15).
- **How it earns.** Advertising revenue from media sold by Amazon Ads; Amazon's filings recognise advertising revenue 'as ads are delivered based on the number of clicks or impressions' and report it only as a company-wide line (FY2025: $68,635 million), with no Twitch break-out (GAM-TWITCH-F02). Pricing is not published: Twitch media needs 'a minimum spend' in Amazon DSP, Premium Video lengths carry 'varying fees', and First Impression Takeover is 'priced on a daily rate' per geo (GAM-TWITCH-F03, GAM-TWITCH-F04); Streaming TV ads bought through sponsored ads are billed per thousand impressions with no minimum budget (GAM-TWITCH-F07). Creators at Affiliate or Partner level 'can receive revenue from running ads on their channel' and control mid-roll timing; the revenue share was not opened (GAM-TWITCH-F05; open question).
- **Supply and data dependencies.** Supply is Twitch's own owned-and-operated service: web, mobile app, console and streaming-TV apps, plus creator channels whose ad monetisation Twitch gates (new creators cannot carry ads; criteria are broadcast recency, unique broadcasts, concurrent viewers, followers and policy adherence) (GAM-TWITCH-F09). Twitch's Terms reserve to Twitch 'the exclusive right to sell, serve, and display advertisements' and bar third-party ad networks from serving through the service (GAM-TWITCH-F20). Video ads are stitched server-side into the broadcast by Twitch's SureStream method (GAM-TWITCH-F16). Data: Amazon DSP buying uses Amazon's first-party signals, and Twitch's Privacy Notice says Twitch data may be combined with Amazon customer data and shared with third-party ad networks and measurement providers (GAM-TWITCH-F19). Suitability signals depend on streamer-applied Content Classification Labels, Twitch's ESRB-rated game database and Twitch moderation (GAM-TWITCH-F09, GAM-TWITCH-F11). Supply-path files: Twitch's ads.txt and app-ads.txt list 21 DIRECT entries across 17 ad systems, while Amazon's APS sellers.json lists Twitch under a different seller ID from the one in Twitch's ads.txt (GAM-TWITCH-F08).
- **Differentiation.** Stream-level suitability built on streamer-applied content labels plus ESRB game ratings, exposed to buyers as three tiers and Contextual Collections, with impression-level content reporting (GAM-TWITCH-F09). Server-side stitched video (SureStream) that reaches console, connected-TV and ad-blocking desktop viewers inside the live broadcast (GAM-TWITCH-F16). Direct connection to Amazon's measurement stack (Amazon Brand Lift, Amazon Marketing Cloud, first-party new-to-brand and detail-page reporting) for Twitch media bought in Amazon DSP (GAM-TWITCH-F12). Creator activations sold alongside media (Bounty Board, Channel Takeover, Brand Partnership Studio, which Amazon says has run more than 1,650 creator-led campaigns since 2020) (GAM-TWITCH-F05, GAM-TWITCH-F25). TAG Brand Safety Certified by independent validation for global operations on TAG's own record (GAM-TWITCH-F15).
- **Risks.** Assurance gap: DV and IAS measure Twitch viewability and IVT, but MRC's accreditation of IAS's Amazon DSP integration explicitly excludes Twitch supply, so accredited counting on Twitch is not established (GAM-TWITCH-F13, GAM-TWITCH-F14). Supply-path inconsistency: Twitch's ads.txt APS seller ID does not resolve in APS's sellers.json, and 17 authorised ad systems sit alongside Terms that bar third-party networks, which buyers may need to reconcile (GAM-TWITCH-F08, GAM-TWITCH-F20). Dated audience claims: headline Twitch audience figures are from 2022-2023 Twitch internal data and are repeated in 2026 material without a period (GAM-TWITCH-F17). Regulatory exposure on data: a reported 2024 Turkish data-protection fine over a platform breach (KVKK record not opened) (GAM-TWITCH-F26). Suitability depends on streamer self-labelling in live, ephemeral content; no independent test of tier accuracy was found (GAM-TWITCH-F09, GAM-TWITCH-F11).
- **Best fit.** Brands buying adult Gen Z and millennial video reach through Amazon DSP that want stream-level suitability tiers and Amazon's first-party outcome reporting (GAM-TWITCH-F09, GAM-TWITCH-F12, GAM-TWITCH-F17). Game publishers and entertainment advertisers promoting launches and trailers through creator bounties or channel takeovers, noting this is paid discovery rather than brand advertising in the paper's objective filter (GAM-TWITCH-F05). Advertisers already buying Prime Video or Streaming TV who want to add a livestreaming line in the same platform (GAM-TWITCH-F07, GAM-TWITCH-F23).
- **Poor fit.** Buyers who require MRC-accredited impression or viewability counting in the specific environment (GAM-TWITCH-F14). Buyers who must buy through a non-Amazon DSP or need multi-DSP deal access (GAM-TWITCH-F07). Campaigns aimed at under-13 audiences (the service is closed to them) or requiring documented teen-specific ad rules (GAM-TWITCH-F18). Buyers wanting in-game placements: Twitch ads render beside or within streams, not inside games (Streamables excepted and unevidenced by named titles) (GAM-TWITCH-F06, GAM-TWITCH-F28).
- **Limits.** No impression-counting definition is published for any Twitch format, and MRC excludes Twitch supply from IAS's accredited Amazon DSP measurement (GAM-TWITCH-F04, GAM-TWITCH-F14). Only one named DSP route (Amazon DSP); no third-party DSP integration is documented, and Twitch's Terms bar third-party ad networks from serving on Twitch (GAM-TWITCH-F07, GAM-TWITCH-F20). No Twitch-only audience figure with an ad-eligible metric inside the scoring period, and no US Twitch audience figure; Amazon's published figures are internal, global and from 2022-2023 (GAM-TWITCH-F17). Suitability tier boundaries are undocumented and tier names differ between Amazon's 2023 and 2026 documents (GAM-TWITCH-F10). No age-based rule for ad delivery, targeting or personalisation was found in the documents opened; Amazon's Twitch ad policy page could not be read (GAM-TWITCH-F18). Creator formats: talking points cannot be modified during a Bounty Board or Channel Takeover campaign, and no third-party verification covers streamer-read ads (GAM-TWITCH-F05, GAM-TWITCH-F13). Pricing and minimum spend for Twitch media are not published (GAM-TWITCH-F03).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | C | Specifications are documented for the media and creator formats opened: Premium Video (placement in live broadcasts across desktop, mobile, tablet, console and streaming TV; lengths 6, 15, 20, 30 and 60 seconds; 1920x1080; bitrate, audio and frame-rate limits; mandatory transcodes; VAST 2.0/3.0; supported third-party tracking events impressions, clicks, quartiles and completes), First Impression Takeover (up to 30 seconds, unskippable, first watched broadcast of the day), Stream Display Ads (desktop and mobile sizes, 50kB, JPG or PNG), Streamables, Bounty Board and Channel Takeover (asset sizes, CTA length, brand colour). Creative guidelines are documented for Stream Display Ads (minimum font sizes, claims no more than 20% of height, border and contrast rules, Mirror-C guidance) and for Channel Takeover (asset guidelines table) (GAM-TWITCH-F04, GAM-TWITCH-F05). No page defines when an impression, view or completion is counted for any Twitch format, and MRC's scope record excludes Twitch supply from IAS's accredited Amazon DSP counting (GAM-TWITCH-F14). [540, 47, 633, 634, 261, 262] |
| Targeting and context | live | C | Against the cohort B2 signal list: (1) content context and (2) contextual segments without personal data: Contextual Collections let buyers 'target or exclude stream content type, such as gaming or lifestyle, or hone in on a specific area of Twitch like Family Friendly games or Sports', available through premium preferred deals in Amazon DSP (GAM-TWITCH-F09, GAM-TWITCH-F07); (3) game category of the content: the October 2023 launch note lists 'popular gaming content, lifestyle, or gaming genre audiences' collections (GAM-TWITCH-F07); (4) content or age rating: Inventory Tiers (Expanded, Standard, Restrictive) exclude progressively more sensitive and mature content using Content Classification Labels and ESRB game ratings from Twitch's game database (GAM-TWITCH-F09). A channel or property list is not documented for media buys (Channel Takeover and Bounty Board select creators through the sales process; Bounty Board targets broadcasters by country). Amazon first-party audience targeting is personal-data targeting and earns no points. [266, 259, 269, 261] |
| Measurement functionality | live | C | Listed functions documented for Twitch buys: (1) viewability, through DoubleVerify and IAS server-to-server integrations that both verifiers announced for Twitch inventory and that Amazon says are 'approved on Twitch' (GAM-TWITCH-F13); (2) reach reporting, as first-party 'unique reach' alongside impressions and video completion rate (GAM-TWITCH-F12); (3) outcome measurement, as first-party detail page views, branded searches and new-to-brand reporting, and Amazon Marketing Cloud measurement 'including sales in the Amazon store' (GAM-TWITCH-F12); (4) brand lift, through Amazon Brand Lift, for which 'All ad campaigns run through the Amazon DSP' are eligible (exposed versus control surveys; US and seven other markets) (GAM-TWITCH-F12). Nielsen, VideoAmp, Lucid, Kantar and 'Twitch RPG' are named on the channel page without documented integration (names only, n/e). Attention is not documented for Twitch (DV's Twitch coverage lists viewability, fraud/IVT and in-geo only). The viewability measured is not MRC-accredited for Twitch: MRC's IAS Amazon DSP letter excludes Twitch supply (GAM-TWITCH-F14). [264, 635, 266, 335, 336, 608, 2] |
| Optimisation and outcomes | live | C | Results with a stated method exist, but outcome objectives and in-flight optimisation for Twitch buys are not documented in the pages opened. Results: dentsu's January 2024 lab study (with Lumen; Twitch collaborated) reports 57% brand recall for simulated livestream ads against a dentsu norm of 38% and a 17% average brand-choice uplift against 7.25%, with brand-choice uplift defined against a control-group baseline (GAM-TWITCH-F22); Amazon's how-to guide footnotes a 52% incremental-reach result for Streaming TV plus Twitch campaigns to Amazon Marketing Cloud data for 21 US and 8 UK advertisers, February 2022 to January 2023 (GAM-TWITCH-F23). A +74% incremental-reach claim for Twitch alongside Prime Video carries no method on the page (GAM-TWITCH-F23). Objectives appear only as marketing benefit statements on the channel page (awareness, reach, traffic, purchase, loyalty). [27, 264, 608, 2] |
| Creative and insertion tools | live | C | All three features are described in vendor documentation. (1) Dynamic insertion: 'Twitch's proprietary SureStream delivery method stitches video creative directly into the broadcast content', and Premium Video specs say part of the inventory 'is SSAI pre-fetch', so creative is inserted server-side into live streams without an app update (GAM-TWITCH-F16). (2) Asset tooling: 'Amazon's creative resizing tool will upscale' a 160x600 display creative to the Mirror-C Stream Display experience and apply colour bands automatically (managed service only) (GAM-TWITCH-F16). (3) AI-assisted generation with controls: the Amazon DSP Gen-AI workflow (30 September 2024) generates Twitch-sized assets from an uploaded image; users 'review all newly generated assets in a flyout', and association with a Twitch line item 'will trigger moderation' before assets serve (GAM-TWITCH-F16). Limit: for Bounty Board and Channel Takeover, talking points 'Cannot be modified during campaign' (GAM-TWITCH-F05). [264, 47, 633, 532, 261, 262] |
| Supply | live | D | Twitch audience figures published by Amazon are Twitch internal data, global, and outside the scoring period, and none is an ad-eligible audience: '105MM average monthly visitors come to the Twitch site' (Twitch internal data, Global, 2023) and 1.3 trillion minutes watched in 2023 on the channel page; 'an average of 35 million users' daily (Twitch internal data, WW, 2022) in the how-to guide; a June 2026 Amazon Ads article repeats '105 million+ monthly visitors' with no period (GAM-TWITCH-F17). The only US figure in the scoring period is Amazon's 200 million+ monthly ad-supported reach of Amazon streaming TV (US, 5 January to 3 February 2026), which is unduplicated across Prime Video, Twitch, Fire TV Channels, Amazon Publisher Direct and third-party exchanges and gives no Twitch share (GAM-TWITCH-F17). No filing or independent measurement of Twitch monthly users was found (Amazon does not break out Twitch; GAM-TWITCH-F02). [2, 264, 608, 260, 199] |
| Coverage | live | E | Not applicable: dimension 7 is outside cohort B2's applicability mask (game-adjacent ads do not render in a game content environment). Coverage tuples are recorded descriptively only; Streamables' in-game tuple is not evidenced by a named title. |
| Programmatic access | live | C | One named DSP: Amazon DSP, self-service and managed service (GAM-TWITCH-F03, GAM-TWITCH-F07). Named deal types: 'non-guaranteed open auction inventory' for Twitch display (launched 14 December 2022, United States); 'Twitch premium preferred deals', a non-guaranteed tier between 'Twitch guaranteed and open marketplace buying' (generally available 25 October 2023, United States among markets); and Twitch guaranteed buys, referenced but not separately documented (GAM-TWITCH-F07). Twitch video is also bookable through Amazon's self-serve Streaming TV ads with no minimum spend (GAM-TWITCH-F07). Twitch's ads.txt authorises 17 ad systems as DIRECT sellers, but no third-party DSP's documentation or any seller page names a non-Amazon DSP route to Twitch, and Twitch's Terms bar third-party ad networks from serving on the service (GAM-TWITCH-F08, GAM-TWITCH-F20). Protocol states are recorded in protocol_states and earn no score step. [2, 258, 259, 260, 605, 257] |
| US availability and coverage | live | C | Available to US buyers: 'US' is listed in the locale lists of every Twitch specification page opened (Premium Video, Stream Display Ads, First Impression Takeover, Streamables, Bounty Board), the December 2022 display launch lists 'North America: United States', the October 2023 premium preferred deals and November 2023 Content Controls notes list the United States, and the channel page's contact form offers the United States (GAM-TWITCH-F21). US coverage is not quantified for Twitch: the only US audience figures are a GWI/Comscore 2022 statistic about viewer overlap with social platforms and Amazon's multi-property streaming-TV reach (GAM-TWITCH-F17). US campaign case studies with results were not found among those opened (Reebok UK and Germany, KITKAT Netherlands, Premier Inn UK). [47, 633, 634, 258, 259, 2] |
| Independent assurance | live | B | (a) MRC: no accreditation of a Twitch metric. MRC's accredited digital list (read 29 September 2026) names Amazon DSP supply sources as Retail and IMDb, Amazon Publisher Services, Fire TV, third-party exchanges and Alexa connected devices, without Twitch; neither the accredited nor the under-review list contains 'Twitch'; and MRC's 7 November 2025 letter accrediting IAS's server-to-server Amazon DSP integration states 'FireTV and Twitch supply sources are excluded' (GAM-TWITCH-F14). This is a scope exclusion, not a denial or lapse. (b) Third-party verification: DoubleVerify (31 October 2023) and IAS (29 May 2024) each announced viewability and invalid-traffic measurement on Twitch inventory through server-to-server integrations, and Amazon says DV and IAS are 'approved on Twitch to measure Invalid Traffic, Viewability, and domain/app-level brand safety' (GAM-TWITCH-F13). No verification is documented for the creator formats (Bounty Board, Channel Takeover, streamer-read ads), and domain/app-level brand safety classifies Twitch as a whole, not the stream. (c) Certification on the certifier's own record: TAG's registry shows Twitch enrolled in Verified by TAG with the Brand Safety Certified seal, achieved 'through independent validation by a third-party auditor', category Direct Seller, applying to 'global operations' (GAM-TWITCH-F15). [15, 16, 332, 337, 335, 336, 266] |
| Transparency | live | C | Property- or content-level reporting is documented as standard: 'content-level reporting, which provides impression-level detail of the content your ads appeared in' is 'available on demand in the Amazon DSP' and is 'part of the standard reporting offering available across Amazon O&Os' (GAM-TWITCH-F09). Supply-path disclosure is published: twitch.tv/ads.txt and twitch.tv/app-ads.txt (identical; OWNERDOMAIN=amazon.com) list 21 entries, all DIRECT, across 17 ad systems, and Amazon's APS sellers.json lists 'Twitch' (twitch.tv) as seller_type publisher (GAM-TWITCH-F08). Inconsistency: the ads.txt APS line uses seller ID 3036, which does not appear in APS's sellers.json (Twitch is seller 650 there). [266, 605, 636, 606] |
| Suitability, age and data controls | live | C | Brand suitability: operational detail. What is set: three Inventory Tiers (Expanded, Standard, Restrictive) and Contextual Collections (target or exclude content types such as Family Friendly games); how applied: signals from streamer-applied Content Classification Labels (mandatory for listed sensitive content, monitored by moderation, AI detection and user reports; mislabelling brings warnings, Twitch-applied labels and label locks) and ESRB ratings in Twitch's game database feed the tiers, with each tier 'excluding progressively more content', plus a monetisation gate (new creators cannot carry ads); scope: Twitch inventory bought in Amazon DSP, with impression-level content reporting (GAM-TWITCH-F09, GAM-TWITCH-F11). Gap: what each tier excludes is not documented, and the 2023 launch note names the tiers Expanded, Moderate (Default), Restrictive (GAM-TWITCH-F10). Age controls: outline for advertising. The service is closed to users under 13 (16 in Australia) and users under the age of majority must be supervised by a parent who accepts the Terms (GAM-TWITCH-F18); for users under 18, content labelled Sexual Themes, Drugs, Violent and Graphic Depictions and Gambling is filtered from discovery by default, adjustable by the user (GAM-TWITCH-F18); optional Parental Controls launched 21 July 2026 (GAM-TWITCH-F18). No document opened states an age-based limit on ad delivery, ad targeting or ad personalisation; Amazon's Twitch ad policy page could not be read (access error). Data practices: operational detail in the Privacy Notice (last modified 12 August 2026): automatically collected data (IP address, device and browser identifiers, interaction data) used to 'provide customized advertisements' and monitor campaign effectiveness; cookie and mobile ad identifiers shared with third-party ad networks, advertisers and ad analytics providers; email address or resettable device identifier shared to match for off-Twitch ads; Twitch data combinable with Amazon customer data; retention for as long as required for stated purposes and deletion on account closure except as required or permitted by law; under-13 (COPPA-defined) personal information deleted when Twitch learns of it (GAM-TWITCH-F19). Regulator findings: a Turkish KVKK fine over a data breach was reported on 16 November 2024 (within three years of the cutoff); it concerns platform security and breach notification, not the advertising product's data practices, and the KVKK record was not opened (GAM-TWITCH-F26). Documentation is not legal compliance. [266, 265, 269, 257, 611, 441, 442, 610] |
Performance evidence: Twitch livestream ads in a simulated livestream environment produced 57% brand recall against a dentsu cross-media norm of 38%, and a 17% average brand-choice uplift against dentsu norms of 7.25%; across all gaming formats in the same study, recall was 21% and below dentsu norms. (commissioned measurement, sponsor dentsu Media US (agency), with Lumen as research vendor; Twitch, Activision Blizzard and Anzu collaborated; whether they paid is not stated) [27]; Amazon says Streaming TV campaigns 'drive 52% incremental reach in the U.S.', footnoted to Amazon Marketing Cloud data for 21 US and 8 UK advertisers that ran both Streaming TV and Twitch campaigns, February 2022 to January 2023. (vendor assertion, sponsor Amazon (owner of Twitch)) [264]; Amazon says campaigns that include Twitch alongside Prime Video deliver +74% incremental reach, 'meaning nearly three-quarters of the audience exposed on Twitch are not reached by Prime Video alone'. (vendor assertion, sponsor Amazon) [608]; A Premier Inn 10-hour sponsored livestream delivered 8.9 million minutes watched, more than 881,000 unique viewers and over 36,000 chat interactions; a Spanish tourism campaign generated 11.1 million viewing minutes with click-through rates '3.6x higher than the national tourism industry standard'. (vendor assertion, sponsor Amazon) [608]; A peer-reviewed Marketing Science paper (published 20 February 2025) finds that live streams of video games by Twitch influencers have a small positive effect on the games' usage and purchases, 'suggesting that most games do not profit from sponsoring live streams by influencers'. (synthesis, sponsor unknown (academic authors Huang and Morozov; funding not visible in the abstract)) [503]
### XBOX Media Solutions (Microsoft) (active)
*Microsoft Corporation (XBOX Media Solutions is a business unit name, not a registered entity found in this review; the contracting entity for XBOX Media Solutions campaigns was not established)*
**SWOT (2026-09-29).** XBOX Media Solutions sells managed brand formats on Microsoft-owned casual inventory, led by Candy Crush Saga, with thin public specifications, no MRC accreditation and no evidenced in-gameplay console product.
- **Strengths.** It sells Rewarded Video, interactive video, Playables, Interactive End Cards and custom integrations it describes as '100% owned and operated' across King, Xbox and Microsoft Casual Games. (vendor documentation) [224, 612, 24] King's app-ads.txt, last modified 29 September 2026, declares activisionblizzardmedia.com as owner domain and as a DIRECT seller. (record) [53] The TAG registry shows Activision Blizzard Media and King.com Ltd. Brand Safety Certified by independent validation, for the United Kingdom only. (record) [390, 391] Microsoft bought Activision Blizzard, including King, for $75.4 billion on 13 October 2023; its FY2026 XBOX revenue was $21,790 million. (record) [522, 198]
- **Weaknesses.** No public specifications, counting rules or creative guidelines were found for its in-game formats; the only spec sheet covers a retired Xbox Home and Store offer. (vendor documentation) [191, 457] MRC lists no gaming accreditation for it, and IAB UK's certified list omits it despite the vendor's 'Gold Standard 2.1 Certified' claim. (record) [15, 16, 352, 223] Headline 98% viewability and 96% completion cite Moat Q1 2023 benchmarks, and the KitKat and Hershey's brand-lift figures name no measurer or method. (vendor claim) [361, 227, 226] Audience is stated only as 'hundreds of millions' of players, and Microsoft's filings do not break out gaming advertising revenue. (record) [223, 198, 247]
- **Opportunities.** Rewarded Portals for Candy Crush Saga entered closed beta in select markets on 12 May 2026, a format the vendor could widen. (inference) [225, 226] An ad-supported cloud streaming test began for Xbox Insiders on 23 July 2026; its ads run before sessions, so it adds console-adjacent, not in-game, supply. (inference) [202] A Microsoft patent application published 3 September 2026, reported secondhand, describes ads at paused gameplay moments; it is a filing, not a product. (inference) [613]
- **Threats.** King's app-ads.txt, dated 29 September 2026, authorises Gameloft, Anzu, Gadsme and Meta routes, so buyers can reach the same inventory without XBOX Media Solutions. (inference) [53, 229] XBOX revenue fell 7% to $21,790 million in FY2026, which may constrain investment in the unit whose revenue sits inside that line. (inference) [198] Microsoft's September 2026 privacy statement bars personalised ads to under-18 accounts while King's policy uses under-13, complicating targeting across its titles. (inference) [21, 234, 228]
- **What is sold.** Directly sold, opt-in in-game brand formats in Microsoft-owned titles: Rewarded Video, In-Game Interactive Video (Video Explorer, Tappable Video, Video Selector), Playables built from six game templates, Interactive End Cards, and custom integrations ('Candy'fied Sponsored Moments in Candy Crush Saga, a Solitaire Event Takeover in Microsoft Casual Games on PC and mobile) (GAM-XBOX-F03). Rewarded Portals, a level-screen format in Candy Crush Saga, is in closed beta in select markets (GAM-XBOX-F04). The same unit also sells console placements outside gameplay (XBOX Experiences, Xbox Landing Experiences, dashboard and store), which belong to the listed row 'Xbox dashboard and store placements', not to this scored product (GAM-XBOX-F13).
- **Who buys it.** Brands and their agencies buying managed campaigns; the documented advertisers inside the scoring period are consumer brands (Hershey's holiday 2025, KitKat/Nestlé across 40 markets) (GAM-XBOX-F09). Buying is through account managers and a contact form; no self-serve or programmatic route is documented (GAM-XBOX-F16).
- **How it earns.** Advertising revenue inside Microsoft's XBOX line. Microsoft's 10-K says XBOX revenue 'is mainly affected by subscriptions and sales of first- and third-party content, as well as advertising'; advertising is not broken out, so the unit's revenue, pricing and growth are not public (GAM-XBOX-F25, GAM-XBOX-F26). Pricing models and CPMs are not published for the scoring period.
- **Supply and data dependencies.** Supply is first-party: King titles (Candy Crush Saga named in current campaigns) and Microsoft Casual Games (Microsoft Solitaire Collection) (GAM-XBOX-F11). The same inventory is authorised to many other sellers in the publishers' app-ads.txt files, including Gameloft (COMBO!), Google, Magnite, PubMatic, Index Exchange, Amazon, Meta and Microsoft's own exchange, with Anzu and Gadsme as resellers (GAM-XBOX-F17, GAM-XBOX-F18). Data: an archived policy says targeting is 'based primarily on 1st party data' (GAM-XBOX-F06); King's privacy policy governs data in King games and opts players into targeted advertising by default where local law allows (GAM-XBOX-F22). Measurement relies on named partners whose scope is not published (GAM-XBOX-F08).
- **Differentiation.** First-party IP integrations inside Candy Crush Saga (Candy'fied Sponsored Moments, Rewarded Portals on the level screen) that third-party networks selling King inventory do not document (GAM-XBOX-F03, GAM-XBOX-F04). One seller across mobile (King), PC (Microsoft Casual Games) and console surfaces (Xbox, outside gameplay), presented as part of Microsoft Advertising (GAM-XBOX-F02, GAM-XBOX-F03). Opt-in, rewarded formats as the default design, with multi-market global campaigns documented in the scoring period (GAM-XBOX-F09).
- **Risks.** Concentration: every in-period campaign found runs in one title, Candy Crush Saga (GAM-XBOX-F11). King inventory is sold through several other routes (COMBO!, SSPs, resellers), which blurs exclusivity and makes reach claims non-additive (GAM-XBOX-F17, GAM-XBOX-F18). Assurance claims on the vendor's page run ahead of certifiers' records (IAB UK Gold Standard not on IAB UK's list; TAG BSC UK-only) (GAM-XBOX-F20, GAM-XBOX-F21). Age rules differ inside one parent (Microsoft accounts under 18 versus King under 13), and an all-ages title carries ads (GAM-XBOX-F22, GAM-XBOX-F23). Rebrand and domain redirect removed public policy documentation; buyers must obtain terms privately (GAM-XBOX-F07).
- **Best fit.** Consumer brands wanting managed, custom, opt-in rewarded and playable activations inside Candy Crush Saga, including multi-market campaigns (GAM-XBOX-F09). Buyers who value title-level control through owned-and-operated game packages and direct takedown on request (GAM-XBOX-F06).
- **Poor fit.** Programmatic buyers who need DSP access or standard deal types (GAM-XBOX-F16). Buyers requiring MRC-accredited in-game measurement or scoped third-party verification (GAM-XBOX-F19). Advertisers seeking in-gameplay placements on console (GAM-XBOX-F13). Plans that need a quantified US audience for the product (GAM-XBOX-F29).
- **Limits.** No public specifications, counting rules or creative guidelines for the in-game formats (GAM-XBOX-F05). No programmatic route, DSP or deal type documented for the in-game formats (GAM-XBOX-F16). No audience figure with metric and period, no ad-eligible share, and no gaming ad revenue in filings (GAM-XBOX-F10, GAM-XBOX-F25). No MRC accreditation or audit; TAG brand-safety seal is UK-only and company-level; third-party verifiers named without scope (GAM-XBOX-F19, GAM-XBOX-F20). Outcome results come only from vendor case studies with no stated method (GAM-XBOX-F09). Console placements documented are outside gameplay; no console title with an in-gameplay placement is named (GAM-XBOX-F13). The detailed brand-safety policy was published on a domain that now redirects; current pages carry marketing language only (GAM-XBOX-F06, GAM-XBOX-F07).
| Dimension | Availability | Evidence | Observed |
|---|---|---|---|
| Format execution | live | D | Formats are named and described on marketing pages: Rewarded Video, three Interactive Video variants, Playables with six named game templates (Bubble Shooter, Catcher, Collector, Glider, Match 3, Ninja Slice), Interactive End Cards (Email Capture, Carousel of 'up to six separate cards', Scratch), video skins, end cards and captions in five languages, plus custom integrations. No specifications (dimensions, durations, file formats), counting approach or creative guidelines are published for the in-game formats; the former ad-specs URL redirects to the hub and its last archived copy found is from 2019. The only spec sheet found (VAST 2.0 video, 380x212 and 428x240 banners) is for the retired Xbox Home and Store programmatic offer, which is outside gameplay and another row. [224, 612, 24, 457] |
| Targeting and context | unconfirmed | C | The archived Activision Blizzard Media Brand Safety page (captured 21 January 2026, inside the scoring period) documents buyer-selectable 'specific packages of our games' so advertisers reach audiences only within games they prefer (a named title or title-list signal), with the example of avoiding age-restricted titles such as Call of Duty; other targeting is 'based primarily on 1st party data' (geography, demographics), which earns no points. King's privacy policy says players who opt out still see ads 'based on criteria such as the Game you are playing, the content of that Game, or the region', which describes delivery criteria, not buyer-selectable targeting. The current Microsoft pages document no targeting controls. [409, 234] |
| Measurement functionality | live | D | The measurement page offers 'first and third-party measurement' and lists eight partners by name only (Nielsen Catalina, Circana, Kantar, GWI, In Market, MRI Simmons, DoubleVerify, Integral Ad Science) with no scope per partner. Brand lift (ad recall, awareness, consideration, purchase intent in percentage points) and viewability are reported to advertisers in case studies (KitKat 2026, Hershey's 2026, Arla 2024), and the ad-solutions page offers measurement 'to understand the brand and sales impact of ad exposure'. No counting rule is aligned to the IIG Guidelines 2.0 or to the IAB/MRC video guideline, and the headline viewability and completion figures are footnoted to 'Moat Analytics Benchmarks, Video Mobile In-App Q1 2023' (Moat was retired on 30 September 2024 per GAM-C-F18). [361, 362, 227, 226, 224] |
| Optimisation and outcomes | live | D | Case studies report brand-lift points and delivery results (KitKat: +29 points ad recall, 500M+ impressions; Hershey's: +55 points ad recall, 1.2x to 2.5x higher lifts 'with Rewarded Portal'; Rewarded Portals blog: 'over 2x incremental brand lift') without a stated method, measurer, sample, control design or period. Creative is optimised before launch through 'player sentiment analysis and ad testing'; no outcome objective a buyer can set and no in-flight optimisation are documented. [227, 226, 225, 361] |
| Creative and insertion tools | live | D | Creative is produced as a managed service: 'Whether the ad experiences are designed and developed by our creative team or feature your captivating content'. Playables come from six templates and custom integrations are built with the game teams, but no buyer-facing tooling, dynamic insertion (asset swap without a game update) or AI-assisted placement with controls is documented. Xbox Activities' 10 templated formats belong to the Xbox Landing Experiences row, not this product. [223, 612, 455] |
| Supply | live | D | Named ad-eligible titles inside the scoring period: Candy Crush Saga (Hershey's holiday 2025, KitKat 2026, Rewarded Portals beta) and Microsoft Solitaire Collection (Solitaire Event Takeover, current page). The only audience statements are vendor wording without a figure, metric period or ad-eligible share ('hundreds of millions of monthly active players across our portfolio'; 'hundreds of millions of players across 200+ countries'). '+400M Call of Duty Mobile downloads' is a download series, not an audience. Microsoft's filings disclose no King, Microsoft Casual Games or Xbox user figure in the documents opened. No console title is named as carrying in-game placements. [223, 191, 24, 227, 226, 228, 198] |
| Coverage | live | D | Two distinct evidenced tuples: mobile x app store x casual title (Candy Crush Saga) and PC x app store (Microsoft Store) x casual title (Microsoft Solitaire Collection). Console is not evidenced as an in-game tuple: 'XBOX Experiences' names no console title and the documented console placements (dashboard, store, Xbox Landing Experiences, Click-to-Engage) sit outside gameplay. Premium PC and console franchises (Call of Duty, Halo, Minecraft, World of Warcraft) are named as portfolio, not as carrying placements. [24, 227, 226, 228, 223] |
| Programmatic access | unconfirmed | E | No DSP, deal type, private marketplace or other programmatic route for the in-game formats is named on the seven Microsoft gaming pages, the case studies or the blog posts opened; buying runs through account managers and a contact form. The only Xbox programmatic document found is the retired Microsoft Invest page for an invite-only private marketplace on Xbox Home and Store banners (outside gameplay; another row). A 2019 package listing called Activision Blizzard Media rewarded video 'not available on Open Exchange' (outside the scoring period). King's and Microsoft Casual Games' app-ads.txt files authorise many SSPs, which shows supply connections but not that this product's brand formats are bought programmatically. [191, 24, 457, 637, 53, 574] |
| US availability and coverage | live | D | No page states that the product is sold to US buyers or quantifies US coverage. Indirect signals: a named US advertiser's campaign (The Hershey Company, holiday 2025, in Candy Crush Saga; the case study does not name the market), US-framed statistics on the hub ('#1 grossing franchise in the U.S. app stores (Candy Crush Saga)'; '58% of U.S. gamers are multiplatform'), and the US Microsoft Store listing for Microsoft Solitaire Collection. The KitKat campaign's five measurement markets exclude the US. TAG's registry shows Activision Blizzard Media's self-attested Certified Against Fraud seal for North America, which is a company-operations record, not a product availability statement. [226, 191, 228, 227, 390] |
| Independent assurance | live | B | (a) MRC: MRC's accredited digital list (read 29 September 2026) has Microsoft Advertising entries only for search clicks and invalid-traffic filtration in desktop and mobile web; no entry for XBOX Media Solutions, Activision Blizzard Media, King or any gaming environment, and MRC's under-review list has none of them. (b) Third-party verification: DoubleVerify and Integral Ad Science are listed by name only, with no scope, and no verifier release covering this product was found; n/e. (c) Other certification on the certifier's record: TAG shows Activision Blizzard Media and King.com Ltd. with Brand Safety Certified by independent third-party validation (Direct Seller) for the United Kingdom only, and Certified Against Fraud by self-attestation (North America and other regions). The vendor's claim of 'IAB UK Gold Standard 2.1 Certified' is not on IAB UK's own current certified and registered list. [15, 16, 390, 391, 352, 223, 361] |
| Transparency | live | B | No title-level reporting to buyers is documented (the archived policy says advertisers 'will know in advance the type of content' through owned-and-operated game packages, which is not reporting). Supply-path disclosure is partial: King's app-ads.txt declares OWNERDOMAIN=activisionblizzardmedia.com and 'activisionblizzardmedia.com, 1, DIRECT', and Microsoft Casual Games' app-ads.txt labels an '# ABM' section, but activisionblizzardmedia.com serves no sellers.json (the path redirects to an HTML page), so the seller of record's own seller types and direct-versus-reseller relationships are not published. [53, 574, 409] |
| Suitability, age and data controls | live | C | Brand suitability: operational detail in the archived Activision Blizzard Media Brand Safety page (captured 21 January 2026): what is set (owned-and-operated inventory only; buyer-chosen 'specific packages of our games', for example to avoid age-restricted titles such as Call of Duty; takedown on request), how it is applied (package selection; takedown 'upon request, within 24 hours' through the account representative) and scope (all inventory, with longer takedown for game integrations and sponsorships). Current Microsoft pages carry only marketing ('100% brand safe'). Age controls: outline. King's privacy policy says it does not knowingly target interest-based ads to anyone under 13 (or a higher age where law requires) or knowingly allow under-13s to play, without describing the age mechanism; Microsoft's Privacy Statement bars personalised ads to Microsoft-account children under 18 by account birthdate, but no document says which rule governs XBOX Media Solutions inventory in King titles; Microsoft Solitaire Collection is listed 'For all ages' with an ad-supported free tier. Data practices: operational detail in King's privacy policy (26 February 2026): device and ad identifiers and gameplay data shared with advertising partners for targeting and measurement, opt-in to targeted advertising by default where local laws allow, device-level, DAA and US state opt-outs, non-targeted ads based on game, content or region after opt-out, retention while the account is active, and deletion of data learned to come from an under-age child. No regulator or court finding within three years of the cutoff: the FTC's 2023 Xbox COPPA settlement was announced on 5 June 2023, outside the window, and concerned Xbox account sign-up, not this product. [409, 234, 21, 228, 425, 224] |
Performance evidence: KitKat campaign in Candy Crush Saga across 40 markets: +29 percentage points ad recall, +7 top-of-mind awareness, +7 consideration, +5 purchase intent (average across UK, France, Germany, India, Australia); 500M+ impressions, 6.9M+ clicks, 496M+ rewards, 97%+ completion. (vendor assertion, sponsor Microsoft (seller of the inventory); measurer not named) [227]; Hershey's Holiday Bells (holiday season 2025) in Candy Crush Saga: +55 points ad recall, +11 top-of-mind awareness, +13 consideration, +7 purchase intent; 1.2x to 2.5x higher lifts with Rewarded Portal; Rewarded Portal delivered 40M+ impressions, 1.67M clicks and a 4.32% CTR. (vendor assertion, sponsor Microsoft (seller); measurer not named) [226]; Rewarded Portals 'early results show over 2x incremental brand lift across ad recall, consideration, and purchase intent when Portals are layered into broader campaigns'. (vendor assertion, sponsor Microsoft) [225]; Older case studies: Maybelline (April 2024) +23 points ad recall; Arla Anchor (December 2024, UK) +30.2 points ad recall and 99.8% viewability; Prada Candy (February 2023) +52 points ad recall and 40,000 samples requested in 24 hours. (vendor assertion, sponsor Microsoft/Activision Blizzard Media (seller); measurer not named) [638, 639, 640]; Rewarded video '96% Completion rate' and '98% Viewability', footnoted to 'Moat Analytics Benchmarks, Video Mobile In-App Q1 2023'. (vendor assertion, sponsor Microsoft) [224, 191]; Microsoft's April 2026 blog: 'According to research we conducted in partnership with Dentsu, 100% of gaming ads are fully viewed'. (vendor assertion, sponsor Microsoft restating a dentsu study in which Activision Blizzard Media collaborated) [456]
## 12. What it means
*The author's playbook said gaming is a media-mix reallocation problem, not a niche test. The public record does not settle that. What it does settle is narrower and more useful: brand buyers in games are buying on sellers' documentation, and the independent checks that exist for other digital media mostly do not exist here yet.*
> **In short.**
> Two hypotheses are supported and three are unresolved. Standards exist, but MRC's record, read on 29 September 2026, shows no accredited intrinsic in-game metric for any product scored in cohorts A and B1 (H2, supported) [15, 16]. Children's and platform rules restrict what in-game ad products can do (H4, supported) [20, 21, 22]. Whether buyers report measurement, safety and planning constraints more often than reach (H1), whether console is in-game inventory at scale (H3), and whether platform ad sales are growing faster than specialists' (H5) cannot be settled from public sources [33, 24, 198]. The playbook's GAP model shares its names with IAB Australia's 2023 categories and cannot be sized [34, 37].
### The hypotheses, coded
| Hypothesis | Outcome | Why |
|---|---|---|
| H1. Buyer constraints | Unresolved | Two surveys not established as US put brand safety first, and neither lead is established as larger than sampling error; no US survey with constraint shares found in the sources screened; the inventory leg holds |
| H2. Standards and assurance | Supported | IIG 2.0 is a final release; MRC's record shows no accredited intrinsic in-game metric for any product scored in cohorts A and B1; interstitial, banner and video ads in game apps fall under the general mobile in-app standards |
| H3. Console | Unresolved | Microsoft's console integrations cannot be placed inside or outside gameplay; at most three EA titles qualify |
| H4. Children's and platform rules | Supported | Rule documents from Roblox, Microsoft, Discord, Epic (Fortnite) and Google Play restrict targeting, personalisation, third-party calls or measurement; COPPA as amended adds separate consent |
| H5. Market structure | Unresolved | At least three specialist companies sold or put into insolvency since 2021, by primary record; comparable revenue series are not disclosed |
| GAP model (question) | Unresolved | No source screened sizes the three lines as mutually exclusive categories |
Each coding follows the rule registered on 28 September 2026 before research began. Unresolved means public sources cannot settle the question; it does not mean the hypothesis is false. Each verdict holds only for its registered measure: H2 for MRC's record and the verifiers' own documentation of the scored products, not for the quality of the measurement; H4 for primary rule documents, not for compliance or enforcement.
### What the playbook got right, and what it did not
The playbook names EMARKETER as the source of its market figures, although its 3.4 billion player count matches a figure dentsu cites [36, 93]. Among the publishers screened by 28 September 2026, EMARKETER is the only one found with public US dollar figures [3, 4, 38, 98], and its public statements disagree about when the series passes $10 billion [99, 92]. The series counts ads placed in games, and most of its dollars are on mobile; EMARKETER also says game developers buy most of the ads served in games, a share of ads rather than of dollars [6, 7]. The series is therefore probably a poor proxy for brand demand; that reading is an inference, since no source splits the dollars by buyer or objective [6, 7, 8]. The playbook's "under 5% of worldwide media investment" most likely traces to dentsu, has no denominator in any accessible text and is not used here [36, 93].
The playbook argued that buyers cannot fit gaming into their planning, safety and measurement systems. The supply-side facts behind the measurement half are well supported: MRC's record lists no accredited intrinsic in-game metric for any scored in-game product, verification rests on announcements, and several sellers' certification claims are missing from the certifiers' own lists [15, 350, 333]. They do not show that buyers cannot fit gaming into their measurement systems, nor that these gaps, rather than audience, hold spend back. In IAB's March 2024 survey of 300 US buyers already investing in games, sponsored by Admazing and Venatus, 86% agreed games advertising is brand safe and 85% said they were confident measuring it [107, 108]. Those who fold games into other channels list "not yet enough scale" first among four unnumbered reasons, a phrase that the report's one illustrative quote for that group ties to the respondent's own spend rather than to audience [107].
The GAP model's three names appeared in IAB Australia's July 2023 report with different contents [34, 91]. Its lines overlap, because the away-from-the-game line includes app installs, an objective [36, 8]. This paper keeps placement and objective as separate fields for that reason.
### For buyers
Treat the category as three buying tasks. Networks selling placements inside other companies' games (cohort A) sell mostly on their own documentation. Platforms and publishers selling their own games (cohort B1) control the inventory and the rules, and in Roblox's case the reporting that settles the invoice: its managed-campaign terms make its own reporting the controlling measurement [1]. Much game-adjacent media (cohort B2) is bought like the rest of digital media: Twitch and Overwolf, for example, name DSP buying routes and document suitability controls [2, 266, 44, 294], while Discord sold its ads only through its sales team during the scoring period and announced an auction-based Ads Manager in managed access on 29 September 2026 [48, 280, 248].
Within those tasks, four buys are common. The table sets out what the record shows about each; the second column is this review's reading of what each is usually bought for, not a finding.
| What is bought | Usually bought for | Buying route documented | Reporting documented | Ask for |
|---|---|---|---|---|
| Intrinsic placements in other companies' games (cohort A) | Repeated exposure inside play, without interrupting it | Direct, private-marketplace and curated deals; few DSPs confirm a network integration on their own sites [118, 176, 119] | The network's own impressions and viewability; MRC's record shows no accreditation for intrinsic measurement [17, 15] | Titles with live dates, each matched to the publisher's app-ads.txt; counting aligned to the IIG guidelines; the verifier's own record for the environment sold |
| Rewarded video and other ad formats in game apps (for example Roblox Rewarded Video) | A completed view the player opted into | On Roblox, sold direct, with programmatic routes documented only in announcements; shown only to users 13 and over [187, 188, 20] | The general mobile in-app standards apply; verifier coverage differs by format [17, 351] | Which verifier measures the unit and in which environment; completion counts; the age limits that apply |
| A branded world or integration on a user-generated platform (Roblox) | Participation in the brand's own space | Portal ads that lead users into an advertiser's experience, and creator integrations registered in Roblox's Ads Manager [208, 210] | In managed campaigns, Roblox's own reporting is the controlling measurement [1] | The age mix of those who took part; how each reported metric is defined; which revenue-share terms apply [43, 57, 190] |
| Streams, communities and creators (Twitch, Discord Quests) | Association with the community around a game | Twitch through Amazon DSP; Discord through its sales team, with Ads Manager announced in managed access [2, 48, 248] | Twitch documents several measurement functions; Discord's Quests terms make impression and reach figures non-binding estimates [264, 280] | The suitability settings and the content labels behind them; what a flat fee guarantees [266] |
When two networks claim the same titles, ask each for the titles it will deliver in and its seller ID in each, and check that ID in the publisher's app-ads.txt, where it is marked DIRECT or RESELLER [53, 76]. Many publisher domains appear in more than one network's sellers.json, so reach that two networks claim over the same titles has to be compared title by title, not added [9, 10, 11, 12, 13, 14].
Ask for what can be checked. Named titles with live dates, matched against publishers' app-ads.txt files, say more than a count of seller accounts, which shows authorised connections, not active supply [53, 14]. MRC's own lists say more than a compliance badge. A verifier's own documentation, scoped by format and environment, says more than a partnership release; DoubleVerify's Roblox coverage, for one, differs by format [351]. For outcomes, ask for a randomised or holdout test with a published method and a named funder; this review found none for an in-game or game-adjacent brand campaign funded by a party without a financial interest, in the sources screened [26, 27, 28].
Children can be a large share of a game audience: of Roblox users who had age-checked, 35% were under 13 globally in the seven days to 30 June 2026, and Roblox cautions that data drawn from age-checked users may not represent its whole user base [43, 246]. Under the amended COPPA Rule, a covered operator needs separate verifiable parental consent before disclosing a child's personal information to third parties, unless the disclosure is integral to the service [23, 421]. Platform rules add limits of their own. None of this is settled by a seller's statement that its product is compliant.
| Buyer question | Test that fits | What the inference needs | Where it fails | What it cannot tell you |
|---|---|---|---|---|
| Is this brand buy or user acquisition? | Ask who buys the product and to what end. Game publishers buying installs, plays or discovery of their own games are user acquisition or paid discovery, whatever the format. | The seller states its buyer mix, or the product's terms name the objective (for example Roblox Sponsored Experiences or Medal Quests). | Published market totals mix both objectives; EMARKETER's US series does not split them. | Split a market-wide total into brand and user-acquisition dollars. [6, 7, 57] |
| Is the supply live, and is it ours to buy? | Ask for named titles with placements live in the last 12 months, then check each publisher's app-ads.txt for a DIRECT line that matches the seller's sellers.json, and for US selling rights. | The files are current when read; a matched line means the publisher authorises that seller. | Seller-account counts, installed SDK counts and audience claims without a metric and period; networks that resell each other's supply. | Show which titles, placements or geographies are actively sold, or deduplicate reach across sellers. [9, 14, 53] |
| Are the impressions counted to a standard, and who checked? | Ask which guideline governs counting (the IIG Guidelines 2.0 for intrinsic placements), then read MRC's own accredited and under-review lists and the verifier's own documentation, scoped by product, format and environment. | The accreditor's and verifier's records are current; an audit counts only when MRC confirms it. | Vendor claims of IIG compliance, certification logos missing from the certifier's list, and verification announced in the future tense. | Show that counts are accurate where no one has audited them. [17, 15, 16, 351] |
| Can the campaign reach players under 18, and on what terms? | Read the platform's own advertising rules and the product's documented age controls; under the amended COPPA Rule, check whether targeting or third-party disclosure of children's data needs parental consent. | The platform's age data and rules apply to the placement bought; documentation describes practice. | Assuming contextual equals compliant, or inferring a user's age. | Establish legal compliance; documentation is not compliance. [20, 23, 422] |
| Did it work? | Ask for a randomised or holdout brand-lift test with a published method, sample, comparator and funding; treat seller-run or seller-funded results as vendor evidence. | The test isolates the placement from other media and the comparator is not the seller's own benchmark. | Benchmarks built mostly from the seller's own campaigns; attention scores offered as outcomes. | Generalise from one title, brand or format to the channel. [27, 26, 28] |
| Is a console placement inside gameplay? | Ask where the ad renders (in the game, the dashboard, the store, a TV app or before a session), the stage (test, beta, launch) with dates, and the named titles. | The platform or publisher documents the placement and its status. | Treating a console device as in-game inventory, or a patent or reported plan as a product. | Show scale where no titles or buying route are named. [202, 24, 25] |
### For sellers
The matrix rewards documentation that can be checked, and the gaps are in the same places for most sellers. Publishing specifications, a counting approach and creative guidelines for each format would move format execution, one of the two weakest dimensions in cohort A along with coverage. Verification documented by the verifier, in the environments sold, and a filing or registry record of US coverage would move Standing and Reach. Naming DSP integrations with deal types, confirmed on the DSP's side, would move programmatic access.
### What would change the thesis
The brief named four developments that would change this paper's reading. In the sources read by 29 September 2026, none had happened.
MRC-accredited intrinsic measurement at scale: MRC's record shows one gaming service with an audit in process, Trivver, which this review lists rather than scores because it found no named live title or campaign for it in the period [16, 114]. Console in-game ad products live and bought programmatically: EA sold its in-game ad units directly at its June 2026 launch, and by 15 September 2026 Digiday reported programmatic buying across EA's mobile games while saying only that more dynamic capabilities were being extended to console and PC, with no DSP or deal type named [25, 86, 193]; Microsoft's cloud ad test runs ads before play [202]. Independent outcome studies for in-game formats: none found for a live brand campaign that meets the test above, and the only study with documented independent funding is a laboratory experiment [26, 27, 28]. A gaming line in a holding company's published media mix: WPP Media forecasts gaming as an ad-revenue line, which is not a planning mix [89, 90].
Dated signals to watch after the cutoff: the outcome of Trivver's MRC audit; whether Xbox's ad-supported cloud streaming moves beyond test; whether EA or Roblox names programmatic deal types and DSPs on the buyer's side; whether Roblox applies the developer revenue share from 30 September 2026, as its Terms say, or from 1 January 2027, as its Creator Hub says, and on 1 October 2026 the two documents still conflicted and no source showed which date applies [57, 190]; whether Zynga Ads+ goes live; and whether a public registration statement from Discord, reported in January 2026 to have filed confidentially for an initial public offering, puts a platform's game-adjacent ad revenue on the record [542, 543].
## 13. Method, limits and interests
*Artificial intelligence (AI) research agents gathered this paper's evidence, separate AI verifiers told to refute it checked 807 of its 807 claims, three outside reviews by another AI model challenged the design, the evidence and the draft, and three internal AI review rounds checked the draft. Hypotheses and scoring rules were fixed before scoring; one full pass scored the products, and separate AI scorers then re-scored 10 rows blind, which tests reproducibility, not independent human judgement. The paper is not peer reviewed, externally audited or independently validated.*
> **In short.**
> Verifier agents checked 807 of the 807 ledger claims: 545 held as stated, 258 held with a qualification, 3 were contradicted and restated, and 1 could not be settled. H2 (standards and assurance) is supported: the Media Rating Council (MRC) record, read on 29 September 2026, shows no accredited intrinsic in-game metric for any scored network or platform product [15, 16]. H4 (children's and platform rules) is supported: Roblox, Microsoft and Discord restrict targeting or personalisation in their own ad products by age or age signal [20, 21, 22]. H1 (buyer constraints), H3 (console) and H5 (market structure) are unresolved [33, 24, 198]. No source screened by 28 September 2026 sizes the author's GAP categories (in-game, around-the-game and away-from-the-game) as mutually exclusive lines [37, 4]. The evidence is public, weighted toward vendor material and frozen on 1 October 2026.
### How the evidence was produced
Parallel research agents worked from one brief, and every output went to a second agent whose job was to refute it. The brief and the shared research rules (version 3, 28 September 2026) fixed the definitions, cohorts, source authority by claim type and evidence classes. The agents covered 11 workstreams, from market size to outcome evidence and agentic buying. Two were added during the research: row evidence for the scored products without a deep profile, and a targeted sweep of buyer surveys, run before writing because the first pass found only one survey that could test H1. A universe sweep classified vendors and market events and proposed additions to the product register.
Deep profiles cover 8 products: Anzu, Frameplay, Gadsme, Roblox's brand ads, XBOX Media Solutions, Twitch, Super League and DoubleVerify's gaming product, which is described without a score. Each profile carries a SWOT (strengths, weaknesses, opportunities and threats) in which every line has a source and a basis; opportunities and threats are the author's reading of dated facts. Other scored products received row evidence, gathered dimension by dimension against the anchors.
Every evidence file records each source's dates, funding and access limits and each finding's passage, evidence class and counterevidence; absence claims rest on logged searches. The register holds 982 sources, each opened or, where blocked, read by another route.
No verifier had seen the research before checking it. It opened the cited source, checked denominator, unit, date, entity and wording, and returned a verdict: supported as stated, supported with qualification, contradicted or unresolved. A restating correction becomes the claim's final wording, which the chapters use. Verifiers of profiles and row evidence also checked every proposed score against its anchor, cap and sources. Streams A to F were verified twice, a day apart, after a workflow re-run; the more severe verdict applies, and on a tie the newer pass. The 32 verifier files record 987 checks. A claim a verifier could not settle stays unresolved and never counts as supported; the ledger records 1 as unresolved.
*Figure fig-pipeline. **Every stage was built to remove things, and the counts are computed, not typed.** The five stages of this project with the actual outputs of each. Every claim in the ledger was checked by an AI verifier instructed to refute it; a claim the verifier could not settle stays unresolved and is never counted as supported. Corrections include score changes and the findings of the three AI-model reviews the author ran and the three internal AI review rounds. Units: counts. Scope: this research project, as of the cutoff. Caveat: Counts describe process volume, not evidence quality; a claim that survived verification is supported by its source, not proven true. Data: data/stats.json*
### Rules fixed before scoring
The hypotheses and the scoring rules were registered on 28 September 2026, before any research agent ran, and neither file changed afterwards. The dates are self-recorded in the files and the project's version history; no third party timestamped them. The hypothesis register holds 14 hypotheses and research questions, each with its population, comparator, outcome and coding rules. Absence of accessible evidence codes as unresolved.
The product register was written the same day and amended as research found products, with the changes logged in the corrections ledger. One product, Medal's game-adjacent advertising, came from the second outside review and was scored after the main pass under the same rules. The register holds 59 products, 28 of them listed and 8 historical. User-acquisition and paid-discovery products are never scored. Other services were classified by what the buyer purchases and where the ad renders, and evidence is never inherited between rows.
The 3 scored cohorts are in-game ad networks across independent titles (cohort A), brand ad products on in-game inventory a platform or publisher controls (B1), and game-adjacent media such as gaming sites, streams and esports (B2). Each cohort has an applicability mask, and n/a (not applicable) comes only from the mask. Cohorts A and B1 apply all 12 dimensions. For game-adjacent media, coverage (the count of evidenced device, distribution and game-environment combinations) is n/a.
Of the 12 dimensions, Depth covers format execution, targeting and context, measurement functionality, optimisation and outcomes, and creative and insertion tools. Reach covers supply, coverage, programmatic access and US availability, which counts documented US availability only. Standing covers independent assurance, transparency, and suitability, age and data controls. Every dimension has written anchors for scores 0 to 5, with numeric thresholds, and some anchors differ by cohort. A 0 needs a source documenting absence or failure; a cell with no basis to score is n/e (no evidence), which never counts as zero.
Each cell carries an evidence grade, from A (filing, regulator or standards-body record) through B (independent), C (vendor technical documentation) and D (vendor marketing) to E (none). The anchor is applied first and the cap second: grade C caps a cell at 4 and grade D at 3, so a 5 needs an official document, filing or independent observation. The caps fold the type of evidence into the score, so a cell's number mixes what is documented with how it is documented: a capability described only in marketing cannot score above 3, however complete the description. The supply anchors treat 20 million monthly and 5 million daily users as equivalent; that is an assumption of the rules, not a measured ratio. Only verified or qualified cells count.
Scoring was one pass on 29 September 2026, after the verifier records were read; score and grade changes that verifiers and the internal review rounds prompted are in the corrections ledger. Of 276 cells, 249 carry a score, 17 are n/e and 10 are n/a. On 1 October 2026, on the frozen evidence, separate AI scorers re-scored blind the 10 banded rows nearest a band line. Exact agreement was 85% of 115 cells, agreement within one point covered 106 of the 106 cells both passes scored, and the quadratic-weighted kappa, a measure of how far two scorers agree, was 0.9. The 5 material disagreements were adjudicated against the sources and changed no score; the two rows whose band would move are marked scorer-sensitive (chapter 11). No human scored any cell, and both passes are AI scorers reading the same rules, so this tests reproducibility, not independent judgement.
### How the Play Quotient is built and tested
The Play Quotient (PQ) summarises documented capability, not performance, within one cohort at a time. A product needs at least four counted numeric cells to enter the matrix.
Each group score is the mean of the row's numeric cells in that group; n/e and n/a are skipped. Raw = 0.45 × Depth + 0.25 × Reach + 0.30 × Standing, renormalised over the groups present. The field mean F is the arithmetic mean of the Raw scores of the cohort's admitted rows. Shrunk = (k × Raw + K × F) / (k + K), where k is the row's number of numeric cells and K, the prior strength, is 4. PQ = 100 × Shrunk / 5; only the displayed value is rounded. Bands come from the unrounded offset 20 × (Shrunk − F): Most documented at +12 or more, Well documented at +1 or more, Partly documented at −8 or more, Least documented below. The names describe how much is documented; the author's earlier papers call the same thresholds Front rank, Strong, Mid and Narrow. A row needs at least 6 numeric cells and one in every group for a band, and a cohort shows bands only with at least 6 band-eligible rows.
The field mean is the cohort's mean Raw score, the weighted group score before shrinkage. The Contextual and Native Quotients in the author's earlier papers pooled group scores instead; the first outside review showed a pooled mean could place six identical rows above the field mean. PQ values are therefore not comparable with those quotients. The prior strength is the author's heuristic, and no statistical model estimates it.
The deletion test turns each numeric cell into n/e in turn and reports quotient, admission and eligibility changes separately. Missing-evidence scenarios fill a row's applicable n/e cells with 1, then 4, with F and other rows frozen, and flag any band that moves; fills of 0 and 5 form a separate hypothetical envelope. Prior strength is rerun at K = 0 (Raw), 2 and 8. Reweightings of Depth, Reach and Standing at equal thirds, 60/20/20, 30/45/25 and 30/20/50 recompute Raw and F. Coverage is published for each row. [The capability matrix and the Play Quotient](#matrix) reports every result.
### Outside and internal reviews
The author ran three reviews with another company's AI model (OpenAI's), and every factual finding was tested against the primary source before anything changed.
The first round reviewed the briefs and rules before research began; its dispositions are dated 28 September 2026. It produced the product and hypothesis registers, explicit masks, the mean-Raw field mean, the K = 0 test, US-only geography and the evidence freeze. A proposed 75% coverage floor was not adopted; coverage is published instead. Its fact checks moved Bidstack to the historical rows after Bidstack Ltd and its parent entered creditors' voluntary liquidation on 18 June 2026 [32, 64].
The second round, on 29 September 2026, spot-checked the evidence. Every finding was upheld in full or in part and applied, tagged R2 in the corrections ledger. The findings restated the console coding as unresolved, fixed a stale H2 denominator, replaced a model-written summary quoted as a source passage, and added Medal.
Three internal review rounds followed, run with AI agents that read the draft chapter by chapter and across chapters for facts, rules, consistency, scoring and disclosure, with rechecks of status-sensitive facts on 30 September and, at the freeze, on 1 October 2026. Each finding was checked against the ledger and, for facts, against the primary source before any change, and each change is logged with its round (IR1 to IR3). The third round aligned this paper with the series' audio paper: band names that describe documentation, flags for provisional bands and the blind second scoring pass. They corrected scores that held some rows to a looser standard than others, a Louisiana law that never took effect and several summaries stronger than their chapters.
The third outside review read the draft. It asked for accreditation to be stated by format, for a buyer rationale for the anchors, for one rule on evidence age, for provisional bands to be marked, for one record per study, for buyer guidance by type of buy and for a shorter abstract; each point was checked against the data and sources, and the changes are tagged R3 in the ledger. The corrections ledger holds 143 entries; 9 carry a contradicted verdict, scoring entries number 5 and withdrawn claims 0.
### How the hypotheses were coded
H2 and H4 are supported and H1, H3 and H5 unresolved under their registered rules, and the GAP question cannot be answered from public sources.
H1 is unresolved [33, 34, 35]. Its survey leg is met only on point estimates: Advertiser Perceptions (200 respondents, July 2024) and IAB Australia (100 respondents, June and July 2023) each put brand safety first, but neither sample is established as US, the first lead is inside sampling error, the second is not established as larger than it, and counting the second rests on the inference that a reach option was offered [33, 34]. No US survey with a base, a period and constraint shares was found in the sources screened [109, 110, 111]. Without that inference only one usable survey remains, which the registered rule codes as unresolved; the inventory leg holds, but the rule needs both legs [33, 34, 35]. [How big it is, and by whose definition](#market) has the detail.
H2 is supported. The Intrinsic In-Game (IIG) Measurement Guidelines 2.0 of the Interactive Advertising Bureau (IAB), IAB Tech Lab and MRC are a final release dated August 2022 [17, 52]. MRC's record, read on 29 September 2026, shows no accredited intrinsic in-game metric and no MRC-confirmed audit for any product scored in cohorts A and B1; the one gaming audit in MRC's process is Trivver's, a listed product [15, 16]. Third-party verification rests mostly on partnership announcements dated 2022 to 2025 [350, 368, 369, 351]. The rule needs six scored products; 9 network and 4 platform products are scored. [Measurement](#measurement) has the detail.
H3 is unresolved under its registered rule [24, 25]. Microsoft's XBOX Media Solutions markets console brand integrations without naming a title, and its documents do not show whether they render inside gameplay [24]. Electronic Arts (EA) names brand integrations in two console titles and in its Madden NFL franchise, a fan database dates one in Madden NFL 26, and EA documents none as an ad placement buyable in the US, so at most three qualifying titles were found; that count is a floor, since other publishers' console titles were not screened [25, 238]. [Console, PC and cloud](#console) has the detail.
H4 is supported [20, 21, 22]. The conflict between Roblox's two revenue-share documents, which the registered rule names as an unresolved condition, concerns developer fees rather than any restriction counted here, so it does not block the verdict [57, 190]. Roblox bars personalised ads for users under 18 and Microsoft for accounts under 18, and Discord does not use its age-assurance signals for targeting [20, 21, 22]. The amended Children's Online Privacy Protection Act (COPPA) Rule took effect on 23 June 2025, with compliance due by 22 April 2026 for most entities [23, 421]. On 25 February 2026 the Federal Trade Commission (FTC) said it will not bring COPPA actions against mixed-audience or general-audience operators over data collected solely to determine age if six conditions hold, the first limiting that data to age verification [422, 423]. Utah's app-store law restricts developers' use and sharing of age-category data, with its duties beginning on 6 May 2027 [432, 433]. Documentation is not compliance. [Suitability, user-generated worlds and children's rules](#safety) has the detail.
H5 is unresolved overall. Part (a) holds on the registered wording: since 1 January 2021, primary records show three specialist companies sold or put into insolvency (Bloxbiz in October 2021, Bidstack from March 2024 and Let's Bounce in January 2026), ; Super League's purchase of the Misfits Ads business line in May 2026 is recorded and not counted, because it was a business line of an esports company [29, 31, 32, 30, 506]. Admix's merger with LandVault, announced in June 2022, is recorded and not counted, because the rule does not list mergers [507]. None of these records says the in-game format caused an exit, although Bidstack's administrators record that a 242-party sale process drew no third-party offer [31, 507]. Part (b) cannot be tested, because comparable platform and specialist revenue series are not disclosed; Roblox calls its ad revenue insignificant without a figure, and Microsoft and Amazon report it only inside broader lines [197, 198, 199]. [Exits, platform growth and the agentic question](#structure) has the detail.
The GAP question tests the author's model of in-game, around-the-game and away-from-the-game advertising [36]. No source screened by 28 September 2026 sizes the three as mutually exclusive, reconcilable categories, so GAP totals stay unresolved and are never added [37, 4, 38].
### Limits
The evidence is public, frozen on 1 October 2026 and weighted toward what vendors say about themselves. The registered scoring period runs to 2 October 2026; the paper was published at the freeze, so its last day is not covered. This review held no interviews and saw no private contract, so a vendor may hold privately what this paper marks n/e. Every absence is bounded to the sources screened by the cutoff. The rules require status-sensitive facts such as ownership and accreditation to be rechecked at the freeze, and later changes to be logged separately.
Access was limited: 103 of the 982 sources could be read only in part, and nothing is quoted from unread parts. EMARKETER's forecast values are for subscribers only [37, 87, 88]. Newzoo's site blocked the research tools, so the author confirmed its forecast of 3.7 billion players worldwide in 2026 and the article date of 25 August 2026 in a browser on 28 September 2026 [105]. In the universe sweep the web-search budget ran out before two broad acquisition queries ran, so the census of exits may be incomplete.
Vendor material dominates. The register holds 503 vendor sources, and the ledger codes 398 of its 807 claims as vendor assertions, against 155 direct records, 20 independent measurements and 19 commissioned measurements. The grade caps stop a vendor page from earning a 5; they add no independent evidence. This review found no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign with documented funding from a party without a financial interest, in the sources screened by 29 September 2026 [26, 27, 28].
Supply files are snapshots. A sellers.json file is where an ad seller declares the accounts it sells for. Adverty's file listed 121 publisher accounts in one version and 119 in the version regenerated at 12:00 Greenwich Mean Time on 28 September 2026 [13]. IION's file, last modified that day, still listed Bidstack as an intermediary seller after Bidstack Ltd had entered liquidation [11]. A declared account shows a supply connection, and nothing about titles, audience or active integrations [9, 10].
This is AI-assisted research. A model checking another model is no external audit: "verified" here means a second agent opened the source and read the passage. The paper is not peer reviewed, not externally audited and not independently validated. Attention metric definitions belong to the companion attention paper, in-game audio to the companion audio paper, and user-acquisition economics to the published mobile app growth paper.
### Interests
The author has worked at Samba (formerly Samba TV) since May 2025. At Verve Group he ran a P&L (profit-and-loss account) spanning mobile-gaming inventory. He co-leads the Signals & Measurement working group for AdCP (the Ad Context Protocol) at AgenticAdvertising.org. No vendor funded, reviewed or saw the paper before publication. Every vendor is scored under the same published rules.
## Appendix: instruments, data and rebuilding the paper
*The vendor universe, the full claim ledger, the corrections ledger and every data file behind the paper, in one place. The numbered source list follows as its own section.*
### Vendor universe
This is every product the review found and classified, scored or not. Each row gives the owner, where the ad renders, how it is bought, its status at the cutoff and why it sits where it does. Scored rows link to the capability matrix in chapter 11. A name that surfaced without an opened source is listed, not researched. User-acquisition and paid-discovery products, SSPs (supply-side platforms), DSPs (demand-side platforms), verification vendors and in-game audio are listed by rule, not by judgment of their quality.
**Vendor universe**
| Company / product | Legal entity / owner | HQ | Scope group | Formats | Buying routes | Main geographies | Status at cutoff | Src |
|---|---|---|---|---|---|---|---|---|
| Anzu in-game ad network | Anzu Virtual Reality Ltd. (private; shareholders not established) | Tel Aviv, Israel | network | intrinsic display and video in third-party games | direct and programmatic; 59 intermediary or both-type sellers in its sellers.json; NBCU US/UK resale reported in 2022 | global; US resale reported 2022 (not rechecked) | active (registry active; sellers.json served) | [510, 641, 9, 164] |
| Frameplay intrinsic in-game network (Frameplay Exchange) | not established | Oakland, California, US (sellers.json contact) | network | intrinsic in-game | direct and exchange; intermediaries AdInMo, Adverty, iion | US | active (sellers.json served) | [10, 628] |
| IION in-game brand network | Iion Pty Ltd / Iion Digital Pty Ltd (shareholders not established) | Sydney, Australia | network | in-game brand placements | sellers.json with 320 entries, 52 of seller_type INTERMEDIARY or BOTH (including Anzu, AdInMo, Adverty, Bidstack, Venatus) | APAC; US via Frameplay announced in future tense (brief) | active (ABN active) | [558, 525, 11] |
| Gadsme Managed Ad Network | Gadsme SAS (shareholders not established) | Paris, France | network | intrinsic display and video; audio to the audio paper | managed and programmatic; intermediaries Digital Turbine, LoopMe, Unity | France/EU; US office reported (headline only) | active (register etat A) | [524, 12, 467] |
| AdInMo in-game ad network (InGamePlay) | AdInMo Ltd (SC648449) | Edinburgh registered office, UK | historical | in-game brand ads in mobile games | listed as intermediary in Frameplay and IION sellers.json | UK/global; US not established | not live at the cutoff (strike-off applied 6 Sep 2026, suspended 17 Sep 2026) | [150, 642, 113, 10] |
| Adverty in-game ad network | Adverty AB (publ) (listed Swedish company) | Stockholm, Sweden | network | in-play ads in third-party games | US test route via Frameplay exchange (Q4 2024) | EU; US via Frameplay | active (issuer releases continue) | [628, 13] |
| Gameloft for brands (COMBO!) | Gameloft SE; Vivendi lists Gameloft among its assets | Paris, France | network | in-app brand ad units in Gameloft and third-party mobile games | direct; DV360 and The Trade Desk named as buying routes by Gameloft for brands | global; US not documented in sources opened | active | [524, 643, 183] |
| Super League in-game advertising | Super League Enterprise, Inc. (Nasdaq: SLE) | Santa Monica, California, US | network | brand placements inside Roblox, Minecraft and Fortnite experiences; Misfits game-portfolio brand rights | direct and managed | US | active; Super Biz platform being deprecated | [506, 511] |
| Zynga Ads+ (third-party publisher network) | Zynga (Take-Two-owned per AdExchanger) | not established | listed | mobile app ad inventory; in-game rendering not documented | DoubleVerify measurement partner; admission rule App Store rating 4.5+ | announced in New York | announced 17 Sep 2026; not shown live | [183, 526] |
| Kidoz kid-safe mobile ad platform | Kidoz Ltd (Israel 514191527); parent not established | Netanya, Israel (registry) | network | mobile ads in kid-safe games (vendor: 40,000+ games) | not screened | global; US not established | active (registry) | [68, 510, 526] |
| Admazing mobile gaming ads for brands | Admazing Co. | not established | network | brand ads in gaming apps | not screened | not established | active (site live) | [644, 526] |
| Roblox brand ads (Immersive Ads and Rewarded Video) | Roblox Corporation | US | platform | Immersive Ads, Rewarded Video | direct; DV360 and other programmatic platforms (per Roblox at PlayFronts); resellers Venatus (UK/FR/DE) and SuperAwesome (under-13, reported) | US and other markets | active | [630, 183, 204, 645] |
| XBOX Media Solutions in-game ads | Microsoft Corporation | US | platform | in-game video and interactive ads, custom integrations in Microsoft-owned games | direct (routes not re-read) | global; US not re-read | active (page live) | [191, 522] |
| Zynga Ads (Zynga's own games) | Zynga (Take-Two-owned per AdExchanger) | not established | platform | in-game ads across Zynga titles (vendor: 150+ games) | not screened | global | active (site live) | [646, 183] |
| Discord Quests | Discord Inc. (private) | San Francisco, California, US | adjacent | sponsored quests (per stream F) | per stream F | US and global | active | [647] |
| Twitch advertising | Amazon.com, Inc. | US | adjacent | stream ads (per stream F) | Amazon DSP (per stream F) | US and global | active | [2] |
| YouTube gaming inventory | Alphabet Inc. | US | listed | gaming-specific YouTube inventory (named product unestablished) | Google Ads / DV360 reservation (per stream F); third-party layer Precisify Gaming | global | no named gaming-specific product confirmed live at the cutoff (YouTube Select Shorts gaming lineup announced 17 Apr 2024) | [255] |
| Overwolf ads (companion apps and gaming websites) | Overwolf, Ltd. (Israel 514330331) | Ramat Gan, Israel | adjacent | display and video in desktop companion apps; Nitro websites; In-Game Audience Map planning tool (Sept 2026) | Deal ID/PMP (per stream F) | global; vendor says top-10 US game publisher (2024) | active | [648, 510, 293, 183] |
| Venatus gaming media | Media Scale Capital Limited (via Project Venus holding chain, from 19 Jun 2026) | London, UK | adjacent | gaming-website takeovers, video and display | direct and programmatic (per stream F) | global; US not established | active; change of control June 2026 | [302, 303, 649, 305] |
| Playwire gaming media | Playwire, LLC (owners not established) | Boca Raton, Florida, US | adjacent | ads on gaming publishers' sites and apps | PMP, programmatic guaranteed, direct (per stream F) | US | active | [650] |
| Enthusiast Gaming media (retained websites) | Enthusiast Gaming Holdings Inc. (TSX: EGLX) | Toronto, Canada | adjacent | gaming websites and tools after the 2025 divestment | web programmatic (per stream F) | US/Canada | active; direct-sales line sold 30 Aug 2025 | [317, 520] |
| ESL FACEIT Group esports media | Savvy Games Group (ESL sale closed 21 Apr 2022) | Cologne, Germany (website operator); ESL Gaming America Inc., New York | adjacent | esports broadcasts and owned channels; event sponsorship descriptive | direct | global; US entity exists | active | [331, 322] |
| IGN advertising (IMAGINE) | not established | US | adjacent | display ads with AI interactive units; audience intelligence platform IMAGINE | not screened | US | active | [183, 526] |
| LIVAD livestream video ads | LIVAD Technologies Inc. | not established | adjacent | video ads in live streams | not screened | not established | active (site live) | [273, 526] |
| Precisify Gaming | Precisify (owner not established) | not established | listed | contextual YouTube campaigns; gaming product named in menu | YouTube | not established | active (site live) | [651, 526, 652] |
| Dexerto | not established | not established | listed | gaming and esports news site (name only) | not screened | not established | sponsor logo only | [526] |
| APEX Gaming Network | not established | not established | listed | not established (name only) | not screened | not established | sponsor logo only | [526] |
| Publisher Collective (Network N) | not established | not established | listed | gaming websites (name only) | seller_type BOTH in IION sellers.json | not established | listed in sellers.json | [11] |
| Arkadium | not established | not established | listed | web games (name only) | seller_type BOTH in IION sellers.json | not established | listed in sellers.json | [11] |
| SuperAwesome (AwesomeAds; Roblox under-13 route) | SuperAwesome management (after Sept 2023 spin-off from Epic, as reported) | not established | listed | kid-safe ad marketplace; Roblox Immersive Ads partner | not screened | not established | active (reported 2024) | [645] |
| Aditude (including CPMStar) | Aditude | US | listed | SSP and direct sales for web games and gaming publishers | SSP | US | active (reported 2024) | [653] |
| GEEIQ | GEEIQ | not established | listed | planning and measurement data for brands in gaming (not a media seller) | n/a | not established | active (site live) | [654, 526] |
| DV360 and The Trade Desk (DSP routes to game inventory) | Alphabet; The Trade Desk, Inc. | US | listed | DSPs | named as buying routes by Gameloft for brands and Roblox | US/global | active | [183] |
| DoubleVerify gaming verification | DoubleVerify Holdings, Inc. | US | listed | verification | measurement partner of Zynga Ads+ | US/global | active | [183] |
| IAS gaming verification | Novacap (per brief) | US | listed | verification | headline 'IAS expands media quality solutions for in-game media buys' on Gadsme newsroom | US/global | active | [467] |
| Gadsme SSP | Gadsme SAS | Paris, France | listed | SSP | SSP | global | active | [524, 12] |
| Gadsme Cross Promo | Gadsme SAS | Paris, France | listed | cross-promotion (user acquisition) | n/a | global | active (per register) | [524] |
| Gadsme in-game audio ads | Gadsme SAS | Paris, France | listed | in-game audio | headline 'Targetspot and Gadsme join forces to launch in-game audio ads' | global | active (per register) | [524, 467] |
| IION gaming-website supply | Iion Pty Ltd / Iion Digital Pty Ltd | Sydney, Australia | listed | gaming-website display and video | Prebid and Amazon TAM (per brief) | global | active | [558, 11] |
| Audiomob and Odeeo in-game audio | not established by U | not established | listed | in-game audio | both appear as intermediaries in IION's sellers.json; Audiomob also in Anzu's | global | active (listed in sellers.json) | [11, 9] |
| Venatus Roblox Rewarded Video resale | Media Scale Capital Limited | London, UK | listed | Roblox Rewarded Video | reseller | UK, France, Germany only | active (announced May 2026) | [204] |
| Admix (admixplay.com) | operator named on site: W Media Ltd / W Media International Ltd (15434064); link to Admix unestablished | London, UK | listed | mobile gaming ads for non-endemic brands (site copy) | embedded seller list of 24 publishers; /sellers.json 404 | not stated | site live | [516, 517] |
| Roblox Sponsored Experiences | Roblox Corporation | US | listed | paid discovery | n/a | not re-read | per register | [630] |
| Fortnite Sponsored Row | Epic Games | US | listed | paid discovery bought by island owners | n/a | not re-read | live 21 Nov 2025 (brief) | |
| Xbox dashboard and store placements; Xbox cloud-streaming ad test; Live TV on PS5 | Microsoft; Sony | US/Japan | listed | console placements outside gameplay; pre-session ads in test; TV inventory | n/a | per stream G | per brief | |
| App-install networks (AppLovin, Unity and others) | various | various | listed | user acquisition | n/a | global | active | [9, 12] |
| Bidstack | Virtual Sport Technology Limited (from 22 Mar 2024); formerly Bidstack Group Plc | London, UK | historical | former in-game ad network | still listed as intermediary in IION sellers.json | n/a | in creditors' voluntary liquidation since 18 Jun 2026; plc dissolved 26 Jun 2025 | [32, 519, 64, 509, 512, 31] |
| Sayollo | Sayollo Media Ltd (Israel 514656453) | Rehovot, Israel | historical | former in-game ad company | n/a | n/a | in liquidation by court; order to open proceedings (undated) | [510] |
| Super Biz (formerly Bloxbiz) | Super League Enterprise, Inc. | Santa Monica, California, US | historical | former Roblox in-game ad platform | n/a | n/a | being deprecated (undated notice) | [29, 511] |
| Misfits Ads | assets bought by Super League Enterprise, Inc. from Esports Now, LLC | US | historical | former Roblox in-game ad business | n/a | n/a | assets sold 1 May 2026 | [506] |
| LandVault (formerly Admix) | Infinite Reality announced acquisition 9 Jul 2024 | London, UK / Dubai (not established) | historical | former in-play placements; later metaverse building | n/a | n/a | merged and rebranded June 2022; acquisition announced July 2024 | [507, 515] |
| Oracle Moat | Oracle Corporation | US | historical | former measurement and verification | n/a | n/a | end of life 30 Sep 2024 | [339] |
| EA Advertising | PIF, Silver Lake and Affinity Partners (Electronic Arts private since August 2026) | US | platform | dynamic ad boards, scoreboards and broadcast-style overlays in select EA SPORTS games; custom brand integrations | direct; mobile programmatic reported September 2026 | US and other | live (launched 15 Jun 2026) | [25, 241] |
| Trivver in-game ads | Not established | US (Newport Beach, per California registration) | listed | intrinsic in-game placements (SDK) | not documented | unconfirmed | no live brand-selling network documented in the period; MRC audit in process | [16] |
| Medal advertising (ad units and sponsored placements) | Not established (Medal B.V.) | Netherlands | adjacent | IAB display and video units, homepage roadblock, sponsored stories, sponsored top 10, branded watermarks, clip contests | direct; programmatic sellers in ads.txt | global; 45% US users (vendor) | live | [252, 290] |
| Medal Quests | Medal B.V. | Netherlands | listed | reward quests with game developers | direct | global | live | [58] |
| ESL FACEIT Group custom brand integrations | Savvy Games Group | Germany | listed | custom in-game placements, Fortnite worlds and maps, event activations | direct | global | offered | [331] |
| Let's Bounce | Super League Enterprise, Inc. | US | historical | brand engagement in gaming and UGC environments | | US | asset acquisition by Super League, January 2026 | [128] |
### Claim ledger
Every claim the paper rests on, with its sources, its evidence class, the verifier's verdict and its final wording after verification and review. Filter by verdict or evidence class. A claim no verifier checked would show as unresolved; the ledger holds 807 claims, of which 807 were checked by a separate verifier.
**Claim ledger**
| ID | Claim (final wording) | Ch. | Evidence class | Review | Confidence | Sources |
|---|---|---|---|---|---|---|
| GAM-B-F10 | Anzu's sellers.json lists 58 INTERMEDIARY sellers, including Unity, Digital Turbine, Liftoff, InMobi, Ogury, Smaato and Audiomob, as well as website and streaming properties (WhoScored, howdidido, WeDoTV, iQIYI), and publishers such as King, Rovio, Zynga, Outfit7, Gameloft and Miniclip authorise anzu.io as a RESELLER under those intermediary IDs; Anzu itself says (9 February 2026) that it bundles intrinsic formats with traditional ad units such as interstitials and rewarded video, so part of Anzu's declared supply path carries impressions that are not intrinsic placements, and part is not game inventory at all. | 1 | independent measurement | supported with qualification | medium | [9, 53, 54] |
| GAM-B-F26 | COMBO! sells in-app ad units rather than intrinsic placements: classic display, rewarded and non-rewarded video, interactive video formats and playable ads, across a stated "+190" games portfolio, with targeting by device settings, demographics, time and location, and game selection; the page's footnotes state that no device data, user IDs or personal data are collected for users under 18, that age and gender are not collected for marketing or advertising from under-18s, and that precise location is not collected for any user. | 1 | vendor assertion | supported with qualification | medium | [55] |
| GAM-G-F20 | adWMG wrote in paid Drum Network content dated 2 December 2025 that its May 2025 partnership with Gadsme unlocked immersive ad formats 'within Xbox and PlayStation titles', but the partnership's own May 2025 announcement (ExchangeWire, 23 May 2025) describes the reverse: adWMG's Xbox and PlayStation clients, meaning console-game advertisers, buying Gadsme's mobile in-game inventory to acquire players. It is user acquisition on mobile inventory and is not evidence of Gadsme placements inside console titles. | 1 | vendor assertion | supported with qualification | low | [40] |
| GAM-G-F24 | Roblox's immersive-ads documentation does not say whether immersive ads (billboards, portals, rewarded video) render on console clients, so that stays unresolved from the sources opened; Roblox's own 4 February 2026 Ads Manager announcement does list 'console' as a device-targeting option, but for Plays-objective campaigns placed on Home and Search, which are paid-discovery placements outside experiences and listed rather than scored. | 1 | vendor assertion | supported with qualification | low | [407] |
| GAM-I-F25 | Unity's and AppLovin's FY2025 10-Ks describe advertising products for app developers (Unity's Grow Solutions; AppLovin's Axon Ads Manager, which AppLovin calls its user acquisition solution) and, in the text searched, name no separately named brand-advertising product sold in games; Unity's 10-K does describe an expansion into programmatic advertising and CTV as 'gaming-adjacent verticals' whose success depends on adoption 'by agencies and brands', and AppLovin's describes expansion to web-based e-commerce advertisers. | 1 | direct record | supported with qualification | medium | [60, 61] |
| GAM-R-F137 | One coverage tuple is evidenced for Adverty in the period (mobile, app store, casual title: Parking Master 3D); its web-game solution is in full production and its SDK supports WebGL and Windows, and its policy admits apps from the Apple, Google, Oculus and Steam stores, but no named live title documents a web, PC or VR tuple. | 1 | synthesis | supported as stated | medium | [79, 80, 81, 82] |
| GAM-GADSME-F27 | Gadsme's audio and Cross Promo products run through the same contracts and supply-path files as the managed network: its formats list includes 'Audio integrations', its terms define an Audio Ad Format counted at audio start and a CPI-priced Cross Promo service, Helium9 Games authorises 'adswizz.com, gadsme, DIRECT', and Gadsme's app-ads.txt carries 45 adswizz.com lines including one for targetspot; these components belong to the audio paper and to the listed Cross Promo row. | 1 | synthesis | supported as stated | high | [95, 59, 96, 97] |
| GAM-ROBLOX-F27 | Roblox also sells brand placements outside gameplay, which this paper's placement field does not count as in-game: Homepage Hero (a 24-hour homepage takeover), Homepage Feature (beta; a template-built branded destination) and Home and Search tiles; its Q2 2026 letter describes Moments as a future 'premium advertising surface', initially for age-checked users over 16 in select markets. | 1 | vendor assertion | supported as stated | medium | [42, 43] |
| GAM-A-F19 | EMARKETER frames a gap between gaming time and ad spend (2024: mobile games 10.7% of US mobile time against 3.8% of mobile ad spend; 2026: more than an hour of gaming a day against 2.3% of digital ad spend) and, in its FAQ, says the gap 'suggests marketers lack confidence in measurement or struggle with fragmented buying processes' without citing a study. Under H1's registered rule, this gap is context and the attribution is not evidence of buyer constraints. | 1, 2 | inference | supported as stated | high | [5, 92] |
| GAM-A-F20 | Sensor Tower's State of Gaming 2026 blog (February 2026) says the majority of mobile game ad spend flows to mobile app networks such as AppLovin and AdMob, while PC and console spend is concentrated in social and video channels, with YouTube taking the largest share; from that context, 'game ad spend' here means spending by game advertisers to promote their games, which the blog does not define. 'Gaming ad spend' therefore has two opposite meanings in published sources (spend placed in games, and spend by game companies), and they must not be mixed. | 1, 2 | vendor assertion | supported with qualification | medium | [39] |
| GAM-A-F08 | No source screened by 28 September 2026 sizes in-game, around-the-game and away-from-the-game advertising as mutually exclusive, reconcilable categories. EMARKETER comes closest: its total US game ad spend series is reported (Digiday, 3 July 2026) to exclude ads on esports, gaming video content and creator sponsored content, and it defines separate mobile in-game, esports and gaming video content series, but most values are paywalled and the esports and gaming video content overlap is not settled. PwC's published video games and esports segment (used by IAB/PwC) combines in-game and esports advertising, although PwC's definitions name separate 'in-app games advertising' and 'esports streaming' components inside internet advertising whose values are not public. WPP Media publishes one undefined gaming line, and no publisher sizes an 'away-from-the-game' category. | 1, 2, 5, 12, 13 | inference | supported with qualification | medium | [37, 4, 38] |
| GAM-A-F09 | The playbook's GAP categories are not mutually exclusive on the placement axis: 'Away-from-the-game' lists app install and re-engagement, which is an objective, while EMARKETER says most in-game ads promote installs of other mobile games, so an install ad shown inside a game fits both 'In-game' and 'Away-from-the-game'. This paper's separate placement and objective fields avoid that overlap. | 1, 2, 12 | inference | supported as stated | medium | [36, 8] |
| GAM-A-F11 | WPP Media's 2026 midyear forecast (16 June 2026) says in its text only that gaming continues to grow within content-driven advertising, and the full forecast is for WPP clients only. The public report page embeds an interactive chart ('Fastest to 100') with a Gaming series in US$ millions indexed by years since the series start, without calendar years, so no calendar-year 2026 midyear gaming figure can be read from the public pages without an unconfirmed anchoring assumption. WPP's forecast carries a named gaming channel as an ad-revenue forecast line, not a published planning media mix. | 1, 2, 12 | forecast | supported with qualification | medium | [89, 90] |
| GAM-A-F18 | The playbook's 'under 5% of worldwide media investment' matches an unattributed line that EMARKETER's January 2026 FAQ repeats twice ('less than 5% of worldwide media investment'), and the figure most likely originates in dentsu's 2025 Gaming Trends Report ('Gaming, Your Marketing Cheat Code', 1 July 2025), which Variety reports as 'less than 5% of global media investment'. The trace is inferred from wording and dates, since neither the playbook nor the FAQ cites dentsu for the figure. None of the accessible texts states a denominator, year or definition of gaming media investment, and dentsu's global gaming lead called it 'only 5%' in the same Variety interview while defining gaming to include apps like Duolingo and crossword puzzles, so the figure cannot be used as a sized share under shared rule 2. | 1, 2, 12 | vendor assertion | supported with qualification | medium | [36, 92, 93, 94] |
| GAM-A-F07 | EMARKETER's forecast database defines three separate US gaming-related ad series: mobile game ad spending (ads placed within games on mobile devices only, excluding esports and gaming video content), esports ad spending (advertising and sponsorship revenues from organised esports competitions on digital platforms, excluding linear TV) and gaming video content ad spending (video ads on recorded and livestreamed game content on digital platforms such as YouTube or Twitch); the values are subscriber-only. | 1, 2, 13 | synthesis | supported as stated | high | [37, 87, 88] |
| GAM-B-F04 | Frameplay and iion announced on 16 June 2025, in the future tense, that iion's 4,000+ gaming titles (vendor-reported) 'will be accessible through the Frameplay Exchange in the US'; the Frameplay Exchange itself opened on 9 October 2024 with Frameplay, Adverty and AdInMo inventory for programmatic and direct buying. | 1, 3 | vendor assertion | supported as stated | medium | [73, 74] |
| GAM-B-F05 | The IION-to-Frameplay resale path is declared in publisher files: Frameplay's sellers.json (last modified 19 August 2025) assigns seller_id 9251981480 to iion as an INTERMEDIARY, and the app-ads.txt files of Miniclip, Outfit7, Zynga and Gameloft (4 of 10 publisher files sampled on 28 September 2026) authorise 'frameplay.gg, 9251981480, RESELLER'; no source found shows US delivery, volume or a go-live date. | 1, 3 | synthesis | supported as stated | medium | [10, 75, 76, 77, 78] |
| GAM-R-F301 | Kidoz is live in the scoring period and fits cohort A: Kidoz Inc.'s FY2025 20-F (filed 29 April 2026) says its SDK is integrated into thousands of mobile apps and that its in-game advertising arrangements are mostly governed by insertion order, with a separate programmatic revenue stream, and its Form 6-K for the half-year to 30 June 2026 (filed 26 August 2026) reports revenue for that period. Its sellers.json, read on 29 September 2026, lists about 270 publisher accounts (268 to 269 depending on the read), most of them third-party developers. | 1, 3 | synthesis | supported with qualification | high | [65, 66, 67] |
| GAM-R-F308 | Kidoz's supply figures are vendor counts that disagree: the site claims 40,000+ kid-safe games and 500M+ monthly reach (and '500M monthly active users' on the Shield page); the 12 June 2026 investor fact sheet, which describes one platform serving both Kidoz (children) and Prado (all ages), claims 100K+ apps and 500M+ global monthly unique users; the SDK README says over 350 million a month; and the 20-F says 'hundreds of million' a month. The only game the vendor names as on its network is Minion Rush (undated alt text). | 1, 3 | vendor assertion | supported with qualification | low | [68, 69, 70, 71, 65, 72] |
| GAM-B-F09 | IION's sellers.json, last modified on 28 September 2026, still lists Bidstack (bidstack.com) as an INTERMEDIARY seller, although Bidstack Ltd is recorded in this paper's product register as in creditors' voluntary liquidation since 18 June 2026. | 1, 3, 10, 13 | independent measurement | supported as stated | high | [11] |
| GAM-B-F01 | In files downloaded on 28 September 2026, six cohort A candidates published sellers.json files listing, as PUBLISHER seller accounts, 1,117 for Anzu, 268 for IION, 223 for Gadsme, 121 for Adverty (119 in the version regenerated at 12:00 GMT on 28 September 2026), 31 for Gameloft (COMBO!) and 30 for Frameplay; these are seller accounts declared by each vendor, not titles, ad-eligible audience or proof of active monetised integrations, and Adverty's count changes from day to day. | 1, 3, 13 | independent measurement | supported with qualification | high | [9, 10, 11, 12, 13, 14] |
| GAM-F-F37 | Venatus packages its offer in the same three categories as the author's GAP model (In the Game, Around the Game, Away from the Game), but its categories overlap with this paper's placement definitions: its in-the-game page includes in-game audio and bespoke experiences, and its away-from-the-game formats embed Twitch live streams in website takeovers. | 1, 5 | inference | supported as stated | medium | [83, 256, 84] |
| GAM-R-F402 | Discord's Quests business mixes objectives: its advertiser FAQ says it is currently focusing on gaming and media and entertainment customers, and of the 21 success stories on its ads page on 29 September 2026, 14 are games (several reporting acquisition or sales outcomes, such as 4.5x net ROAS for Mouse P.I. For Hire), 4 are film or TV releases and 3 are non-entertainment consumer brands (Mountain Dew, Wendy's, Walmart, the last described as the first brand to use Video Quests); Paramount used a Play Quest for a film premiere. | 1, 5 | synthesis | supported with qualification | medium | [62, 48, 63] |
| GAM-R-F502 | Inference: Overwolf's app ads are rendered by the Overwolf Ads SDK inside a third-party companion app's window, which the app developer places and which Overwolf describes as full-screen and 'in-game', so the ad is drawn by the companion app over or beside the game rather than by the game; this review keeps the product in cohort B2 as a recorded judgment, and excludes Scalibur (open-web audience extension) and the custom Minecraft and Fortnite activations from this row. | 1, 5 | inference | supported as stated | medium | [44, 45, 46] |
| GAM-NFA-F02 | As served on 29 September 2026, Medal B.V.'s medal.tv/grow page sells placements rendered in its own gameplay-clip app and website, not inside games: IAB-standard ad units including a homepage roadblock, sponsored stories, sponsored top 10, branded watermarks and clip contests. Its ads.txt lists direct programmatic sellers (Google, OpenX, Magnite, PubMatic), and it claims 10M MAU, 45% US users and 4.5B monthly impressions (vendor assertions; base and period unstated). Its only case study (Paddington, undated) reports game-promotion outcomes (the game's concurrent users and time in game), and its Quests are partnerships with game developers, so they are listed as paid discovery. | 1, 5 | vendor assertion | supported as stated | medium | [252, 655, 58, 290] |
| GAM-C-F01 | The Intrinsic In-Game Advertising Measurement Guidelines 2.0 are a final joint release of the IAB, IAB Tech Lab and MRC: first issued September 2009, opened for public comment on 14 June 2022 until 15 July 2022, released as version 2.0 in August 2022 (IAB release page dated 31 August 2022), and listed by MRC on 28 September 2026 as last updated August 2022. | 1, 6 | direct record | supported as stated | high | [17, 50, 51, 52] |
| GAM-C-F02 | The guidelines cover impression, viewable-impression and audience measurement of in-game ad placements served by an ad server with a corresponding ad session, in desktop web and application, mobile web and application, and console environments (including AR/VR headsets); hard-coded branded elements are not directly within scope because they generally lack an ad session (future versions may consider them, and any measurement of them is expected to follow this guidance); interstitials, web banners and in-stream or outstream video are excluded and fall under the existing IAB and MRC desktop, mobile web and mobile in-app guidelines, although the guidelines say their in-game-specific guidance should also be considered for those placements; interaction, engagement and outcomes are not covered. | 1, 6 | direct record | supported with qualification | high | [17] |
| GAM-F-F38 | Two B2 products reach console players without rendering inside console gameplay: Twitch Premium Video runs in the Twitch app on consoles (not clickable there), and Discord Play Quests reward play that can happen on console while the Quest itself renders in Discord, which says it has native integrations with Xbox and PlayStation. | 1, 8 | inference | supported as stated | medium | [47, 48, 49] |
| GAM-A-F01 | EMARKETER's FAQ of 10 January 2026 says US game ad revenues "will surpass $10 billion by 2029" and that in-game advertising will be 2.3% of overall US digital ad spending in 2026; it does not define game ad revenues or split brand from user-acquisition spend. | 2 | forecast | supported as stated | high | [92] |
| GAM-A-F02 | EMARKETER's article of 26 December 2025 says US game ad revenues "will reach nearly $10 billion in the US by 2029" and that in-game advertising will be 2.3% of overall digital ad spending in 2026, declining to 2% by 2029. | 2 | forecast | supported as stated | high | [99] |
| GAM-A-F10 | WPP Media's This Year Next Year end-of-year forecast (10 December 2025) puts global gaming ad revenue at $8.5 billion in 2025, up 29.5%, the fastest-growing content channel and 0.7% of total global ad revenue; Marketing Brew reported the same forecast at $10.7 billion for 2026. WPP Media's public pages give no definition of gaming. | 2 | forecast | supported as stated | medium | [38, 102] |
| GAM-A-F12 | PwC's Global Entertainment & Media Outlook 2025-29 release (24 July 2025) puts total global video games revenue at $224 billion in 2024, rising to nearly $300 billion in 2029 (whole-industry revenue including consumer spending), and says 'advertising in video games' will rise from 32.8% in 2024 to 38.5% in 2029 without stating the base of the percentage. PwC's Outlook 2026-30 release (issued by 24 June 2026; the BizTechReports reproduction read here is dated 9 July 2026) gives no video games advertising figure in its reproduced text. | 2 | forecast | supported with qualification | medium | [103, 104] |
| GAM-A-F13 | The IAB/PwC Internet Advertising Revenue Report for full-year 2025 (April 2026), a report on US advertising revenue, says 'video games and esports digital and non-digital advertising revenue grew 22.0% in 2025, accelerating from 8.7% in 2024', sourcing the line to PwC's Entertainment and Media Outlook; the chart labels neither geography nor Outlook edition, and whether the 2025 value is a reported actual or a model estimate is not stated. The report publishes no dollar value for this line, and it defines a 'Gaming (In-App, PC & Console)' survey format but reports no revenue for that format. | 2 | synthesis | supported with qualification | medium | [4] |
| GAM-A-F14 | EMARKETER forecasts that 59.2% of the US population will be digital gamers in 2026 (October 2025 forecast); it defines digital gamers as internet users of any age who play online or internet-updated games on any device at least once a month. Its August 2023 forecast gave 54.6% for 2024. | 2 | forecast | supported as stated | high | [92, 100, 5] |
| GAM-A-F15 | ESA's 2026 Essential Facts reports that 212.3 million Americans aged 5 to 90 play video games and that 67% of the US population play at least one hour a week, based on a YouGov online survey of 13,545 respondents fielded 11-25 February 2026; ESA's 2025 page gave more than 205 million. | 2 | commissioned measurement | supported as stated | medium | [3, 545] |
| GAM-S-F01 | Advertiser Perceptions' 2024 Gaming Advertising Report, a survey of 200 marketer and agency contacts from its AdPROs community fielded 8-17 July 2024, lists five biggest obstacles to gaming advertising: brand safety concerns 29%, can reach the same consumers more effectively elsewhere 26%, difficult to measure 26%, lack of scale or inadequate reach 26%, and organisational resistance 25%. | 2 | vendor assertion | supported as stated | medium | [33, 35] |
| GAM-S-F02 | Advertiser Perceptions' own summary of its 2024 gaming report (9 September 2024) says brand-safety concerns about gaming are fading while the perception that the same audiences can be reached more effectively elsewhere is a growing challenge; MediaPost's column on the report says brand safety is less likely to be the top concern than the year before. No 2023 shares are published, so the change cannot be quantified. | 2 | vendor assertion | supported as stated | medium | [101, 33] |
| GAM-S-F03 | IAB Australia's Game Advertising State of the Nation 2023 (wave 3; 100 advertising decision makers in agencies and brands, fieldwork June-July 2023) asked for the three biggest barriers to more investment in game advertising: brand safety 47%, measuring campaign success 36%, brand suitability 36%, evidence of effectiveness for brand and/or ROI 34%, and how it can work in combination with other media 32%. No reach or scale barrier appears among the five shown. | 2 | commissioned measurement | supported as stated | high | [34, 91] |
| GAM-C-F08 | On 28 September 2026 MRC's list of accredited digital services uses only four environment codes (desktop, mobile web, mobile in-app, connected TV), contains no gaming, in-game or console entry, and does not list any of the ten cohort A or B1 products in the product register (Anzu, Frameplay, IION, Gadsme, AdInMo, Adverty, Gameloft, Super League, Roblox, XBOX Media Solutions). | 2, 6 | direct record | supported as stated | high | [15] |
| GAM-A-F06 | EMARKETER's March 2024 forecast put the year US game ad revenues pass $10 billion at 2026, but its later public statements give no single revised crossing year: the 26 December 2025 article says revenues 'will reach nearly $10 billion in the US by 2029' and the 10 January 2026 FAQ says they 'will surpass $10 billion by 2029', although both link to the same EMARKETER forecast series ('US Total Game Ad Spending'), so the two statements conflict and the forecast vintage behind each is unverified. EMARKETER data reported by Digiday on 3 July 2026 (vintage not stated) put total US game ad spend, excluding esports, gaming video content and creator sponsored content, at $9.21 billion in 2026, $9.7 billion in 2027 and $10.2 billion in 2028; the underlying forecast values are subscriber-only, and the public texts do not say whether the changes reflect lower expected spend, a changed definition or restated history. | 2, 12 | synthesis | supported with qualification | medium | [5, 6, 108, 99] |
| GAM-A-F16 | Published player counts use different definitions and none is an ad-eligible audience: EMARKETER counts monthly internet gaming (59.2% of the US population in 2026, forecast), ESA counts play of at least one hour a week among ages 5 to 90 (212.3 million, 67%, 2026 survey), Newzoo forecasts 3.7 billion players worldwide in 2026, and dentsu cites over 3.4 billion gamers worldwide (the same Variety article's opening says more than 3.5 billion). | 2, 12 | synthesis | supported as stated | high | [100, 3, 93] |
| GAM-A-F17 | IAB's 'Changing the Game' study (26 March 2024), a survey by Advertiser Perceptions of 300 brand and agency decision makers who already invest in games advertising, found 86% agree games advertising is brand safe, 85% are confident in measuring it and 79% find it relatively easy to buy, though games advertising 'still falls behind other digital media channels concerning ease of buying and planning'; 40% planned to increase spend. | 2, 12 | commissioned measurement | supported as stated | medium | [656, 108] |
| GAM-S-F06 | IAB Australia's July 2023 report classes game advertising into 'in the game' (gameplay visuals or audio, skins, sponsored content), 'around the game' (during the gaming experience but outside gameplay, such as in-app rewarded video, interstitials and overlaid audio ads) and 'away from the game' (streaming, esports, influencer and content marketing). These are the same three names as the author's GAP model, but with different contents: the playbook puts rewarded formats in 'In-game' and app install in 'Away-from-the-game' (GAM-A-F09). | 2, 12 | synthesis | supported as stated | high | [34, 91] |
| GAM-S-F08 | In IAB's March 2024 study (300 US current buyers), 86% agreed gaming is brand safe (36% strongly), second to digital audio at 89%; 79% agreed games advertising is easy or uncomplicated to buy, tied with CTV/OTT (79%), above online video (75%) and below digital audio (82%), social media (84%) and digital display (91%), with the lowest 'strongly agree' share of the six channels (30%). | 2, 12 | commissioned measurement | supported as stated | high | [107] |
| GAM-S-F09 | IAB's March 2024 study (with Advertiser Perceptions; sponsored by Admazing and Venatus) found 51% of advertisers treat games as part of another media channel and 49% as a channel of their own; among those who treat games as part of other channels, the four reasons listed, unnumbered and without shares, begin with 'not yet enough scale', followed by 'has same ad formats as other media', 'runs on multiple devices / can't buy it as a single channel' and 'not yet cost effective on its own'. | 2, 12 | commissioned measurement | supported as stated | medium | [107] |
| GAM-A-F22 | Newzoo forecasts 3.7 billion players worldwide in 2026, up 4.4% (article dated 25 August 2026); this is a global player forecast, not US reach and not an ad-eligible audience. | 2, 13 | forecast | supported as stated | medium | [105] |
| GAM-S-F17 | Under simple-random-sampling assumptions, the 95% margin of error on a single share is about ±6 percentage points for a 26-29% share on Advertiser Perceptions' base of 200 and about ±9-10 points for a 32-47% share on IAB Australia's base of 100. Whether a gap between two answers to the same multi-response question is distinguishable depends on how many respondents chose both, which neither publisher reports. Brand safety's 3-point lead over lack of scale (29% vs 26%) is very unlikely to be a distinguishable rank. IAB Australia's 15-point lead of brand safety over the fifth-ranked barrier (47% vs 32%) has a 95% half-width of about ±7 to ±17 points depending on that overlap (about ±13 points if the choices were independent), so it is not established as a distinguishable rank either. | 2, 13 | inference | supported with qualification | medium | [33, 34] |
| GAM-S-F19 | In the sources screened by 28 September 2026, this review found no US-specific survey that publishes a base, a fieldwork period and shares for constraints on brand advertising in games, and no 2025 or 2026 survey on the subject with public constraint shares; Advertiser Perceptions' 2025 gaming results, IAS's 2026 Industry Pulse and IAB's August 2026 'Business Case for Gaming' could not be read in full. | 2, 13 | synthesis | supported with qualification | medium | [109, 110, 111] |
| GAM-B-F03 | Cohort A networks declare one another as INTERMEDIARY sellers: Frameplay's sellers.json lists iion, Adverty and AdInMo as INTERMEDIARY sellers, and IION's sellers.json lists Anzu, AdInMo, Adverty and Bidstack as INTERMEDIARY sellers. These declarations authorise resale; they do not show which inventory is actively resold, so a buyer of one network may be buying another network's supply (an inference). | 3 | independent measurement | supported as stated | high | [10, 11] |
| GAM-B-F07 | The latest public list of Frameplay DSP integrations found is from December 2021 (Magnite, Adelphic, Yahoo, InMobi and The Trade Desk, plus its own ad server; buys then through private marketplaces or direct deals, as reported by AdExchanger); Frameplay's current advertiser page quotes GumGum as continuing to offer Frameplay inventory to its clients, but this review found no DSP's own documentation confirming a current Frameplay or Frameplay Exchange integration in the sources screened by 29 September 2026. | 3 | synthesis | supported with qualification | low | [177, 119] |
| GAM-B-F08 | IION's public scale claims carry no metric or period and do not describe its own ad-eligible audience: its home page says "60,000+ games", its June 2025 release said "4,000+ gaming titles", and its advertisers page says it is "Home to a premium audience of 3.6 billion", labelled "Unique Audience" in a block about gaming in general; the figure matches common estimates of the world's player population (this review's inference), and IION does not claim it as its own reach. Its offer is organised as in-game, around the game and away from the game, of which only the in-game line falls in cohort A. | 3 | vendor assertion | supported with qualification | high | [115, 116, 73] |
| GAM-B-F11 | Anzu's direct publisher relationships are confirmed from both sides for seven sampled publishers: Miniclip, Zynga, Rovio, Vector Unit, Illusion Labs, Gameloft and Helium9 Games each authorise an anzu.io DIRECT seller ID that Anzu's sellers.json assigns to that publisher. | 3 | synthesis | supported as stated | high | [75, 77, 54, 78, 96, 9] |
| GAM-B-F15 | Google's Display & Video 360 API v4 Exchange enum (last updated 10 June 2026) lists 90 values (89 named exchanges plus EXCHANGE_UNSPECIFIED), none of which names Anzu, Frameplay, Gadsme, Adverty, AdInMo, IION or Gameloft, so none of these networks is a named DV360 exchange in that record; several intermediaries in their sellers.json files (Unity, Digital Turbine/Fyber, InMobi, Smaato) are named DV360 exchanges, as is GumGum, which Frameplay quotes as a reseller. | 3 | independent measurement | supported with qualification | high | [119] |
| GAM-B-F16 | Gadsme states that all its publisher relationships are direct and exclusive, with "100% exclusive inventory from 200+ game studios" and "No reseller stacking", but its own sellers.json lists three INTERMEDIARY sellers (Digital Turbine at fyber.com, LoopMe, Unity); King, Miniclip, Rovio, Gameloft and Zynga authorise gadsme.com as a RESELLER under the Digital Turbine ID; and 60 of its publisher domains also appear in Anzu's sellers.json, including Gameloft and Miniclip, which Gadsme names among the publishers of its "more than 1,600 mobile games" (reported October 2025). | 3 | synthesis | supported with qualification | medium | [118, 170, 12, 53, 75] |
| GAM-B-F17 | Sports Interactive, the game's publisher, announced on 29 September 2025 an exclusive three-year partnership under which Gadsme's SDK is integrated into the PC/Mac edition of Football Manager 26, which is publisher-confirmed selling rights and exclusivity for one premium PC title; a Gadsme-delivered programmatic campaign for Easy Cash in France ran in FM26 match cutscenes by 11 December 2025, according to partner content co-produced by Gadsme and the agency, so live placements inside the scoring period rest on a commissioned source, with no US statement. | 3 | vendor assertion | supported with qualification | medium | [140, 467] |
| GAM-B-F19 | Gadsme documents three buying routes (its own DSP, market SSPs and DSPs, and private marketplaces) across mobile, PC and console, and its programmatic page names three deal types (private marketplace deals, curated deals and direct programmatic partnerships); one DSP, Eskimi, documents a Gadsme integration on its own site (page updated 6 August 2025, content from 2022), and no other DSP's documentation of a Gadsme integration was found. | 3 | vendor assertion | supported with qualification | medium | [118, 176] |
| GAM-B-F21 | Adverty reported net sales of SEK 12.9 million for Q2 2026 (KSEK 12,928 against 7,228), up 79% year on year, and SEK 20.7 million for January to June 2026 against SEK 13.5 million a year earlier (KSEK 20,730 against 13,495; the CEO statement rounds the prior year to 13.4), with its first positive quarterly result before tax (KSEK 436); the interim report, which states it was reviewed by the company's auditor, says sales and business development also take place in the USA but gives no US revenue, live game count or Frameplay-routed revenue. | 3 | direct record | supported with qualification | high | [79] |
| GAM-B-F22 | Publisher files and Adverty's own sellers.json disagree: Miniclip authorises adverty.com seller IDs 997 and 2063 as DIRECT and Helium9 Games authorises ID 1581 as DIRECT, but none of these IDs appears in Adverty's sellers.json (121 entries in the version last modified 27 September 2026; 119 in the version of 28 September 2026). A different kind of mismatch affects IION: Miniclip, Zynga and Gameloft all declare "iion.io, 10329, DIRECT", but IION's sellers.json assigns 10329 to another publisher, Big Cake (bigcakegame.com). | 3 | synthesis | supported with qualification | high | [75, 96, 13] |
| GAM-B-F25 | Gameloft states that COMBO! reaches 1.3 billion monthly active players worldwide, 57 million monthly users in North America and an audience 27% under 18 (Gameloft proprietary data, March 2026), but its four regional figures sum to about 1.0 billion, and 1.3 billion was already the headline at COMBO!'s launch on 20 September 2022. | 3 | vendor assertion | supported as stated | medium | [117, 158] |
| GAM-B-F32 | No cohort A vendor publishes a market share or concentration figure with a stated denominator: Gameloft calls COMBO! 'the largest premium in-game ad network', Frameplay's president called its exchange 'by far, the biggest pool that you can access' while declining to share figures, and Anzu calls itself 'the leading intrinsic in-game ad solution', none with a base, period or source. | 3 | synthesis | supported as stated | medium | [55, 74, 122] |
| GAM-R-F103 | AdInMo's web and ad-serving infrastructure stopped responding in mid-2026: Internet Archive captures of www.adinmo.com returned HTTP 200 up to 10 July 2026 and HTTP 525 from 11 July 2026 through 27 September 2026; on 29 September 2026 the site returned 525 and the hostnames rtb.adinmo.com and sdktools.adinmo.com, which the archive had captured in May to July 2026, did not resolve (non-existent domain). | 3 | independent measurement | supported as stated | medium | [152, 113] |
| GAM-R-F105 | Classification: the AdInMo in-game ad network was live at the start of the scoring period but is not live at the cutoff (its operator declared no trading in the three months to 6 September 2026, and its site and ad-serving hosts are offline), so this review proposes moving the row from cohort A to 'historical' and scoring no cells; if the freeze recheck shows the suspended strike-off still unresolved and no successor operator, 'historical' stands, and the fallback is 'listed' only if the author requires a completed dissolution. | 3 | inference | supported as stated | medium | [112, 153, 113] |
| GAM-R-F116 | IION's cohort A supply is documented mainly through Voodoo: a case listed as 20 November 2025 describes a direct integration delivering 70 to 80 million monthly impressions and names Mob Control, Epic Plane and Paper.io 2 as Voodoo's hit titles (it does not say those titles carried IION placements), and Voodoo's app-ads.txt authorises 'iion.io, 10231, DIRECT', matching IION's sellers.json entry for Voodoo SAS. In the IION pages screened through 29 September 2026 (home, advertisers, platform, about, case studies and the blog posts listed in stream R), this review found no ad-eligible audience figure with a metric and period for IION's in-game supply: '60,000+ games' has no metric or period, the advertisers page's '3.6 billion' 'Unique Audience' has no period and is not described as IION's own reach, and the 1.2 million average daily active users IION reports for Arkadium (October 2025) cover game-website supply, not in-game supply. | 3 | synthesis | supported with qualification | medium | [138, 167, 11, 115] |
| GAM-R-F118 | IION names two DSPs, DV360 and The Trade Desk, as routes for programmatic deals on its Voodoo (February 2025) and Azur Games (March 2025) inventory; its October 2025 Arkadium post mentions 'programmatic deals on their preferred DSP' without names, no deal type is specified, and the DSPs' own documentation was not found to confirm the integrations. | 3 | vendor assertion | supported as stated | low | [178, 179, 657] |
| GAM-R-F119 | IION documents no direct US sales presence: the June 2025 Frameplay release describes iion as 'serving brands and publishers across APAC and EMEA' and says its titles 'will be accessible through the Frameplay Exchange in the US'; publisher files (Voodoo, Lihuhu) authorise frameplay.gg under iion's reseller ID, which shows the path is authorised, not that US delivery occurs; IION's about page lists the United States among 12 remote team locations. | 3 | synthesis | supported as stated | medium | [73, 167, 162, 166] |
| GAM-R-F207 | Gameloft gives different COMBO! network sizes: 1.2 billion active users (27 January 2026 page) and 1.3 billion monthly active players (March 2026 data; the same figure used at the September 2022 launch and in 2023 and 2025 materials), whose four regional figures, under four metric names, sum to 1,004 million, including 57 million monthly users in North America. The offer page shows six franchise and 11 publisher logos without dates; the only titles Gameloft names with campaign dates (Asphalt Legends Unite for Hyundai, April 2025; Asphalt Legends for Samsung, October 2024) fall before the scoring period. | 3 | vendor assertion | supported with qualification | medium | [117, 658, 659, 55] |
| GAM-R-F208 | COMBO!’s third-party supply connection is confirmed from both sides for five publishers on 29 September 2026: King, Zynga, Miniclip, Rovio and Outfit7 authorise gameloft.com DIRECT under seller IDs that Gameloft’s sellers.json assigns to them (31 PUBLISHER accounts in total); King, Zynga and Miniclip also authorise "gameloft.com, 198101, DIRECT" and Rovio "Gameloft.com, 198102, DIRECT", IDs that do not appear in Gameloft’s sellers.json. | 3 | synthesis | supported as stated | high | [14, 53, 77, 75, 54, 76] |
| GAM-R-F217 | Trivver describes its 11 June 2026 release, a PlayCanvas SDK on Google Cloud Marketplace that lets developers add optional branded "AI Gateways" to browser games, as its "first commercial deployment" of in-gameplay brand infrastructure, with Unreal Engine and Unity support "in development"; this is Trivver's own framing, since it announced an Unreal Engine SDK as available in March 2019, and its use-case page says it integrates with Unity, Unreal and PlayCanvas. | 3 | vendor assertion | supported with qualification | high | [155, 156] |
| GAM-R-F220 | Classification (inference): Trivver in-game ads does not meet the register’s condition for scoring (a live brand-selling network documented in the scoring period); it is a measurement and monetisation SDK first sold commercially in June 2026 for one browser engine, with no named live title or campaign, so this review proposes "listed", with brand-placement availability "unconfirmed", while its MRC audit in process stays recorded as H2 evidence. | 3 | inference | supported with qualification | medium | [155, 16] |
| GAM-R-F222 | Admazing’s supply path is visible only from the publisher side: four major publishers list "exchange.admazing.co, AM3, RESELLER", but Admazing serves no sellers.json at admazing.com, www.admazing.co or exchange.admazing.co, so the seller of record behind account AM3 cannot be resolved. | 3 | independent measurement | supported as stated | high | [75, 54, 77, 147] |
| GAM-R-F228 | Admazing quantifies its supply only as cumulative delivery (7.3 billion+ impressions served, 10.8 thousand+ campaigns, 270+ or 250+ clients in different page variants) with no period, names no live title, and cites "191 million US mobile gamers and a global monthly audience of 2.5 billion", which this review reads as market-wide figures rather than Admazing's reach (inference). | 3 | vendor assertion | supported with qualification | medium | [130, 129] |
| GAM-R-F312 | Kidoz sells to US buyers. Its FY2025 20-F reports North American revenue (by impressions served) of $9,041,962 of $18,433,172 total and says most accounts receivable are concentrated in the United States among a small number of customers. Separately, two customers, whose country is not stated, each contributed more than 10% of 2025 revenue ($6.99 million, or 38%, and $2.22 million). Kidoz's site lists two vice presidents of sales for North America. No US-only audience or title list was found. | 3 | direct record | supported with qualification | high | [65] |
| GAM-R-F348 | Zynga's app-ads.txt (last modified 23 September 2026) authorises gadsme.com, anzu.io, gameloft.com (seller ID 'ZYNGA'), iion.io and pubmatic.com as DIRECT sellers and frameplay.gg, adverty.com and kidoz.net as resellers, and declares Gadsme and Odeeo as inventory partners; Gameloft's sellers.json lists nine Zynga-labelled publisher accounts, and PocketGamer.biz reported on 23 October 2025 that Gadsme had partnered with Zynga to integrate ads into unnamed select titles. These records document authorised supply relationships, not which Zynga titles or placements each network actively sells or where; the view that Zynga Ads shares sellable supply with these networks is this review's inference, and reach must never be summed across them. | 3 | synthesis | supported with qualification | high | [139, 77, 14, 200] |
| GAM-ANZU-F06 | Anzu names five DSPs as onboard for its click-enabled intrinsic formats (The Trade Desk, StackAdapt, Basis, Smadex, Adikteev) and claims integration with '40+ programmatic partners including all major DSPs'; the DSP side confirms only StackAdapt (its own page, modified 27 August 2026) and Smadex (a Smadex partner post of 5 February 2026). | 3 | vendor assertion | supported as stated | medium | [122, 123, 120, 175] |
| GAM-ANZU-F08 | Anzu's sellers.json is not a game-only file: a narrow keyword screen by this review found at least 118 of its roughly 958 unique publisher domains to be TV, streaming, radio, news or weather properties (for example roku.com, samsungtvplus.com, cnn.com, accuweather.com, cumulusmedia.com), and it also lists children's video properties such as toongogglesjunior.com and kidzbop.com; its intermediaries include WhoScored, WeDoTV and iQIYI. | 3 | inference | supported with qualification | medium | [9] |
| GAM-ANZU-F10 | In the scoring period (2 October 2025 to 2 October 2026) Anzu names six titles as carrying its placements: UFL on PC (13 May 2026), Gangsta Island: Crime City on web (19 November 2025), Gameloft's Asphalt 8: Airborne and Dragon Mania Legends (26 November 2025), and APPS's Fashion Battle and Drill & Collect (11 August 2026). | 3 | vendor assertion | supported with qualification | medium | [132, 133, 134, 135] |
| GAM-ANZU-F27 | Anzu calls its inventory 'exclusive' (February 2026 release; a WPP quote on its developer page), but no publisher- or placement-level exclusivity is documented, and its Marketplace Terms grant publishers a non-exclusive licence to its technology. | 3 | synthesis | supported as stated | medium | [122, 619, 171] |
| GAM-FRAMEPLAY-F06 | Publisher app-ads.txt files read on 29 September 2026 confirm matching DIRECT authorisations of frameplay.gg for at least five of Frameplay's 30 publisher accounts: Miniclip (53945695), Helium9 Games (39338461), Wizard Games (70894634), FUN-GI Games (4923326410) and Castry (5828510305). Seven sampled accounts carry no frameplay.gg line (Vector Unit, Illusion Labs, Game Masons, Cerberus Studios, Nautilus Mobile, RacoonDigi, DoubleTap). Gameloft authorises Frameplay only via iion's ID 9251981480 rather than its own account 3150595190. Miniclip, Outfit7, Zynga, Gameloft and dartsoffury.com authorise frameplay.gg as RESELLER under iion's ID, and Miniclip, Helium9 Games, Castry and dartsoffury.com authorise it as RESELLER under Adverty's ID 2018037567. | 3 | synthesis | supported with qualification | medium | [75, 96, 78, 76, 77, 10] |
| GAM-FRAMEPLAY-F08 | Inference: the "5,000+ mobile games" figure cannot be reconciled with Frameplay's own supply records and plausibly includes supply resold from other networks (iion alone was announced as bringing 4,000+ titles), so it must not be added to IION's, Adverty's or AdInMo's reach. | 3 | inference | supported as stated | low | [136, 73, 10] |
| GAM-FRAMEPLAY-F26 | Frameplay's SDK guide names Google DV360 as demand for video-capable ad spaces, but DV360's own exchange enum (updated 10 June 2026) lists 89 named exchanges and none is Frameplay, and GumGum, which appears in an undated testimonial on Frameplay's advertiser page, lists no Frameplay entry in its sellers.json; the route by which DV360 buys Frameplay inventory is undocumented. | 3 | synthesis | supported as stated | medium | [173, 119, 137, 174] |
| GAM-GADSME-F06 | Gadsme markets its inventory as exclusive ('All our publisher relationships are direct and exclusive', '100% exclusive inventory from 200+ game studios', 'No multi-SSP exposure', 'No reseller stacking', 'monetized exclusively by Gadsme'), but its own Publisher General Terms restrict exclusivity to 3D and audio ads, apply it per player or game session, and let a publisher A/B test a competing 3D network provided Gadsme runs for at least 50% of players. | 3 | synthesis | supported as stated | high | [118, 170, 95, 59] |
| GAM-GADSME-F12 | Gadsme's aggregate supply claims carry no metric period and do not agree with each other: 'over 1,500 games' (December 2024) and 'more than 1,500 games' (July 2025), 'more than 1,600 mobile titles' (October 2025), 'thousands of games' and '200+ game studios' (advertisers page, undated), and 'more than 100' game developers (Gadsme boilerplate in ARROVA's release, 13 May 2026). | 3 | vendor assertion | supported as stated | medium | [159, 139, 118, 160] |
| GAM-SUPERLEAGUE-F05 | Part of Super League's supply is resold or shared. The Q2 2026 10-Q says media is delivered via the Super Biz Roblox platform, the Roblox Immersive Ads platform and other platforms. The FY2025 10-K names Anzu, Bidstack, AdInMo and Frameplay as examples of competing in-game ad-tech providers and says that in certain instances Super League partners with such firms by including their inventory in its programs, without saying which firms' inventory it has used; it reports revenue gross except where it uses a reseller partner. It also has an exclusive partnership with AdArcade for playables in rewarded mobile video inventory. | 3 | direct record | supported with qualification | high | [30, 128] |
| GAM-SUPERLEAGUE-F08 | The executed Exclusive Brand Partnership Agreement (1 May 2026) gives Super League exclusive rights to present Misfits Games (Roblox games published or co-published under the Misfits brand) for partnership opportunities that Misfits has not already discussed with others. The rights exclude games Misfits does not solely control and require Misfits' prior written approval before any brand is approached. Misfits keeps sole and absolute discretion over terms and pricing, and Super League may not make changes to Misfits Games. Third-party owners are not bound, exclusivity for an opportunity lapses if terms are not agreed within two weeks, and the initial term is one year. | 3 | direct record | supported as stated | high | [172] |
| GAM-SUPERLEAGUE-F10 | Under its Strategic Properties initiative, Super League holds 15.0% equity in the Roblox experience Hide or OOF! (agreement of 5 January 2026). That stake comes with 15.0% of the game's gross revenue, a 12.0% fee on certain direct brand transactions and exclusive rights to certain brand-partnership opportunities; the filing says the game was visited more than 570 million times in 16 months. In April 2026 Super League also announced an ownership interest in My Avatar!, which the vendor says has over 800,000 daily active users. The filing describes these experiences as generating 'differentiated inventory for playable advertising'. | 3 | direct record | supported as stated | medium | [30, 146, 660] |
| GAM-R-F317 | Kidoz Inc.'s reported revenue ($18.4 million in 2025) mixes three engines (programmatic, partner channels, direct brand relationships), performance units (cost per install) and the all-ages Prado division, so no public figure isolates brand revenue on the kid-safe network. | 3, 4 | direct record | supported as stated | high | [65] |
| GAM-R-F216 | MRC’s own record on 29 September 2026 lists Trivver as "Pre-audit reviewed 7/2022, Audit in process as of 4/2026" for Impressions, T-Views and Product Actions/Actions/Engagements across desktop, mobile web, mobile in-app and gaming consoles; the entry is absent from the Internet Archive capture of 5 February 2026 and present in the capture of 11 April 2026, and Trivver is not on MRC’s accredited list. | 3, 6 | direct record | supported as stated | high | [16, 154, 661] |
| GAM-B-F31 | This review found no public documentation, in the sources screened by 28 September 2026, of any cohort A network supporting an agentic buying protocol (AdCP, AAMP or other), so all five protocol states under shared rule 12 are unrecorded for every cohort A network; the only supply-chain standard observed in use is the IAB Tech Lab sellers.json format, version 1.0, published by six of them. | 3, 10 | synthesis | supported as stated | medium | [9, 10, 11, 12, 13, 14] |
| GAM-I-F26 | No in-game seller in this paper's product register shows an implemented AdCP connector, interoperability test, production use or transaction in the sources screened: AdCP's latest stable release is v3.1.24 (23 September 2026, with v3.2.0-rc.7 on 27 September 2026), and the public agent registry listed 34 agents on 28 and 29 September 2026, none operated by Roblox, Microsoft, Anzu, Frameplay, Gadsme, AdInMo, Adverty, Gameloft, Super League or IION. At the schema level AdCP can represent in-game products by channel: its v3.1.24 channels enum includes 'gaming' ('In-game advertising across platforms', present since v3.0.0), which its media-channel taxonomy defines to include in-game display, rewarded video and playable ads (its migration guide adds intrinsic in-game), while its canonical creative formats defer in-game ('Out of scope until cross-engine standards land') and list 'playable' only as a custom format shape. | 3, 10 | direct record | contradicted | medium | [184, 185, 121] |
| GAM-R-F101 | AdInMo Ltd (SC648449), the operator named in AdInMo's own privacy policy, had its strike-off application signed on 6 September 2026 by its sole remaining director, who declared that none of the circumstances in sections 1004 and 1005 of the Companies Act 2006 existed. Section 1004 bars an application if the company traded or otherwise carried on business at any time in the previous three months. The application was filed on 7 September 2026, gazetted on 15 September 2026, and suspended on 17 September 2026 because an objection was received; on 29 September 2026 the register showed 'Active — Active proposal to strike off'. | 3, 10 | direct record | supported as stated | high | [112, 149, 151, 150] |
| GAM-U-F03 | Bidstack Ltd (09835625) and its parent Virtual Sport Technology Limited (15133297) each entered creditors' voluntary liquidation with winding up commencing on 18 June 2026, with the same liquidator appointed that day. | 3, 10, 13 | direct record | supported as stated | high | [32, 64] |
| GAM-B-F27 | King, Rovio, Outfit7, Zynga and Miniclip each authorise gameloft.com as a DIRECT seller in their app-ads.txt files under seller IDs that Gameloft's sellers.json lists as their PUBLISHER accounts (files read 28 to 29 September 2026). This documents authorised supply relationships with COMBO!'s operator, not which titles, placements or geographies COMBO! actively sells, so the view that some King inventory is offered both through COMBO! and through Microsoft's XBOX Media Solutions is this review's inference, unverified at title or placement level; Gameloft is in turn listed as a PUBLISHER in Anzu's, Gadsme's and Frameplay's sellers.json, with matching DIRECT lines in its own app-ads.txt for Anzu and Gadsme but not for Frameplay. | 3, 12 | synthesis | supported with qualification | high | [53, 54, 76, 14, 662, 78] |
| GAM-E-F17 | Under Roblox's managed-campaign advertising terms (effective 30 April 2026), neither party exchanges personal data. Advertisers receive only aggregate reporting, and Roblox's reporting is the sole and controlling measurement for all campaigns. An advertiser's approved third-party verification partner may dispute invalid-traffic determinations only where the discrepancy exceeds 10%. | 4 | vendor assertion | supported as stated | high | [1] |
| GAM-I-F07 | Roblox's documented in-experience brand formats are Immersive Ads (video billboards up to 30 seconds, click-to-play or autoplay; static image ads; portal ads that teleport users to an advertiser's experience) and Rewarded Video (full-screen, 6 to 30 seconds, shown only to users 13 and over), with counting rules stated in its Creator Hub documentation. | 4 | vendor assertion | supported as stated | high | [208, 207, 20] |
| GAM-I-F12 | Roblox's Creator Hub page on ad integrations (accessed 28 September 2026) says the revenue share starts 1 January 2027 for campaigns beginning on or after that date and is charged per 1,000 qualified visits (US $1.50 for the first 28 days, $0.10 thereafter, a $400,000 cap per campaign in 2027); in Roblox's public source repository for these pages the 1 January 2027 date first appears in a commit of 17 June 2026, after the Terms' 1 May 2026 effective date, and neither date appears in the April 2026 versions. | 4 | vendor assertion | supported as stated | high | [190, 219] |
| GAM-I-F16 | Roblox sells brand placements outside gameplay as well: the Homepage Feature (closed beta announced 6 January 2026; described as launched in the 5 February 2026 letter, but still labelled 'Homepage Feature (Beta)' on Roblox's brands page on 29 September 2026), a 24-hour homepage takeover (Homepage Hero), Homepage and Search tiles, and Moments, which the Q2 2026 letter describes as a future 'premium advertising surface' initially for age-checked users over 16 in select markets. | 4 | vendor assertion | supported with qualification | medium | [188, 201, 42, 43] |
| GAM-I-F17 | XBOX Media Solutions, sold under Microsoft Advertising, documents in-game Rewarded Video, In-Game Interactive Video (Video Explorer, Tappable Video, Video Selector), Playables and custom integrations: 'Candy'fied Sponsored Moments in Candy Crush Saga, a Solitaire Event Takeover in Microsoft Casual Games on PC and mobile, and console 'XBOX Experiences' described as native display formats and branded content; it says its custom integrations 'range from immersive in-game moments to XBOX dashboard & store placements' and calls them '100% owned and operated'. The pages do not say which console placements, if any, render inside gameplay. | 4 | vendor assertion | supported with qualification | high | [191, 224, 24] |
| GAM-I-F19 | The four XBOX Media Solutions gaming pages opened on 28 September 2026 name no DSP, private marketplace, deal type or other programmatic route for the in-game formats, and Microsoft Advertising's MCP Server page describes access to Microsoft Advertising campaigns, ad groups, ads and keywords without mentioning gaming inventory. | 4 | inference | supported as stated | medium | [192, 191] |
| GAM-I-F20 | Microsoft does not break out gaming advertising: its FY2026 10-K reports XBOX revenue of $21,790 million (FY2025 $23,455 million; FY2024 $21,503 million), a line that includes advertising alongside content, Game Pass and cloud gaming, and Search advertising revenue of $15,176 million (FY2025 $13,878 million), which covers Bing, Copilot, Microsoft News and Edge rather than games. | 4 | direct record | supported as stated | high | [198] |
| GAM-I-F21 | Take-Two, a publisher that sells ads in its own games (a cohort B1 candidate), reports Advertising net revenue of $494.3 million for the year ended 31 March 2026, $466.1 million for FY2025 and $656.1 million for FY2024, or 7.4%, 8.3% and 12.3% of total net revenue (derived); its 10-K says this advertising is primarily in mobile titles and includes banner, interstitial, offer, watch-to-earn, branded virtual item and sponsorship formats. | 4 | direct record | supported as stated | high | [200] |
| GAM-R-F321 | Electronic Arts has been privately owned by PIF, Silver Lake and Affinity Partners since the acquisition closed (Form 8-K and Form 25-NSE filed 4 August 2026). EA filed a Form 15-12G on 14 August 2026 to end its SEC registration, so the 'NASDAQ: EA' descriptor in the launch release is out of date, and filings will no longer provide new EA facts. | 4 | direct record | supported with qualification | high | [241] |
| GAM-R-F332 | No title-level reporting to buyers is documented for EA Advertising, and ea.com/sellers.json returned HTTP 403 (an access error). EA does publish an app-ads.txt for its apps (OwnerDomain ea.com) with DIRECT and RESELLER declarations; it names Anzu as manager domain ('managerdomain=anzu.io,in-game') and authorises Anzu and IION as direct sellers. That is supply-path disclosure for EA's mobile inventory, which Digiday (15 September 2026) places inside EA's ad offer. None is published for the console ad units. | 4 | inference | supported with qualification | low | [240, 25] |
| GAM-R-F340 | Zynga Ads is live and fits cohort B1. It sells brand formats and integrations inside Zynga's own mobile games, with in-period activations naming titles (Survivor's 50th season in five Zynga games, March 2026, commercial terms not stated; the Hyundai NEXT Cup Tour in Top Eleven, June 2026). Take-Two's FY2026 10-K says the Zynga label earns revenue from in-game sales and advertising and that Take-Two contracts with ad networks, agencies, brokers and directly with advertisers. Take-Two reports group Advertising net revenue of $494.3 million for FY2026, a figure that covers all ad types, not brand ads alone. | 4 | synthesis | supported with qualification | high | [194, 242, 200] |
| GAM-R-F342 | Zynga Ads names seven brand offers (non-skip video, rewarded video, interactive video, playables, shoppables, integrations, sponsorships) with performance claims (85-95% video completion, 96% viewable) but publishes no specifications or counting approach in the pages readable; the 2026 media kit link returned HTTP 404 on 29 September 2026. | 4 | vendor assertion | supported as stated | medium | [244, 245] |
| GAM-ROBLOX-F11 | Roblox's Advertising Standards (last updated 5 June 2026) state that ads may not be personalised for users under 18 and that rewarded video may be offered only to users 13 and over, and ban reward mechanisms for under-13 ad engagement; a Restricted Advertising Categories for U13 Users section effective 4 May 2026 bans whole categories when the intended audience includes users under 13 (including all food and beverage and personal care, cosmetic, fitness and diet products), and age-banded entertainment rules (no R, TV-MA or M-rated content for ages 9 to 12, plus no PG-13, TV-14 or Teen-rated content for ages 5 to 8) were added on 5 June 2026; Roblox's help pages say personalised ads are served only to users 18 and over. | 4 | vendor assertion | supported with qualification | high | [20, 213, 214] |
| GAM-ROBLOX-F15 | Counterparty documents confirm three Roblox supply-side partnerships but no DSP-side documentation: Google's blog (1 April 2025) calls its Ad Manager partnership 'foundational' to Roblox's immersive ads; Magnite (6 January 2026) says its integration 'will make' Roblox Rewarded Video available to advertisers, for campaigns 'in mobile in-app environments', and invites buyers to activate through their Magnite account manager; PubMatic (10 April 2024) says Roblox video inventory 'will ultimately be made available' via programmatic guaranteed, PMPs, auction packages and open exchange. No Amazon DSP or Liftoff page about Roblox was found in the searches logged. | 4 | vendor assertion | supported with qualification | medium | [212, 211, 189] |
| GAM-ROBLOX-F22 | Roblox's own documents still conflict on 29 September 2026 over the developer revenue share for Advertising Integrations: the Terms (created 27 March 2026, last edited 30 April 2026, effective 1 May 2026) say it begins 30 September 2026 per thousand visits while an integration is present, at unstated pricing; the Creator Hub page, live on 29 September 2026, says it applies to campaigns beginning on or after 1 January 2027 per 1,000 qualified visits (US $1.50 for the first 28 days, then $0.10; $400,000 cap per 2027 campaign); and Roblox's 20 March 2026 forum announcement and a Reuters report of 27 March 2026 already placed the start in 2027. | 4 | vendor assertion | supported as stated | high | [57, 190, 215, 218] |
| GAM-ROBLOX-F23 | Roblox's product is resold through named partners in markets outside the US: Venatus as 'Official Reseller for Rewarded Video' in the UK, France and Germany (May 2026), and Playwire, whose partnership Roblox says 'expands programmatic access to Roblox’s premium gaming inventory across select key APAC markets' (October 2025); SuperAwesome sells under-13 inventory as exclusive partner. | 4 | vendor assertion | supported as stated | medium | [204, 205, 206] |
| GAM-XBOX-F11 | Candy Crush Saga is named in two dated in-period campaigns (Hershey's holiday 2025; KitKat, published 18 August 2026); Solitaire is named only through an undated 'Solitaire Event Takeover' on a page served in the scoring period (identifying it as Microsoft Solitaire Collection is the reviewer's inference, supported by the ad-supported free tier on the Store listing), and no console title is named as carrying in-gameplay placements. | 4 | synthesis | supported with qualification | medium | [226, 227, 24, 228] |
| GAM-XBOX-F18 | King inventory is authorised to or sold through routes other than XBOX Media Solutions' managed sales: Gameloft's COMBO! ('gameloft.com, KING, DIRECT'), Anzu and Gadsme reseller paths, Meta Audience Network (nine facebook.com DIRECT lines; a Meta story citing Q4 2020 data), and Microsoft Advertising's own gaming deals (described on 26 September 2024, in-period status not established); audience, titles or impressions across these routes must never be summed. | 4 | synthesis | supported with qualification | medium | [53, 229] |
| GAM-XBOX-F23 | Microsoft's Privacy Statement (last updated September 2026) says Microsoft does not deliver personalised advertising to children whose Microsoft account birthdate identifies them as under 18, a higher threshold than King's under-13 rule, and no document read states which rule governs XBOX Media Solutions inventory in King titles; Microsoft Solitaire Collection is listed on the US Microsoft Store 'For all ages' with an ad-supported free tier. | 4 | vendor assertion | supported as stated | medium | [21, 228] |
| GAM-E-F15 | Roblox's Advertising Integrations Terms, effective 1 May 2026, cover ads that developers place directly in their experiences rather than through Roblox's own ad products. The developer must ensure that neither the integration nor the experience makes programmatic calls to third-party advertising and measurement services, and may not share personally identifiable engagement information with advertisers. From 30 September 2026 the developer owes Roblox a revenue share charged per thousand visits while the integration is present, at a rate the Terms do not state. | 4, 7 | vendor assertion | supported as stated | high | [57] |
| GAM-ROBLOX-F12 | Roblox announced on 20 March 2026 that from 4 May 2026 age-appropriate formats including billboard ads could be shown to all users including under-13s, and its own announcement FAQ says all under-13 ads 'will be contextual and sold directly via Roblox or creators' and not through programmatic partners such as Google; per AdExchanger (4 June 2026), a Roblox spokesperson said under-13 ads exclude rewarded video, are targeted mainly by experience context but also by generalised demographics (country, gender, age range), and that SuperAwesome is the exclusive partner; Roblox's brand-safety page, live on 29 September 2026, still says users under 13 are not eligible to see ads. | 4, 7 | vendor assertion | supported with qualification | medium | [215, 216, 206, 217] |
| GAM-XBOX-F15 | Whether Microsoft sells in-gameplay console inventory stays unresolved at 29 September 2026: TechCrunch reported on 18 April 2022, from Business Insider's unnamed sources, a program for ads in free-to-play Xbox games 'set to launch by the third quarter' (a Microsoft spokesperson told Insider 'we don't have anything further to share'); Anzu's undated page claims to be 'the sole officially licensed in-game ad provider for Xbox'; a Microsoft patent application published on 3 September 2026 (US 2026/0260258 A1, read only through a secondary report) describes in-gameplay ad pauses, which is a filing, not a product; and a sponsored article produced by The Drum Studios in January 2026, built on an interview with the unit's general manager, describes Xbox's approach as 'the idea of game-adjacent rather than in-game' (the article's narration, not a quotation). In the Microsoft sources screened through that date (the XBOX Media Solutions gaming-solutions pages, case studies and blog posts, Xbox Wire and Microsoft Learn, listed in the Xbox profile), this review did not establish a live in-gameplay console product: the custom-integrations page says Microsoft integrates brands 'into gameplay across mobile, PC, and console' but names no console title, date or in-gameplay console format. | 4, 8 | synthesis | supported with qualification | low | [231, 232] |
| GAM-G-F21 | Electronic Arts launched EA Advertising on 15 June 2026, saying advertisers 'can now' reach players through native ad units (digital ad boards, scoreboards and brand broadcast overlays) dynamically served in select, unnamed EA SPORTS games from a proprietary ad server and SDK built for the Frostbite engine, with impression measurement EA describes as aligned to IAB standards and viewability and measurement 'in partnership with Integral Ad Science'; EA says it works directly with brands and agencies on bespoke integrations and that 'expanded buying capabilities' are planned, and the release does not state the buying route for the ad units. EA has been privately owned by PIF, Silver Lake and Affinity Partners since the acquisition closed (Form 8-K filed 4 August 2026). | 4, 8, 10 | vendor assertion | supported with qualification | high | [25] |
| GAM-I-F01 | Roblox does not disclose advertising revenue: its 10-K for FY2025 (filed 11 February 2026) and its 10-Q for Q2 2026 (filed 30 July 2026) both describe revenue from advertising and licensing as 'insignificant' and report no advertising line. | 4, 10 | direct record | supported as stated | high | [197, 246] |
| GAM-R-F350 | Zynga has a programmatic sales team and urges buyers to buy premium publishers 'through programmatic pipes'. In July 2026 it announced a partnership with the SSP PubMatic: Zynga integrates PubMatic's OpenWrap SDK, and advertisers can buy Zynga inventory through PubMatic's Activate and AgenticOS. Zynga's app-ads.txt authorises pubmatic.com as a direct seller. No DSP, deal type or protocol state is named. | 4, 10 | vendor assertion | supported with qualification | medium | [195] |
| GAM-I-F08 | Roblox's programmatic routes are documented only in announcements and marketing, by Roblox and by its supply-side partners: PubMatic announced on 10 April 2024 that Roblox video inventory would 'ultimately' be available to its buyers via programmatic guaranteed, private marketplaces, auction packages and open exchange; on 1 April 2025 Roblox announced Rewarded Video through Google Ad Manager and direct insertion orders 'in the coming weeks', with Billboards and Google Authorized Buyers to follow; on 6 January 2026 it announced expansion 'this year' with the DSPs Amazon DSP and Liftoff and the SSPs Index Exchange, Magnite and PubMatic, and Magnite's own post that day said its integration 'will make' Rewarded Video inventory available to its buyers; Roblox's brands page (accessed 29 September 2026) lists 'Magnite, Index Exchange, and Display & Video 360' as 'major DSPs' and limits programmatic targeting to 18-plus inventory. No DSP's own documentation of a live Roblox integration was found in the sources screened. | 4, 10, 12 | vendor assertion | supported with qualification | medium | [187, 188, 42] |
| GAM-R-F329 | At launch (15 June 2026) EA Advertising's console integrations and ad units were sold directly and bespoke, with expanded buying planned, and on 3 July 2026 EA's advertising VP said EA was not yet working with programmatic partners such as The Trade Desk or Criteo. By 15 September 2026 Digiday reported that EA already offers programmatic buying across its mobile games and is extending more dynamic capabilities to its console and PC titles. No DSP or deal type is named. | 4, 10, 12 | vendor assertion | supported with qualification | high | [25, 86] |
| GAM-I-F24 | Comparable platform and specialist revenue series are not disclosed, so part (b) of H5 is unresolved from public sources: Roblox calls its ad revenue insignificant without a figure; Microsoft reports any gaming advertising only inside its XBOX revenue line, which also combines hardware, content, subscriptions and cloud gaming, while its Search advertising line is not gaming; Amazon reports only company-wide advertising services; and the series that are disclosed (Take-Two advertising in USD by fiscal year to March, Adverty net sales in SEK, Super League media and advertising revenue including influencer work) differ in metric, currency, period and scope. | 4, 10, 13 | synthesis | supported with qualification | high | [197, 198, 199, 200, 143, 128] |
| GAM-I-F13 | Roblox's Q1 2026 letter (30 April 2026) says creators running independent brand integrations must register campaigns in Ads Manager and migrate to native formats such as Rewarded Video and immersive billboards, under advertising policies taking effect the following month; Roblox's own Advertising Integrations Terms and Creator Hub page nonetheless keep a registered, Roblox-approved Advertising Integrations programme, subject to a revenue share, for integrations that do not substitute for Roblox's native formats, so the letter describes a channelling rule rather than a ban on creator-sold integrations. | 4, 12 | direct record | supported with qualification | high | [210, 57] |
| GAM-F-F01 | Through the scoring period Discord sold Quests, rewarded ad units rendered inside the Discord client, only through its sales team, and its advertiser pages still said so on 30 September 2026; no programmatic route is documented. On 29 September 2026 Discord announced Discord Ads Manager, its own auction-based campaign tool, available with managed access to a select group of advertisers and planned as fully self-serve in early 2027, and a new mobile-first Video Quests format sold through it; the earlier Video Quests became Video Quests Reserve, a reservation-based offering (GAM-NFA-S11). Its advertiser pages read on 29 September 2026 list two formats, Play Quests and Video Quests (which Discord calls 100%-viewable, on mobile and desktop), but Discord's own releases announced Instant Play Quests (cloud-streamed trials, 9 March 2026, stated as forthcoming), Mission Quests (22 July 2026, first campaign live from 23 July 2026) and Play Quest+ (20 August 2026, first campaign with EA's Battlefield 6 in September 2026). Discord says Quests can run in all regions where Discord is available, with country targeting, and that it currently focuses on advertisers in gaming and media and entertainment. | 5 | vendor assertion | supported with qualification | high | [48] |
| GAM-F-F03 | Discord documents three measurement functions for Quests, all through partners: Kantar brand lift studies (partnership announced 28 May 2025 and restated 16 July 2025), AppsFlyer as its first mobile measurement partner for Video Quests on Mobile, and a Gamesight server-to-server integration for PC and console publishers that Discord describes as in beta (2 October 2025). | 5 | vendor assertion | supported with qualification | high | [340, 281] |
| GAM-F-F04 | Discord's published audience figures are its own: 200 million or more monthly active users who play 1.9 billion hours of games a month (releases of 20 March and 2 October 2025) and 90 million or more daily active users as of Q4 2025 (ads page), all global with no US split and no ad-eligible share. | 5 | vendor assertion | supported as stated | medium | [281, 49, 48] |
| GAM-F-F05 | Discord's privacy policy (effective 29 September 2025) says it uses user information to surface sponsored content such as Quests, lets users limit that personalisation in settings, and may share information with analytics and measurement partners when a user engages with sponsored content; its Quests help article (updated 31 August 2026) says users aged 13 to 17 can complete Quests but no more than three per day, and that Quests are personalised by interest or shown contextually, with personalisation that users can switch off; its 22 September 2026 post places accounts globally into adult (18+) and teen (13-17) age groups and, as read, says nothing about ads. | 5 | vendor assertion | supported with qualification | medium | [663, 289] |
| GAM-F-F13 | Twitch's creator and sponsorship routes, grouped by Amazon as Twitch Creator Activations, include Creator Sponsorships, Bounty Board (a marketplace where creators accept paid sponsorships from their dashboard to play a game or show a trailer on stream), Channel Takeover (channel skins, 30-60 second streamer-read ads, follower promotions and sponsored subscriptions), premium sales packages and the Twitch Brand Partnership Studio for custom branded content; Bounty Board is listed for nine locales including the US. | 5 | vendor assertion | supported with qualification | high | [261, 262, 2] |
| GAM-F-F15 | The only named gaming-specific YouTube buying product found in Google's own sources is a gaming lineup of YouTube Select Shorts, announced on 17 April 2024 alongside sports, entertainment, beauty, food and automotive lineups for the US, Canada, the UK, Australia, Korea and India; YouTube Select lineups are reservation products covering the top 1% of channels in a market. | 5 | vendor assertion | supported as stated | medium | [254, 255] |
| GAM-F-F16 | MRC's accredited list (read 29 September 2026) covers Google's video buying tools and YouTube data: Google Ads rendered video impressions and TrueView views on desktop, mobile web, mobile app and CTV, and video viewability on desktop, mobile web and mobile app only; Google YouTube Brand Safety content-level processes on desktop, mobile web and mobile app; and DoubleVerify, IAS and Zefr third-party reporting of YouTube data from Google Ads, DV360 and YouTube Reserve. None of these entries is gaming-specific. | 5 | direct record | supported with qualification | high | [15] |
| GAM-F-F18 | Overwolf's app ads render inside third-party desktop companion apps that run over the game (Overwolf says within the app's boundaries, during downtime in the game or when the player opens the app, with placement set by the app developer under Overwolf's guidelines, and describes them as served in-game on the game screen); it also sells ads on 500+ gaming websites. Buyers can deal direct or programmatically through Deal ID or PMP with named DSPs and SSPs including The Trade Desk, Magnite, OpenX, Xandr, Adform, PubMatic, Index Exchange, FreeWheel, Yahoo and Basis, and can choose specific games and apps. | 5 | vendor assertion | supported with qualification | high | [44, 291] |
| GAM-F-F19 | Overwolf documents standard IAB video and display formats with sizes and durations, reporting by geography, game and app and GPU/CPU, and says its inventory is monitored by HUMAN, MOAT, DoubleVerify, Protected Media and Pixalate with GeoEdge scanning; the MOAT reference is out of date because Oracle ended Moat on 30 September 2024. | 5 | vendor assertion | supported as stated | medium | [664, 44] |
| GAM-F-F21 | Overwolf announced on 8 May 2024 that it had acquired NitroPay, renamed Nitro and run as a division alongside CurseForge and Tebex; Overwolf said Nitro works with over 300 publishers running 500+ gaming websites with over eight billion ad impressions across over a billion pageviews. | 5 | vendor assertion | supported as stated | medium | [293] |
| GAM-F-F22 | MRC's digital accredited list read on 28 September 2026 names Amazon DSP accreditation only for the supply sources Retail and IMDb, Amazon Publisher Services, Fire TV, third-party exchanges and Alexa connected devices, and contains no entry naming Twitch, Discord, Overwolf, Playwire, Venatus, Enthusiast Gaming or ESL FACEIT Group. | 5 | direct record | supported as stated | medium | [15] |
| GAM-F-F24 | Media Scale Capital Limited, a UK holding company incorporated on 1 June 2026 whose registered person with significant control is To Media Groep (a Dutch BV), became the person with significant control of Project Venus Topco Limited, the top of the holding chain above Venatus Media Limited, on 19 June 2026, holding 75% or more of shares and voting rights; on the same date the previous controlling PSC, the private equity fund Livingbridge 7 Global LP, and two individual PSCs ceased, and the board changed. Venatus announced the acquisition, with a new CEO and CFO, on 14 July 2026. | 5 | direct record | supported with qualification | high | [302, 303, 304, 305] |
| GAM-F-F26 | Venatus sells gaming-site inventory as away-from-the-game media (high-impact takeovers including embedded Twitch Takeovers, cross-device video and banners) on a network it says reaches 400 million or more monthly active users, naming OP.GG, FUTBIN and PlanetMinecraft, and also claims exclusive representation of EA's in-app advertising, an in-game route outside cohort B2. | 5 | vendor assertion | supported as stated | medium | [84, 256] |
| GAM-F-F28 | Playwire sells 'PC In-Game Video' (200M+ monthly video impressions, 80% completion, geo and contextual targeting, available via PMP, programmatic guaranteed or direct IO) and 'Minecraft' in-game video (non-skippable, audio-on, 50M+ monthly impressions), but its pages do not document where either ad renders: inside the game client, a launcher, a companion app, a web page or a map. | 5 | vendor assertion | supported as stated | medium | [314, 313] |
| GAM-F-F29 | Playwire's own material places its Minecraft rewarded video on third-party Minecraft multiplayer servers: server operators such as Vortex Games (a Playwire partner since 2018) offer players in-game items or boosts for watching a video, and Playwire's February 2025 blog describes the core Minecraft server audience as young teenagers aged 13 to 15. | 5 | vendor assertion | supported as stated | medium | [315, 316] |
| GAM-F-F30 | Playwire claims 74M+ total US gaming reach and, for kids, 80% monthly reach of US children aged 2 to 12 with COPPA-compliant delivery across CTV, mobile and gaming, without a metric, measurement provider or period for either figure. | 5 | vendor assertion | supported as stated | low | [313, 665] |
| GAM-F-F31 | Enthusiast Gaming sold its direct sales business line, including Omnia Media Inc. and GameCo Esports Canada Inc., to Vertiqal Studios Corp. for C$900,000 in cash plus assumed liabilities (net liability about US$2.6 million at 30 June 2025), closing on 30 August 2025, and kept U.GG, Icy-Veins, TheSimsResource, Addicting Games, PocketGamer and other sites, which it says carry its web programmatic advertising revenue. | 5 | vendor assertion | supported as stated | high | [317] |
| GAM-R-F405 | The targeting Discord documents for Quests is audience-based rather than contextual: countries, players matched on interests and the games they already play (Discord first-party data), broad age ranges from self-declared age, and player-based mobile targeting; Quests are shown contextually (for example to everyone in a server) only when a user has opted out of personalisation, and no buyer-selectable contextual signal is documented; its Ads Policy bans targeting on sensitive categories. | 5 | vendor assertion | supported with qualification | medium | [62, 286, 666, 22, 281, 287] |
| GAM-R-F407 | Discord's case study for Bethesda's DOOM: The Dark Ages reports a 7-day Video Quest in the US and Canada whose brand impact was measured through a Kantar Brand Lift Study (+11.3-point ad awareness, +3.4-point play intent at 90% confidence), with the headline tiles sourced to 'Discord Internal Data, 2025' and no sample or control design stated; for Play Quest+ the buyer sets in-game events as the objective through Discord's Events API, and no in-flight optimisation is documented. | 5 | vendor assertion | supported as stated | medium | [341, 284] |
| GAM-R-F412 | No independent assurance record for Discord Quests was found: MRC's digital accredited-services list (29 September 2026) has no Discord entry, a TAG Registry search for 'Discord' returned no company (the control query 'Twitch' returned a certified entry), and Discord's addendum says Quests use no third-party ad server, which leaves no documented route for third-party verification tags. | 5 | direct record | supported as stated | medium | [15, 333, 280] |
| GAM-R-F414 | Discord documents age and data controls for Quests in operational detail (users aged 13 to 17 can complete at most three Quests per 24 hours; age-assurance signals are used only to keep adult ads from teens and never to target ads; two user settings govern personalisation from Discord activity and from advertiser or third-party data, with a US-state opt-out and a non-personalised fallback), but its brand-suitability documentation is an ad-content policy, with Discord keeping control of placement and no advertiser-side placement exclusion documented. | 5 | vendor assertion | supported as stated | high | [285, 22, 288, 287, 280] |
| GAM-R-F415 | LIVAD's livestream overlay ads appear live in the scoring period on present-tense trade reporting (Tubefilter, 18 September 2026, naming overlays and sidebars as its core business for clients such as Adidas, KFC and L'Oreal), an ads.txt last modified on 3 August 2026 and a 2026 site copyright; no named campaign inside the scoring period was found, and LIVAD's own blog has no post after 2023. Its CEO presented at IAB PlayFronts in New York on 17 September 2026, announcing Virtual Product Placements, which had not been used in a live campaign. | 5 | vendor assertion | supported with qualification | medium | [273, 274, 272] |
| GAM-R-F417 | LIVAD documents specifications for five stream formats in a spec sheet hosted by its reseller DigitalsCom (file dated 12 November 2024): Video 640x360, Side Bar 300x1080, Snipe 1920x250 and Rich Media 1920x1080 (roughly 6 to 30 seconds, webm/mp4/gif/jpeg/png, up to 100MB, a chat sponsor message of up to 150 characters and an optional streamer script of up to 500 characters) and a 60-second Poll with up to five answers; its streamer contract defines a view as the count of viewers during each branded overlay display session, billed per view or per thousand. | 5 | vendor assertion | supported as stated | medium | [667, 668, 275] |
| GAM-R-F422 | LIVAD's ads.txt (last modified 3 August 2026) authorises Adform (two direct lines), Improve Digital and Anzu as direct sellers plus Anzu's reseller chain; Adform's sellers.json lists LIVAD Medya Teknoloji Anonim Şirketi (livad.stream) as a publisher under both Adform seller IDs, and Anzu's sellers.json lists "livad stream" (www.livadstream.com) as a publisher, so supply connections through Adform and Anzu are documented on both sides; Improve Digital's sellers.json has no matching entry, and no DSP, deal type or buyer-side route is documented. | 5 | vendor assertion | supported with qualification | medium | [276, 9, 278] |
| GAM-R-F424 | LIVAD's US footprint is indirect: LIVAD Technologies Inc. (Delaware, company 6813534) holds the site copyright but is listed as inactive by OpenCorporates (the Delaware registry itself was not checked), while the streamer contract, privacy policy and Adform's sellers.json name LIVAD Medya Teknoloji A.Ş. of Beyoğlu, Istanbul (Turkish law); no US campaign, US sales contact or US buyer terms were found, and its CEO's presentation at IAB PlayFronts in New York on 17 September 2026 is the only US commercial signal in the sources screened. | 5 | synthesis | supported with qualification | medium | [279, 275, 669, 272] |
| GAM-R-F425 | No independent assurance record for LIVAD was found: MRC's digital accredited list has no LIVAD entry and a TAG Registry search for 'LIVAD' returned no company on 29 September 2026; its reseller's brand-safety slide shows IAB and IAB Tech Lab member logos, which are memberships, not assurance. | 5 | direct record | supported as stated | medium | [15, 333, 670] |
| GAM-R-F426 | IGN advertising is sold by IGN Entertainment, Inc., a Delaware subsidiary of Ziff Davis, Inc. (Exhibit 21.1 to Ziff Davis's 10-K filed 24 February 2026), through sales teams in the US, UK and Australia, and was live in the scoring period: its ad-specification site (copyright 2026) lists units for eight named sites, and its ads.txt was updated on 25 September 2026. | 5 | direct record | supported as stated | high | [295, 253, 296, 297] |
| GAM-R-F428 | IGN publishes specifications for 24 ad units (IAB standard, custom and video) with dimensions, durations, file sizes, third-party serving rules (VAST 3.0 or higher), production lead times, downloadable templates and live demos, available across eight named sites (IGN, Eurogamer, GamesIndustry.biz, Nintendo Life, Pure Xbox, Push Square, Rock Paper Shotgun, VG247) in the US, UK and Australia, plus creative policies (LEAN, CPU limits, no auto-play audio); no counting approach for impressions or views is documented. | 5 | vendor assertion | supported as stated | high | [297, 298, 299, 671] |
| GAM-R-F430 | IGN documents measurement only as support for buyers' own vendors: third-party tags must come from Google Ad Manager's recognised vendors, monitoring tags are supported but blocking tags are not (because gaming content is often falsely flagged as unsafe), YouTube pre-roll can be measured only by YouTube-certified vendors, and Ziff Davis's UK guidelines say it does not itself implement ad verification tools; IMAGINE lists a 'Measure' module whose documentation was not accessible. | 5 | vendor assertion | supported as stated | medium | [299, 298, 672, 300, 301] |
| GAM-R-F434 | IGN's IMAGINE platform was announced on 24 September 2025 with a beta for select partners from Q4 2025 and full rollout planned for early 2026; trade reporting in September 2026 says IGN is using it to revamp display advertising, but the IMAGINE site was behind a sign-in on 29 September 2026, so its general availability at the cutoff is unconfirmed. | 5 | vendor assertion | supported as stated | medium | [300, 183, 301] |
| GAM-R-F436 | No independent assurance record for IGN was found: MRC's digital accredited list has no Ziff Davis or IGN service (the only 'IGN' string matches are inside the word 'Campaign'), and TAG Registry searches for 'Ziff', 'IGN' and 'Gamer Network' returned no company on 29 September 2026. | 5 | direct record | supported as stated | medium | [15, 333] |
| GAM-R-F437 | IGN sells pre-roll across its own sites and also on its YouTube channels and IGN1 channels, so part of IGN's video inventory overlaps with YouTube's gaming inventory and must be flagged as shared supply against any YouTube row. | 5 | inference | supported as stated | medium | [298] |
| GAM-R-F510 | Overwolf documents direct and programmatic buying through Deal ID or PMP (and Direct, PMP and PG for Gamer Grid segments, per a vendor announcement) and names DSPs among its partners: its FAQ lists The Trade Desk, Xandr, Adform, Yahoo, Basis and Simpli.fi within a list that mostly names SSPs, and its legal provider list (updated 29 December 2025) names The Trade Desk, Xandr, Yahoo, Adform and Zeta Global as "Advertising / Demand Partner"; no DSP-side confirmation and no protocol support were found, and whether Xandr is still a live buy-side route at the cutoff was not established. | 5 | vendor assertion | supported with qualification | medium | [44, 294, 46] |
| GAM-R-F527 | Venatus's claim of "exclusive inventory on sites like OP.GG, FUTBIN and PlanetMinecraft" does not hold for OP.GG as a global or US claim: OP.GG's ads.txt (last updated 28 September 2026, managed by RevIQ) authorises nitropay.com and overwolf.com as DIRECT sellers under seller ID scl26 in a section labelled Scalibur and does not contain the seller ID Venatus's sellers.json assigns to OP.GG, while op.gg also appears as a publisher in Playwire's and Nitro's sellers.json and planetminecraft.com in Playwire's; the same ads.txt does name venatus.com as the manager domain for the UK, France, Germany, Australia and New Zealand, so Venatus appears to represent OP.GG in those five markets, where exclusivity is neither confirmed nor contradicted by the file. | 5 | synthesis | supported with qualification | medium | [84, 307, 308, 309] |
| GAM-R-F528 | Venatus offers buyers a curated gaming-publisher PMP and direct sales with guaranteed impressions and fixed pricing, but no DSP is named on any Venatus page opened. | 5 | vendor assertion | supported as stated | medium | [249, 306] |
| GAM-R-F540 | Playwire does sell an advertiser-facing gaming media product: its advertiser pages offer 'Access 74M passionate gamers across 200+ premium communities' naming U.GG, FextraLife and Destructoid, and its advertising terms (publication date 21 November 2025) govern insertion orders with advertisers under the AAAA/IAB Standard Terms 3.0; the same pages bundle mobile in-app games, PC in-game video, Minecraft video and branded games, which are in-game placements outside this B2 row. | 5 | vendor assertion | supported as stated | medium | [310, 311, 312, 313] |
| GAM-R-F547 | Playwire's gaming supply is shared with other rows: its sellers.json lists 22 Enthusiast Gaming entries (including "EG - U.GG and Probuildstats (O&O)" and "EG - Fextralife (N)"), 34 "Overwolf LTD." entries, OP.GG and PlanetMinecraft, and Kidoz LTD typed as PUBLISHER; by root domain it shares roughly 66 to 67 domains with Venatus's sellers.json and 82 to 84 with Nitro's (the count depends on how domains are normalised), which contradicts "exclusive" in its "700+ exclusive premium publishers" claim at least for the shared domains. | 5 | synthesis | supported with qualification | medium | [309, 311, 308] |
| GAM-R-F602 | After the 30 August 2025 divestment, Enthusiast Gaming's filings say its inventory is sold programmatically and only 'to a much lesser extent' through direct sales, with direct campaigns expected 'through third parties' that are not named. The corporate site has no advertising page, but retained properties publish buyer-facing advertising pages: Icy Veins (updated 30 March 2026) documents its ad-unit layout and routes advertising inquiries to ads@enthusiastgaming.com, and Addicting Games offers 'Advertising Opportunities' with an inquiry contact. | 5 | direct record | supported with qualification | medium | [318, 250, 673, 674] |
| GAM-R-F603 | The programmatic path for Enthusiast's retained sites runs through outsourced intermediaries: its filings through August 2026 name only Playwire (partnership dated 11 March 2024 in the MD&A, announced 28 March 2024 per the AIF), but on 29 September 2026 the ads.txt files of all five retained sites checked were Freestar-hosted with MANAGERDOMAIN=freestar.com and OWNERDOMAIN=enthusiastgaming.com (headers dated 25 to 29 September 2026), while three of them still carry conflicting legacy owner or manager declarations (including managerdomain=playwire.com on Icy Veins) and a 2 May 2026 capture of u.gg/ads.txt showed Playwire as manager; Freestar's sellers.json lists 'Enthusiast Gaming Inc.' (seller 64) and Playwire's lists seven Enthusiast entries. | 5 | synthesis | supported with qualification | medium | [250, 318, 319, 320, 675, 676, 309] |
| GAM-R-F612 | This review found no documentation of measurement functions (viewability, attention, lift, reach and frequency, outcomes), outcome objectives or results, creative tooling, or independent verification for Enthusiast's retained inventory, in the corporate site pages, FY2025 AIF, Q2 2026 MD&A and ads.txt files screened; MRC's digital accredited list names neither Enthusiast nor its intermediaries. | 5 | synthesis | supported as stated | medium | [15, 250] |
| GAM-R-F620 | ESL FACEIT Group's advertising offer was live at the cutoff: its Ad Solutions page (published 28 July 2026, modified 4 September 2026) sells premium media on live and on-demand esports broadcasts and video and display on owned and operated channels, alongside custom brand integrations (including in-game placements and custom Fortnite worlds and maps) and content; only the premium media line is the cohort B2 product, the integrations and event activations being descriptive. | 5 | vendor assertion | supported as stated | high | [323] |
| GAM-R-F623 | EFG documents delivery reporting only: its case metrics are impressions, unique players, click-to-site rate and video view-through rate, and its claims of 'clear measurement across every touchpoint' and 'industry leading viewability' carry no method, metric definition or measurement provider. | 5 | vendor assertion | supported as stated | medium | [323] |
| GAM-R-F627 | Much of EFG's media audience sits on third-party platforms: the December 2025 channel page listed 374 Twitch channels (16M followers), 277 YouTube channels (78.1M subscribers), 105 Facebook, 32 Instagram, 144 Twitter and 7 TikTok channels, and the current page names YouTube, Twitch and Meta as distribution, so EFG's reach overlaps the Twitch and YouTube supply recorded in other rows and must not be summed with it. | 5 | vendor assertion | supported as stated | medium | [328, 323] |
| GAM-R-F628 | EFG's ad pages name no DSP, deal type or programmatic route, but its web inventory is offered programmatically through managed intermediaries: faceit.com and csstats.gg ads.txt files ('Publift Fuse', updated May and July 2026) declare MANAGERDOMAIN=publift.com and OWNERDOMAIN=faceit.com, and Publift's sellers.json lists 'FACEIT' as PUBLISHER for both domains; mobalytics.gg's ads.txt (updated 2 September 2026) declares MANAGERDOMAIN=nitropay.com, and NitroPay's sellers.json lists Mobalytics (seller 332) as PUBLISHER. | 5 | synthesis | supported with qualification | medium | [251, 324, 325] |
| GAM-R-F629 | EFG's US availability is shown only indirectly: its imprint names ESL Gaming America Inc. (New York), a US company (Kolex, Inc., Santa Monica) sponsored a FACEIT mission with partner advertising during the scoring period, and EFG broadcast IEM Atlanta from the US in May 2026, but no US sales contact, US buyer terms or US audience figure for the ad product was found. | 5 | synthesis | supported with qualification | medium | [331, 323, 327] |
| GAM-R-F631 | FACEIT documents an age rule for advertising, 'We do not provide personalized advertisements to players under the age of 18 years', in a mission privacy notice updated 22 July 2026 that also sets out, for one sponsored mission, the data exchanged with the partner, the recipients and deletion after 12 months without login; no brand-suitability controls for advertisers are documented, and FACEIT's Content Standards govern user content, not ad placement. | 5 | vendor assertion | supported as stated | medium | [329, 330] |
| GAM-R-F633 | The FACEIT and Mobalytics platforms that EFG lists as ad properties are group-owned: ESL FACEIT Group Limited (UK 07751649) is named as controller for FACEIT data, and Esports Insider reported on 17 March 2025 that EFG acquired the North American analytics platform Mobalytics, to operate as a standalone business. | 5 | synthesis | supported as stated | medium | [329, 327, 323] |
| GAM-R-F701 | Medal's ad units and sponsored placements were live inside the scoring period and render in Medal's own gameplay-clip app and website, not inside games. On 29 September 2026 medal.tv/grow offered ad units that 'utilize IAB standards', including an 'IAB-standard synced Medal homepage roadblock' (case study), plus sponsored stories, a sponsored top 10, branded watermarks and clip contests. Medal's help centre says ads appear for free users inside the Medal app, near the comment section or on clip pages; the US Google Play listing declares that the Android app contains ads, while the iOS App Store privacy label declares no tracking or advertising use (developer self-declarations, not verified by the stores). medal.tv/ads.txt, served 29 September 2026, carries a record annotated 20251118, and PubMatic's sellers.json (last updated 29-09-2026) lists medal.tv. In November 2025 Medal's CEO told GamesBeat that ads are Medal's main revenue driver and that he would like most ads on Medal to be about new games. On 3 September 2026 Green Man Gaming announced 'Buy Now' buttons on Medal clip pages (a commerce integration with no disclosed payment terms). The product fits cohort B2 (game-adjacent media); its demand mix between brands and game marketers is unquantified. | 5 | synthesis | supported with qualification | high | [252, 677, 290, 572, 678, 679] |
| GAM-R-F711 | No independent assurance record for Medal's ad products was found in the sources screened by the cutoff. The MRC's accredited and under-review digital pages contain neither 'Medal' nor 'Ferox'. The TAG Registry returned no company for Medal, Medal.tv or Ferox, while control searches returned records. No third-party viewability, invalid-traffic or suitability verification is documented, and the 90%+ viewability claim names no measurer. | 5 | direct record | supported as stated | medium | [334, 15, 252] |
| GAM-TWITCH-F03 | Amazon Ads sells Twitch media (video and display on the Twitch homepage, discovery pages and streamer channels) self-service through Amazon DSP or as managed service with an unpublished minimum spend, sells creator activations through account executives, and says advertisers need not sell on Amazon to buy Twitch ads. | 5 | vendor assertion | supported as stated | high | [2] |
| GAM-TWITCH-F06 | Twitch Streamables is sold in the Twitch catalogue but renders inside 'Twitch Partnered' mobile games as an opt-in, full-screen, non-skippable 30-second livestream, with delivery reporting; no partnered game is named. | 5 | vendor assertion | supported as stated | medium | [263] |
| GAM-TWITCH-F07 | Twitch media is bought programmatically only through Amazon DSP in the documents opened: display as non-guaranteed open-auction inventory (launched 14 December 2022, United States); video through premium preferred deals, a non-guaranteed tier between Twitch guaranteed and open marketplace buying with contextual collections (generally available 25 October 2023 to self-service advertisers); desktop Vertical Video Ads through open auction, programmatic guaranteed and private auction deals (mobile Vertical Video Ads open auction only); and through Amazon's Streaming TV ads, which include Twitch and carry no minimum budget when bought through sponsored ads in the Ads Console (US businesses), but a $10,000 recommended campaign minimum (self-service) or $50,000 minimum (managed service) when bought through Amazon DSP. | 5 | vendor assertion | supported with qualification | high | [258, 259, 260, 680] |
| GAM-TWITCH-F09 | Amazon's Twitch brand-safety guide (published 22 February 2026) documents first-party suitability tools for Amazon DSP buyers: three Inventory Tiers (Expanded, Standard, Restrictive) that exclude progressively more sensitive content, Contextual Collections to target or exclude content types such as gaming, lifestyle, Family Friendly games or Sports, signals from mandatory Content Classification Labels and ESRB ratings in Twitch's game database, standard impression-level content reporting, a monetisation gate under which new creators cannot carry ads, DV and IAS approval for IVT, viewability and domain/app-level brand safety, and a TAG Brand Safety Certified seal. | 5 | vendor assertion | supported as stated | high | [266] |
| GAM-TWITCH-F10 | Amazon's 6 November 2023 launch note names the Twitch Premium Video content tiers 'Expanded, Moderate (Default), Restrictive' and says they replaced the Content Ratings Filter, while the February 2026 guide names them Expanded, Standard and Restrictive; neither document states what each tier excludes. | 5 | vendor assertion | supported as stated | high | [269, 266] |
| GAM-TWITCH-F11 | Since 20 June 2023 Twitch streamers must apply whichever Content Classification Labels fit their stream; the six launch labels (Mature-Rated Games; Sexual Themes; Drugs, Intoxication, or Excessive Tobacco Use; Violent and Graphic Depictions; Significant Profanity or Vulgarity; Gambling) were joined in November 2024 by Politics and Sensitive Social Issues, and Twitch's current help article and API reference list seven. A Mature-rated game label is applied automatically for ESRB M-rated games, and from 20 July 2023 repeated mislabelling leads to labels applied to the channel for a set period. | 5 | vendor assertion | supported as stated | high | [265] |
| GAM-TWITCH-F13 | DoubleVerify (31 October 2023) and IAS (29 May 2024) each announced server-to-server measurement of Twitch inventory: DV for viewability, fraud/IVT and in-geo via the Twitch Ad Server and/or Amazon DSP, and IAS for viewability and IVT on Twitch. | 5 | vendor assertion | supported as stated | high | [335, 336] |
| GAM-TWITCH-F14 | No MRC record opened names Twitch as an accredited supply source. MRC's 7 November 2025 letter accrediting IAS's server-to-server Amazon DSP integration states that 'FireTV and Twitch supply sources are excluded'; MRC's 29 May 2026 letter on Amazon DSP's own first-party measurement lists Amazon Retail, IMDb, APS-enrolled web and app publishers, third-party exchanges, Fire TV and Alexa connected devices and does not mention Twitch; and neither the accredited nor the under-review digital list contains 'Twitch' (29 September 2026). | 5 | direct record | supported with qualification | high | [332, 15, 16] |
| GAM-TWITCH-F15 | TAG's registry records Twitch as enrolled in Verified by TAG with the Brand Safety Certified seal, achieved through independent validation by a third-party auditor in the Direct Seller category, applying to Twitch's global operations (read 29 September 2026). | 5 | direct record | supported as stated | high | [337] |
| GAM-TWITCH-F28 | Twitch Premium Video reaches console players inside the Twitch app on consoles, where the ads are not clickable, which is a game-adjacent placement on a console device rather than an in-game placement. | 5 | inference | supported as stated | medium | [47, 264] |
| GAM-R-F401 | Discord Quests were live throughout the scoring period (2 October 2025 to 2 October 2026) as rewarded ad units inside the Discord client. Discord's advertiser pages read on 29 September 2026 list two formats (Play Quests and Video Quests); its releases added Video Quests on Mobile and an Arena Quests alpha (2 October 2025), announced Instant Play Quests as forthcoming (9 March 2026; the Help Center calls Cloud Play Quests experimental), announced Mission Quests on 22 July 2026 after a beta that was running by December 2025 (the Resident Evil campaign from 23 July 2026 was the first Quest with an in-person activation, not the first Mission Quest), and announced Play Quest+ on 20 August 2026 with a first campaign for EA's Battlefield 6 planned for September 2026. | 5, 8 | vendor assertion | supported with qualification | high | [62, 281, 282, 681, 283, 284, 63] |
| GAM-F-F33 | Savvy Gaming Group, 100% owned by Saudi Arabia's Public Investment Fund, bought ESL (from MTG) and FACEIT and merged them into ESL FACEIT Group; the deal was announced on 24 January 2022 and MTG said it closed the ESL sale on 21 April 2022. | 5, 10 | vendor assertion | supported as stated | high | [321, 322] |
| GAM-R-F404 | Discord sells Quests on an insertion order with Discord Inc. under the IAB Standard Terms and Conditions v3.0 as amended by its own addendum (updated 12 May 2026), which treats Quests as flat-fee deliverables that use no third-party ad server, gives Discord 'ultimate authority and control' over the placement and frequency of Quest prompts, provides only summary reports of estimated metrics after the midpoint and end of each Quest, and treats impression and reach numbers as non-binding estimates with no makegoods, under New York law. | 5, 12 | vendor assertion | supported as stated | high | [280] |
| GAM-C-F04 | An in-game viewable impression requires at least 50% of the intended creative in view for one continuous second (two continuous seconds for video), a visible ad area of at least 1.5% of the screen, an ad angle no greater than 55 degrees on any one axis (recommended, with empirically supported alternatives allowed), segregated reporting of occluded impressions, polling every 200 ms (with a 3-of-5 majority option), and TV or monitor power state measured and considered for viewability. | 6 | direct record | supported as stated | high | [17] |
| GAM-C-F05 | The guidelines strongly encourage, but do not require, independent third-party audits of in-game ad-serving applications, and they require measurers to vet and periodically review SDK and API inputs to measurement. | 6 | direct record | supported as stated | high | [17] |
| GAM-C-F06 | The IAB Gaming Measurement Framework, released by the IAB alone on 26 June 2025 with no version number, is a reporting menu: it lists gaming ad formats by media type (display, video, audio, custom) and sets baseline (must-have) and additional (nice-to-have) metrics for each format, including viewability as a baseline metric, without counting thresholds; its display matrices refer the intrinsic in-game column to the Intrinsic In-Game guidelines, and its closing page calls those guidelines 'the primary guideline for gaming platforms and providers', although their own scope is ad-served intrinsic placements only. | 6 | direct record | supported with qualification | high | [348, 347, 354] |
| GAM-C-F10 | MRC's current accreditation letters for DoubleVerify (13 April 2026) and IAS (9 July 2026; IAS also holds a 7 November 2025 letter for third-party reporting of Amazon DSP) cover desktop, mobile web, mobile in-app and, for some metrics, connected TV; none names a gaming or console environment or lists the Intrinsic In-Game guidelines among the standards the accredited statistics comply with. IAS's separate In-Game Measurement service, which MRC had listed in pre-audit for desktop, mobile and gaming console environments, was removed from MRC's process in January 2026 after IAS withdrew. | 6 | direct record | supported with qualification | high | [18, 19] |
| GAM-C-F16 | On 29 September 2026 the TAG registry shows DoubleVerify and IAS holding Certified Against Fraud and Brand Safety Certified globally (independent validation) plus Certified for Transparency (North America and Europe, by TAG validation); Activision Blizzard Media and King.com Ltd. each hold Brand Safety Certified for the United Kingdom only (independent validation, Direct Seller) and Certified Against Fraud by self-attestation (Direct Seller) for North America, Asia Pacific, Africa, Central America, the Middle East, Oceania and South America (King also Russia); Gadsme is enrolled in Verified by TAG with no seal; no record was returned for Anzu, Frameplay, AdInMo, Adverty, Gameloft, Super League, Roblox, IION, Trivver or Xbox. | 6 | direct record | supported with qualification | high | [333] |
| GAM-C-F17 | Anzu announced on 12 August 2021 that it had earned the TAG Brand Safety Certified seal (independent validation by ABC), the TAG Certified Against Fraud seal and IAB UK Gold Standard 2.0; on 28 September 2026 neither the TAG registry (queries for Anzu, Anzu.io, Anzu Virtual Reality and Anz) nor IAB UK's certified and registered list showed Anzu. | 6 | synthesis | supported as stated | medium | [382, 333, 352] |
| GAM-C-F18 | On 28 September 2026 Anzu's media-value page still lists MOAT among its measurement partners, although Oracle's own FAQ states that all Oracle Advertising products, including Moat Analytics and Moat Reach, reached end of life on 30 September 2024 (a product retirement, with customer notices sent the week of 17 June 2024). | 6 | synthesis | supported as stated | high | [364, 339] |
| GAM-C-F19 | Anzu's measurement guide (22 August 2024) misstates the guidelines: it says an ad must be viewed for at least one second at under 55 degrees 'to be considered an impression' and that 'having the ad load no longer counts', whereas the guidelines count an impression at load and begin-to-render, define a viewable impression separately, and set the 55-degree limit as a recommendation with empirically supported alternatives. | 6 | synthesis | supported as stated | high | [359, 17] |
| GAM-C-F20 | Two cohort A vendors make compliance claims with no accreditor record behind them: Anzu's boilerplate (January 2024) calls it 'compliant with the industry measurement standards outlined by the MRC and IAB', and Adverty's home page (accessed 28 September 2026) says its tracking is 'verified by leading industry stakeholders and 100% in line with the latest IAB and MRC measurement guidelines' without naming a verifier; neither vendor appears on MRC's accredited or under-review lists. | 6 | synthesis | supported as stated | high | [384, 353, 15, 16] |
| GAM-C-F22 | For game-adjacent media, MRC's accreditations touch YouTube only in general terms, with no gaming-specific scope: Google YouTube Brand Safety content-level processes (desktop, mobile web, mobile in-app); DoubleVerify, IAS and Zefr third-party reporting of YouTube Ads Data Hub data (desktop, mobile web, mobile in-app; DV's excludes YouTube TV, YouTube Kids and YouTube Shorts); Zefr's third-party reporting of YouTube content-level brand safety (desktop, mobile web, mobile in-app, CTV); and TrueView views within Google Ads, DV360 and Ads Data Hub. Amazon DSP's MRC accreditation names its supply sources without Twitch. Twitch holds TAG Brand Safety Certified for global operations; Venatus holds TAG Brand Safety Certified for the UK and is IAB UK Gold Standard certified (Sell Side Support); no MRC record was found for Twitch, Discord or Venatus. | 6 | direct record | supported with qualification | high | [15, 333, 352] |
| GAM-C-F23 | Attention measurement in games is offered or claimed by several parties, each by vendor statement only: DoubleVerify said on 1 April 2025 that advertisers could measure attention with DV Authentic Attention on Roblox Rewarded Video (its billboard coverage named only fraud/IVT and viewability), while DV's MRC accreditation for Authentic Attention covers display and video in desktop, mobile web and mobile in-app and does not name Roblox or in-game environments; Frameplay announced an 'Intrinsic Time-in-View' metric on 5 April 2022 that it says Lumen and eye square each validated against their own eye tracking on one campaign (McCormick's Frank's RedHot in Frameplay's Basketball Battle); and dentsu and Lumen released eye-tracking gaming research on 16 January 2024 covering Twitch livestreams, Activision Blizzard rewarded video and Anzu intrinsic ads. | 6 | vendor assertion | supported with qualification | medium | [351, 18, 356, 384] |
| GAM-D-F12 | IAS's own in-game announcements name only viewability and invalid-traffic functions: Roblox (12 September 2024; 3D viewability and IVT for display and video on mobile web and in-app), Anzu (28 July 2022 for mobile, extended on 11 July 2023 to 2D and 3D environments on mobile, PC and console), Gadsme (15 December 2022) and Frameplay (5 October 2023, IVT and impression counts). Roblox's own announcement of 1 April 2025 says DoubleVerify and IAS 'will offer' coverage including brand safety and suitability on Roblox, and that IAS brand-safety and suitability pre-bid solutions are available for programmatically bought Roblox media through enabled DSPs; no IAS document opened describes a suitability classification delivered inside an in-game placement. | 6 | vendor assertion | supported with qualification | medium | [371, 350] |
| GAM-R-F204 | Gameloft reports network-level viewability ("Valid & Viewable" 87.9% for display and 93.3% for video in mobile in-app, against a 67.9% benchmark) without a source, and elsewhere an 80% completed-viewability rate under a combined source line that includes an "Integral Ad Science x Gameloft for Brands (S1 2025) Data Study"; it labels COMBO! "IAS compliant". No measurement function is documented as available to buyers, and no IAS record of the relationship was found. | 6 | vendor assertion | supported with qualification | medium | [55, 377] |
| GAM-R-F325 | EA claims IAB-aligned impression measurement and viewability and invalid-traffic measurement "in partnership with Integral Ad Science"; IAS's published sitemaps (4,656 URLs, read 29 September 2026) contain no page about Electronic Arts, while they do contain in-game pages for Anzu and Gadsme. | 6 | synthesis | supported as stated | medium | [25, 376] |
| GAM-DOUBLEVERIFY-F05 | On 6 August 2026 DoubleVerify agreed to be acquired by Neptune BidCo US Inc., parent company of Nielsen Company (US), LLC and an affiliate of funds managed by Elliott and Brookfield, for $13.60 per share in cash (enterprise value about $2.15 billion); the deal is expected to close by Q1 2027 subject to a stockholder vote and regulatory approvals. The preliminary proxy (11 September 2026) says both parties filed HSR notifications on 20 August 2026 (a DoubleVerify customer presentation filed on 25 August 2026 gives 19 August) and that the waiting period would expire at 11:59 p.m. ET on 21 September 2026 unless terminated early, lengthened by a pull-and-refile or extended by a Second Request; no public record read by 29 September 2026 shows which occurred. The outside date is 6 May 2027 (extendable to 6 August 2027), and termination fees are $60 million (payable by DoubleVerify) or $144 million (payable by Parent) in specified circumstances. | 6 | direct record | supported with qualification | high | [380, 381, 584] |
| GAM-DOUBLEVERIFY-F26 | Synthesis: in the sources screened by 29 September 2026, DoubleVerify documents gaming coverage in three places: Roblox (billboard IVT and viewability through a custom integration; Rewarded Video functions framed prospectively), Microsoft Casual Games (named in DoubleVerify's 15 October 2025 release as one of the Microsoft properties where brand-suitability, viewability and invalid-traffic measurement can be activated for Microsoft Audience ads, with format placement and devices not stated), and Zynga (a reported DoubleVerify attention figure, with DoubleVerify's logo on Zynga's ad-solutions page). No DoubleVerify record was found for any cohort A intrinsic network (Anzu, Frameplay, Gadsme, AdInMo, Adverty, IION, Gameloft for brands, Super League), and the other gaming mentions (XBOX Media Solutions, Gadsme, Overwolf) are the counterparties' own words without scope. | 6 | synthesis | supported with qualification | medium | [351, 203, 361, 398, 375] |
| GAM-FRAMEPLAY-F10 | Frameplay documents its own impression-counting approach for developers, capturing player view metrics for up to 60 seconds per displayed ad and aggregating them into one impression weighted by on-screen time, screen share, ad-space share, orientation, skew and obstruction, and its 2022 release says viewability uses Frameplay's proprietary requirements; no source opened states alignment with the IAB/MRC Intrinsic In-Game Guidelines 2.0. | 6 | vendor assertion | supported as stated | medium | [355, 356, 357] |
| GAM-ROBLOX-F08 | Compared by this review with the IIG Guidelines 2.0 (August 2022), Roblox documents image-ad impression thresholds with the same headline numbers as the guideline's viewable display criteria (at least 50% of pixels visible, at least 1 second looked at, 1.5% of the viewport, an angle of up to 55 degrees), but its wording does not establish equivalence. Roblox does not say the second is continuous, that the 1.5% refers to the visible portion of the ad relative to the screen (IIG section 3.2.2), or how repeat exposures within a session are counted (IIG counts one viewable impression per user session), so alignment is unverified. Separately, its autoplay video impression (0.5 seconds) is shorter than the guideline's 2 continuous seconds. Roblox applies these definitions to the Impressions column of its Ads Manager (beta) reports and does not claim alignment with the guidelines. | 6 | synthesis | supported with qualification | medium | [17, 208] |
| GAM-XBOX-F08 | XBOX Media Solutions' measurement page offers 'first and third-party measurement' and names eight partners (Nielsen Catalina, Circana, Kantar, GWI, In Market, MRI Simmons, DoubleVerify, Integral Ad Science) without stating which formats, environments or metrics each measures, and its headline 98% viewability and 96% completion figures are footnoted to Moat mobile in-app benchmarks from Q1 2023. | 6 | vendor assertion | supported as stated | high | [361, 362, 224] |
| GAM-XBOX-F21 | XBOX Media Solutions' about-us page claims 'IAB UK Gold Standard 2.1 Certified', but IAB UK's own Gold Standard certified and registered list, read on 29 September 2026, does not include XBOX Media Solutions, Activision Blizzard Media, King or Microsoft, although XBOX Media Solutions is listed in IAB UK's member directory. | 6 | synthesis | supported as stated | medium | [223, 352, 682] |
| GAM-C-F09 | On 29 September 2026 MRC's Services Under Review (Digital) lists one gaming-environment service: Trivver, pre-audit reviewed July 2022 and audit in process as of April 2026, for Impressions, T-Views and Product Actions/Actions/Engagements across desktop, mobile web, mobile in-app and gaming consoles (including 2D, 3D and extended, augmented, virtual and mixed realities); no other in-game service is under review. Until January 2026 MRC also listed 'IAS In-Game' in pre-audit (display and video rendered and viewable impressions, inclusive of GIVT filtration, in desktop, mobile and gaming console environments); MRC removed it from in-process status in January 2026 because IAS did not intend to engage in a third-party audit and withdrew from the accreditation process. | 6, 12 | direct record | supported with qualification | high | [16, 17] |
| GAM-D-F02 | Roblox's Advertising Standards (5 June 2026) use content ratings as suitability rules: self-serve and independent ad integrations may not promote Mature-rated games or R and TV-MA content; Roblox-managed ads may promote them only to 18+ audiences; ads to users aged 9 to 12 may not promote M or PEGI 18 games; and ads to users aged 5 to 8 may also not promote Teen or PEGI-12 games or PG-13 and TV-14 content. | 7 | vendor assertion | supported as stated | high | [20] |
| GAM-D-F03 | Roblox gates which experiences can carry its immersive ads: the game must be public, the owner ID-verified with two-factor authentication, the Maturity and Compliance Questionnaire approved by Roblox, and at least 2,000 unique visitors a month maintained; the Advertising Standards separately bar publishers from misclassifying experience metadata, including the experience guideline rating, to increase ad delivery, and from using PolicyService API to show ads to users not allowed to see them, including users under 13. | 7 | vendor assertion | supported as stated | high | [407, 20] |
| GAM-D-F07 | SuperAwesome says it is the sole third-party vendor of Roblox advertising for under-13s worldwide and that its in-house staff manually play each title and evaluate it against 11 brand-suitability flags (examples given: mild or severe violence, realistic weapons); it separately offers 80+ custom contextual categories for audience targeting; its release says Roblox will keep direct sales of younger-audience inventory to selected strategic partners in North America. The flags are not enumerated, and how manual play covers the roughly 15,000 new games SuperAwesome says are uploaded daily is not documented. | 7 | vendor assertion | supported with qualification | medium | [415, 206] |
| GAM-D-F08 | Roblox's Content Maturity Labels (Minimal, Mild, Moderate, Restricted) rest on creator-completed questionnaires with creator-determined content descriptors, and Restricted experiences are available only to age-checked 18+ users; for games offered to users under 16 through Roblox Kids and Roblox Select, Roblox said on 13 April 2026 that it will assign the label after a review. Roblox announced on 3 September 2025 that IARC ratings (ESRB in the US) would replace these labels 'when implemented'; on 13 and 30 April 2026 it said the transition would start 'later this year', and its help page (edited 28 May 2026) and Creator Hub documentation (updated 25 September 2026) still describe Roblox's own four labels. | 7 | vendor assertion | supported with qualification | medium | [412, 413] |
| GAM-D-F10 | Anzu's brand-safety policy (version April 2023, linked from its live site on 28 September 2026) documents advertiser allow and block lists by publisher, app or game, country and platform for open- and private-auction campaigns, a pre-approved allow list for insertion-order and private-auction buys, a manual human check of all direct campaign creatives, a category block list, third-party verification tags, suitability to each game's PEGI rating on a reasonable-efforts basis, and takedown within one Israeli business day. | 7 | vendor assertion | supported as stated | high | [397] |
| GAM-D-F16 | Epic's Fortnite Developer Rules (last updated 29 April 2026) act as suitability controls on brands in islands: all developers must obtain IARC age ratings on publishing, content must not exceed Fortnite's maximum age rating, content targeting children under 13 as its primary audience is not accepted, paid commercial content must be disclosed and may be published only by individual developers aged 18 or older, sponsored commercial content may not promote categories including tobacco, alcohol, marijuana and firearms, and island promotion must not target under-13s. | 7 | vendor assertion | supported as stated | high | [220] |
| GAM-D-F18 | Discord's Ads Policy (updated 9 September 2026) applies to all native and third-party ads on Discord, including ads in Activities, and to the ad, landing page, username and reward; it prohibits ads intended solely for children under 13 (Discord's minimum age) and categories including alcohol, adult sexual content and real-money gaming, but the policy page documents no advertiser-side placement suitability control. | 7 | vendor assertion | supported as stated | medium | [288] |
| GAM-D-F19 | Twitch documents advertiser suitability controls built on mandatory content labels: since 20 June 2023 streamers must apply whichever Content Classification Labels fit their stream (six at launch; seven in Twitch's current help article and API reference, after a Politics and Sensitive Social Issues label was added in November 2024), mislabelling brings warnings and Twitch-applied labels, and a Mature-rated game label is applied automatically for ESRB M-rated games. Per Amazon's Twitch brand-safety guide (dated 22 February 2026), these label signals inform three Inventory Tiers (Expanded, Standard, Restrictive), Contextual Collections to target or exclude content types such as Family Friendly games, and on-demand impression-level content reporting in Amazon DSP. | 7 | vendor assertion | supported with qualification | high | [265, 266] |
| GAM-D-F21 | The World Federation of Advertisers discontinued GARM, the Global Alliance for Responsible Media created in 2019, on 9 August 2024, saying allegations had drained its resources; on 29 July 2026, announcing a settlement of the GARM litigation with X Corp., WFA stated it will not form or restart GARM or a similar initiative. | 7 | direct record | supported as stated | high | [402, 418] |
| GAM-D-F22 | IAB Tech Lab's Content Taxonomy page (last updated 11 February 2026) lists version 3.1 (December 2024; open for public comment until 24 January 2025) as current. Its Content Taxonomy 3.0 implementation guidance, both the September 2021 public-comment draft and the version on IAB Tech Lab's GitHub (implementation.md, last changed 5 August 2025), maps platform game genres to taxonomy categories under a Game form-factor vector and defines brand-safety floor and suitability risk levels by reference to the September 2020 GARM and 4A's APB Brand Safety & Suitability Framework; WFA announced on 9 August 2024 that GARM would discontinue its activities, the notice does not say the framework was withdrawn, and the GitHub guidance a year later still tells buyers to familiarise themselves with it. | 7 | direct record | supported with qualification | high | [683, 419, 402, 403] |
| GAM-D-F26 | AdExchanger reported on 4 June 2026 that Roblox faced at least 130 child-safety lawsuits in federal court plus suits by several states; the federal cases are centralised as a multidistrict litigation (MDL No. 3166, Northern District of California, JPML transfer order of December 2025), not a class action. Alabama and West Virginia announced settlements with Roblox on 21 April 2026, and Roblox's Q2 2026 Form 10-Q (30 July 2026) lists settlement agreements with Alabama, Mississippi, Nevada, South Dakota and West Virginia and $91 million accrued in the first half of 2026. The sources screened document no advertiser pulling spend from Roblox over these matters. | 7 | synthesis | supported with qualification | medium | [216] |
| GAM-D-F28 | Content ratings are the common suitability signal across the documents opened: Roblox restricts ad content by ESRB and PEGI levels per audience age band and announced IARC adoption; Twitch auto-labels ESRB M-rated streams and holds ESRB ratings in its game database; Epic requires IARC ratings for every island; Adverty lets advertisers include or exclude PEGI levels; Gadsme checks PEGI at onboarding; and Anzu targets suitability to each game's PEGI rating. | 7 | synthesis | supported as stated | high | [20, 266, 220, 81, 398, 397] |
| GAM-D-F29 | Coded against the dimension 12 suitability subcriterion, 4 of the 8 cohort A seed networks document suitability controls in operational detail in vendor documentation (Anzu, Gadsme, Adverty, and AdInMo, the last only in an archived copy of 7 October 2025 because its live page returned HTTP 525), 2 document them in outline (Gameloft for brands in a live brand-safety policy page; Frameplay in its Terms of Use of 1 January 2023), and 2 (iion, Super League) show marketing language only in the pages opened; in cohort B1, Roblox documents controls in operational detail and XBOX Media Solutions in marketing only; in cohort B2, Twitch documents advertiser controls in operational detail, Discord documents an ad-content policy but no placement control, and six seeds were not screened. | 7 | synthesis | supported with qualification | medium | [397, 398, 81, 408, 266, 224] |
| GAM-E-F04 | Rule text: the amended COPPA Rule lists serving contextual advertising and capping ad frequency among the activities that support internal operations (§ 312.2; the Commission also places ad attribution within the listed activities); it exempts from parental consent only the collection of a persistent identifier, with no other personal information, used solely for that support and disclosed in the § 312.4(d)(3) notice (§ 312.5(c)(7)); it bars using information collected for those activities for behavioural advertising or profiling; and it requires separate verifiable parental consent for disclosures to third parties that are not integral to the service (§ 312.5(a)(2)). The Commission's accompanying statement says the Rule does not prohibit collecting personal information for targeted advertising where the parent gives opt-in consent, which must be off by default and cannot be a condition of the child's access, and gives third-party advertising as an example of a non-integral disclosure; read together (an inference), contextual in-game ads in child-directed games can run without parental consent under those conditions, while behaviourally targeted ads need prior parental opt-in consent, plus separate consent where the child's data goes to third parties. | 7 | direct record | supported with qualification | high | [23] |
| GAM-E-F08 | The 25 February 2026 statement creates no substantive rights, keeps the FTC's right to bring individual cases, and lasts until final rule amendments or withdrawal. It never mentions advertising. Its first condition, however, bars any use of age-verification data other than verifying age, so an operator relying on the statement cannot use that data to target or measure ads. This last point is a reading of the text. | 7 | inference | supported as stated | high | [422] |
| GAM-E-F12 | The FTC's COPPA Safe Harbor page, read on 29 September 2026, lists 20 related case entries (the first page of the FTC's list of cases tagged Children's Privacy + COPPA), each dated by its case page's last-updated date rather than its filing date. Classified by complaint, 16 are COPPA Rule actions filed between 24 April 2019 and 29 September 2025: four against game companies (HyperBeard, filed 4 June 2020; Epic Games, 19 December 2022; Microsoft, 5 June 2023; Cognosphere, 17 January 2025) and one against an ad-technology company (OpenX, 15 December 2021). Four are not COPPA Rule actions: Epic's separate administrative billing matter, Miniclip's Section 5 case over a false claim of safe-harbor membership, Retina-X and Illuminate Education. None of the 20 entries is an action against an in-game ad network or game-ad vendor, so in that list this review found no FTC COPPA action against such a vendor filed from January 2023 to 29 September 2026. | 7 | synthesis | supported with qualification | medium | [427] |
| GAM-E-F20 | Roblox's Q2 2026 shareholder letter states that, as of the seven days ended 30 June 2026, 35% of users who had age-checked globally were under 13 and 38% were aged 13-17, so 73% were under 18. In the US, O18 users were about one-third of age-checked DAUs, so roughly two-thirds were under 18 (derived; the letter gives no US under-13 share). Under Roblox's Advertising Standards none of these under-18 users can receive personalised ads. These shares apply only to the age-checked population, which Roblox says may not represent the whole user base. | 7 | inference | supported with qualification | medium | [43, 20] |
| GAM-E-F21 | Microsoft's Privacy Statement (September 2026) says it does not deliver personalised advertising to children whose Microsoft account birthdate shows them to be under 18. King's privacy policy (26 February 2026), which covers Candy Crush and the other King games where XBOX Media Solutions sells in-game ads, says King does not knowingly target interest-based ads to anyone under 13, or older where local law requires. Where local laws allow, King opts players into targeted advertising by default. | 7 | vendor assertion | supported as stated | high | [21, 234] |
| GAM-E-F24 | Google's Play Age Signals terms (last modified October 2025) allow developers to use age signals only to provide age-appropriate content and experiences in compliance with laws. They expressly forbid use for advertising, marketing, user profiling or analytics, on pain of losing API access or having apps removed. The API began returning signals for Brazil on 17 March 2026 and for eligible Texas users whose accounts were created after 28 May 2026. | 7 | vendor assertion | supported as stated | high | [221] |
| GAM-E-F27 | Louisiana HB 570 (2025 Regular Session, Act 481) never took effect: Act 185 of 2026 (HB 977), signed on 15 May 2026, provides that Act 481 shall not become effective and re-enacts app-store rules from 1 July 2027, under which developers may not sell age category data to any person. As enacted in 2025, the law would have required developers to check age category and, for minors, obtain verifiable parental consent through the app store, barred sharing age category data, and set Attorney General enforcement with civil fines of up to $10,000 per violation. | 7 | direct record | supported as stated | high | [429] |
| GAM-E-F28 | Texas SB 2420, the App Store Accountability Act, was due to take effect on 1 January 2026. A federal district judge in Austin preliminarily enjoined it in December 2025. The Fifth Circuit entered a temporary administrative stay by 28 May 2026 and then a stay pending appeal (reported by 10 June 2026). On 6 July 2026 the Supreme Court declined to reinstate the injunctions, so the law can be enforced while the appeal continues. Google returns Texas age signals for accounts created after 28 May 2026, and Apple has applied the law to new Texas accounts since 4 June 2026. | 7 | synthesis | supported with qualification | medium | [428, 431, 221] |
| GAM-E-F29 | California's Digital Age Assurance Act (AB 1043, Chapter 675 of 2025, approved 13 October 2025; amended by AB 1856, Chapter 184 of 2026, approved 10 September 2026) becomes operative on 1 January 2027. A developer that receives an age-bracket signal from an operating system or app store is deemed to have actual knowledge of the user's age range, including on its own websites used to create or log into the account, unless it holds internal clear and convincing information of a different age, which then governs. It must use the signal to comply with the law and may not share it with third parties for any purpose the Act does not require (Civ. Code § 1798.501(d) as amended). | 7 | direct record | supported with qualification | high | [430] |
| GAM-E-F30 | Maryland's Online Data Privacy Act (Chapter 455 of 2024, effective 1 October 2025) forbids a controller to process personal data for targeted advertising, or to sell personal data, where it knew or should have known the consumer is under 18. Its definition of targeted advertising excludes ads based on the context of a current search, website or app visit, ads based on activity within the controller's own sites or apps, and processing solely to measure ad frequency, performance or reach. | 7 | direct record | supported as stated | high | [436] |
| GAM-E-F35 | Gadsme's privacy policy (updated 24 February 2026) says its services collect device advertising identifiers (IDFA, Google Advertising ID), location data and ad-view history, and app data that includes a 'COPPA flag'. It says Gadsme does not knowingly collect personal information from children under 13. The policy does not say how traffic carrying the COPPA flag is handled. | 7 | vendor assertion | supported as stated | medium | [444] |
| GAM-R-F231 | For dimension 12, Admazing documents data practices in operational detail (device identifiers, IP and geolocation collected; used for delivery, retargeting and frequency; information "typically" deleted quarterly), states in outline that its services are not directed to under-18s and that it collects no personal information from them, and offers only marketing language on brand safety plus advertiser content guidelines that were not found on the site. | 7 | vendor assertion | supported with qualification | medium | [547, 129, 548] |
| GAM-R-F315 | Kidoz documents all three suitability subcriteria: brand suitability (publisher vetting and SDK-based environment classification before inventory becomes buyable; human and programmatic review of every creative), data practices (no advertising identifiers; short-lived session identifier; approximate location and identifiers deleted after at most two months), and age limits applied network-wide (every bid request treated as non-consenting; separation of the all-ages Prado platform; listed in Google Play's Families self-certified ads SDKs). | 7 | synthesis | supported as stated | medium | [69, 181, 563, 65, 564] |
| GAM-R-F333 | EA's company-wide privacy policy (last updated 23 June 2025) documents ad-serving data practices and an account-level opt-out of targeted advertising on PC and console. It also states an age-based limit: for players who identify as children under the minimum age of digital consent, identifiers are used only for internal operations such as contextual ads or frequency capping, with parental consent where required and console parental controls respected. No brand-suitability controls for EA Advertising are documented. | 7 | vendor assertion | supported with qualification | medium | [410] |
| GAM-R-F354 | Zynga Ads claims a 100% brand-safe owned-and-operated environment; the Take-Two privacy policy (February 2026) documents sharing of identifiers, IP address and profile inferences with advertising providers and a targeted-ads opt-out, and outlines that child-directed or all-ages services may use an age gate; no suitability mechanism or age-based ad limit specific to Zynga Ads is documented. | 7 | vendor assertion | supported as stated | medium | [244, 411] |
| GAM-ROBLOX-F20 | Roblox's advertiser-side suitability controls are named but not documented in operating detail for the scoring period: its brands pages list 'Suitability Tiers, Age and Geo controls' and Advertiser Select ('Our hand curated list of brand safe experiences'), and the only operational tier definitions found are in a January 2024 post keyed to experience 'age recommendations' that says the tiers 'may change over time'; Roblox's standards separately bar publishers from misclassifying the experience guideline rating to increase ad delivery. | 7 | vendor assertion | supported as stated | medium | [42, 217, 408, 20] |
| GAM-XBOX-F06 | Activision Blizzard Media's Brand Safety page, as captured by the Internet Archive on 21 January 2026, stated that it runs inventory only in games 100% owned and operated by the Activision Blizzard King group, offers audience targeting based primarily on first-party data (geography, demographics), lets advertisers buy specific packages of games (for example to avoid age-restricted titles such as Call of Duty), and will make all reasonable efforts to take down or remedy unsuitable delivery on request within 24 hours, longer for game integrations and sponsorships. | 7 | vendor assertion | supported as stated | medium | [409] |
| GAM-D-F11 | DoubleVerify's post of 1 April 2025 says its Roblox coverage differs by format: for in-game billboard image and video ads it extended only fraud/IVT and viewability measurement through a custom integration on Roblox's direct-sold inventory, while for Rewarded Video, then about to launch, advertisers 'will be able to' measure brand safety and suitability and use DV's pre-bid brand-suitability controls through DV tags. No later DV or Roblox record opened in this check confirms the Rewarded Video suitability function in operation. | 7, 12 | vendor assertion | supported with qualification | high | [351] |
| GAM-ROBLOX-F05 | In its Q2 2026 letter Roblox reported that 57% of global DAUs had age-checked by the end of the quarter (US above 70%; under-18 penetration 60% globally and 75% in the US); of age-checked users, 35% were under 13, 38% were 13 to 17 and 27% were 18 or over globally in the seven days ended 30 June 2026, and users 18 or over were about one-third of age-checked DAUs in the US. | 7, 12 | direct record | supported as stated | medium | [43, 246] |
| GAM-E-F01 | The FTC announced its amended COPPA Rule on 16 January 2025. The amendments were published in the Federal Register on 22 April 2025 (90 FR 16918) and took effect on 23 June 2025. Most regulated entities had until 22 April 2026 to comply; the safe-harbor provisions in § 312.11(d)(1), (d)(4) and (g) carry their own deadlines. | 7, 13 | direct record | supported as stated | high | [23, 421] |
| GAM-E-F07 | The FTC's enforcement policy statement of 25 February 2026, adopted 2-0, applies only to mixed-audience and general-audience operators. Services primarily directed to children must still treat every user as a child. The FTC will not bring COPPA actions over personal information collected, used or disclosed solely to determine age if six conditions are met: use only for age verification, disclosure only to third parties that give written assurances, prompt deletion, notice in the privacy policy, reasonable security, and reasonable steps to ensure accurate results. The operator must also comply with COPPA in every other respect. 'Age verification' in the statement covers estimation, verification and inference tools. | 7, 13 | direct record | supported as stated | high | [422, 423] |
| GAM-G-F01 | On 23 July 2026 Xbox began a limited-time test, open to Xbox Insiders, in which players stream games they own free with ads; ads play before sessions begin, sessions are limited to one hour, and Microsoft states that the core game experience is unchanged and that it does not interrupt the flow of gameplay, so the placement renders outside gameplay. | 8 | vendor assertion | supported as stated | high | [202] |
| GAM-G-F02 | Variety reported on 23 July 2026 that test players see a two-minute block of multiple ads from globally recognisable, market-specific brands before joining the cloud queue, then get a one-hour session without ad breaks; eligible streams are games the player owns or free-to-play games, with Forza Horizon 6, 007 First Light, Call of Duty: Warzone, Genshin Impact and War Thunder named as examples, which are eligible titles, not titles carrying ads. | 8 | vendor assertion | supported as stated | medium | [459] |
| GAM-G-F03 | On 3 September 2026 Xbox announced that from November 2026 Game Pass plans will carry monthly cloud-gaming hour limits (Ultimate 15, Premium 10, Essential 5) and that players without Game Pass will be able to buy cloud playtime hours to stream games they own; the announcement does not mention the ad-supported test, and no dated general-availability plan for ad-supported streaming was found in the sources screened by 28 September 2026. | 8 | vendor assertion | supported as stated | medium | [460] |
| GAM-G-F04 | XBOX Media Solutions markets custom integrations on console that range from 'immersive in-game moments' to Xbox dashboard and store placements, and a console product it calls XBOX Experiences built on native display formats and branded content; the page names no console title, launch date, specification, US availability or programmatic route, and whether XBOX Experiences render inside gameplay cannot be established from it. | 8 | vendor assertion | supported as stated | medium | [24, 458] |
| GAM-G-F06 | In April 2022 TechCrunch and XDA, relaying Business Insider's unnamed sources, reported that Microsoft was building a program for brands to advertise in free-to-play Xbox games (for example as rendered billboards in racing games) through a private marketplace, set to launch by the third quarter of 2022; Microsoft said it had nothing further to share, and no product-specific evidence that the program launched was found in the sources screened by 28 September 2026, so its status stays unresolved. | 8 | synthesis | supported as stated | medium | [231, 461] |
| GAM-G-F07 | Microsoft Technology Licensing's patent application US 2026/0260258 A1, filed 23 April 2026 as a continuation of an application filed 1 May 2023 (now abandoned) and published 3 September 2026, claims a system that suspends a user's interaction with a game on a promoted-content trigger, presents promoted audiovisual content, and grants a credit that delays the next suspension; a published application is not a product, plan or launch. | 8 | direct record | supported as stated | high | [449] |
| GAM-G-F09 | Live TV on PS5 launched for PS5 players in the US on 31 August 2026 as a free ad-supported service from Sony Pictures Television with more than 100 live channels in the console's Media section; it is TV inventory on a console, rendered outside gameplay, with Canada under consideration. | 8 | vendor assertion | supported as stated | high | [446] |
| GAM-G-F10 | Variety reported on 31 August 2026, in an interview with Sony Pictures Entertainment's Peter Wood, that Sony will work with Publica and PubMatic, two companies Variety describes as focused on connected-TV advertising, for Live TV on PS5; no Sony, Publica or PubMatic release was found, and deal types and DSP access are not described, so treating the buying route as a CTV route is an inference. The same report cites Sony Interactive Entertainment data of more than 95 million PS5 units sold worldwide as of 30 June 2026, a hardware figure that is not an ad-eligible audience. | 8 | vendor assertion | supported with qualification | medium | [468] |
| GAM-G-F11 | Digiday reported on 5 August 2026 that Sony Interactive Entertainment Media Solutions posted a global director role in mid-July 2026 to scale a client-partnership sales organisation across the PlayStation ecosystem in programmatic, FAST and CTV and 'emerging ad solutions'; hiring is a signal, not a product, and Digiday says none of Sony's earlier attempts, including a 2022 in-game system reported by Bloomberg, became a durable, scaled ads business. | 8 | synthesis | supported as stated | medium | [451] |
| GAM-G-F12 | Sony Interactive Entertainment's patent application US 2026/0138036 A1, filed 15 November 2024 and published 21 May 2026, describes in its abstract a gaming console that determines that an appropriate time for alternate content is about to begin during gameplay, estimates how long that time will last, and shows alternate content on at least part of the screen for a shorter duration (its figure shows a 'waiting in queue' screen); the abstract says 'alternate content', not advertising, and a published application is not a product or launch. | 8 | direct record | supported with qualification | high | [450] |
| GAM-G-F13 | No Nintendo advertising product or public Nintendo policy text on in-game advertising was found in the sources screened by 28 September 2026; AUTOMATON reported on 20 May 2025 that Nintendo Switch store review flagged an indie game's parody commercials as in-game advertising 'against the platform's rules' and that the developer had to remove them. | 8 | synthesis | supported as stated | low | [452] |
| GAM-G-F14 | NVIDIA's GeForce NOW FAQ says free members have seen up to two minutes of video ads before each free session, while in queue, since 5 March 2024, that ads do not interrupt gameplay, that members under 16 cannot opt in to personalised ads, and, at access on 28 September 2026, that NVIDIA is moving to a new advertising provider so free members may see fewer or no ads. | 8 | vendor assertion | supported as stated | high | [447] |
| GAM-G-F15 | An Android Authority explainer dated 6 May 2025 says Amazon Luna shows ads while connecting to a game but not while the game loads or during gameplay; no Amazon document on Luna ads was found, and CNBC reported on 23 July 2026 that Amazon was adding Luna to Prime Video without mentioning ads. | 8 | synthesis | supported as stated | low | [448, 458] |
| GAM-G-F16 | Valve's Steamworks documentation states that Steam contains no paid advertising and does not support advertising models in games distributed on Steam, bars requiring players to watch ads to play and rewarding ad engagement, and allows real brands as product placement when not disruptive. | 8 | vendor assertion | supported as stated | high | [453] |
| GAM-G-F17 | The Epic Games Store requires developers to disclose paid advertisements and promotional content in an offer, including product placement, logos and slogans for real-world brands, and labels a declared offer as promotional content; no Epic Games Store ad product for advertisers was found in the sources screened. | 8 | vendor assertion | supported as stated | medium | [454] |
| GAM-G-F19 | On 13 May 2026 Anzu announced a partnership with Strikerz Inc. to sell in-game advertising (LED perimeter boards, pitch-side assets, branded scoreboards, referee kits) in the free-to-play football game UFL on PC only, although UFL launched on console in December 2024; UFL's 7 million players worldwide is a cumulative player claim, not an ad-eligible audience. | 8 | vendor assertion | supported as stated | medium | [469] |
| GAM-G-F22 | EA's 15 June 2026 release names brand integrations in EA SPORTS FC 26 (Xfinity and Peacock; Comcast's 17 November 2025 release dates dynamic LED ad-boards in Ultimate Team matches in the US), EA SPORTS College Football 26 (Mountain Dew's 'DEW University', live from 15 July 2025, and Lowe's on the Stadium Pulse broadcast overlay per Lowe's 11 April 2025 release) and Madden NFL 26 (Lowe's Ultimate Team program, released 27 September and 18 October 2025 per the fan database MUT.GG); the release reports engagement as games played, matches played and challenges completed, not impressions. | 8 | vendor assertion | supported with qualification | medium | [25] |
| GAM-R-F322 | EA Advertising renders brand placements inside gameplay: Comcast dated dynamic in-stadium LED ad-boards in EA SPORTS FC 26 Football Ultimate Team matches in the US (17 November 2025). FC 26 runs on PlayStation 5, Xbox Series X\|S, PC, Nintendo Switch and Switch 2, but no source states which platforms carry the ad-boards. The same partnership added offers for US players of EA SPORTS FC Mobile. EA also names College Football 26 and Madden NFL integrations, and its brand page lists FC, Madden NFL, College Football, NHL, skate. and The Sims. | 8 | synthesis | supported with qualification | medium | [235, 236, 25, 237] |
| GAM-R-F330 | EA Advertising is available to US buyers: US advertisers (Xfinity and Peacock, Lowe's, Mountain Dew) are named and Comcast places FC 26 ad-boards "in the US"; no US ad-eligible audience or US title list is published. | 8 | vendor assertion | supported as stated | medium | [235, 25] |
| GAM-ANZU-F12 | UFL's PC version, which Anzu sells, is distributed through the game's own dedicated launcher from its official website (full PC release 5 December 2025 after early access from November 2025, per Operation Sports), not through Steam: on 29 September 2026 its Steam page said the game is not yet available on Steam, release date to be announced, with XTEN Limited as developer and publisher. | 8 | independent measurement | supported with qualification | medium | [471, 132] |
| GAM-ANZU-F13 | Anzu calls itself 'the sole officially licensed in-game ad provider for Xbox' on an undated page and announced Xbox Tools and Middleware Program participation in November 2019; AdExchanger relayed the licence claim in March 2022, but this review found no Microsoft record confirming it and no console title named as carrying Anzu placements in the scoring period. | 8 | vendor assertion | supported as stated | medium | [232, 462, 164] |
| GAM-FRAMEPLAY-F27 | Frameplay's SDK supports Unity, Unreal and Cocos2dx and its developer page claims availability on desktop, console and mobile, but the Unity changelog's newest release is 3.8.1 of 16 October 2024, console support appears as "PS5 Ad Loading Support (Experimental)" in November 2023, and no live console or PC title was found; the Unity quick start page was last updated on 8 October 2025. | 8 | vendor assertion | supported as stated | high | [464, 465, 466] |
| GAM-GADSME-F14 | Gadsme markets 'mobile, PC and console' placements, and adWMG's paid member article (2 December 2025) says their May 2025 partnership unlocked formats 'within Xbox and PlayStation titles', but no console title is named in any source opened, and Gadsme's own press page describes the same deal as adWMG tapping 'Gadsme's mobile in-game inventory for console-focused advertisers'. | 8 | synthesis | supported as stated | medium | [170, 40, 467] |
| GAM-GADSME-F30 | Gadsme's pages describe monetisation of 'mobile, PC and console' games, but the only environments evidenced by named or documented live integrations in the scoring period are PC (Football Manager 26 via Steam) and mobile games from named publishers, giving two coverage tuples. | 8 | inference | supported as stated | medium | [170, 141, 472] |
| GAM-ROBLOX-F06 | Roblox's 10-K says the Roblox client runs on iOS, Android, Windows, Mac, Xbox, PlayStation, Chromebook and select VR hardware, and that creators insert immersive ads (Rewarded Video, Billboards and Portals) as ad units for Roblox to serve programmatically; no Roblox document opened states on which devices these ads are served. | 8 | direct record | supported as stated | high | [197, 207] |
| GAM-XBOX-F13 | XBOX Media Solutions' documented console products render outside gameplay or are unspecified: 'XBOX Experiences' uses 'native display formats and branded content' without naming a title, the custom-integrations range extends to 'XBOX dashboard & store placements', Xbox Activities launched on 30 June 2025 inside Xbox Landing Experiences, the April 2026 blog lists Xbox Landing Experiences and Click-to-Engage on Xbox, and Microsoft's retired Invest page describes an invite-only private marketplace for Xbox Home and Store banners. | 8 | synthesis | supported as stated | medium | [24, 455, 456, 457] |
| GAM-F-F27 | A Venatus-issued release of 18 March 2026 says an analysis by Lumen Research found Venatus video audiences nearly twice as likely to view an ad, viewing for 8 seconds on average against a 2-second benchmark, with up to five times higher attention per 1,000 impressions, but it gives no sample, period, geography, benchmark source or funding. | 9 | vendor assertion | supported as stated | low | [395] |
| GAM-H-F02 | dentsu's Play Attention report (January 2024, with Lumen) found gaming ads more viewable and more attended than dentsu's norms (99% viewability, 6,736 attentive seconds per 1,000 impressions, $3.38 attention cost per 1,000), but brand recall of 21%, which dentsu says is highly variable across gaming environments and lower than its norms. | 9 | commissioned measurement | supported as stated | high | [27, 384] |
| GAM-H-F04 | In the dentsu study, Activision Blizzard mobile rewarded video had 10,043 attentive seconds per 1,000 impressions against 6,100 for dentsu's online video norm, and a 100% on-screen impression rate against a dentsu norm of 87%. Intrinsic in-game ads with Anzu had 3,442, above online display (1,416) but below social (3,975) and dentsu's audio norm (10,126). | 9 | commissioned measurement | supported as stated | high | [27] |
| GAM-H-F05 | Sources for the dentsu study disagree on two figures for Anzu. The in-game attention figure is 3,442 seconds in dentsu's report and Lumen's release but 3,422 in Anzu's blog. Anzu's on-screen visibility is 98% in dentsu's report but 99% in Lumen's release, and Anzu's blog reports both a 98% viewability score and a 99% on-screen visibility score. Anzu's blog is the only source found for the sample: 200 in-person gamers aged 18+ in Las Vegas and 250 online control respondents. | 9 | synthesis | supported as stated | high | [481, 27, 384] |
| GAM-H-F07 | Anzu's releases on its studies with Lumen report different benchmarks in different years. In October 2022 (Lumen eye-tracking across a series of Anzu mobile and PC campaigns over the previous 12 months), players viewed in-game ads for 2.9 seconds against a Lumen benchmark of 1.5 seconds, and 83% of ads were viewed against 62%. In June 2023 (a 'two-year study'), in-game ads were viewed for 3.1 seconds against a 'digital ad norm' of 2.9 seconds, with 98% viewability against 78%. The only sample figure is Lumen managing director Mike Follett's 2022 statement that Anzu 'has conducted attention tests amongst 5,000 respondents globally'; neither release gives geography or a sample for the specific results. | 9 | vendor assertion | supported with qualification | high | [482, 483] |
| GAM-H-F09 | An Anzu and Happydemics study (released 26 March 2024) of 110 intrinsic in-game brand campaigns reports that intrinsic in-game ads were 5 percentage points above the online media average on purchase consideration, 13 points above the digital media average on ad recall and 11 points above on brand identification. The comparator is Happydemics' database of over 600 advertisers, not a control group. | 9 | vendor assertion | supported as stated | high | [474] |
| GAM-H-F11 | Frameplay's release of 27 October 2025 says Happydemics' in-game benchmark draws on 173 in-game campaigns, of which Frameplay represents nearly 90 studies. It says in-game had the highest ad recall of any digital format at 32%, and 52% brand attribution, 9 points above the media average. Frameplay itself scored 54% on attribution and likeability and 24% on consideration, 2 points above the in-game benchmark. | 9 | vendor assertion | supported as stated | high | [475] |
| GAM-H-F12 | Frameplay supplied just under 52% of the in-game campaigns in Happydemics' benchmark set (nearly 90 of 173), so its statement that it beat the in-game benchmark compares Frameplay with a benchmark about half made up of its own campaigns, unless Happydemics excluded them, which the release does not say (inference). | 9 | inference | supported with qualification | medium | [475] |
| GAM-H-F15 | Roblox's own brand lift study, run with Latitude from 20 to 27 March 2024 among 2,100 US monthly Roblox users (300 control, 1,800 exposed, ages 13-49), reports statistically significant increases in brand awareness, ad awareness, opinion, consideration and recommendation, averaged across three brands, without magnitudes. Exposed users reported higher intended actions, from +10 points for visiting the brand's Roblox experience to +3 points for searching for brand information. | 9 | commissioned measurement | supported as stated | high | [28] |
| GAM-H-F16 | Digiday reported on 12 May 2025 that a MediaScience study for Roblox measured 140 US Roblox users aged 13-34 with eye tracking and skin-conductance sensors. Roblox figures showed 100 times more attention to branded content on Roblox than to social ads and 35 times more than to streaming-TV ads, and unaided recall 211% higher than social. For recall against streaming TV, the reports conflict: Digiday says 95%, Tubefilter 38%. | 9 | vendor assertion | supported as stated | medium | [484, 485] |
| GAM-H-F17 | Digiday reported on 12 May 2025 that agency marketers said Roblox would need to explore other metrics, such as direct sales conversions, to more firmly convince brands; Carat's Boaz Burns said attention scores are 'a very small piece of the decision-making matrix'. In the same article, gaming leads at dentsu, TMA and Omnicom's Zero Code said clients would find the research compelling enough to potentially increase spending. Roblox said it planned to expand first- and third-party measurement to connect activations to sales outcomes. | 9 | synthesis | supported with qualification | medium | [484] |
| GAM-H-F18 | Roblox's 18 June 2026 release reports that its 'Immersive Is Effective' survey, 'conducted in partnership with Ipsos' (fielded 6-27 May 2026; bases of 200 Roblox users aged 13-34, 200 streaming-platform users aged 13-34 and 200 social-platform users whose age range the footnote does not state; survey geography not stated), found that 52% of the Roblox users say they pay more attention to ads on Roblox than they usually do, 13 points above the streaming base and 11 above the social base, and that 63% of respondents say product information they see on Roblox is memorable. The full report sits behind a registration form and was not opened, and the release does not say who paid for the study. | 9 | vendor assertion | supported with qualification | medium | [486, 487] |
| GAM-H-F20 | Discord's case study for Bethesda's DOOM: The Dark Ages says the brand impact of a 7-day Video Quest in the US and Canada was measured through a Kantar Brand Lift Study, and reports a +11.3-point lift in ad awareness, +3.4 points in play intent (at 90% confidence) and +5.7 points in top-of-mind awareness among frequent FPS players. The page attributes the headline lift figures to 'Discord Internal Data, 2025', and its disclaimer says 'Based on Discord data'. | 9 | vendor assertion | supported with qualification | high | [341] |
| GAM-H-F22 | Discord's World of Warcraft x Mountain Dew case study, one of three non-entertainment consumer brands among the 21 Quest success stories listed on 29 September 2026 (Discord calls Mountain Dew the first FMCG brand to run a Quest), reports a 16-point increase in brand favorability among surveyed exposed users and a 6-point lift in positive brand perception. It cites 'Discord Internal Data, 2025' and names no research partner or method. | 9 | vendor assertion | supported with qualification | high | [490] |
| GAM-H-F23 | Microsoft's April 2026 gaming ad blog states that, per research done in partnership with dentsu, '100% of gaming ads are fully viewed', against online video at 86% and social at 77%. dentsu's report gives 99% viewability for gaming overall and a 100% on-screen impression rate for rewarded video only, and Microsoft's own October 2024 post scoped the dentsu result to Candy Crush Saga video ads (99% viewability, 100% of viewers). The same April 2026 paragraph says the research shows 'immersion predicts ... sales with 83% accuracy'; that figure does not appear in dentsu's report, and Microsoft's October 2024 post attributes its immersion findings to an 'Activision Blizzard Media Immersion Study via Alter Agents, US, March 2024', a seller-commissioned study with no public method or accuracy figures. A February 2025 post cites a +17% ad-recall lift for mobile rewarded video with no named study, sample or date. | 9 | vendor assertion | supported with qualification | high | [456, 580, 27, 384] |
| GAM-H-F24 | In an independent peer-reviewed eye-tracking study (Journal of Eye Movement Research, December 2025), 79 adults played a mobile tennis game with four embedded static brand ads for ten minutes. Only 21 (26.6%) reported noticing any advertisement unaided, and whether the ad sat in a task-relevant zone predicted recall better than fixation time did. | 9 | independent measurement | supported as stated | high | [26] |
| GAM-H-F25 | A peer-reviewed meta-analysis of 38 data sets (Journal of Advertising, 2021) found that consumers have a more positive attitude toward advergames themselves than toward other types of advertising, and that advergames seem persuasive and drive choice behaviour, but that brand and product information seems less likely to be remembered than with other advertising and advergames are less likely to be recognised as advertising, with younger consumers more susceptible to their persuasive effect. | 9 | synthesis | supported with qualification | medium | [498] |
| GAM-H-F26 | Two peer-reviewed experiment papers by Ingendahl and colleagues find that brands placed in video games gain in attitude only when met during positive in-game experiences, not negative ones. The 2022 paper describes the overall effect of brand placements on brand attitudes as small and varying substantially between studies. | 9 | synthesis | supported as stated | medium | [499, 500] |
| GAM-H-F28 | A peer-reviewed Marketing Science paper (February 2025) used high-frequency data on 60,000 Twitch streamers, sampled every 10 minutes for eight months, and plausibly exogenous within-day variation in top influencers' broadcast hours. It estimates an elasticity of 0.027 of a game's concurrent players to its live-stream viewers, an effect the authors call moderate and short-lived, and back-of-the-envelope calculations suggest that only about one sixth of games profit from sponsored live streams. | 9 | synthesis | supported with qualification | low | [503] |
| GAM-H-F30 | The purchase-type outcomes found for in-game brand activations are vendor claims without a published method: Super League's case-study tiles ('+6.3% lift in purchase rate' for a Lunchables 'Playables virtual world' designed to keep children coming back, and a 'Highest-ever mobile sales day' for Chipotle's Roblox maze), Anzu's statement that a December 2024 Claritas beta for a major US advertiser found 'an incremental lift in conversions' (no figure), and Anzu's April 2025 claim that intrinsic ads beat advertisers' CPA goals by 21% on average. This review found no public randomised or holdout sales-lift or incrementality study with published design and results for in-game brand advertising in the sources screened by the cutoff. | 9 | vendor assertion | supported with qualification | medium | [496, 488] |
| GAM-H-F31 | Several cohort A and game-adjacent sellers publish attention or lift figures with no method. Gadsme's advertiser page states an average 35% ad recall and +25 points on brand image and consideration from 'brand lift surveys conducted on Gadsme inventory'. iion's Garnier case (4 March to 13 April 2024, 674,343 impressions, measured with Amplified Intelligence) reports 7.38 seconds of attention per video impression against a 1.10-second benchmark, and gives two different display figures (1.48 and 2.4 seconds of active attention). | 9 | vendor assertion | supported as stated | high | [118, 495] |
| GAM-H-F32 | Digital Turbine and Amplified Intelligence reported on 3 April 2023 that, in a two-week study in June 2022 of seven brands, viewers actively attended to Digital Turbine's in-game mobile video ads for a combined 22 seconds, against 1.6 seconds for mobile web video and 2.5 seconds for social video. The ads scored 160 on Amplified's STAS brand-choice index, against a benchmark of 100. | 9 | vendor assertion | supported as stated | high | [494] |
| GAM-H-F33 | Gameloft for brands' 'Attention & Emotion in Play' study, presented in paid content on The Drum (1 June 2026), reports 18% more positive attention for an in-game interactive video ad than for a YouTube pre-roll and 24% higher emotional engagement by facial coding. The sample, period and funding details are not given, and the full report is behind a form. | 9 | vendor assertion | supported as stated | high | [492, 493] |
| GAM-R-F326 | EA says brands can tailor placements to objectives and optimise campaigns, and it reports integration engagement as games, matches and challenges played. The only method-backed result found is Ipsos's GameTrack exposed-versus-unexposed analysis (Q1 2026, 37,000 interviews, 16 European countries, run on behalf of Video Games Europe). It found an average +31% uplift on 'a brand for people like me' across nine in-game campaigns that include, but are not stated to be limited to, EA SPORTS FC integrations. The per-brand uplifts for EAFC integrations were Beats +97%, Pepsi +26% and Visa +18%, against +6% for brands that did not advertise in the games. | 9 | synthesis | supported with qualification | medium | [25, 477] |
| GAM-GADSME-F22 | Gadsme's advertisers page reports brand-lift averages ('35% in ad recall', '+25 pts on both brand image and consideration') from 'Brand lift surveys conducted on Gadsme inventory', an IVT rate below 0.5%, and that 'According to IAS and DoubleVerify, in-game campaigns consistently outperform traditional advertising', none with a named provider, sample, period or method. | 9 | vendor assertion | supported as stated | low | [118] |
| GAM-ROBLOX-F24 | The outcome evidence for Roblox brand ads found by this review is commissioned or vendor-reported: Roblox's own Latitude brand lift study (March 2024, 2,100 US monthly users, 300 control and 1,800 exposed, ages 13 to 49) reports 'significant' lift in brand metrics without magnitudes or a stated significance test; its Ipsos study (May 2026, 200 users per platform) measures self-reported attention; EDO's pilot publishes no figures; and a Roblox executive cited, at IAB PlayFronts in September 2026, a Sam's Club rewarded video result (18% lift in sales per cardholder, 6% lift in conversions) attributed to Cint, a named Roblox measurement partner, with no published design. This review found no public study with an independent funder and published design and results for Roblox ads, in the sources screened by 29 September 2026. | 9 | synthesis | supported with qualification | medium | [629, 486, 488] |
| GAM-TWITCH-F22 | dentsu's January 2024 Play Attention report (dentsu with research partner Lumen) found that livestream ads in a simulated livestream environment, charted as Twitch, produced 57% brand recall against a dentsu cross-media norm of 38% and a 17% average brand-choice uplift against dentsu norms of 7.25%, with uplift defined against a control-group baseline; the report gives no sample size, fieldwork dates or funding, and does not say whether Twitch took part. | 9 | commissioned measurement | supported with qualification | medium | [27] |
| GAM-XBOX-F09 | The two XBOX Media Solutions case studies inside the scoring period report brand lift without method: KitKat in Candy Crush Saga across 40 markets (+29 percentage points ad recall, averaged over the UK, France, Germany, India and Australia; 500M+ impressions; published 18 August 2026) and Hershey's holiday season 2025 in Candy Crush Saga (+55 points ad recall, +7 purchase intent); neither names a measurer, sample, control design or period. | 9 | vendor assertion | supported as stated | low | [227, 226] |
| GAM-I-F14 | Roblox reported Rewarded Video in over 400 experiences on 6 January 2026 and in over 1,200 games on 20 May 2026, when it also began testing a pre-entry 'Join with Reward' placement and a drag-and-drop Studio plugin; MediaPost reported on 15 September 2026, citing Roblox's RDC presentation, a Studio plug-in with drag-and-drop 2D and 3D rewarded-video templates 'expected to launch in Q4 2026' and an upcoming rewarded surface that runs before gameplay, plus a multi-reward rotation option creators will 'eventually' see, and said the presentation described these placement options as no-code and using AI; no primary Roblox document for the Q4 timing or the AI element was found. | 10 | vendor assertion | supported with qualification | medium | [188, 209, 508] |
| GAM-I-F15 | The only AI asset-generation feature documented for Roblox advertising in the sources screened is Ads Manager's 'AI generate' option, which creates three thumbnail variations from a text prompt for Sponsored games shown on the Home page and in search, a paid-discovery product that this paper lists rather than scores; no AI generation or AI-assisted placement is documented for Immersive Ads or Rewarded Video in Roblox's documentation read. | 10 | vendor assertion | supported as stated | medium | [527, 529, 530] |
| GAM-I-F23 | Adverty AB reported net sales of SEK 33.7 million in 2025 (KSEK 33,684), against KSEK 31,841 in 2024, 31,418 in 2023 and 16,685 in 2022, and KSEK 20,730 for H1 2026 against 13,495 for H1 2025. | 10 | direct record | supported as stated | high | [143, 79] |
| GAM-I-F27 | IAB Tech Lab's agentic work screened has no game-specific object: ARTF v1.0 was released on GitHub on 20 July 2026 while its standard page still says public comment ended 15 January 2026 and that it 'will be finalized soon'; AAMP 2.0 (23 April 2026) released buyer and seller agent SDKs, with the seller agent at v2.4.2 (18 August 2026) connecting to Google Ad Manager, FreeWheel, PubMatic, Index Exchange and Magnite; and Agentic Mobile v1.0.0-draft, a draft for public comment, includes a REWARDED placement type for mobile in-app inventory. | 10 | direct record | supported as stated | medium | [533, 534, 535, 536, 537] |
| GAM-I-F28 | Super League's FY2025 10-K (31 March 2026) states 'We currently do not incorporate AI/ML into our products', while its Q2 2026 10-Q (14 August 2026) reports a January 2026 expansion of its exclusive partnership with AdArcade introducing an 'AI-powered Creative-as-a-Service' that converts brands' video ads into playables for rewarded mobile video inventory. | 10 | direct record | supported as stated | medium | [128, 30, 528] |
| GAM-R-F341 | Zynga Ads+ was presented at IAB PlayFronts (17 September 2026) as the launch of a third-party publisher network that brings about 100 leading casual mobile games into Zynga's ad offering. It admits apps rated 4.5 or higher and uses DoubleVerify as measurement partner. Zynga describes the result as combined reach (its own games plus Ads+) of 'up to' (IAB) or 'more than' (PocketGamer.biz) 100 million unduplicated monthly consumers. No participating app is named and no product documentation was found, so its status is unconfirmed and it stays listed, not scored. A Zynga-led IAB Brand Academy session on it is scheduled for 21 October 2026. | 10 | synthesis | supported with qualification | medium | [183, 203] |
| GAM-U-F01 | Bidstack Group Plc (UK company 04466195), an AIM-listed holding company, entered administration on 22 March 2024. The same day its joint administrators completed a pre-packaged sale of its business and assets, including its IP and its shareholdings in subsidiaries, for £804,848 to Virtual Sport Technology Limited, a vehicle incorporated and owned by the plc's then management team (a connected-party sale). The sale followed a marketing process that contacted 242 parties and produced no third-party offer. The register records Virtual Sport Technology Limited as Bidstack Ltd's controlling owner (75% or more) from that date. | 10 | direct record | supported with qualification | high | [31, 512, 519] |
| GAM-U-F04 | Bidstack's administrators report, from unaudited consolidated management accounts, group turnover of £3.1m and a loss of £9.7m (including £3.8m of exceptional restructuring costs) for the year to December 2023, against turnover of £5.3m and a loss of £6.1m in FY22. They attribute the administration to liquidity constraints after further drawdowns on an October 2023 convertible loan note facility from Irdeto could not be made, because its conditions (including City Code requirements) could not be met. They also record that a February 2024 sale process contacting 242 parties produced no offer. Parties cited strategic fit, timelines, uncertainty over future revenues and the outlook for the media, advertising and gaming sectors. | 10 | direct record | supported with qualification | medium | [31] |
| GAM-U-F06 | Israel's Registrar of Companies dataset, last updated on 28 September 2026, records Sayollo Media Ltd (private company 514656453, incorporated 16 August 2011, Rehovot) as 'in liquidation by court' with the sub-status 'order to open proceedings' and 2021 as the last year of a filed annual report, and gives no order date. | 10 | direct record | supported as stated | high | [510] |
| GAM-U-F07 | Admix, an in-play product-placement company founded in 2018, announced a merger with metaverse builder LandVault in a release dated 21 June 2022, under which the merged entity would be branded LandVault; the release gives the growth of its metaverse revenue as context. | 10 | vendor assertion | supported as stated | high | [507] |
| GAM-U-F08 | On 9 July 2024 Infinite Reality announced a $450 million all-stock acquisition of Landvault, which it described as an immersive-technology and digital-twin company, in a release filed with the SEC under Form 425. | 10 | vendor assertion | supported as stated | high | [515] |
| GAM-U-F10 | Super League Gaming bought substantially all the assets of Bloxbiz Co., a Roblox in-game ad platform, plus its founders' goodwill, under an asset purchase agreement signed and closed on 4 October 2021 (Form 8-K filed 7 October 2021, Items 1.01, 2.01, 3.02, 8.01 and 9.01). It paid $6.0 million at closing ($3.0 million cash and $3.0 million in stock) plus up to $11.5 million of contingent consideration tied to revenue milestones through fiscal 2023, a maximum of $17.5 million. Its release of 6 October 2021 says Bloxbiz reached more than 25 million monthly active Roblox users across more than 75 games (vendor figures). | 10 | direct record | supported with qualification | high | [29] |
| GAM-U-F11 | As read on 28 September 2026, Super League's Super Biz site says the Super Biz platform is being deprecated: new sign-ups are closed, the self-serve dashboard is being wound down and no new campaigns or integrations run through it, and work continues with a small group of studios on branded content and 3D character integrations. | 10 | vendor assertion | supported as stated | high | [511] |
| GAM-U-F12 | Super League Enterprise closed its purchase of the Misfits Ads business assets from Esports Now, LLC on 1 May 2026 (agreement 16 March 2026; stockholders approved the share issuance on 30 April 2026). At closing it paid $1.5 million in cash, 26,768 shares, a pre-funded warrant for 509,682 shares and a warrant for 536,450 shares at $18.00. It owes a fixed $300,000 delayed cash payment on the first anniversary. A contingent earn-out of up to $1.2 million in cash and 105,571 shares is tied to first-year gross-profit milestones and to Super League's market capitalisation at the first and second anniversaries. | 10 | direct record | supported with qualification | high | [506] |
| GAM-U-F13 | On 19 June 2026 Media Scale Capital Limited (17252129), a UK holding company incorporated on 1 June 2026 and itself controlled (75% or more of voting rights) by To Media Groep, a Dutch B.V., became the controlling owner (75% or more) of Project Venus Topco Limited. That company controls Venatus Media Limited (07183957, media representation services) through Project Venus Midco 1, Midco 2 and Bidco Limited. The sellers were Livingbridge 7 Global LP, a private-equity fund holding a majority since September 2021, and two individuals, all of whom ceased as PSCs on 19 June 2026. Venatus announced the acquisition in July 2026, describing the buyer as an investment fund. | 10 | direct record | supported with qualification | high | [302, 303, 649, 305] |
| GAM-U-F15 | Enthusiast Gaming Holdings Inc. closed the sale of its direct-sales business line, including Omnia Media Inc. and GameCo Esports Canada Inc., to Vertiqal Studios Corp. on 30 August 2025 for C$900,000 in cash plus the assumption of its liabilities (net liability about US$2.6 million at 30 June 2025), and continued to issue quarterly results through Q2 2026. | 10 | vendor assertion | supported as stated | high | [317, 520] |
| GAM-U-F16 | Oracle's Advertising EOL FAQ (version 3.0, 12 August 2024) states that Oracle would end all Oracle Advertising products and support, including Moat Analytics and Moat Reach, effective 30 September 2024, with end-of-life notices emailed in the week of 17 June 2024. | 10 | direct record | supported with qualification | high | [339] |
| GAM-U-F21 | Overwolf announced on 8 May 2024 that it had acquired NitroPay, renamed Nitro and run as a division, a website ad-tech company serving more than 500 sites including Maxroll.gg, Liquipedia.net and VLR.gg. | 10 | vendor assertion | supported as stated | high | [293] |
| GAM-GADSME-F26 | Gadsme's SDK serves ads dynamically at run time (tags call and display ads, formats auto-refresh, and 'Ad Content Pooling' selects content from several demand sources in real time), is described as compatible with Unity, Unreal and C++ proprietary engines, and was extended in October 2024 to Unreal, Godot, Cocos2d and custom C++ engines across mobile, home console and PC; no AI-assisted placement or buyer-side asset tooling was found in the pages opened. | 10 | vendor assertion | supported as stated | medium | [59, 531, 596, 118] |
| GAM-ROBLOX-F16 | On 29 September 2026 Roblox's app-ads.txt listed 15 authorised sellers (DIRECT: three Google accounts, two PubMatic, Magnite/rubiconproject, Index Exchange, Smart, Sharethrough, InMobi and Meta; RESELLER: LoopMe, Smaato, PubNative and Verve), and SSP sellers.json files label Roblox as a PUBLISHER on PubMatic (seller 164509, roblox.com) and as reached through Google's exchange bidding on Magnite (seller 19318, 'ROBLOX - Exchange Bidding') and Index Exchange (seller 214070, 'Roblox Corporation (via Google OB)'). | 10 | independent measurement | supported as stated | high | [538, 572, 539, 573] |
| GAM-SUPERLEAGUE-F02 | On 18 August 2026 Super League signed subscription agreements with Metaplanet Holdings, Inc. and Evo Fund. On closing, expected in Q4 2026 subject to stockholder approval at the 16 October 2026 annual meeting (after the cutoff), Metaplanet would deliver 2,100 bitcoin and $2.5 million, the company would be renamed Superplanet, Inc., and Metaplanet would hold about 95.7% of common stock. The proxy says the existing business is expected to continue under current management. | 10 | direct record | supported as stated | high | [513, 514] |
| GAM-TWITCH-F16 | Amazon documents three creative-insertion features for Twitch: server-side stitching of video creative into live broadcasts (SureStream, SSAI pre-fetch), an automatic resizing tool that turns a 160x600 display creative into the Mirror-C Stream Display format, and an Amazon DSP Gen-AI workflow (30 September 2024) that generates Twitch-sized assets from an uploaded image with user review and creative moderation before serving. | 10 | vendor assertion | supported as stated | medium | [264, 47, 633, 532] |
| GAM-XBOX-F02 | The former Activision Blizzard Media sells as XBOX Media Solutions, which it says operates as part of Microsoft Advertising; the rename was announced on its LinkedIn company page on 4 February 2026, stating that access to 'the same gaming advertising solutions across Xbox, King, and Microsoft Casual Games' continues and that the page name, visual identity and website URLs changed. | 10 | vendor assertion | supported as stated | high | [222, 223] |
| GAM-XBOX-F28 | On 29 September 2026 the AgenticAdvertising.org public agent registry listed 34 agents, none operated by Microsoft, Xbox, King or Activision, and Microsoft Advertising's MCP Server page describes access to Microsoft Advertising campaigns, ad groups, ads and keywords without mentioning gaming inventory, so no protocol state is recorded for this product. | 10 | synthesis | supported as stated | medium | [121, 192] |
| GAM-E-F03 | Under amended § 312.5(a)(2), a covered operator must obtain separate verifiable parental consent before disclosing a child's personal information to third parties, unless the disclosure is integral to the service. The FTC's Statement of Basis and Purpose says disclosures for advertising purposes are not integral, so a child-directed game needs that separate consent before disclosing children's personal information to advertising or measurement partners, except partners that only provide support for internal operations (for example contextual ad serving or frequency capping, see GAM-E-F04), which the Rule does not treat as third parties. | 12 | direct record | supported with qualification | high | [23, 421] |
| GAM-F-F06 | Discord Inc. is privately held: its EDGAR record (CIK 0001763840) shows only Form D private-placement notices, the latest on 15 March 2023, and EDGAR full-text search found no S-1, F-1 or public draft registration statement filed by Discord by 28 September 2026, although Game Developer, citing Bloomberg, reported on 7 January 2026 that Discord had confidentially filed for an initial public offering. | 12 | direct record | supported with qualification | high | [542, 684, 543] |
| GAM-R-F218 | This review found no named game title, no named brand campaign and no published sellers.json, ads.txt or app-ads.txt for Trivver in the scoring period, in the sources screened by 29 September 2026. Trivver's news page is stale (latest item a Forbes piece of 12 May 2025; the June 2026 release is not listed), and its home page shows Google Cloud, ByteDance and NVIDIA under "Trusted by" with a disclaimer that the logos imply no endorsement or affiliation. A July 2026 partnership with SCCG Management, an iGaming consultancy, was reported but not opened and names no title in the headline. | 12 | synthesis | supported with qualification | medium | [114, 685, 157, 155] |
| GAM-S-F07 | IAB's 'Changing the Game' study (March 2024) surveyed 300 US brand and agency decision makers who already invest in games advertising, was sponsored by Admazing and Venatus, publishes no fieldwork dates or methodology page, and asks no question about barriers; it therefore cannot rank constraints for H1. | 12 | commissioned measurement | supported as stated | high | [107] |
| GAM-E-F26 | Utah's App Store Accountability Act (SB 142, 2025) set app-store and developer duties to begin on 6 May 2026 and barred developers from sharing age category data with anyone. The 2026 amendment (HB 498) moves those duties to 6 May 2027 and adds that developers may use app-store age category data only to enforce their own age restrictions, comply with law and apply their own safety features. | 13 | direct record | supported as stated | high | [432, 433] |
| GAM-A-F04 | EMARKETER said that most in-game ads promote installs of other mobile games (May 2023 report abstract), that game developers buy most of the ads served in games (May 2024 report abstract and June 2024 article) and that most in-game ad spend goes to mobile (June 2024 article); the first two describe shares of ads, the third a share of spend by platform, and none gives a quantified dollar split by buyer or between brand and user-acquisition objectives. This paper's reading of EMARKETER's US game ad revenue series as mixed objective and probably mostly user acquisition, not brand spend, is therefore the author's inference, not an EMARKETER figure. | abstract, 1, 2, 12 | synthesis | supported with qualification | medium | [6, 7, 8] |
| GAM-B-F02 | Across the six readable sellers.json files downloaded on 28 September 2026, 1,619 per-vendor unique publisher domains collapse to 1,381 unique domains (1,380 with Adverty's later file of the same day): 176 domains appear in two or more vendors' files, 51 in three or more, 10 in four or more, and miniclip.com in five (Anzu, IION, Gadsme, Gameloft, Frameplay); summing vendor supply therefore counts 238 surplus domain appearances across those 176 shared domains. | abstract, 1, 3, 12 | synthesis | supported with qualification | high | [9, 11, 12, 10, 13, 14] |
| GAM-A-F05 | EMARKETER's March 2024 forecast put US mobile game ad revenues at $7.77 billion of $8.59 billion total US game ad revenues in 2024, so mobile made up 90.4%, and put mobile game ad revenues at 2.8% of total US digital ad spending in 2024. | abstract, 2 | forecast | supported as stated | medium | [5] |
| GAM-A-F21 | Among the publishers screened by 28 September 2026, EMARKETER is the only one found with public US dollar figures for game advertising: IAB/PwC publishes a growth rate only, PwC a share only (global), WPP Media global figures only, no Magna gaming line was found, Omdia's pages could not be opened, and the Circana series in ESA's 2026 report is consumer spending ($60.7 billion in 2025), not advertising. | abstract, 2, 12 | synthesis | supported as stated | medium | [3, 4, 38, 98] |
| GAM-S-F18 | Coding H1's survey leg under its registered rule on point-estimate ranks: Advertiser Perceptions 2024 (n=200, July 2024; geography not stated) puts brand safety (29%) above lack of scale (26%) and reaching the same consumers more effectively elsewhere (26%), with measurement tied with scale at 26%. IAB Australia 2023 (n=100, June-July 2023, Australia) charts brand safety 47%, measurement 36%, suitability 36%, effectiveness evidence 34% and fit with other media 32% as its top five. A reach-and-scale option offered under identical wording in the 2022 wave (16%, ninth of thirteen) does not appear, so it scored below 32% if it was offered in 2023. The survey leg is therefore met on the rule's wording only on point estimates and on the inference that the reach option was offered in 2023; without that inference only one survey qualifies and the leg is unresolved. It is weak: neither sample is established as US or limited to brand advertising; the Advertiser Perceptions lead is inside sampling error; that firm says the reach-elsewhere perception is growing; and in the same Advertiser Perceptions survey the top two benefits cited are reach-related (hard-to-reach audiences 43%, incremental reach 38%). | abstract, 2, 13 | inference | supported with qualification | low | [33, 34, 35] |
| GAM-E-F36 | H4 is supported on its registered rule. For at least five products or platforms, a primary rule document restricts a targeting, personalisation, third-party-call or measurement function the product offers elsewhere: Roblox (no personalisation under 18; no third-party ad or measurement calls in developer-placed ads; aggregate-only reporting), Microsoft (no personalised ads for accounts under 18), Discord (age-assurance signals not used for targeting), Fortnite (no targeting of under-13s with island advertising) and Google Play (age signals barred from advertising use); Apple's developer agreement also limits age-range data to age-appropriate content or legal compliance. COPPA § 312.5(a)(2) adds separate parental consent for advertising disclosures in child-directed games. The conflict between Roblox's two revenue-share documents, which the registered rule cites as an example of an unresolved condition, concerns developer fees rather than any restriction counted here, so it does not block this verdict. | abstract, 4, 5, 7, 12, 13 | synthesis | supported with qualification | high | [20, 213, 57, 21, 22, 220, 221, 23] |
| GAM-C-F25 | Preliminary coding of H2 (inference): the in-game guidelines are a final release, and on 29 September 2026 MRC's records show no accreditation or MRC-confirmed audit for any of the 13 cohort A and B1 candidates in the product register (nine networks, four platforms); the one gaming-environment audit in process is Trivver's, a listed product, and IAS In-Game Measurement left MRC's process in January 2026 after IAS withdrew. H2's population is the products actually scored in cohorts A and B1, which admission has not yet fixed, so both shares are to be computed on that scored set at the cutoff (third-party verification currently rests on announcements dated 2022 to 2025 for Anzu, Gadsme, Frameplay and Roblox, and EA Advertising and Admazing state IAS or DoubleVerify partnerships without a verifier record), and H2 codes as unresolved if fewer than six products are scored. | abstract, 4, 6, 11, 12, 13 | inference | supported with qualification | medium | [52, 15, 16, 350, 368, 369, 351] |
| GAM-G-F25 | Coding of H3 (inference) on the sources screened by 28 September 2026: H3 is unresolved under its registered rule. Microsoft's XBOX Media Solutions markets console brand integrations 'into gameplay' and 'XBOX Experiences' that 'leverage blockbuster gaming titles', names no title, and its own documents do not establish whether these render inside gameplay or what their status is; no console platform documents a named, generally available in-game brand ad product with a buying route, and every other platform placement found sits outside gameplay (Xbox dashboard and store, Live TV on PS5, pre-session ads in the Xbox cloud test and on GeForce NOW, Luna's connection screen) or is in test. EA's release names brand integrations in two console titles and its Madden NFL franchise: EA SPORTS FC 26 (in-match LED ad-boards in the US, per Comcast and EA's announcement of 17 November 2025), Madden NFL 26 (named and dated only by a fan database) and EA SPORTS College Football 26 (live from July 2025, before the window); EA documents none of them as an ad placement buyable in the US, so at most three qualifying titles were found, and that count is a floor because other publishers' console titles were not screened. | abstract, 4, 8, 13 | synthesis | supported with qualification | medium | [25, 202, 24, 446, 447, 469] |
| GAM-U-F35 | From 1 January 2021 to 28 September 2026, primary records show three specialist companies sold or put into insolvency: Bloxbiz, whose assets Super League bought in October 2021; Bidstack, whose plc entered administration with a management pre-pack sale in March 2024 and whose operating company entered creditors' voluntary liquidation in June 2026; and Let's Bounce, Inc., a gaming and UGC brand-engagement company whose assets Super League bought in January 2026. This meets H5 part (a)'s threshold of three. Recorded and not counted: Super League's purchase of the Misfits Ads business line from an esports company in May 2026; Admix's merger with LandVault, announced in June 2022, because the registered rule does not list mergers; the retirement in progress of Super Biz; Sayollo's undated court liquidation; and AdInMo Ltd's suspended voluntary strike-off. None of the records says the in-game format caused an exit, but Bidstack's administrators record that a 242-party sale process drew no third-party offer amid doubts about future revenue and sector outlook. | abstract, 4, 10, 13 | synthesis | supported with qualification | high | [29, 507, 31, 32, 506, 511, 510, 150] |
| GAM-DOUBLEVERIFY-F27 | Inference: because MRC accredits DoubleVerify's mobile in-app impression and viewability measurement inclusive of OM SDK traffic, standard display and video ads served through OM SDK in mobile game apps plausibly fall within accredited measurement, while rendered 3D intrinsic placements, UGC-platform billboards and console environments fall outside it; MRC's record names no game, so this reach is inferred, not attested. | abstract, 6 | inference | supported as stated | low | [18, 15, 366] |
| GAM-DOUBLEVERIFY-F11 | Roblox's Advertising Terms (Managed Campaigns), effective 30 April 2026, make Roblox's reporting 'the sole means of tracking and reporting delivery' and 'the controlling measurement for all campaigns', exclude impressions Roblox's own systems flag as invalid traffic from reporting, provide that an advertiser may raise an invalid-traffic discrepancy through the invoice dispute process if its approved third-party verification partner finds a discrepancy exceeding 10%, and give advertisers aggregate reporting only. | abstract, 6, 12 | vendor assertion | supported with qualification | high | [1] |
| GAM-H-F01 | This review found no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign whose funding is documented as coming from a party without a financial interest, in the sources screened by 29 September 2026 (including a verifier re-screen). Every field result for a specific live brand campaign was published by the seller of the measured inventory, by a measurement firm selling the method (including Ipsos's August 2026 exposed-versus-unexposed GameTrack analysis for Europe, run on behalf of the trade body Video Games Europe), or by an agency that trades on its norms. The one peer-reviewed field-data study found (Huang and Morozov 2025, Twitch game streams, funding unknown) concerns game promotion, not brand campaigns. The only study with documented independent funding is a laboratory experiment (Büyükakgül, Yüce and Katırcı, no external funding); the other laboratory studies found were run by or with sellers or do not state their funding. | abstract, 9, 12, 13 | synthesis | supported with qualification | medium | [26, 27, 28] |
### Corrections ledger
Every substantive change logged during and after the verification passes, including the findings of the three AI-model reviews the author ran and of the internal AI review rounds, register changes, restated claims and the scoring adjustments. Each entry records what was challenged, what was checked, the verdict, the change and how it was verified.
**Corrections ledger**
| ID | Original claim or score | Challenge | Evidence checked | Verdict | Change made | Reason | Affected |
|---|---|---|---|---|---|---|---|
| C01 | Brief, research question 1: EMARKETER FAQ (10 January 2026) forecasts in-game advertising at 2.3% of US digital ad spending in 2026 (falling to 2% by 2029). | Workstream A found the decline to 2% by 2029 only in the article of 26 December 2025; the FAQ gives only the 2026 share. | EMARKETER FAQ (10 Jan 2026) and article (26 Dec 2025), both opened via Internet Archive snapshots; verifier A (two passes) supported the finding. [92, 99] | supported with qualification | The brief attributes the 2% figure to the 26 December 2025 article and labels the mostly-user-acquisition reading of the series as an inference. | Cite each dated forecast in its own words (brief landmine). | RESEARCH-BRIEF.md research question 1 |
| C02 | GAM-I-F26: AdCP v3.1.24 documentation contains no gaming, in-game, intrinsic, rewarded or playable terms. | Verifier I (refute mode). | AdCP v3.1.24 channels enum lists gaming (In-game advertising across platforms); the media-channel taxonomy defines it to include in-game display, rewarded video and playable ads. [184, 185, 121] | contradicted | Restated: AdCP can represent in-game products at channel level (schema representation only); no in-game seller in the register shows an implemented connector, interoperability test, production use or transaction, and none is among the 34 agents in the public registry. | A keyword search missed the enum and taxonomy files. | evidence/research/I.json GAM-I-F26; agentic_map rows for AdCP |
| C03 | GAM-ANZU-F11: no Anzu ad-eligible audience figure with a metric and period was found. | Verifier for the Anzu profile. | Anzu product-update blog, 9 September 2025: 250 million monthly active users and more than 3 billion global impressions; NBCUniversal Anzu page: 47M+ US monthly average users (modified June 2023). [123, 132, 617] | contradicted | The profile records both vendor figures, outside the scoring period; the supply cell is unchanged at 3. | Missed sources. | evidence/profiles/anzu.json |
| C04 | GAM-GADSME-F07: Gadsme sellers.json has no entry naming SEGA or Sports Interactive. | Verifier for the Gadsme profile. | Gadsme sellers.json (29 Sep 2026) contains a PUBLISHER entry Sports Interactive Games (sigames.com). [12, 9] | contradicted | Restated with the Sports Interactive entry; no entry names SEGA. | Counting error in the profile. | evidence/profiles/gadsme.json |
| C05 | Product register: YouTube gaming inventory in cohort B2. | Workstream F found no named gaming-specific YouTube buying product confirmed live at the cutoff. | Google sources name only a YouTube Select Shorts gaming lineup announced 17 April 2024, status at the cutoff unverified. | structural | Renamed YouTube Select gaming lineup and moved to listed. | The register scores only a named gaming-specific product. | data/product-register.json |
| C06 | Product register seed (28 September 2026): 8 cohort A, 2 B1, 8 B2 candidates. | Workstreams U, C and G found further live products. | EA Advertising launch (15 June 2026), Zynga Ads, Kidoz, Admazing, IGN, LIVAD; Trivver in MRC process. | structural | Added EA Advertising and Zynga Ads (B1); Kidoz and Admazing (A); IGN and LIVAD (B2); Trivver (A, later listed); plus listed and historical records (Zynga Ads+, SuperAwesome, Precisify, Aditude, GEEIQ, DSPs, LandVault, Let's Bounce, ESL FACEIT custom integrations). | Discovery by the vendor-universe sweep, before scoring. | data/product-register.json |
| C07 | AdInMo in-game ad network in cohort A. | Workstream U and stream R part 1 found the operator applied for strike-off and the service withdrawn. | Companies House DS01 (6 Sep 2026) and suspension (17 Sep 2026); site HTTP 525 since 11 Jul 2026; ad-serving hostnames not resolving on 29 Sep 2026. [150, 642, 113] | supported as stated | Moved to historical; not scored; recheck at the freeze. | Only live and beta products are scored. | data/product-register.json |
| C08 | Trivver in-game ads in cohort A (provisional). | Stream R part 2. | No named live title, campaign or supply file in the scoring period; June 2026 developer SDK; MRC audit in process. | supported with qualification | Moved to listed; its MRC entry stays evidence for H2. | Register condition for scoring not met. | data/product-register.json |
| C09 | Stream R proposed moving Enthusiast Gaming media to listed (no buyer-facing product documented). | Verifier R part 6. | Icy Veins Advertise page (updated 30 March 2026) with unit specs and an ads inquiry address; Addicting Games and The Sims Resource advertiser pages. | contradicted | The row stays in cohort B2. | Buyer-facing product documented at property level. | data/product-register.json |
| R2-01 | See GAM-E-F04 (claim-ledger final_wording; research/E.json finding) | OpenAI round 2 (P0): The closing sentence could be read as a categorical ban on behavioural advertising to children. | FTC final rule, 90 FR 16918 (read as the Federal Register raw text, https://www.federalregister.gov/documents/full_text/text/2025/04/22/2025-05904.txt, which is the same document as https://www.federalregister.gov/documents/2025/04/22/2025-05904/childrens-online-privacy-protection-rule). Statement of Basis and Purpose, pp. 16935-16936: the exception 'permits operators to collect persistent identifiers for contextual advertising purposes without parental consent as long as they do not also collect other personal information'. p. 16936: 'the Commission declines to modify the COPPA Rule's treatment of contextual advertising.' p. 16948: 'a separate consent requirement for non-integral disclosure ... | supported as stated | Restated: Rule text: the amended COPPA Rule lists serving contextual advertising and capping ad frequency among the activities that support internal operations (§ 312.2; the Commission also places ad attribution within the listed activities); it exempts from parental consent only the collection of a persistent identifier, with no other personal information, used solely for that support and disclosed in the § 312.4(d)(3) notice (§ 312.5(c)(7)); it bars using information collected for those activities for behavioural advertising or profiling; and it requires separate verifiable parental consent for disclosures to third parties that are not integral to the service (§ 312.5(a)(2)). The Commission's accompanying statement says the Rule does not prohibit collecting personal information for targeted advertising where the parent gives opt-in consent, which must be off by default and cannot be a condition | The Federal Register text confirms both of the reviewer's quotes. The Commission says explicitly that it did not intend a categorical ban on behavioural advertising to children when a parent consents, so the current closing sentence ('but not on behavioural targeting') can be read as a ban the Rule does not impose. | GAM-E-F04 (claim-ledger final_wording; research/E.json finding) |
| R2-02 | See GAM-A-F04 (claim-ledger final_wording); RESEARCH-BRIEF.md research question 1 (the EMARKETER objective sentence); RESEARCH-BRIEF.md Landmines, second bullet | OpenAI round 2 (P1): "Shares of ads" also covered a statement about spend; the mostly-user-acquisition reading needs an inference label everywhere. | EMARKETER, Sara Lebow, 'US game ad revenues growth slows', 10 June 2024 (read in full at https://web.archive.org/web/20260908201342id_/https://www.emarketer.com/content/us-game-ad-revenues-growth-slows), 'Beyond the chart', second bullet: 'Most in-game ad spend goes to mobile, and most of the ads are bought by game developers.' The first half is about spend, split by platform. The second half is about ads. EMARKETER report abstract, 30 May 2024, topic US (https://web.archive.org/web/20260512090736id_/https://www.emarketer.com/content/us-in-game-ad-revenues-forecast-2024): 'Game developers still buy most of the ads served in games. But new, more standardized formats have lured brands that are ... | supported as stated | Restated: EMARKETER said that most in-game ads promote installs of other mobile games (May 2023 report abstract), that game developers buy most of the ads served in games (May 2024 report abstract and June 2024 article) and that most in-game ad spend goes to mobile (June 2024 article); the first two describe shares of ads, the third a share of spend by platform, and none gives a quantified dollar split by buyer or between brand and user-acquisition objectives. This paper's reading of EMARKETER's US game ad revenue series as mixed objective and probably mostly user acquisition, not brand spend, is therefore the author's inference, not an EMARKETER figure. | The June 2024 sentence is explicitly about spend, but by platform; the ad-buying statements concern ads, not dollars. So neither the current 'shares of ads' summary nor the brief's 'spend ... bought by game developers' paraphrase matches the source. No EMARKETER page opened gives a brand-versus-UA dollar split, so 'mostly user acquisition' must be labelled as inference everywhere, the Landmines bullet included. | GAM-A-F04 (claim-ledger final_wording); RESEARCH-BRIEF.md research question 1 (the EMARKETER objective sentence); RESEARCH-BRIEF.md Landmines, second bullet |
| R2-03 | See GAM-A-F06 (claim-ledger final_wording); support GAM-V-A-S01; new source record for the EMARKETER forecast series page; data/estimates.json; RESEARCH-BRIEF.md Landmines, first bullet | OpenAI round 2 (P1): "No earlier than about 2029" conflicts with the FAQ's "surpass by 2029"; the pieces do not establish one crossing year. | The March 2024 vintage has the crossing in 2026. EMARKETER 10 June 2024 (https://web.archive.org/web/20260908201342id_/https://www.emarketer.com/content/us-game-ad-revenues-growth-slows): 'US game ad revenues won't cross the $10 billion mark until 2026' ('per our March 2024 forecast'). EMARKETER, Marisa Jones, 26 December 2025 (https://web.archive.org/web/20251229162956id_/https://www.emarketer.com/content/game-ad-revenues-grow--spending-falls-behind-consumer-attention): 'Game ad revenues will reach nearly $10 billion in the US by 2029, per our forecast. ... It will represent 2.3% of overall digital advertising spending in 2026 and decline to 2% by 2029.' EMARKETER FAQ, Peter Clark, 10 Janua ... | supported with qualification | Restated: EMARKETER's March 2024 forecast put the year US game ad revenues pass $10 billion at 2026, but its later public statements give no single revised crossing year: the 26 December 2025 article says revenues 'will reach nearly $10 billion in the US by 2029' and the 10 January 2026 FAQ says they 'will surpass $10 billion by 2029', although both link to the same EMARKETER forecast series ('US Total Game Ad Spending'), so the two statements conflict and the forecast vintage behind each is unverified. EMARKETER data reported by Digiday on 3 July 2026 (vintage not stated) put total US game ad spend, excluding esports, gaming video content and creator sponsored content, at $9.21 billion in 2026, $9.7 billion in 2027 and $10.2 billion in 2028; the underlying forecast values are subscriber-only, and the public texts do not say whether the changes reflect lower expected spend, a changed definition or | The reviewer is right that the current wording contradicts the FAQ's 'surpass ... by 2029', that it is contradicted by the Digiday-reported $10.2 billion in 2028, and that the two pieces do not establish one crossing year. It is only partly right that the wording 'reverses by 2029': 'no earlier than about 2029' is a correct reading of the December article's 'nearly $10 billion'. Both pieces link to the same forecast series, which turns a supposed revision into a wording conflict, while the vintage each piece used stays unverified. | GAM-A-F06 (claim-ledger final_wording); support GAM-V-A-S01; new source record for the EMARKETER forecast series page; data/estimates.json; RESEARCH-BRIEF.md Landmines, first bullet |
| R2-04 | See GAM-U-F23 (support source); GAM-U-S08; product-register row 'Anzu in-game ad network' (operating_entity); also GAM-U-F29, GAM-U-F31 and the vendor-universe rows for Anzu, Kidoz and Overwolf that cite GAM-U-S08 | OpenAI round 2 (P1): Anzu's registry facts cited a query URL that returns Sayollo. | Live registry reads on 29 Sep 2026 (CKAN datastore, resource f004176c-b85f-4542-8901-7b3176f9a054). (1) https://data.gov.il/api/3/action/datastore_search?resource_id=f004176c-b85f-4542-8901-7b3176f9a054&q=515647808 returns total 1: _id 559416, "מספר חברה": 515647808, "שם באנגלית": "ANZU VIRTUAL REALITY LTD.", "סוג תאגיד": "ישראלית חברה פרטית", "סטטוס חברה": "פעילה" (active), "תאריך התאגדות": "24/04/2017", "שנה אחרונה של דוח שנתי (שהוגש)": 2024, "שם עיר": "תל אביב - יפו". (2) The URL recorded for GAM-U-S08, ...&q=514656453, returns total 1: _id 461470, "SAYOLLO MEDIA LTD", status "בפרוק ע~י בימ~ש", incorporated "16/08/2011", Rehovot. The same URL does not return the Overwolf or Kidoz records ... | supported with qualification | GAM-U-F23 support[0]: change source_id GAM-U-S08 to GAM-ANZU-S33 (URL ...&q=515647808) and set locator to 'datastore record for 515647808 (_id 559416)'. Do NOT overwrite GAM-U-S08's URL, because it correctly retrieves Sayollo for GAM-U-F06 and GAM-U-F35. Product-register Anzu row, operating_entity: replace 'GAM-U-S08' with 'GAM-ANZU-S33'. Same defect elsewhere: GAM-U-F29 (Overwolf 514330331) and GAM-U-F31 (Kidoz 514191527) cite GAM-U-S08 for records its URL does not return. Either split GAM-U-S08 into per-record sources with q=514330331 and q=514191527 (both confirmed active on 29 Sep 2026), or list the per-record query URLs in GAM-U-S08's reliability_notes. Apply the same change to the vendor-universe rows for Anzu, Kidoz and Overwolf. Company facts need no change: the verifier's current wording for GAM-U-F23 matches the record. | The reviewer is right that the only URL recorded for the Anzu facts returns Sayollo, and that q=515647808 confirms the facts. The fix belongs on GAM-U-F23's citation (repoint it to GAM-ANZU-S33), not on GAM-U-S08's URL, which the Sayollo findings need, and the same misattribution affects the Overwolf and Kidoz citations. | GAM-U-F23 (support source); GAM-U-S08; product-register row 'Anzu in-game ad network' (operating_entity); also GAM-U-F29, GAM-U-F31 and the vendor-universe rows for Anzu, Kidoz and Overwolf that cite GAM-U-S08 |
| R2-05 | See GAM-D-F15 (knock-on: GAM-D-F29, stream D summary, D data tables suitability_controls and suitability_documentation_by_cohort) | OpenAI round 2 (P1): Frameplay documents controls in outline in its Terms, not only marketing language. | Frameplay Terms & policies, https://frameplay.com/terms/ ('Updated on January 01, 2023', read 29 Sep 2026). Publishers: 'your dashboard with allow you to create a new game and enter its title, genre, development stage, game rating, content restrictions and whether the game is compliant with ... (COPPA)'. Also: 'Game publishers are solely responsible for determining content restrictions. Frameplay expressly disclaims any liability related to setting improper content restrictions for a specific game.' Advertisers: 'You will also be able to choose your audience by location, language, gender, age range, content restrictions and state whether the ad is COPPA compliant.' The Terms do not say how t ... | supported as stated | Restated: In the live pages opened by 29 September 2026, three seed products show only marketing language on brand safety, with no documented suitability control: iion ('Brand Safe' premium placements), XBOX Media Solutions ('100% brand safe' rewarded video; its measurement page names DoubleVerify and IAS as partners without stating what they do) and Super League (a March 2024 partnership with Common Sense Networks). Two others document controls in outline, as vendor documentation whose effectiveness and current implementation are unverified: Frameplay's Terms of Use (updated 1 January 2023) let publishers set each game's rating and content restrictions, for which publishers are 'solely responsible', and let advertisers choose audience by 'location, language, gender, age range, content restrictions' (its 5 October 2023 IAS partnership covers invalid traffic and impression counts only); Gameloft fo | Frameplay's Terms, as quoted, document publisher-set and advertiser-set content restrictions. Those are controls in outline, not marketing language. The pass-2 verifier made this correction, but the claim ledger kept the pass-1 wording that lists Frameplay as marketing-only. | GAM-D-F15 (knock-on: GAM-D-F29, stream D summary, D data tables suitability_controls and suitability_documentation_by_cohort) |
| R2-06 | See GAM-D-F19; GAM-TWITCH-F11 (Twitch profile) | OpenAI round 2 (P1): Twitch now lists seven content labels, not six. | 1. Twitch Help, 'Content Classification Labels', https://help.twitch.tv/s/article/content-classification-labels?language=en_US (rendered 29 Sep 2026). 'Streamers need to apply Content Classification Labels to indicate ... Mature-Rated Games; Sexual Themes; Drugs, Intoxication, or Excessive Tobacco Use; Violent and Graphic Depictions; Significant Profanity or Vulgarity; Politics and Sensitive Social Issues; or Gambling.' The label table defines Politics and Sensitive Social Issues as 'Discussions or debates about politics or sensitive social issues such as elections, civic integrity, military conflict, and civil rights.' On mislabelling: 'they will receive a warning and a label will be applie ... | supported as stated | Restated: Twitch documents advertiser suitability controls built on mandatory content labels: since 20 June 2023 streamers must apply whichever Content Classification Labels fit their stream (six at launch; seven in Twitch's current help article and API reference, after a Politics and Sensitive Social Issues label was added in November 2024), mislabelling brings warnings and Twitch-applied labels, and a Mature-rated game label is applied automatically for ESRB M-rated games. Per Amazon's Twitch brand-safety guide (dated 22 February 2026), these label signals inform three Inventory Tiers (Expanded, Standard, Restrictive), Contextual Collections to target or exclude content types such as Family Friendly games, and on-demand impression-level content reporting in Amazon DSP. | Twitch's current help article and its API list seven labels, including Politics and Sensitive Social Issues, added in November 2024. The present-tense count of six in GAM-D-F19 and GAM-TWITCH-F11 is out of date. | GAM-D-F19; GAM-TWITCH-F11 (Twitch profile) |
| R2-07 | See GAM-D-F22 (claim-ledger final_wording; research/D.json finding) | OpenAI round 2 (P1): WFA discontinued GARM's activities; the notice does not withdraw the framework. | WFA, 9 August 2024 (https://wfanet.org/knowledge/item/2024/08/09/wfa-discontinues-garm): 'Today we announce that GARM will discontinue its activities.' and 'WFA therefore is making the difficult decision to discontinue GARM activities.' The notice describes 'GARM's toolset includes the Brand Safety Floor and the Adjacency Standards Framework' and says nothing about withdrawing or retiring the framework. IAB Tech Lab implementation.md (raw file at https://raw.githubusercontent.com/InteractiveAdvertisingBureau/Taxonomies/develop/implementation.md; per the GitHub API, the last commit to the file is 2025-08-05) says: 'Buyers should familiarize themselves with the GARM/4A's APB Brand Safety & Sui ... | supported as stated | Restated: IAB Tech Lab's Content Taxonomy page (last updated 11 February 2026) lists version 3.1 (December 2024; open for public comment until 24 January 2025) as current. Its Content Taxonomy 3.0 implementation guidance, both the September 2021 public-comment draft and the version on IAB Tech Lab's GitHub (implementation.md, last changed 5 August 2025), maps platform game genres to taxonomy categories under a Game form-factor vector and defines brand-safety floor and suitability risk levels by reference to the September 2020 GARM and 4A's APB Brand Safety & Suitability Framework; WFA announced on 9 August 2024 that GARM would discontinue its activities, the notice does not say the framework was withdrawn, and the GitHub guidance a year later still tells buyers to familiarise themselves with it. | WFA's notice discontinues GARM's activities and does not withdraw the framework, but the current relative clause ('framework, which WFA discontinued') attaches the discontinuation to the framework. The guideline's continued use of the framework after August 2024 is what matters here, and it should be kept separate from the initiative's closure. | GAM-D-F22 (claim-ledger final_wording; research/D.json finding) |
| R2-08 | See GAM-E-F12 (claim-ledger final_wording; research/E.json finding); source-register GAM-E-S08 | OpenAI round 2 (P1): The FTC list holds COPPA-related entries dated by last update, not distinct COPPA Rule actions by filing date. | GAM-E-F12 relies on GAM-E-S08, the Safe Harbor page (https://www.ftc.gov/enforcement/coppa-safe-harbor-program), not on the tag list. Its 'Cases' block (read 29 September 2026) has 20 entries, from 'Illuminate Education, Inc., In the Matter of (June 5, 2026)' to 'Unixiz, Inc. doing business as i-Dressup.com (April 24, 2019)', then 'View More Cases'. That link (https://www.ftc.gov/enforcement/cases-proceedings/terms/1421+875) is headed 'Cases Tagged with Children's Privacy + Children's Online Privacy Protection Act (COPPA)', and its first 20 entries are the same. Each entry's date matches its case page's 'Last Updated' date, not the filing date. 1923203-epic-games-matter: 'Last Updated June 2 ... | supported as stated | Restated: The FTC's COPPA Safe Harbor page, read on 29 September 2026, lists 20 related case entries (the first page of the FTC's list of cases tagged Children's Privacy + COPPA), each dated by its case page's last-updated date rather than its filing date. Classified by complaint, 16 are COPPA Rule actions filed between 24 April 2019 and 29 September 2025: four against game companies (HyperBeard, filed 4 June 2020; Epic Games, 19 December 2022; Microsoft, 5 June 2023; Cognosphere, 17 January 2025) and one against an ad-technology company (OpenX, 15 December 2021). Four are not COPPA Rule actions: Epic's separate administrative billing matter, Miniclip's Section 5 case over a false claim of safe-harbor membership, Retina-X and Illuminate Education. None of the 20 entries is an action against an in-game ad network or game-ad vendor, so in that list this review found no FTC COPPA action against such | The FTC pages confirm that four of the 20 entries are not COPPA Rule actions and that each displayed date is the case page's update date (both Epic entries show 25 June 2025 for matters filed on 19 December 2022). The finding relied on the Safe Harbor page (GAM-E-S08), not the tag list. Reclassifying and re-dating leaves the bounded absence unchanged, because no entry at any date is against an in-game ad vendor. | GAM-E-F12 (claim-ledger final_wording; research/E.json finding); source-register GAM-E-S08 |
| R2-09 | See GAM-ROBLOX-F08 (and the Roblox profile summary line citing it) | OpenAI round 2 (P1): Matching headline numbers do not establish equivalence with the IIG image-ad criteria. | IIG Guidelines 2.0 (August 2022), https://mediaratingcouncil.org/sites/default/files/Standards/In-Game%20Advertising%20Measurement%20Guidelines%2008-31-22.pdf: - Section 3.2, time requirement: 'greater than or equal to one continuous second, post ad render'. - Note 3 under 3.2: 'Once an ad qualifies as a valid Viewable Impression, it should only be counted as one Viewable Impression within that user session, regardless of subsequent exposures.' - Section 3.2.1: '2 continuous seconds of the video advertisement is played'. - Section 3.2.2: 'The 1.5% should represent the total coverage of the visible portion of the ad relative to the screen real estate. In other words, an ad that is 2% of a use ... | supported as stated | Restated: Compared by this review with the IIG Guidelines 2.0 (August 2022), Roblox documents image-ad impression thresholds with the same headline numbers as the guideline's viewable display criteria (at least 50% of pixels visible, at least 1 second looked at, 1.5% of the viewport, an angle of up to 55 degrees), but its wording does not establish equivalence. Roblox does not say the second is continuous, that the 1.5% refers to the visible portion of the ad relative to the screen (IIG section 3.2.2), or how repeat exposures within a session are counted (IIG counts one viewable impression per user session), so alignment is unverified. Separately, its autoplay video impression (0.5 seconds) is shorter than the guideline's 2 continuous seconds. Roblox applies these definitions to the Impressions column of its Ads Manager (beta) reports and does not claim alignment with the guidelines. | The IIG text requires the visible portion to cover 1.5% of the screen, one continuous second, and one count per session. Roblox's wording states none of these, so matching headline numbers does not establish equivalence. The video gap stands. | GAM-ROBLOX-F08 (and the Roblox profile summary line citing it) |
| R2-10 | See GAM-G-F25 (and its claim-ledger final_wording); H3 coding; the stream G summary; table console_pc_named_titles; the counterevidence field of GAM-R-F322 | OpenAI round 2 (P1): "H3 supported" bypasses the registered unresolved rule and overstates the qualifying title count. | H3 registered rules (data/hypotheses.json). Supported: "Fewer than ten named console or cloud titles are documented, by the network, the platform or the game's publisher, as carrying live in-game brand placements buyable in the US in the 12 months to the cutoff, and no console platform documents a generally available in-game brand ad product; placements that do exist sit outside gameplay (dashboard, store, TV apps, pre-session ads) or are in test." Unresolved: "A placement's location inside or outside gameplay, or its status, cannot be established from the platform's, network's or publisher's own documents." Microsoft (https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions ... | supported with qualification | Restated: Coding of H3 (inference) on the sources screened by 28 September 2026: H3 is unresolved under its registered rule. Microsoft's XBOX Media Solutions markets console brand integrations 'into gameplay' and 'XBOX Experiences' that 'leverage blockbuster gaming titles', names no title, and its own documents do not establish whether these render inside gameplay or what their status is; no console platform documents a named, generally available in-game brand ad product with a buying route, and every other platform placement found sits outside gameplay (Xbox dashboard and store, Live TV on PS5, pre-session ads in the Xbox cloud test and on GeForce NOW, Luna's connection screen) or is in test. EA's direct-sold program names brand integrations in three console titles: EA SPORTS FC 26 (in-match LED ad-boards in the US, per Comcast and EA's announcement of 17 November 2025), Madden NFL 26 (dated only | The reviewer is right that 'supported' cannot be reached: the supported rule needs every existing platform placement to sit outside gameplay or be in test, and Microsoft's own page leaves its console integrations unplaced, which triggers rule_unresolved. It is also right that the three titles are not all qualifying ones. It is wrong that the EA release leaves the titles unnamed, since its partner list names FC 26, College Football 26 and Madden NFL (only the ad units' 'select EA SPORTS games' are unnamed), and wrong to hold H3 open 'pending qualifying title-specific evidence': fewer qualifying titles would satisfy H3's under-ten clause, not block it. | GAM-G-F25 (and its claim-ledger final_wording); H3 coding; the stream G summary; table console_pc_named_titles; the counterevidence field of GAM-R-F322 |
| R2-11 | See GAM-C-F25 (and its claim-ledger final_wording); H2 coding note | OpenAI round 2 (P1): The H2 share used a stale denominator of ten register products. | Product register (data/product-register.json), as read 29 Sep 2026: - Cohort 'network' has nine rows: Anzu, Frameplay, IION, Gadsme, Adverty, Gameloft for brands (COMBO!), Super League in-game advertising, Kidoz, Admazing. - Cohort 'platform' has four rows: Roblox brand ads, XBOX Media Solutions in-game ads, EA Advertising, Zynga Ads. - 'Trivver in-game ads' is cohort 'listed'. data/scoring-rules.json maps 'network' to "Cohort A: in-game ad networks" and 'platform' to "Cohort B1: platform and publisher in-game ads". H2 (data/hypotheses.json): - Population: "every product scored in cohorts A and B1 at the cutoff". - rule_unresolved includes "fewer than six products are scored in cohorts A a ... | supported as stated | Restated: Preliminary coding of H2 (inference): the in-game guidelines are a final release, and on 29 September 2026 MRC's records show no accreditation or MRC-confirmed audit for any of the 13 cohort A and B1 candidates in the product register (nine networks, four platforms); the one gaming-environment audit in process is Trivver's, a listed product, and IAS In-Game Measurement left MRC's process in January 2026 after IAS withdrew. H2's population is the products actually scored in cohorts A and B1, which admission has not yet fixed, so both shares are to be computed on that scored set at the cutoff (third-party verification currently rests on announcements dated 2022 to 2025 for Anzu, Gadsme, Frameplay and Roblox, and EA Advertising and Admazing state IAS or DoubleVerify partnerships without a verifier record), and H2 codes as unresolved if fewer than six products are scored. | The register now holds 13 A/B1 candidates (9 network, 4 platform), and Trivver is listed, so '4 of the 10 register products' is stale. H2 is defined on scored products, not candidates, and the new candidates add vendor verification claims, so both numerators must be rebuilt. The accreditation arm (zero) holds for any denominator, which means H2 still waits only on the six-scored condition. | GAM-C-F25 (and its claim-ledger final_wording); H2 coding note |
| R2-12 | See GAM-B-F27; GAM-R-F348; propagated wording in evidence/profiles/anzu.json, gadsme.json and xbox.json; GAM-B-F03 | OpenAI round 2 (P1): Reciprocal seller declarations establish authorisation, not active buyable supply. | https://www.king.com/app-ads.txt (Last-Modified 29 Sep 2026 21:45:56 GMT on my read): line 688 'gameloft.com, KING, DIRECT, 9df283b2158909d5'; lines 77 and 787 'gameloft.com, KING, DIRECT'; line 1 'OWNERDOMAIN=activisionblizzardmedia.com'. https://gameloft.com/sellers.json (32 entries): {'seller_id': 'KING', 'seller_type': 'PUBLISHER', 'name': 'King', 'domain': 'king.com'}, plus ZYNGA and eight further Zynga-named PUBLISHER entries (GRAM, PEAK, SOCIALPOINT, POPCORE, ROLLICGAMES, DRAGONCITYGAME, MONSTERLEGENDSGAME, WORDLIFEGAME). IAB Tech Lab ads.txt 1.1 (https://iabtechlab.com/wp-content/uploads/2022/04/Ads.txt-1.1.pdf): the standard lets content owners declare which advertising systems are ... | supported with qualification | Restated: GAM-B-F27: King, Rovio, Outfit7, Zynga and Miniclip each authorise gameloft.com as a DIRECT seller in their app-ads.txt files under seller IDs that Gameloft's sellers.json lists as their PUBLISHER accounts (files read 28 to 29 September 2026). This documents authorised supply relationships with COMBO!'s operator, not which titles, placements or geographies COMBO! actively sells, so the view that some King inventory is offered both through COMBO! and through Microsoft's XBOX Media Solutions is this review's inference, unverified at title or placement level; Gameloft is in turn listed as a PUBLISHER in Anzu's, Gadsme's and Frameplay's sellers.json, with matching DIRECT lines in its own app-ads.txt for Anzu and Gadsme but not for Frameplay. / GAM-R-F348: Zynga's app-ads.txt (last modified 23 September 2026) authorises gadsme.com, anzu.io, gameloft.com (seller ID 'ZYNGA'), iion.io and pubmat | The reviewer is right: under the IAB specs, an app-ads.txt DIRECT line matched to a sellers.json PUBLISHER entry establishes authorisation, not that any title, placement or US impression is buyable, so 'therefore reachable' and 'is sold by' go beyond the evidence. The modification keeps the verified matches the reviewer's text drops (Zynga and Miniclip, and Gameloft's position in the Anzu, Gadsme and Frameplay files), adds Gameloft's own COMBO! member-list assertion, and recodes GAM-B-F27 from independent_measurement to synthesis. | GAM-B-F27; GAM-R-F348; propagated wording in evidence/profiles/anzu.json, gadsme.json and xbox.json; GAM-B-F03 |
| R2-13 | See GAM-H-F18; source GAM-H-S14 (alias GAM-ROBLOX-S22); registry GAM-REG-07; data/study-evidence.json row 'Roblox / Ipsos Immersive Is Effective' | OpenAI round 2 (P1): A study known only from the seller's release was coded commissioned_measurement. | Roblox newsroom, Allison McDuffee, 18 June 2026 (https://about.roblox.com/newsroom/2026/06/roblox-ipsos-edo-immersive-effective-measurement-partnerships-cannes-lions): 'As revealed in Immersive Is Effective, a new study conducted in partnership with Ipsos, advertisements on Roblox ...'. Footnote 1: 'Source: Roblox's "Immersive Is Effective" in partnership with Ipsos. Base: n=200 13-34 Roblox users, n=200 13-34 streaming platform users, n=200 social platform users. 50% female, 50% male. Fielded: May 6-27, 2026.' Body: '52% of surveyed 13- to 34-year-old Roblox users spend more time paying attention to ads on Roblox than they usually do, outperforming streaming by 13 percentage points and soci ... | supported as stated | Restated: Roblox's 18 June 2026 release reports that its 'Immersive Is Effective' survey, 'conducted in partnership with Ipsos' (fielded 6-27 May 2026; bases of 200 Roblox users aged 13-34, 200 streaming-platform users aged 13-34 and 200 social-platform users whose age range the footnote does not state; survey geography not stated), found that 52% of the Roblox users say they pay more attention to ads on Roblox than they usually do, 13 points above the streaming base and 11 above the social base, and that 63% of respondents say product information they see on Roblox is memorable. The full report sits behind a registration form and was not opened, and the release does not say who paid for the study. | The only opened source is Roblox's own release. 'In partnership with Ipsos' names a research partner, not a payer, and neither the release nor the gated report page states who funded the study. Under the shared rules the finding is a vendor_assertion, with funding recorded as unknown. | GAM-H-F18; source GAM-H-S14 (alias GAM-ROBLOX-S22); registry GAM-REG-07; data/study-evidence.json row 'Roblox / Ipsos Immersive Is Effective' |
| R2-14 | See GAM-R-F309 (also GAM-R-F353 and GAM-U-F23, listed below) | OpenAI round 2 (P1): Comparisons of vendor-published sellers.json and app-ads.txt files were coded direct_record. | https://kidoz.net/sellers.json (ext.time_last_generated 'September 29, 2026 09:59:31'; contact_address Netanya, Israel): 281 entries (269 PUBLISHER, 11 INTERMEDIARY, 1 BOTH) on 252 domains on my read, including {'seller_id': 15286, 'name': 'sybo games_COPPA', 'domain': 'sybogames.com', 'seller_type': 'PUBLISHER'} and 15288 'sybo games_NON-COPPA'. https://sybogames.com/app-ads.txt line 1385 'kidoz.net, 15286, DIRECT, a109366414b7335e' and line 1386 'kidoz.net, 15288, DIRECT, a109366414b7335e'. SHARED-RESEARCH-RULES.md defines direct_record as a filing, registry or official document. An official document is one issued by a regulator, standards body, registry, court or accreditor, or a company' ... | supported as stated | GAM-R-F309 evidence_class: direct_record → synthesis. Provenance: Kidoz's sellers.json and the SYBO, Outfit7 and TutoTOONS app-ads.txt files (declarations published by those companies), read and counted by the author on 29 September 2026. Funding: n/a (the author's reading). Confidence stays high for what the files declare; the existing limitation already says the files do not show live delivery or scale. The claim wording (verifier correction) stays. Other findings coded direct_record that rest on sellers.json or app-ads.txt readings: (1) GAM-R-F353 (Zynga's app-ads.txt, a zynga.com/sellers.json 404 and a vendor page) → synthesis. (2) GAM-U-F23 (registry facts plus the author's count of Anzu's sellers.json) → split the registry facts (direct_record) from the sellers.json count, or code the whole finding synthesis. Correctly direct_record, with the ad file used only incidentally: GAM-U-F | The rules reserve direct_record for filings, registries and documents issued by regulators, standards bodies, courts or accreditors. sellers.json and app-ads.txt are declarations that commercial parties publish under an IAB format. A comparison across them is therefore synthesis with vendor-published provenance, and it does not validate active integrations. | GAM-R-F309 (also GAM-R-F353 and GAM-U-F23, listed below) |
| R2-15 | See GAM-R-F301; new ytd candidate finding for the Kidoz row; source register (add the 6-K) | OpenAI round 2 (P1): A half-year result sat inside a finding labelled actual. | The EDGAR index https://www.sec.gov/Archives/edgar/data/1318482/000149315226040250/0001493152-26-040250-index.htm reads: 'Form 6-K - Report of foreign issuer', Filing Date 2026-08-26, Period of Report 2026-06-30; EX-99.1 is ex99-1.htm. In https://www.sec.gov/Archives/edgar/data/1318482/000149315226040250/ex99-1.htm, the 'Unaudited Condensed Interim Consolidated Statements of Operations and Comprehensive Loss' show columns 'Six Months ended June 30, 2026 \| Six Months ended June 30, 2025 \| Three Months ended June 30, 2026 \| Three Months ended June 30, 2025' and the line 'Revenue: $ 6,281,576 $ 5,168,519 $ 3,334,835 $ 2,430,216'. Net loss is $(2,372,907) against $(1,111,340) for the six months. ... | supported with qualification | Restated: Kidoz is live in the scoring period and fits cohort A: Kidoz Inc.'s FY2025 20-F (filed 29 April 2026) says its SDK is integrated into thousands of mobile apps and that its in-game advertising arrangements are mostly governed by insertion order, with a separate programmatic revenue stream, and its Form 6-K for the half-year to 30 June 2026 (filed 26 August 2026) reports revenue for that period. Its sellers.json, read on 29 September 2026, lists about 270 publisher accounts (268 to 269 depending on the read), most of them third-party developers. | The reviewer is right: a six-month result inside the current fiscal year is a year-to-date figure, and F301 carries it under claim_status 'actual'. Each finding has one claim_status, so the split is done by removing the figure from F301 and filing it as its own ytd finding, noting that it is unaudited, consolidated and company-wide. | GAM-R-F301; new ytd candidate finding for the Kidoz row; source register (add the 6-K) |
| R2-16 | See GAM-SUPERLEAGUE-F09 (claim-ledger final_wording); source GAM-SUPERLEAGUE-S10 measurement_period | OpenAI round 2 (P1): "None has a stated period" was wrong for the campaign count and the monthly metric. | Super League release, 18 March 2026 (https://ir.superleague.com/news-events/press-releases/detail/291/super-league-signs-definitive-agreement-to-acquire-misfits), body: 'Over the past three years, the Misfits Ads Division has executed more than 150 brand partnership programs across some of the largest and most popular games in Roblox, many of the most successful Minecraft servers, and in partnership with several of the most recognized gaming creators in the world.' Body: '...Misfits' significant owned-and-operated game portfolio reaching more than 100 million monthly active users, which will remain within Misfits as a standalone business.' About Misfits Gaming Group: 'With a portfolio of gam ... | supported as stated | Restated: Super League's 18 March 2026 release says the Misfits Roblox game portfolio, which 'will remain within Misfits as a standalone business', reaches 'more than 100 million monthly active users' (repeated in the 6 May 2026 completion release) and includes Knockout, 'a current Top 10 game on all of Roblox', and that the Misfits Ads Division executed more than 150 brand partnership programs 'over the past three years' across Roblox games, Minecraft servers and gaming creators. The audience figure names a monthly metric but no reference month, measurement source or deduplication across games, the ranking names no metric or window, and the three-year period has no start or end date; all are vendor assertions found only in releases (the completion release was furnished, not filed, as Exhibit 99.1 to the 6 May 2026 Form 8-K), not in the 10-K, 10-Q or transaction agreements, and none is an ad-eligi | The release gives the campaign count a stated, though undated, three-year period and gives the audience a monthly metric, so 'none has a stated period' is wrong. What is missing is a reference month, measurement source and deduplication for the MAU, and a metric and window for the ranking; all three stay vendor assertions and none is ad-eligible reach. | GAM-SUPERLEAGUE-F09 (claim-ledger final_wording); source GAM-SUPERLEAGUE-S10 measurement_period |
| R2-17 | See product-register.json: Misfits Ads (us_available) | OpenAI round 2 (P1): Misfits Ads us_available "no (assets sold)" treated an asset purchase as a retirement. | Super League release (https://ir.superleague.com/news-events/press-releases/detail/297/super-league-completes-acquisition-of-misfits-ads-division, 6 May 2026): "The Misfits Ads Division adds programmatic revenue, proprietary technology, and proven execution with leading brands." It also says "Proprietary rewarded video and monetization capabilities increase control over how programs are delivered". Super League Form 10-Q for Q2 2026 (https://www.sec.gov/Archives/edgar/data/1621672/000143774926028066/slgg20260630_10q.htm): "Disclosure of revenue and net loss for Misfits on a stand-alone basis for the applicable periods presented is not practical due to the integration of Misfits, including s ... | supported as stated | 1. Misfits Ads, us_available: "unconfirmed: Super League bought the Misfits Ads business assets on 1 May 2026 and reports Misfits' sales and products as integrated with its own (Form 10-Q for Q2 2026); continuing capabilities are assessed under the Super League in-game advertising row, and no source opened establishes whether the legacy Misfits Ads offering remains separately available to US buyers." 2. Misfits Ads cohort stays 'historical'; operating_entity and owner are unchanged. 3. For consistency, apply the same treatment to the 'Let's Bounce' row: us_available 'unconfirmed: asset acquisition by Super League, January 2026; activities integrated per the Q2 2026 10-Q; separate availability not established'. | An asset acquisition is not a retirement (shared rule 11). The buyer's filing records integration of the Misfits sales and products, not cessation, so 'no' asserts more than the record shows. | product-register.json: Misfits Ads (us_available) |
| R2-18 | See product-register.json and vendor-universe.json: new rows 'Medal advertising' (cohort adjacent, B2 candidate) and 'Medal Quests' (listed); new finding for stream U or R | OpenAI round 2 (P1): Medal's brand-facing media offer was missing from the register. | Medal grow page (https://medal.tv/grow, served 29 Sep 2026, footer © 2026). - Pitch: "Reach your audience where they play"; "By aligning your brand or activation with the daily habits of gamers". - Placements: "Sponsored stories: Communicate your brand's value to drive purchase intent, wishlist adds, or clicks"; also Sponsored top 10, Branded watermarks, Clip contest. - Ad units: "Ads on Medal are served front and center with billions of impressions: they're high visibility and utilize IAB standards"; "4.5B monthly impressions served"; "90%+ Viewability". - Audience: "10M MAU", "3M DAU", "45% US". The base and period are unstated; the page elsewhere says "over 12 million gamers". - Case stud ... | supported with qualification | Restated: As served on 29 September 2026, Medal B.V.'s medal.tv/grow page sells placements rendered in its own gameplay-clip app and website, not inside games: IAB-standard ad units including a homepage roadblock, sponsored stories, sponsored top 10, branded watermarks and clip contests. Its ads.txt lists direct programmatic sellers (Google, OpenX, Magnite, PubMatic), and it claims 10M MAU, 45% US users and 4.5B monthly impressions (vendor assertions; base and period unstated). Its only case study (Paddington, undated) reports game-promotion outcomes (the game's concurrent users and time in game), and its Quests are partnerships with game developers, so they are listed as paid discovery. | The reviewer is right that Medal sells brand-facing, game-adjacent media and fits B2 as a candidate, with Quests kept as listed paid discovery; its ads.txt and SSP sellers.json entries confirm live supply. It overstates the evidence in two places. The 'delivered paid-media case' is an undated game-promotion case measured by the game's concurrent users, not a brand campaign in the window. And '45% US' is a user share with no stated base, while the operator is a Dutch B.V. with no documented US purchasing terms. | product-register.json and vendor-universe.json: new rows 'Medal advertising' (cohort adjacent, B2 candidate) and 'Medal Quests' (listed); new finding for stream U or R |
| R2-19 | See GAM-C-F06 support (the GAM-C-S07 passage); source-register GAM-C-S07 | OpenAI round 2 (P1): A model-generated page summary was quoted as a source passage. | The IAB guideline page (https://www.iab.com/guidelines/gaming-measurement-framework/, fetched with curl on 29 September 2026, HTTP 200) shows 'IAB Gaming Measurement Framework' and 'June 26, 2025'. It then reads: 'Developed in collaboration with industry leaders, the framework outlines standard gaming ad formats by media type (Display, Video, Audio, and Custom) and explains baseline (must-have) and additional (nice-to-have) metrics to help advertisers evaluate performance and make smarter investment decisions.' The list of benefits breaks off at 'Support brand investment by . . .', followed by 'Claim your free account to continue reading. Log In or Create Account'. The ungated text contains ... | supported as stated | GAM-C-F06 support: replace the GAM-C-S07 entry with a verbatim passage: 'Developed in collaboration with industry leaders, the framework outlines standard gaming ad formats by media type (Display, Video, Audio, and Custom) and explains baseline (must-have) and additional (nice-to-have) metrics to help advertisers evaluate performance and make smarter investment decisions.' Locator: 'page header (June 26, 2025) and opening paragraph, ungated portion, read 2026-09-29'. Update the claim-ledger 'passage' field to match. GAM-C-F06 limitations: replace 'The S07 support is a model summary, not a verbatim passage' with 'The guideline page is gated after its opening paragraphs; the ungated text names no co-issuer and does not mention MRC, accreditation or audit (bounded to that text); substantive framework claims rest on the full PDF (S05).' Source-register GAM-C-S07: access_limits 'partial: read | The quoted S07 passage is a model summary presented as support, which breaks the rule that support must be an exact passage. The page is also gated after its opening lines, so any absence of an MRC reference can be claimed only for the readable part, and the substantive claims already rest on the PDF. | GAM-C-F06 support (the GAM-C-S07 passage); source-register GAM-C-S07 |
| R2-20 | See GAM-R-F116; GAM-XBOX-F15 | OpenAI round 2 (P1): Two absence statements were not bounded to the sources screened. | IION home page, https://www.iion.io/ (read 29 Sep 2026): '60,000+ games. Unmatched scale and precision.' It also says '41% \| Unique Gaming Reach \| Reach untapped audiences through standard programmatic', with no base. IION advertisers page, https://www.iion.io/advertisers: 'Home to a premium audience of 3.6 billion.' and '3.6 BILLION \| Unique Audience', with no period or definition. IION blog, 23 Oct 2025, https://www.iion.io/blog/iion-unlocks-access-to-1-2m-daily-gamers-through-premium-arkadium-partnership: Arkadium brings '1.2 million average daily active users spending 45 minutes per session across a portfolio of 300+ popular casual games' and 'over 1 billion monthly available ad reques ... | supported with qualification | Restated: GAM-R-F116: IION's cohort A supply is documented mainly through Voodoo: a case listed as 20 November 2025 describes a direct integration delivering 70 to 80 million monthly impressions and names Mob Control, Epic Plane and Paper.io 2 as Voodoo's hit titles (it does not say those titles carried IION placements), and Voodoo's app-ads.txt authorises 'iion.io, 10231, DIRECT', matching IION's sellers.json entry for Voodoo SAS. In the IION pages screened through 29 September 2026 (home, advertisers, platform, about, case studies and the blog posts listed in stream R), this review found no ad-eligible audience figure with a metric and period for IION's in-game supply: '60,000+ games' has no metric or period, the advertisers page's '3.6 billion' 'Unique Audience' has no period and is not described as IION's own reach, and the 1.2 million average daily active users IION reports for Arkadium (Octo | Both statements break rule 1 and need the named search scope. The reviewer's bounded IION wording also needs scoping to in-game supply, because IION itself publishes a figure with a metric and period (1.2 million average DAU) for its Arkadium web-game supply. The Microsoft marketing sentence the reviewer cites should be acknowledged, since it names no console title or format and so does not establish a product. | GAM-R-F116; GAM-XBOX-F15 |
| C10 | When two verifier passes gave the same verdict rank, the ledger kept the first file in name order, so pass-1 wording won ties for streams A to F. | Found while checking R2-05 and R2-12 (pass-1 wording survived for GAM-D-F29 and GAM-B-F27). | build/build-claims.mjs keeps the first verdict unless a later one is more severe; A-pass1.json sorted before A.json. | structural | Pass-1 verifier files renamed .pass1.json so the newer pass-2 wording wins ties; a more severe verdict still wins either way. | Pass 2 ran a day later and holds newer facts. | evidence/verify/A-F pass-1 files; data/claim-ledger.json |
| S01 | Proposed cell scores written by the profile and row-evidence agents before verification. | Refute-mode verifiers checked every proposed cell against the anchors, caps and sources. | evidence/verify/.json and evidence/verify/R-part*.json, cells_checked | supported with qualification | In the single scoring pass (29 September 2026) 19 proposed scores were changed to the verifier's suggested score and 3 evidence grades corrected: superleague/supply 3→2; iion-in-game-brand-network/optimisation_outcomes 3→2; iion-in-game-brand-network/programmatic_access 3→2; iion-in-game-brand-network/suitability_age_data 1→2; gameloft-for-brands/measurement_function 2→3; gameloft-for-brands/coverage 2→ne; admazing-mobile-gaming-ads-for-brands/targeting_context 3→2; kidoz-kid-safe-ad-network/format_execution 2→1; kidoz-kid-safe-ad-network/measurement_function 3→1; ea-advertising/programmatic_access 0→1; ea-advertising/transparency ne→3; ea-advertising/suitability_age_data 2→3; zynga-ads/programmatic_access 1→2; livad-livestream-video-ads/us_availability 2→1; venatus-gaming-media/format_execution 2→1; playwire-gaming-media/targeting_context 3→2; playwire-gaming-media/programmatic_access 2→ne; enthusiast-gaming-media/format_execution 1→2; esl-faceit-group-esports-media/us_availability 1→2. Grades: iion-in-game-brand-network/programmatic_access D→B; adverty-in-game-ad-network/supply A→D; ea-advertising/transparency E→B. IION us_availability relabelled live (was announced). | Only verified or qualified cells count; a qualified cell is scored against the qualified claim. | data/scores.json, data/matrix-extra.json |
| IR1-G1-01 | Among the 49% who treat games as part of other channels rather than a channel of their own, the first reason listed is "not yet enough scale" | Internal review round 1 (blocking, fact): The split is reversed. In IAB's March 2024 study, 51% treat games as part of another channel and 49% treat games as their own channel. The ledger note behind this line is also reversed, so a reader who checks the ledger gets the same wrong number. | GAM-S-S01, https://www.iab.com/wp-content/uploads/2024/03/IAB_AP_Changing_The_Game_March_2024.pdf, p. 19, donut chart '% ADVERTISERS THAT CONSIDER GAMES ITS OWN MEDIA CHANNEL'. The chart labels read 'PART OF ANOTHER CHANNEL 51%' and 'OWN CHANNEL 49%'. The page was rendered and viewed as an image on 29 Sep 2026. The claim-ledger GAM-S-F09 review_note says 'p. 19 shows 51% own channel and 49% part of another channel', which is the reverse of the chart. The ledger note also needs correcting. | supported as stated | 02-market.md: '49% who treat games as part of other channels' changed to '51%' (merged with the G1-08 wording). | I rendered p.19 of the IAB PDF and read the text coordinates. 51% sits beside 'PART OF ANOTHER CHANNEL' and 49% beside 'OWN CHANNEL', so the chapter had the split reversed. The ledger's GAM-S-F09 review_note has the same reversal, but it is outside my edit scope and the orchestrator should correct it. | chapters/02-market.md |
| IR1-G1-02 | Primary records show four specialist exits since 2021, and none attributes an exit to the format | Internal review round 1 (blocking, consistency): The count of four contradicts the ledger's final wording and chapter 10. Both say at least three specialist companies. The fourth event, Misfits Ads, was a business line sold by an esports company, not a specialist company. Two of the four events (Bloxbiz and Misfits Ads) were purchases by one buyer, Super League, which the ledger calls consolidation rather than exits. | claim-ledger GAM-U-F35 final_wording: 'primary records show at least three specialist in-game ad companies sold, merged or put into insolvency ... Further events are Super League's purchase of the Misfits Ads business line from an esports company'. The same review_note says 'Misfits Ads was a business line of Esports Now, LLC, an esports company, not a specialist company' and 'Two of the listed exits (Bloxbiz and Misfits Ads) are consolidation by one acquirer'. Chapter 10, line 65: 'the three companies alone meet the threshold'. | supported as stated | 00-abstract.md finding 7: 'four specialist exits' changed to 'at least three specialist in-game ad companies sold, merged or put into insolvency'. | The GAM-U-F35 final_wording and chapter 10 (line 65) both say 'at least three specialist in-game ad companies'. Misfits Ads was a business line of an esports company, so it is not a specialist company. | chapters/00-abstract.md |
| IR1-G1-03 | On console, the ad placements found sit outside gameplay or are in test, except where a publisher sells its own games. | Internal review round 1 (major, consistency): This is the overstatement that the ledger label explicitly corrects. Microsoft's XBOX Media Solutions markets console integrations 'into gameplay', and its own documents do not show whether they render inside or outside gameplay. That uncertainty is the reason H3 is unresolved. The in-short summary leaves it out, so it contradicts finding 5 in the same abstract and chapter 8. | claim-ledger GAM-G-F25 review_note, Label: 'The stream summary's opening sentence (every ad placement ... renders outside gameplay or is in test) overstates the evidence ... should say renders outside gameplay, is in test, or, for Microsoft's console custom integrations, cannot be placed from the platform's own documents'. GAM-G-S04 (https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/custom-integrations): 'integrating brands into gameplay across mobile, PC, and console'. Chapter 8, line 6: 'codes H3 ... as unresolved because Microsoft markets console integrations "into ga | supported with qualification | 00-abstract.md in-short: 'On console, the platform placements found sit outside gameplay or are in test, except Microsoft's console integrations, whose location its own documents do not establish; brand placements inside console gameplay are found only where a publisher sells its own games.' | The GAM-G-F25 label calls the outside-gameplay summary an overstatement, and GAM-G-S04 markets integrations 'into gameplay across mobile, PC, and console'. I shortened the reviewer's fix and kept the in-short free of citations. | chapters/00-abstract.md |
| IR1-G1-04 | the amended COPPA Rule has applied since 23 June 2025 (H4, supported) | Internal review round 1 (major, rule): This merges the effective date with the compliance date, which shared rule 8 forbids. The amended Rule took effect on 23 June 2025, but most regulated entities did not have to comply until 22 April 2026. The abstract's wording also contradicts chapter 7. COPPA is not expanded anywhere in the abstract. | GAM-E-S01, Federal Register API (https://www.federalregister.gov/api/v1/documents/2025-05904.json): 'Effective date: The amended Rule is effective June 23, 2025.' hypotheses.json H4 population: 'COPPA as amended (effective 23 June 2025, compliance by 22 April 2026)'. Chapter 7, line 53: 'it took effect on 23 June 2025, and most regulated entities had until 22 April 2026 to comply'. | supported with qualification | 00-abstract.md finding 4: 'the amended Children's Online Privacy Protection (COPPA) Rule took effect on 23 June 2025, with compliance due for most regulated entities by 22 April 2026'. | The Federal Register API gives effective_on 2025-06-23. Its DATES section says 'Except with respect to §312.11(d)(1), (d)(4), and (g), regulated entities have until April 22, 2026 to comply.' I used the same wording as chapter 7 (line 8) so the two chapters match. | chapters/00-abstract.md |
| IR1-G1-05 | but neither is established as US and one lead is inside sampling error; H1 is unresolved | Internal review round 1 (major, fact): 'One lead is inside sampling error' implies that the other lead, IAB Australia's, is a real difference. The ledger's final wording says IAB Australia's lead is not established as a distinguishable rank either. The same wording appears in the chapter 2 in-short (see G1-06), in the chapter 12 table (line 15) and in chapter 13 (line 61). | claim-ledger GAM-S-F17 final_wording: 'Brand safety's 3-point lead over lack of scale (29% vs 26%) is very unlikely to be a distinguishable rank. IAB Australia's 15-point lead of brand safety over the fifth-ranked barrier (47% vs 32%) has a 95% half-width of about ±7 to ±17 points ... so it is not established as a distinguishable rank either.' The review_note adds: 'GAM-S-F18 and the summary lean on the exceeds either share's margin reading; their confidence should not rest on it.' | supported as stated | 00-abstract.md finding 8: 'one lead is inside sampling error' changed to 'neither lead is shown to exceed sampling error'. | GAM-S-F17 says Advertiser Perceptions' lead is 'very unlikely to be a distinguishable rank' and that IAB Australia's lead is 'not established as a distinguishable rank either'. 'One lead' therefore implies the other lead holds, which the ledger does not support. The same wording remains in chapter 12 (line 15) and chapter 13 (line 61), which are outside my scope. | chapters/00-abstract.md |
| IR1-G1-06 | neither is established as US, one lead is inside sampling error, and the same publisher says the belief that the audience can be reached more cheaply elsewhere is growing | Internal review round 1 (major, fact): The source says the same audiences can be reached 'more effectively' elsewhere, not 'more cheaply'. Effectiveness and cost are different constraints, and the chapter body (line 49) uses the correct wording. The sentence also carries the misleading 'one lead is inside sampling error' (see G1-05). | GAM-S-S03, https://www.advertiserperceptions.com/gaming-advertising-2024-top-takeaways-from-our-latest-report/ (datePublished 2024-09-09): 'the rise in perception that the same audiences can be reached more effectively elsewhere is a growing challenge'. GAM-S-S02, MediaPost: 'Can reach the same consumers more effectively elsewhere—26%'. claim-ledger GAM-S-F02 final_wording: 'reached more effectively elsewhere'. | supported as stated | 02-market.md in-short: 'neither lead is shown to exceed sampling error' and 'the same audiences can be reached more effectively elsewhere'. | I fetched the Advertiser Perceptions post (GAM-S-S03), which reads 'the same audiences can be reached more effectively elsewhere is a growing challenge'. 'More cheaply' is wrong. I also carried over the G1-05 correction. | chapters/02-market.md |
| IR1-G1-07 | IAB's March 2024 study of 300 US buyers already investing in games asked no barrier question | Internal review round 1 (major, rule): The study is presented as IAB's alone. In fact it was run with Advertiser Perceptions and sponsored by Admazing and Venatus, both of which this paper scores (cohort A and cohort B2). Shared rule 5 requires naming the party behind a number when that party has an interest. The 86% brand-safety figure is used as US counterevidence without that label. | GAM-S-S01, https://www.iab.com/wp-content/uploads/2024/03/IAB_AP_Changing_The_Game_March_2024.pdf: cover 'SPONSORED BY:' with Admazing and Venatus in the back matter; foreword 'This study, in collaboration with Advertiser Perceptions'; chart footers 'IAB/Advertiser Perceptions 2024, among 300 U.S. advertisers'. The source-register record reads 'IAB and Advertiser Perceptions; sponsored by Admazing and Venatus'. capability-matrix.json lists 'Admazing mobile gaming ads for brands' (cohort A) and 'Venatus gaming media' (cohort B2). | supported as stated | 02-market.md: 'IAB's March 2024 study with Advertiser Perceptions, sponsored by Admazing and Venatus (both scored in this paper), surveyed 300 US buyers ... and asked no barrier question'. | The PDF cover says 'SPONSORED BY:', p.38 is headed 'Our sponsors' and lists Admazing and Venatus, and p.2 says 'in collaboration with Advertiser Perceptions'. Both sponsors are rows in capability-matrix.json (network and adjacent cohorts), so rule 5 requires naming them. | chapters/02-market.md |
| IR1-G1-08 | the first reason listed is "not yet enough scale" (GAM-S-S01) claim GAM-S-F09 | Internal review round 1 (major, fact): 'Not yet enough scale' is used as US evidence of a reach constraint pointing the other way. The report's own quote for this group ties scale to the advertiser's own spend, not to audience or inventory size. The four reasons are also unnumbered boxes. The ledger calls this weaker counterevidence than presented. The fix below also carries the 51% correction from G1-01. | GAM-S-S01, p. 20, panel 'WHY ADVERTISERS CONSIDER GAMES PART OF OTHER CHANNELS Top 4 Reasons', with the quote 'We are not spending enough yet to take advantage of the scale that exists' attributed to Brand, Technology. claim-ledger GAM-S-F09 review_note: 'The report's own quote for this subgroup ties scale to the advertiser's own spend level ... an ambiguous inference'. GAM-S-F18 review_note: 'so it is weaker counterevidence than presented'. | supported with qualification | 02-market.md: 'Among the 51% ... the first of four unnumbered reasons listed is "not yet enough scale", but the report's illustrative quote for that group ties scale to the advertiser's own spend rather than to audience size, so it is weak evidence of a reach constraint'. | The rendered p.20 shows four unnumbered boxes, and the quote under the part-of-other-channels panel is 'We are not spending enough yet to take advantage of the scale that exists' (Brand, Technology). That matches the review_notes for GAM-S-F09 and GAM-S-F18. I called it the 'illustrative quote' because it is one respondent's quote, not the report's finding. | chapters/02-market.md |
| IR1-G1-09 | at $224 billion in 2024, most of it consumer spending | Internal review round 1 (major, fact): The PwC release does not say that most of the $224 billion is consumer spending. 'Most' can only be inferred by assuming that the 32.8% is advertising's share of that total, and the next clause of the same sentence says the base of the 32.8% is not stated. The phrase is an unlabelled inference (rule 9) and is stronger than the ledger's final wording. | GAM-A-S18, PR Newswire (24 July 2025): 'Total revenues were $224 billion in 2024, with the industry expected to grow to nearly $300 billion in 2029 at a CAGR of 5.7%' and 'advertising in video games (rising from 32.8% in 2024 to 38.5% in 2029)'. The release has no statement on the consumer share of video games revenue. claim-ledger GAM-A-F12 final_wording: '(whole-industry revenue including consumer spending)'. | supported as stated | 02-market.md: 'most of it consumer spending' changed to 'a whole-industry total that includes consumer spending'. | The PR Newswire release (GAM-A-S18) says only 'Total revenues were $224 billion in 2024' and gives 32.8% with no base. It makes no statement about the consumer share, and the ledger's final wording is 'whole-industry revenue including consumer spending'. | chapters/02-market.md |
| IR1-G1-10 | The IAB/PwC Internet Advertising Revenue Report for 2025 says US "video games and esports" advertising revenue grew 22.0% in 2025, after 8.7% in 2024, and publishes no dollar value | Internal review round 1 (major, rule): The text calls the figure 'US' and then says the chart does not label the geography. It also presents 22.0% as reported growth. The ledger's final wording says it is not stated whether the 2025 figure is an actual or a model estimate. When the report was published, PwC's latest Outlook still treated 2025 as a forecast year. Shared rule 9 requires every number to be labelled as actual or forecast. | GAM-A-S20, https://iabhongkong.com/sites/default/files/2026-04/IAB_PwC_Internet_Ad_Revenue_Report_Full_Year_2025_April_2026.pdf, p. 27: 'Video games and esports digital and non-digital advertising revenue grew 22.0% in 2025, accelerating from 8.7% in 2024'. claim-ledger GAM-A-F13 final_wording: 'the chart labels neither geography nor Outlook edition, and whether the 2025 value is a reported actual or a model estimate is not stated'. The same entry's label says: 'Code the 2025 growth rate as estimate, vintage unstated'. | supported with qualification | 02-market.md: 'The IAB/PwC ... Report for 2025, a report on US advertising revenue, says "video games and esports" advertising revenue grew 22.0% ... and the report does not say whether the 2025 figure is a reported actual or a model estimate'. | The IAB/PwC PDF (p.28) reads 'Video games and esports digital and non-digital advertising revenue grew 22.0% in 2025, accelerating from 8.7% in 2024' and labels no geography or vintage. The GAM-A-F13 label says to code 2025 as an estimate of unstated vintage. I kept the quoted segment name and made the report, rather than the chart, the subject of the not-stated clause. | chapters/02-market.md |
| IR1-G1-11 | The estimates table lists each figure with its publisher, vintage, metric, geography and what it includes, excludes or leaves unstated. | Internal review round 1 (major, consistency): The the estimates instrument table behind this sentence has two errors. First, it leaves out the EMARKETER figures reported by Digiday ($9.21bn for 2026, $9.7bn for 2027, $10.2bn for 2028), which the chapter text and fig-estimates both use, so the table does not list 'each figure'. Second, its IAB/PwC 2025 +22.0% row is coded claim_status 'actual', against the ledger label (see G1-10). | data/estimates.json has 34 rows and none with 9.21, 9.7 or 10.2. The IAB/PwC row reads {'value': '+22.0%', 'period': '2025', 'claim_status': 'actual'}. scripts/derivations.mjs, line 16, adds the Digiday points to fig-estimates. GAM-R-S322 (alias GAM-V-A-S01), Digiday, 3 July 2026: 'is forecast to reach $9.21 billion by the end of 2026, rising 5.3% to $9.7 billion by 2027, and another 4.7% to $10.2 billion in 2028, according to data from eMarketer'. claim-ledger GAM-A-F13 label: 'The data table rows for this line carry the same issue.' | contradicted | No change. | The finding is correct but its fix is outside my scope. data/estimates.json has 34 rows, none with 9.21, 9.7 or 10.2. The IAB/PwC +22.0% row has claim_status 'actual', and scripts/derivations.mjs line 16 adds the Digiday points only to the figure. The fix belongs in data/estimates.json, which I may not edit (edit scope is chapters 00-02 only). The chapter sentence is right once the data is fixed. | chapters/02-market.md |
| IR1-G1-12 | Format specifications, creative tooling and independent assurance are the thinnest dimensions. | Internal review round 1 (major, consistency): This contradicts data/capability-matrix.json. Coverage has the lowest mean score of any dimension in both in-game cohorts (1.50 in A and 1.50 in B1, with 3 n/e cells in A) and is left out. Independent assurance is not among the lowest means (2.00 overall; 2.25 in B2). It is thin because it has the most unscored cells (7 n/e). | Means computed from capability-matrix.json rows (numeric cells only). Overall: coverage 1.50 (10 scored, 3 n/e, 10 n/a), creative_insertion 1.77, format_execution 1.83, independent_assurance 2.00 (16 scored, 7 n/e). Cohort A: coverage 1.50, format 1.56, creative 1.89, assurance 2.00. Cohort B1: coverage 1.50, creative 1.50, format 1.75, assurance 1.75. | supported with qualification | 00-abstract.md finding 9: 'Across the matrix, coverage of devices and environments, format specifications and creative tooling have the lowest mean scores, and independent assurance has the most cells with no evidence to score.' | I recomputed from capability-matrix.json (0 unverified cells). Overall means: coverage 1.50 (3 n/e), creative 1.77, format 1.83, assurance 2.00 with 7 n/e, the most of any dimension. In cohort A the lowest three are coverage 1.50, format 1.56 and creative 1.89. I added 'Across the matrix' and a short gloss on coverage. Chapter 11's standfirst (line 4) still says 'format specifications, creative to | chapters/00-abstract.md |
| IR1-G1-14 | mostly on mobile and mostly bought by game developers | Internal review round 1 (major, fact): 'Mostly bought by game developers' is attached to the dollar series in parallel with 'mostly on mobile', which is a dollar share (90.4%). EMARKETER's statement concerns the share of ads served, not dollars, and it is not attributed here. The ledger says no source splits the dollars by buyer. Chapter 2's standfirst makes a similar unattributed statement ('most of those ads are bought by game developers promoting other games'). | claim-ledger GAM-A-F04 final_wording: 'the first two describe shares of ads, the third a share of spend by platform, and none gives a quantified dollar split by buyer'. GAM-A-S03 (EMARKETER, 10 June 2024): 'Most in-game ad spend goes to mobile, and most of the ads are bought by game developers.' | supported as stated | 00-abstract.md in-short: 'mostly on mobile; EMARKETER says game developers buy most of the ads served in games, and nobody publishes a figure for brand spend'. 02-market.md standfirst: 'and EMARKETER says most of those ads are bought by game developers promoting other games'. | GAM-A-F04 says the developer statements describe shares of ads, not dollars, and GAM-A-S03 reads 'most of the ads are bought by game developers'. The abstract attached this to the dollar series without attribution. I also added the missing attribution to chapter 2's standfirst. | chapters/00-abstract.md |
| IR1-G1-15 | Roblox's Plays-objective campaigns run on Home and search, outside experiences (GAM-D-S02) claim GAM-G-F24 | Internal review round 1 (major, fact): The cited source does not support the sentence. GAM-D-S02 is the Creator Hub page on immersive ads, which covers billboard, video and portal ad setup and never mentions the Plays objective or Home and Search placement. The passage the ledger relies on is Roblox's DevForum announcement of 4 February 2026, which is registered as GAM-V-G-S10. | GAM-D-S02, https://create.roblox.com/docs/production/monetization/immersive-ads, fetched 29 Sep 2026: contains no 'Plays', 'Search' or 'Home' placement text; its 'Home' hits are the Studio toolbar tab. GAM-V-G-S10, https://devforum.roblox.com/t/ads-manager-ui-targeting-enhancements/4333400 (created 2026-02-04): 'Right now, all campaigns focus on Plays' and 'Soon, you will also be able to choose whether your ad appears on Home, Search, or both.' claim-ledger GAM-G-F24 review_note cites this announcement. | supported as stated | 01-defining.md: (GAM-D-S02) changed to (GAM-V-G-S10) on the Plays-objective sentence. | The fetched GAM-D-S02 markdown has no Plays, Search or Home placement text (its only 'Home' hits are the Studio toolbar). The devforum post of 2026-02-04 (GAM-V-G-S10) reads 'all campaigns focus on Plays' and 'choose whether your ad appears on Home, Search, or both'. GAM-V-G-S10 is in the source register. | chapters/01-defining.md |
| IR1-g1-minor | Minor wording, label and voice issues in 02-market.md, 00-abstract.md, 01-defining.md | Internal review round 1, minor findings | See review record | supported with qualification | Applied or rejected per finding: G1-13 (applied); G1-16 (applied); G1-17 (applied); G1-18 (applied); G1-19 (applied); G1-20 (applied_modified); G1-21 (applied); G1-22 (applied); G1-23 (applied); G1-24 (applied); G1-25 (applied_modified); G1-26 (applied); G1-27 (applied_modified); G1-28 (applied); G1-29 (applied_modified) | Wording, labels, bounded absences and voice | chapters/02-market.md, chapters/00-abstract.md, chapters/01-defining.md |
| IR1-g2-01 | The figure matches common estimates of the world's players, and IION does not claim it as its own reach; the match is this review's inference | Internal review round 1 (major, fact): The chapter says IION does not claim 3.6 billion as its own reach and quotes only the 'Unique Audience' block. The hero of the same page says 'Home to a premium audience of 3.6 billion.' directly above 'Explore the immersiion Platform', which reads as IION claiming that audience. So the text is kinder to the vendor than the page is, and it drops the one line that makes the claim. | GAM-B-S35 https://www.iion.io/advertisers, read with curl on 29 Sep 2026: 'Reach the Unreachable Audience \| Home to a premium audience of 3.6 billion. \| Explore the immersiion Platform' and, lower down, 'Why Smart Brands Choose Gaming ... 3.6 BILLION \| Unique Audience \| Reach an audience unavailable anywhere else'. The ledger GAM-B-F08 review_note quotes the same hero line. | supported as stated | 03-networks.md audience paragraph: now quotes the hero line and the 'Unique Audience' block, and labels the reading of 3.6 billion as a player-population figure, not a measured reach of IION's supply, as this review's inference. | Confirmed by curl of iion.io/advertisers on 29 Sep 2026. The hero reads 'Home to a premium audience of 3.6 billion.' above the separate '3.6 BILLION / Unique Audience' counter, and the GAM-B-F08 passage quotes the same hero line. Saying IION 'does not claim it as its own reach' was kinder to the vendor than the page is. | chapters/03-networks.md |
| IR1-g2-02 | authorises Anzu and IION as direct sellers and names Anzu as manager domain | Internal review round 1 (major, consistency): EA's only IION DIRECT line uses seller ID 10329. IION's sellers.json assigns that ID to another publisher, Big Cake, and has no Electronic Arts entry. Chapter 3 treats this same ID as a mismatch, not a confirmed relationship. Presenting EA's IION line on a par with its Anzu lines, which do match, overstates EA's supply link to IION. | ea.com/app-ads.txt (GAM-R-S330) fetched 29 Sep 2026: line 2 'managerdomain=anzu.io,in-game'; 'anzu.io,5c5bcfb8fbabb940bd369e15,DIRECT' and 'anzu.io,62306a59e5886b16e80964eb,DIRECT'; the only IION DIRECT line is 'iion.io,10329,DIRECT,013a29748465dc57' (line 1576). Anzu sellers.json (GAM-B-S01) assigns those two IDs to 'Electronic Arts' and 'Playdemic' (domain ea.com). IION sellers.json (GAM-B-S03) has {"seller_id":"10329","name":"Big Cake","domain":"bigcakegame.com","seller_type":"PUBLISHER"} and no EA entry. Chapter 3 line 53 flags the same 10329 mismatch for Lihuhu. | supported as stated | 04-platforms.md EA paragraph: says the Anzu DIRECT IDs match EA and Playdemic, and that the one IION DIRECT line uses 10329, which IION assigns to Big Cake; cites GAM-B-S01 and GAM-B-S03 and adds claim GAM-R-F332. | Verified on ea.com/app-ads.txt (29 Sep): managerdomain=anzu.io, two anzu.io DIRECT lines, and one iion.io DIRECT line (10329). anzu.io/sellers.json assigns those two IDs to Electronic Arts and Playdemic (ea.com). iion.io/sellers.json assigns 10329 to Big Cake (bigcakegame.com) and has no EA entry. The old wording presented the IION line as equal to the matched Anzu lines. | chapters/04-platforms.md |
| IR1-g2-03 | IION's case study on Voodoo reports 70 to 80 million monthly impressions after a direct integration | Internal review round 1 (major, fact): The 70 to 80 million monthly impressions come from a publisher-monetisation integration across rewarded video, interstitial, banner and native formats. They are not intrinsic or brand placements inside the game. In a chapter about brand placements rendered inside games, the unqualified number reads as in-game brand supply. This breaks the rule on keeping delivery metrics apart and stating the unit. | GAM-R-S121 https://www.iion.io/case-studies/how-voodoo-achieved-7x-revenue-growth-through-direct-integration, read 29 Sep 2026: 'Migration from intermediary integration to direct integration across rewarded video, interstitial, banner, and native formats ... 7X revenue increase following direct integration \| 70-80 million monthly impressions delivered'. The ledger GAM-R-F116 review_note says: 'The case is a publisher-monetisation case ... not a brand-placement case'. | supported with qualification | 03-networks.md: the Voodoo sentence now names the four formats and calls the 70 to 80 million a delivery count for the publisher's monetisation, not for placements drawn into the game's scenes. | The case study (curl, 29 Sep) says 'direct integration across rewarded video, interstitial, banner, and native formats' and '70-80 million monthly impressions delivered'. The GAM-R-F116 note calls it a publisher-monetisation case. The fix's phrase 'standard in-app units' is avoided because COMBO! sells rewarded and interstitial units inside cohort A, so the edit names the unit as a monetisation de | chapters/03-networks.md |
| IR1-g2-04 | StackAdapt's own page and a Smadex partner post confirm, both for performance buying | Internal review round 1 (major, fact): StackAdapt's page confirms a partnership with Anzu for non-clickable blended ads measured by brand lift. It does not confirm performance buying or the click-enabled formats. The DSP list also appears only in the PRWeb copy of Anzu's release, and the ledger asks for it to be attributed that way. The sentence as written says 'Anzu names'. | GAM-B-S23 https://www.stackadapt.com/in-game-advertising, read 29 Sep 2026 (dateModified 2026-08-27): 'Blended in-game ads are non-clickable and non-disruptive ... StackAdapt is partnered with Anzu.io', 'Measure how exposure influences brand awareness with Brand Lift and Engagement Lift studies'. The GAM-ANZU-F06 review_note says: 'The DSP list appears only in the PRWeb wire copy ... Say "per the PRWeb release"'. | supported as stated | 03-networks.md: the sentence is now attributed to 'Anzu's PRWeb release'. StackAdapt's page is described as a partnership for non-clickable blended ads (modified 27 August 2026), and only the Smadex post as a performance partnership. | Re-fetched stackadapt.com/in-game-advertising (dateModified 2026-08-27). It says 'StackAdapt is partnered with Anzu.io' and describes the ads as 'non-clickable and non-disruptive'. It does not confirm performance buying. The GAM-ANZU-F06 note says the DSP list is only in the PRWeb copy. The Smadex post does describe a performance, clickable partnership. | chapters/03-networks.md |
| IR1-g2-05 | none operated by Anzu, Frameplay, Gadsme, AdInMo, Adverty, Gameloft, Super League or IION | Internal review round 1 (major, rule): The absence claim covers every cohort A network, but the evidence behind it screened only some of them. GAM-B-F31's searches covered the six networks with sellers.json files. The registry sentence names eight operators, including AdInMo, which is no longer in cohort A, and leaves out Kidoz and Admazing, which are. The claim is therefore not bounded to the vendors actually screened. The same wording appears in the inshort. | GAM-B-F31 review_note: searches for 'AdCP, Ad Context Protocol and agentic terms with each network name' cover the six sellers.json networks. The product register lists 'Kidoz kid-safe ad network' and 'Admazing mobile gaming ads for brands' as cohort 'network'. GAM-I-S36 https://agenticadvertising.org/api/registry/agents, re-read 29 Sep 2026: 34 agents and zero string matches for 'kidoz', 'admazing', 'anzu', 'frameplay', 'gadsme', 'adverty', 'gameloft', 'super league' or 'iion'. | supported as stated | 03-networks.md protocol paragraph: screening date moved to 29 September 2026 with GAM-I-S36 added to the first sentence. The registry list now names Kidoz and Admazing, with AdInMo kept last. | Re-read the public AdCP registry (29 Sep): count 34, and zero string matches for kidoz, admazing, anzu, frameplay, gadsme, adverty, gameloft, super league, iion and adinmo. The GAM-B-F31 searches predate Kidoz and Admazing joining cohort A, so the absence claim is now also bounded by the registry source. | chapters/03-networks.md |
| IR1-g2-06 | COMBO! and Super League claim programmatic routes, and no source opened names a DSP or deal type for either | Internal review round 1 (major, fact): A source the review opened and cites in chapter 4 contradicts this absence claim. AdExchanger quotes Gameloft for brands' senior vice president naming DV360 and The Trade Desk as ways to buy the network. The remark is a pitch framing, not a documented integration, but 'no source opened names a DSP' is false as written. | GAM-R-S348 AdExchanger, 18 Sep 2026 https://www.adexchanger.com/gaming/gaming-wants-to-prove-its-just-like-other-media-channels-while-also-owning-how-its-different/, read 29 Sep 2026: 'it can be hard to decide whether to play the gaming card or say, "I'm just a video network like any other, and you can buy me through DV360 or The Trade Desk," said Jean-Baptiste Godinot, SVP of Gameloft for brands'. | supported with qualification | 03-networks.md programmatic paragraph: adds that at IAB PlayFronts (September 2026) Gameloft for brands' Jean-Baptiste Godinot weighed telling advertisers 'you can buy me through DV360 or The Trade Desk', which names DSPs without documenting an integration. No DSP documentation or deal type was found for either network; cites GAM-R-S348. | WebFetch of the AdExchanger article confirms Godinot's line about deciding whether to say 'you can buy me through DV360 or The Trade Desk'. That makes 'no source opened names a DSP' false. The article body calls him SVP but its photo caption and Gameloft's own post (GAM-R-S209) say EVP, so the edit gives no title. It also avoids 'framed the pitch', because the quote is posed as a choice he weighs. | chapters/03-networks.md |
| IR1-g2-07 | Football Manager 26 (PC and Mac), confirmed by its publisher | Internal review round 1 (major, rule): The column lists titles dated inside the scoring period (2 October 2025 to 2 October 2026). The publisher's confirmation is dated 29 September 2025, before the period, and it announces an integration into a game not yet released. The only in-period evidence of live placements is the Minted piece co-produced by Gadsme, which the row does not cite. 'Confirmed by its publisher' therefore overstates the in-period live status under the Q2 rule. | GAM-B-S29 register: 'page "Posted: 29.09.25" ... Announces integration into the then-upcoming FM26; it does not itself show live placements' (ledger GAM-B-F17 review_note; the page itself was not re-opened in this review). The GAM-B-F17 final_wording says in-period live placements 'rest on a commissioned source' (GAM-GADSME-S17, Minted, 11 Dec 2025, co-produced with Gadsme). The capability-matrix gadsme supply cell says: 'The live-status evidence for FM26 is vendor-involved (grade D, cap 3)'. | supported as stated | 03-networks.md table Gadsme cell: 'exclusive announced by its publisher on 29 September 2025, just before the period; in-period delivery rests on partner content co-produced by Gadsme (December 2025)'; adds GAM-GADSME-S17. | Per the GAM-B-S29 register and the GAM-B-F17 note, the publisher announcement is dated 29 Sep 2025, before the period, and does not show live placements. In-period delivery rests on the Minted piece co-produced by Gadsme (GAM-GADSME-S17, 11 Dec 2025). The capability-matrix gadsme supply cell grades this as vendor-involved evidence. | chapters/03-networks.md |
| IR1-g2-minor | Minor wording, label and voice issues in 03-networks.md, 04-platforms.md | Internal review round 1, minor findings | See review record | supported with qualification | Applied or rejected per finding: g2-08 (applied); g2-09 (applied); g2-10 (applied); g2-11 (applied_modified); g2-12 (applied); g2-13 (applied); g2-14 (applied); g2-15 (applied); g2-16 (applied); g2-17 (applied); g2-18 (applied_modified); g2-19 (applied_modified); g2-20 (applied); g2-21 (applied); g2-22 (applied); g2-23 (applied); g2-24 (applied); g2-25 (applied); g2-26 (applied_modified) | Wording, labels, bounded absences and voice | chapters/03-networks.md, chapters/04-platforms.md |
| IR1-g3-01 | no verifier, MRC or certifier record for Zynga Ads was found, though the Trustworthy Accountability Group (TAG) registry was not checked for it | Internal review round 1 (blocking, fact): False on both counts. The TAG registry was checked for Zynga (the paper's own register records the query), and it returns a Zynga record: enrolled in Verified by TAG, with no seal. So the text wrongly says no certifier record exists and wrongly says the registry was not queried. | TAG registry endpoint https://hubgloo.opxar.com/search?q=Zynga, read 29 Sep 2026: "Zynga is currently enrolled in the Verified by TAG program." (no seal listed; same pattern as the Gadsme control). source-register GAM-V-R2-S12 is titled "TAG registry endpoint re-queries (Gameloft SE, TAG-ID 9df283b2158909d5; controls King, Zynga, Gadsme)". claim-ledger GAM-R-F213 review_note says "the same endpoint returns records for King, Zynga and Gadsme ('enrolled in the Verified by TAG program')". | supported as stated | 06: Zynga sentence now reads 'no verifier or MRC record ... and on 29 September 2026 the TAG registry listed Zynga as enrolled in Verified by TAG with no seal', citing GAM-C-S24 and GAM-V-R2-S12 as well. | I queried hubgloo.opxar.com/search?q=Zynga live and it returned 'Zynga is currently enrolled in the Verified by TAG program.' with no seal. King and DoubleVerify return seals on the same endpoint. GAM-V-R2-S12 and the GAM-R-F213 note record the query, so the old text was wrong on both counts. | chapters/06-measurement.md |
| IR1-g3-02 | Its "74M+ total US gaming reach" has no metric, measurement provider or period | Internal review round 1 (major, fact): Contradicted by the cited page. On that page the same 74M+ figure is headed "Monthly Active Gamers", so a metric is stated. What is missing is a measurement provider and a dated period. The ledger's own review_note quotes the heading. | GAM-F-S42 https://www.playwire.com/game-advertising, read 29 Sep 2026: "Reach 74M+ Monthly Active Gamers Our gaming ecosystem of best-in-class publishers delivers: 74M+ total US gaming reach". claim-ledger GAM-F-F30 review_note: "Game advertising page: 'Reach 74M+ Monthly Active Gamers ... 74M+ total US gaming reach'." | supported as stated | 05: 'Its 74M+ figure, labelled both "Monthly Active Gamers" and "total US gaming reach", has no measurement provider or dated period'. | The live playwire.com/game-advertising page reads 'Reach 74M+ Monthly Active Gamers ... 74M+ total US gaming reach'. A metric label is stated. No provider and no dated period are given. | chapters/05-adjacent.md |
| IR1-g3-03 | "total US gaming reach"; no metric or period | Internal review round 1 (major, consistency): The table row repeats the 'no metric' error: the source labels this figure 'Monthly Active Gamers'. It is also inconsistent with the Overwolf row, which records the same label ('monthly active gamers') as a metric and marks only 'no period'. | GAM-F-S42 (read 29 Sep 2026): "Reach 74M+ Monthly Active Gamers ... 74M+ total US gaming reach". Same table, Overwolf row: "monthly active gamers, both shown; no period or deduplication". | supported as stated | 05 table, Playwire row: '"monthly active gamers", also called "total US gaming reach"; no provider or period'. | Same source as g3-02. The fix also makes the row consistent with the Overwolf row. | chapters/05-adjacent.md |
| IR1-g3-04 | The only programmatic route in the documents opened is Amazon DSP | Internal review round 1 (major, fact): This 'only' claim is narrower than the ledger's final_wording for the same claim marker (GAM-TWITCH-F07). The ledger also records Amazon's streaming TV ads, which include Twitch and which US brands can buy through sponsored ads with no minimum spend, outside the DSP. The chapter drops that route and states exclusivity. | GAM-TWITCH-F07 final_wording: "...and through Amazon's Streaming TV ads, which include Twitch and carry no minimum budget when bought through sponsored ads in the Ads Console (US businesses)...". GAM-TWITCH-S16 https://advertising.amazon.com/solutions/products/streaming-tv-ads, read 29 Sep 2026: ads "can appear on Prime Video, Twitch, live sports..."; "available to ... brands in the U.S., whether or not they sell on the Amazon store, with no minimum spend requirement". | supported with qualification | 05: 'Amazon DSP is the only demand-side platform named in the documents opened: ... (GAM-TWITCH-S11, GAM-TWITCH-S12). Amazon's streaming TV ads, which Amazon says can appear on Twitch, are a further route open to US brands (GAM-TWITCH-S16) claim GAM-TWITCH-F07.' | The exclusivity claim was narrower than the GAM-TWITCH-F07 final wording. The live GAM-TWITCH-S16 page says the ads 'can appear on Prime Video, Twitch' and are available to 'brands in the U.S.'. It no longer describes a separate 'sponsored ads' route: it says Sponsored TV is now 'Streaming TV' within Display, Video, and Audio campaigns in Amazon Ads Agent. It also says 'no minimum spend requiremen | chapters/05-adjacent.md |
| IR1-g3-05 | Oracle ended all its advertising products, Moat Analytics and Moat Reach included, on 30 September 2024 | Internal review round 1 (major, consistency): States as a completed retirement something the cited source only announced. The Oracle FAQ of 12 August 2024 uses the future tense. Chapter 10 codes the same event as 'Retirement (announced)' and says 'this review opened no record of the retirement itself'. The two chapters contradict each other, and the sentence breaks rule 8 (announcement is not retirement). | GAM-C-S28 https://www.oracle.com/a/ocom/docs/oracle-advertising-end-of-life-faqs.pdf (read 29 Sep 2026): "August 12, 2024 Version 3.0 ... Oracle will End-Of-Life (EOL) all Oracle Advertising products and support effective September 30, 2024." chapters/10-structure.md line 51: "states that Oracle would end all its advertising products ... on 30 September 2024, and this review opened no record of the retirement itself"; ledger GAM-U-F16. | supported as stated | 06: 'Oracle's end-of-life FAQ of 12 August 2024 set 30 September 2024 as the end date for all its advertising products, Moat Analytics and Moat Reach included (GAM-C-S28).' | The GAM-C-S28 FAQ (v3.0, 12 Aug 2024) uses the future tense ('Oracle will End-Of-Life ...'). GAM-U-F16 and chapter 10 treat this as an announced retirement, so the completed-tense sentence contradicted them. | chapters/06-measurement.md |
| IR1-g3-06 | although Oracle ended Moat on 30 September 2024 (GAM-F-S26, GAM-F-S27) | Internal review round 1 (major, consistency): This has the same announced-versus-done problem as g3-05, which contradicts chapter 10. The citation is also wrong: GAM-F-S27 is Overwolf's Standard Media page, not Oracle's end-of-life FAQ (GAM-C-S28), so nothing cited supports the Moat date. | source-register GAM-F-S27 = "Standard Media - High-Impact Display & Video Ads (Overwolf)". GAM-C-S28 Oracle FAQ v3.0 (12 Aug 2024): "Oracle will End-Of-Life (EOL) all Oracle Advertising products and support effective September 30, 2024"; Moat Analytics and Moat Reach are listed. Ledger GAM-F-F19 note cites Oracle's FAQ for the date. | supported as stated | 05: 'although Oracle's end-of-life FAQ set 30 September 2024 as Moat's end date (GAM-F-S26, GAM-C-S28)'. | GAM-F-S27 is Overwolf's Standard Media page, not Oracle's document, and the tense had the same problem as g3-05. | chapters/05-adjacent.md |
| IR1-g3-07 | and 3 are non-entertainment consumer brands (Mountain Dew, Wendy's, Walmart) | Internal review round 1 (major, fact): This leaves out a qualification that the ledger review_note requires. Two of the three consumer-brand Quests were tied to entertainment properties. Without that, the sentence overstates how far Quests reach beyond games and entertainment, and it is the evidence the paragraph uses for 'Quests mix objectives'. | GAM-F-S01 https://discord.com/ads/quests, read 29 Sep 2026 (21 success-story links counted): "Wendy's serves spooky-season fandom with Netflix's Wednesday"; "World of Warcraft x Mountain Dew: Crown of the Violet Rose ... Mountain Dew pioneered the first Discord Quest by a consumer food & beverage brand, partnering with World of Warcraft". Ledger GAM-R-F402 review_note: "add that the Wendy's and Mountain Dew Quests were tied to entertainment IP (Netflix's Wednesday; World of Warcraft)". | supported as stated | 05: added ', two of them tied to entertainment properties (World of Warcraft; Netflix's Wednesday)' after the three consumer brands. | The GAM-R-F402 review_note requires this qualification. The live discord.com/ads/quests page confirms 'Wendy's serves spooky-season fandom with Netflix's Wednesday' and 'World of Warcraft x Mountain Dew: Crown of the Violet Rose'. | chapters/05-adjacent.md |
| IR1-g3-08 | streamers must apply Content Classification Labels (six at launch, seven in Twitch's current help article) | Internal review round 1 (major, rule): The citation does not support the sentence. It cites only the 2023 Twitch blog (GAM-D-S27), which says the labels 'should be used', a weaker claim than 'must'. No cited source supports the count of seven. The ledger says 'must' rests on the Amazon guide (GAM-D-S28). The count of seven is in Twitch's API reference (GAM-NFA-S02), which is in the register but not cited. | Ledger GAM-TWITCH-F11 review_note: "The blog says 'should be used'; the 2026 Amazon guide calls the labels 'mandatory'. The finding's 'must' rests on the guide." GAM-D-S28 (read 29 Sep 2026): "Content Classification Labels (CCLs), and they are mandatory". GAM-NFA-S02 https://dev.twitch.tv/docs/api/reference/ (read 29 Sep 2026) lists seven label ids: DebatedSocialIssuesAndPolitics, DrugsIntoxication, Gambling, MatureGame, ProfanityVulgarity, SexualThemes, ViolentGraphic. | supported with qualification | 05: '(six at launch, seven in Twitch's current help article and API reference)', with the citation changed to (GAM-D-S27, GAM-D-S28, GAM-NFA-S01, GAM-NFA-S02). | The blog cited (D-S27) says 'should be used'. The 'mandatory' wording is in GAM-D-S28, which I confirmed live. The count of seven is backed by the help article GAM-NFA-S01, which is in the register but was not cited; it would not render for me. It is also backed by the API reference GAM-NFA-S02, where I confirmed the seven label ids live. I kept both, matching the ledger final wording ('current he | chapters/05-adjacent.md |
| IR1-g3-minor | Minor wording, label and voice issues in 06-measurement.md, 05-adjacent.md | Internal review round 1, minor findings | See review record | supported with qualification | Applied or rejected per finding: g3-09 (applied); g3-10 (applied_modified); g3-11 (applied); g3-12 (applied); g3-13 (applied_modified); g3-14 (applied); g3-15 (applied); g3-16 (applied); g3-17 (applied); g3-18 (applied); g3-19 (applied); g3-20 (applied); g3-21 (applied); g3-22 (applied); g3-23 (applied); g3-24 (applied); g3-25 (applied); g3-26 (applied_modified) | Wording, labels, bounded absences and voice | chapters/06-measurement.md, chapters/05-adjacent.md |
| IR1-g4-01 | Among in-game ad networks (cohort A), Anzu, Gadsme, Adverty and AdInMo document controls in operational detail, AdInMo only in an archived copy of 7 October 2025 (GAM-D-S15, GAM-D-S16, GAM-D-S18, GAM-R-S110). | Internal review round 1 (blocking, consistency): Puts AdInMo in cohort A. The product register moved AdInMo out of cohort A into the historical rows on 29 September 2026, and chapter 3 lists the cohort without it. The sentence also presents a partial list as the whole cohort: Kidoz and Admazing are cohort A rows in the matrix and are not mentioned. A reader would take AdInMo for a live cohort A network documenting controls today. | data/product-register.json, AdInMo in-game ad network: cohort "historical", "Moved from cohort A to historical on 29 September 2026 ... the director signed a strike-off application on 6 September 2026 ... adinmo.com has returned HTTP 525 since 11 July 2026". chapters/03-networks.md line 8: "The cohort A networks are Anzu, Frameplay, IION, Gadsme, Adverty, Gameloft for brands (COMBO!), Super League, Kidoz and Admazing. AdInMo left the cohort". data/capability-matrix.json cohort network rows include Kidoz kid-safe ad network (suitability_age_data v=3) and Admazing (v=2). | supported with qualification | 07 para 1: cohort A with operational detail is now Anzu, Gadsme, Adverty and Kidoz [+GAM-R-S304;GAM-R-S302, claim GAM-R-F315]; Super League and Admazing are marketing only [+GAM-R-S234, claim GAM-R-F231]; AdInMo gets its own sentence as moved out of the cohort on 29 Sep 2026, archived copy of 7 Oct 2025 [GAM-R-S110]. | Confirmed. product-register.json lists AdInMo as cohort "historical" since 29 September 2026, and chapter 3 lists Kidoz and Admazing as cohort A. The suggested line "Chapter 11 codes the other two" was not accurate, because chapter 11's prose does not code their suitability. Kidoz and Admazing were therefore coded directly from ledger GAM-R-F315 and the matrix: Kidoz brand suitability in operation | chapters/07-safety.md |
| IR1-g4-02 | and iion and Super League show marketing language only in the pages opened (GAM-D-S20, GAM-R-S202, GAM-B-S20, GAM-FRAMEPLAY-S05, GAM-D-S23, GAM-D-S21) | Internal review round 1 (major, consistency): Contradicts the matrix and the source register. The review opened iion's Inventory Partner General Terms (in the register), and the matrix codes iion's suitability subcriterion in operational detail from them. So "marketing language only in the pages opened" is false for iion. | data/capability-matrix.json, IION in-game brand network, suitability_age_data v=2: "Verifier (supported with qualification): Missed source: Inventory Partner General Terms ... Suitability in operational detail: network-wide Content Guidelines state which property content is excluded, how it is enforced ... and the scope (all Enabled Properties)". Register GAM-V-R1-S03, https://www.iion.io/inventory-partner-general-terms-conditions (accessed 2026-09-29). Re-read today: "Inventory partners must adhere to these guidelines and will not deliver any Advertisement to any Enabled Property that contain | supported with qualification | 07 para 1: iion moved out of the marketing-only group into a new sentence: only its inventory-partner terms document an operational control, which bars delivery to properties with excluded content, including content directed to under-13s (GAM-D-S23, GAM-V-R1-S03). | Confirmed by re-reading GAM-V-R1-S03 today. It says partners "will not deliver any Advertisement to any Enabled Property that contains ... Any content directed to children under the age of 13" and warns of "iion blocking ads ... or suspending or terminating your account". The matrix verifier codes this as suitability in operational detail. iion now has its own sentence, so the rule it sets is stat | chapters/07-safety.md |
| IR1-g4-03 | Among in-game ad networks screened, Anzu, Gadsme, Adverty and AdInMo (AdInMo only in an archived copy) document suitability controls in operational detail, and iion and Super League only in marketing language in the pages opened (GAM-D-S15, GAM-D-S16, GAM-D-S18, GAM-R-S110, GAM-D-S23, GAM-D-S21). | Internal review round 1 (major, consistency): The In short summary repeats both errors. It treats AdInMo as a current network (it is a historical row) and puts iion at marketing-only, although the matrix codes its suitability in operational detail from its partner terms. | data/product-register.json AdInMo cohort "historical" (moved 29 September 2026). data/capability-matrix.json IION suitability_age_data verifier note: "Suitability in operational detail: network-wide Content Guidelines" (GAM-V-R1-S03). | supported with qualification | 07 In short: "Anzu, Gadsme, Adverty and Kidoz ... in operational detail, iion only in its inventory-partner terms, and Super League and Admazing only in marketing language ...; AdInMo, since moved out of the cohort, did so in a copy archived in October 2025", with citations updated to match. | Same errors as g4-01 and g4-02, so the In short summary was aligned with the corrected paragraph. Kidoz and Admazing were added so the cohort list is not partial. | chapters/07-safety.md |
| IR1-g4-04 | Among platform sellers (cohort B1), Roblox documents controls in operational detail and XBOX Media Solutions in marketing language only (GAM-D-S04, GAM-D-S22). | Internal review round 1 (major, consistency): The sentence contradicts the matrix on both B1 rows. It also contradicts this chapter's own later paragraph, which says Roblox names its advertiser-side controls "without operating detail". The matrix codes Roblox brand suitability in outline for the scoring period, with advertiser tiers documented operationally only in the January 2024 post. It codes XBOX brand suitability in operational detail, from the Activision Blizzard Media page archived on 21 January 2026. Neither row is coded as the sen | data/capability-matrix.json, Roblox brand ads, suitability_age_data v=4: "two subcriteria (age controls, data practices) documented in operational detail and the third (brand suitability) in outline for the scoring period ... advertiser-selectable tiers are documented operationally only in the 2024 post". XBOX Media Solutions in-game ads, v=4: "two subcriteria in operational detail (brand suitability, data practices) ... Brand suitability is in operational detail only on the archived 21 January 2026 capture" (GAM-XBOX-S17). Ledger GAM-ROBLOX-F20 final_wording: "Roblox's advertiser-side suitabi | supported with qualification | 07 para 1 B1 sentence: Roblox's ad-content and publisher rules are operational but its advertiser tiers are defined only in a January 2024 post; XBOX was operational only on a page archived 21 Jan 2026 that is no longer served, and its live pages use marketing language (GAM-D-S01, GAM-D-S02, GAM-D-S04, GAM-XBOX-S17, GAM-D-S22) claim GAM-XBOX-F06. | Confirmed. The matrix codes Roblox brand suitability in outline for the scoring period, with advertiser tiers defined only in the January 2024 post (GAM-ROBLOX-F20). XBOX brand suitability is operational only on the archived 21 January 2026 capture (GAM-XBOX-F06), and the live URL now redirects. "Screened" was added because EA and Zynga, the other B1 rows, were not screened for this subcriterion. | chapters/07-safety.md |
| IR1-g4-05 | In the sources screened by 29 September 2026, this review found no DoubleVerify record for any cohort A network (GAM-C-S12, GAM-DOUBLEVERIFY-S30). | Internal review round 1 (major, fact): The absence claim is contradicted by a source this chapter cites. Gadsme's brand-safety policy (GAM-D-S16, cited in paragraph 2) names DoubleVerify tag wrapping, and Gadsme's advertisers page names DoubleVerify among its verification partners. Only DoubleVerify's own site shows no record of a cohort A network. As written, readers will take the sentence to mean that no cohort A network has any documented DoubleVerify link. | GAM-D-S16, https://www.gadsme.com/brand-safety (re-read 29 Sep 2026, 'Updated February 2026'): "An advertiser can wrap their tags with IAS or Double Verify to ensure greater brand safety should they choose." https://www.gadsme.com/advertisers: "Collaboration with leading verification and industry organisations including IAS, DoubleVerify, IAB and TAG ensures transparency, measurement accuracy and brand safety across campaigns". | supported with qualification | 07 verifier para: "In DoubleVerify's own records screened on 29 September 2026, including its sitemap, this review found none naming a cohort A network; Gadsme's policy says only that advertisers can wrap their tags with IAS or DoubleVerify (GAM-C-S12, GAM-DOUBLEVERIFY-S30, GAM-D-S16)." | Confirmed. Gadsme's brand-safety page (Updated February 2026) says "An advertiser can wrap their tags with IAS or Double Verify". The absence claim is now bounded to DoubleVerify's own records. I also re-checked DV's sitemap today: 462 URLs, and no Anzu, Frameplay, Gadsme, Adverty, iion, Gameloft, Super League, Kidoz or Admazing string, including in image captions. "Their tags" is used because the | chapters/07-safety.md |
| IR1-g4-06 | and the conflict between Roblox's revenue-share dates concerns developer fees, so it does not block the verdict (GAM-E-S01, GAM-E-S10, GAM-E-S11) | Internal review round 1 (major, rule): The registered H4 rule names this exact conflict as its example of an unresolved condition. The text sets it aside without saying that the registered rule named it. A reader who compares the text with hypotheses.json sees an undisclosed departure from the pre-registered coding. The departure may be defensible, but it has to be stated as the author's reading. | data/hypotheses.json H4 rule_unresolved: "Rule documents conflict (for example Roblox's two revenue-share dates) or cannot be opened, or the product's function cannot be read from its documentation." chapters/13-method.md does not mention the revenue-share conflict. | supported with qualification | 07 H4 para: "...and the registered rule names the conflict between Roblox's two revenue-share documents as an example of an unresolved condition [...]. On this review's reading, that conflict concerns developer fees rather than any restriction counted here, so it does not block the verdict (GAM-E-S10, GAM-E-S11)." | Confirmed. The hypotheses.json H4 rule_unresolved names "Roblox's two revenue-share dates" as its example. The ledger's GAM-E-F36 final_wording already discloses this. The text now states the registered example first and then the review's reading, in two sentences. | chapters/07-safety.md |
| IR1-g4-07 | Per Amazon's guide of 22 February 2026, the labels inform three inventory tiers and impression-level content reporting in Amazon's demand-side platform (DSP), but the guide does not say what each tier excludes (GAM-D-S28, GAM-D-S29). | Internal review round 1 (major, fact): The live guide was modified on 29 September 2026, before the freeze, and now places content-level reporting in Amazon Ads Agent, not Amazon DSP. The sentence no longer matches the cited source. Chapter 5 should be checked for the same point. | GAM-D-S28, https://advertising.amazon.com/library/guides/twitch-brand-safety, fetched 29 Sep 2026, page metadata "datePublished":"2026-02-22", "dateModified":"2026-09-29": "content-level reporting, which provides impression-level detail of the content your ads appeared in. The report is available on demand in the Amazon Ads Agent". Ledger GAM-D-F19 review_note (earlier read): "content-level reporting 'available on demand in the Amazon DSP'". | supported with qualification | 07 Twitch para: content reporting "which the guide, as modified on 29 September 2026, places in Amazon Ads Agent; the guide does not say what each tier excludes"; the verifier para now defines "demand-side platforms (DSPs)". | Confirmed. The live guide's metadata reads datePublished 2026-02-22 and dateModified 2026-09-29, and the text says the report "is available on demand in the Amazon Ads Agent". The review's earlier reading is not described, because that is process detail. Removing "(DSP)" here left the later "DSPs" in the chapter undefined, so the definition moved to that sentence. Chapter 5 was checked and does no | chapters/07-safety.md |
| IR1-g4-08 | the federal cases are centralised as a multidistrict litigation (MDL No. 3166) (GAM-D-S07). | Internal review round 1 (major, fact): The only source cited, AdExchanger, calls the cases a "federal class action" and gives no MDL number, so the citation contradicts the sentence it supports. The MDL fact rests on the JPML transfer order, which is in the register and not cited. | GAM-D-S07, https://www.adexchanger.com/gaming/roblox-opens-up-advertising-to-kids-under-13/ (re-read 29 Sep 2026): "Roblox currently faces at least 130 child safety lawsuits that have been combined into a federal class action in the Northern District of California, in addition to other suits brought by state governments". Register GAM-V-D-S17: "Transfer Order, In re: Roblox Corporation Child Sexual Exploitation and Assault Litigation, MDL No. 3166". | supported with qualification | 07 litigation para: citation changed to (GAM-D-S07, GAM-V-D-S17). | Confirmed. AdExchanger says "combined into a federal class action" and gives no MDL number. The JPML transfer order, re-read, is captioned "MDL No. 3166" and centralises the cases in the Northern District of California. The citation was added without repeating AdExchanger's inaccurate label, as the ledger GAM-D-F26 advises ("not a class action"). | chapters/07-safety.md |
| IR1-g4-09 | Apple's Declared Age Range API returns the Texas law's age brackets for new Texas accounts from 4 June 2026, and Apple's developer agreement, updated 18 August 2026, allows requests only for age-appropriate content or features or to comply with laws that require age-range information (GAM-E-S18, GAM-E-S19, GAM-E-S17, GAM-E-S20). | Internal review round 1 (major, fact): None of the cited sources supports the 4 June 2026 date or the 18 August 2026 agreement. GAM-E-S19 (23 December 2025) says Apple paused its Texas plans, which contradicts the sentence for a reader who opens it. The supporting records, Apple's 3 June 2026 post and the DPLA, are in the register but not cited. | Apple Developer news RSS (https://developer.apple.com/news/rss/news.rss, read 29 Sep 2026): 3 June 2026 'Update for Apps Distributed in Texas': "These changes will go into effect starting June 4, 2026. Developers can request age category data for these Apple Accounts through the Declared Age Range API" (register GAM-V-E-S02). 23 Dec 2025 (GAM-E-S19): "Apple will pause previously announced implementation plans". Register GAM-V-E-S03: Apple Developer Program License Agreement § 3.3.3(O), pub 2026-08-18. | supported as stated | 07 Apple sentence: citations changed to (GAM-E-S18, GAM-V-E-S02, GAM-E-S20, GAM-V-E-S03). | Confirmed. Apple's news RSS entry of 3 June 2026 (GAM-V-E-S02) says "These changes will go into effect starting June 4, 2026. Developers can request age category data ... through the Declared Age Range API". The GAM-E-S19 post of 23 Dec 2025 says Apple "will pause" its plans. DPLA §3.3.3(O), with Schedule 1 last updated August 18, 2026, limits requests to "age-appropriate content and/or features . | chapters/07-safety.md |
| IR1-g4-19 | EA's release names brand integrations in three console titles (GAM-G-S25) claim GAM-G-F22. | Internal review round 1 (major, fact): The release names only two titles, EA SPORTS FC 26 and EA SPORTS College Football 26. Madden appears only as the franchise, "Madden NFL". The Madden NFL 26 edition comes from the fan database MUT.GG, not from EA. So EA does not document three titles, which matters because H3's registered rule counts titles documented by the network, platform or publisher. | GAM-G-S25, https://news.ea.com/press-releases/press-releases-details/2026/Electronic-Arts-Introduces-EA-Advertising-Launching-Brands-Directly-Into-Gameplay-and-Live-Experiences/ (re-read 29 Sep 2026): "Xfinity and Peacock ... in EA SPORTS FC 26"; "Mountain Dew's 'DEW University,' a fully playable team experience in EA SPORTS College Football 26"; "Lowe's, which integrated into EA SPORTS FC, Madden NFL, and College Football through Ultimate Team challenges". Ledger GAM-G-F22 review_note: "The Madden edition can be named as Madden NFL 26, though the dates rest on a fan database, not EA." | supported with qualification | 08: "EA's release names brand integrations in two console titles, EA SPORTS FC 26 and EA SPORTS College Football 26, and in its Madden NFL franchise"; the Madden sentence now reads "MUT.GG, a fan database, places a Lowe's Ultimate Team program in Madden NFL 26 ...; EA's release names only the franchise". | Confirmed. The EA release names "EA SPORTS FC 26" and "EA SPORTS College Football 26", but only "Madden NFL" for Madden. The edition is attributed to MUT.GG in the following sentence rather than repeated here. Re-checked MUT.GG today: "Madden 26 Ultimate Team Lowe's ... First Release Sep 27th, 2025 Latest Release Oct 18th, 2025". | chapters/08-console.md |
| IR1-g4-20 | EA names brand integrations in three console titles and documents none as an ad placement buyable in the US (GAM-G-S25). | Internal review round 1 (major, fact): Same error in the In short summary: EA's release names two titles and the Madden NFL franchise, not three titles. | GAM-G-S25 passages as in g4-19: titles named are "EA SPORTS FC 26" and "EA SPORTS College Football 26"; Madden appears only as "Madden NFL". | supported with qualification | 08 In short: "EA names brand integrations in two console titles and in its Madden NFL franchise, and documents none as an ad placement buyable in the US". | Same error as g4-19. The fan-database detail is left in the body so the In short summary stays short. | chapters/08-console.md |
| IR1-g4-21 | The unit's general manager described its console approach in January 2026 as "game-adjacent rather than in-game" (GAM-G-S07, GAM-B-S26, GAM-V-XBOX-S03) claim GAM-XBOX-F15. | Internal review round 1 (major, fact): The quotation is misattributed. In The Drum piece, "game-adjacent rather than in-game" is the writer's narration, not words in quotation marks from the general manager, Greg Carroll, and it describes Xbox's approach generally, not console. The piece is sponsored content produced by The Drum Studios, which the text does not disclose. The register records funding as "sponsorship not stated on the page", which is wrong. Two of the three citations (GAM-G-S07, TechCrunch 2022; GAM-B-S26, Anzu page) d | GAM-V-XBOX-S03, https://www.thedrum.com/news/when-everybody-s-a-player-it-makes-no-sense-for-brands-to-still-be-on-the-sidelines (read 29 Sep 2026): "Sponsored by: What's this? Sponsored content is created for and in partnership with an advertiser and produced by the Drum Studios team"; "according to Greg Carroll, general manager of Xbox Media Solutions"; narration: "That's where the idea of game-adjacent rather than in-game comes in. 'You can support, you can reward, you can unlock experiences,' he says." Register GAM-V-XBOX-S03 funding: "unknown (interview format; sponsorship not stated on t | supported with qualification | 08: "A sponsored article produced by The Drum Studios in January 2026, built on an interview with the unit's general manager, Greg Carroll, describes Xbox's approach through "the idea of game-adjacent rather than in-game" (GAM-V-XBOX-S03) claim GAM-XBOX-F15." | Confirmed. The Drum page is labelled "Sponsored content ... produced by the Drum Studios team". The phrase is narration ("That's where the idea of game-adjacent rather than in-game comes in") about Xbox's approach generally, not a quote from Carroll about console. GAM-G-S07 and GAM-B-S26 do not support this sentence, and both are cited elsewhere in the chapter. Out of scope: chapter 4 line 51 repe | chapters/08-console.md |
| IR1-g4-22 | Three is a floor, because this review did not screen other publishers' console titles and networks may run console campaigns they do not publicise (GAM-G-S25). | Internal review round 1 (major, consistency): The paragraph says "at most three" titles qualify, that EA documents none as buyable in the US, and that College Football 26 is only inferred inside the window. It cannot then call three a floor. What is a floor is the market count, not three. As written, the sentence asserts that at least three qualifying titles exist, which the evidence does not show. Chapters 4 (line 59) and 12 repeat the construction. | Same paragraph: "At most three EA console titles qualify"; "EA documents none of the three as an ad placement buyable in the US"; "College Football 26's integrations went live in July 2025, before the twelve-month window opened, so their presence inside it is inferred". data/hypotheses.json H3 rule_supported: titles must be "documented, by the network, the platform or the game's publisher, as carrying live in-game brand placements buyable in the US". | supported with qualification | 08: "Any count from this screen is a floor for the market, because this review did not screen other publishers' console titles and networks may run console campaigns they do not publicise". | Confirmed. "At most three" and "Three is a floor" contradict each other, because the floor applies to the market count, not to EA's three. Out of scope: chapters 4 and 12 repeat the construction. | chapters/08-console.md |
| IR1-g4-minor | Minor wording, label and voice issues in 07-safety.md, 08-console.md | Internal review round 1, minor findings | See review record | supported with qualification | Applied or rejected per finding: g4-10 (applied_modified); g4-11 (applied); g4-12 (applied); g4-13 (applied); g4-14 (applied); g4-15 (applied_modified); g4-16 (applied); g4-17 (applied); g4-18 (applied_modified); g4-23 (applied); g4-24 (applied); g4-25 (applied); g4-26 (applied_modified); g4-27 (applied_modified); g4-28 (applied); g4-29 (applied_modified); g4-30 (applied); g4-31 (applied) | Wording, labels, bounded absences and voice | chapters/07-safety.md, chapters/08-console.md |
| IR1-g5-01 | Bloxbiz, Admix and Bidstack are companies, and Misfits Ads was an in-game ad business line that an esports company sold, so the three companies alone meet the threshold | Internal review round 1 (blocking, rule): Part (a) of H5 is coded as met by counting Admix as a company that was sold or put into insolvency. The registered rule (a) lists only 'entered insolvency, liquidation or dissolution, retired their product, or been sold'. It does not list mergers. The Admix release shows Admix as the continuing party, not the one sold: Admix's founder and chief executive led the merged company, LandVault's founders left, and Admix says it chose to 'acquire resources to enter the metaverse in-house'. The product | GAM-U-S09, https://www.thedrum.com/profile/admix/article/metaverses-biggest-builder-merges-with-game-tech-firm-to-bring-brands-into-web3-1 (21 June 2022): 'that revenue stream quickly overtook the in-play revenue ... which prompted Admix to acquire resources to enter the metaverse in-house' and 'The merged company will be led by Sam Huber, founder, and CEO of Admix, with LandVault's founders moving on to a new venture.' data/hypotheses.json H5 rule_supported: '(a) at least three specialist in-game ad companies have entered insolvency, liquidation or dissolution, retired their product, or been | supported with qualification | 10-structure: the in-short and Part (a) now count Bloxbiz, Let's Bounce and Bidstack, with Admix recorded as a merger, rebrand and pivot and not counted. The Admix paragraph now says Admix's CEO would lead the merged company and LandVault's founders would leave. GAM-B-S45 added to the count citations. | Confirmed against GAM-U-S09. The release says 'The merged company will be led by Sam Huber, founder, and CEO of Admix, with LandVault's founders moving on', and describes Admix acquiring 'resources to enter the metaverse in-house'. Admix was not sold, and the H5 rule_supported wording does not list mergers. GAM-B-S45 (10-Q) confirms 'On January 5, 2026 ... the Company acquired Let's Bounce, Inc.', | chapters/10-structure.md |
| IR1-g5-02 | Two of these events, Bloxbiz and Misfits Ads, were purchases by one buyer, Super League, which reads as consolidation as much as departure; that reading is an inference | Internal review round 1 (major, consistency): The chapter leaves out Super League's January 2026 acquisition of Let's Bounce, Inc. It is missing from the event table, the Super League paragraph and the consolidation sentence. The ledger's final_wording for GAM-U-F35 lists it ('and of Let's Bounce, Inc. in January 2026'), and the product register holds it as a historical row. The event is a specialist sale shown by a primary filing, so it bears on the H5(a) count (see g5-01). With it, Super League made three of the purchases, not two. | GAM-B-S45, Super League Form 10-Q for Q2 2026, https://www.sec.gov/Archives/edgar/data/1621672/000143774926028066/slgg20260630_10q.htm: 'On January 5, 2026 ("Bounce Effective Date") the Company acquired Let's Bounce, Inc. ("Bounce"), a marketing technology company focused on enabling scalable, measurable brand engagement inside gaming and UGC environments. The total purchase price for Bounce, which was structured as an asset acquisition, was $200,000'. data/claim-ledger.json GAM-U-F35 final_wording: 'Further events are Super League's purchase of the Misfits Ads business line ... and of Let's B | supported as stated | 10-structure: added a Let's Bounce row to the table (5 January 2026, Form 10-Q) and a Let's Bounce sentence to the Super League paragraph. The consolidation sentence now names three Super League purchases. | Confirmed in GAM-B-S45: on 5 January 2026 Super League acquired Let's Bounce, Inc., 'a marketing technology company focused on enabling scalable, measurable brand engagement inside gaming and UGC environments', in an asset acquisition priced at $200,000. The GAM-U-F35 final_wording lists it too. | chapters/10-structure.md |
| IR1-g5-03 | The table at the end of this chapter holds stat studies study records; some record the same study from a second source. | Internal review round 1 (major, rule): The 36 rows include three studies recorded twice from the same source, not from a second one. Rows 31 to 33 ('Happydemics in-game benchmark', 'Happydemics meta-analysis', 'Comscore brand lift') cite GAM-FRAMEPLAY-S20/S21/S22, which source-aliases.json maps to GAM-H-S08/S09/S10, the sources of rows 5 to 7. That breaches shared rule 13 ('One study, one record') and inflates the published count, which is 33 unique studies. The same chapter's figure fig-studies counts 31 (figure-data.json studies_to | data/study-evidence.json rows 31-33 src GAM-FRAMEPLAY-S20, -S21, -S22. data/source-aliases.json: 'GAM-FRAMEPLAY-S20' -> 'GAM-H-S08', 'GAM-FRAMEPLAY-S21' -> 'GAM-H-S09', 'GAM-FRAMEPLAY-S22' -> 'GAM-H-S10'. The rendered table cites [450], [554] and [522] twice each. data/figure-data.json studies_total = 31. data/claim-ledger.json GAM-H-F01 review_note: 'The Ipsos GameTrack analysis is the largest-sample field comparison found ... It should enter the study_evidence table.' GAM-R-S328 https://www.ipsos.com/en-uk/game-advertising-significantly-improves-brand-affinity: 'Across nine campaigns analyse | supported with qualification | 09-outcomes: the table sentence now says the table holds the outcome review's rows plus five rows from Frameplay's vendor profile, three of which repeat Frameplay studies from the same releases, and that Ipsos GameTrack is not in the table. | Confirmed. study-evidence.json rows 32-34 (stream 'frameplay'; GAM-FRAMEPLAY-S20/S21/S22) alias to GAM-H-S08/S09/S10, the sources of rows 6-8. They are the same releases, not a second source. The figure counts only the outcome review's 31 rows. Ipsos GameTrack (GAM-R-S328) is absent from the table. The data fix (dedupe rows 32-34, add an Ipsos row, recompute stats) is in data/study-evidence.json, | chapters/09-outcomes.md |
| IR1-g5-04 | ### What independent research finds | Internal review round 1 (major, rule): The heading calls a section 'independent research' although most of what it covers has unknown funding. The meta-analysis, both Ingendahl papers, Min et al., Huang and Morozov, Zagni et al. and Hatakeyama are all 'funding unknown', and the section itself says so. This contradicts the chapter's own coding rule and the Q8 rule that a study counts as independent only with documented disinterested funding. It also contradicts the shared rule 'If funding is unknown, independence is unknown, not assum | data/hypotheses.json Q8 rule_supported: 'a study counts as independent only when its funding is documented as coming from a party without a financial interest'. Chapter text: 'Its subject is game promotion through creators, and its funding is unknown, as is the funding of most academic papers read only as abstracts (GAM-H-S35, GAM-H-S31, GAM-H-S34)'. data/figure-data.json studies: 'Academic, funding not stated' n=7; 'Academic, no outside funding' n=1. | supported as stated | 09-outcomes: the heading 'What independent research finds' becomes 'What academic research finds'. | Every academic study except GAM-H-S30 has 'funding unknown' in study-evidence.json. Q8's rule counts a study as independent only with documented disinterested funding, so 'independent' in the heading was wrong. No chapter links to the old heading anchor. | chapters/09-outcomes.md |
| IR1-g5-05 | Seller recall levels of 32% to 49% use other measures, such as spontaneous or prompted recall, so they sit in tension with that result without contradicting it (inference) | Internal review round 1 (major, fact): The reconciliation misdescribes the measures. The laboratory paper reports its 26.6% under 'Unaided Brand Recall' and defines unaided recall as spontaneous. Anzu's 36% is 'spontaneous advertising recall', the same kind of unaided measure. So seller figures do not all 'use other measures'. The real differences lie in setting, exposure, question wording and undisclosed samples. | GAM-H-S30 https://pmc.ncbi.nlm.nih.gov/articles/PMC12733859/: '3.1. Unaided Brand Recall ... When asked whether they noticed any advertisements during gameplay, 21 participants (26.6%) reported seeing at least one advertisement' and 'Brand recall refers to an individual's ability to spontaneously (unaided) or prompted (aided) remember a brand'. GAM-H-S05 https://www.anzu.io/news/lumen-study: 'Across the studies average spontaneous advertising recall was 36%'. GAM-H-S04 https://www.anzu.io/news/attention-report-pr: '49% prompted recall'. | supported with qualification | 09-outcomes: the reconciliation sentence now says the seller recall figures come from live campaigns with unpublished recall questions and timing, that Anzu's 36% is the same unaided kind and its 49% is prompted, and that the figures differ in setting, exposure and wording (inference). | Confirmed. The GAM-H-S30 section '3.1. Unaided Brand Recall' reports the 26.6% and defines unaided recall as spontaneous. Anzu's 2022 release gives 'average spontaneous advertising recall was 36%', and the 2023 release gives '49% prompted recall'. I left out the reviewer's claim that samples are not published, because the Anzu 2022 release quotes 5,000 respondents for its attention tests. I kept o | chapters/09-outcomes.md |
| IR1-g5-06 | The report does not say whether Twitch took part, and not all ad categories shown were gaming-specific (GAM-H-S01). | Internal review round 1 (major, fact): The sentence suggests Twitch's participation is unknown. Two sources the chapter cites say otherwise. Lumen's release of the report, cited in the same section, is titled and published 'with Twitch Ads, Anzu and Activision Blizzard' and carries an 'About Twitch Ads' company description beside Anzu's. Microsoft's October 2024 post credits the study to dentsu 'in partnership with Lumen Research, Twitch Ads, Activision Blizzard Media, and Anzu'. Leaving out that evidence understates the seller invol | GAM-C-S32 https://lumen-research.com/blog/dentsu-and-lumen-release-in-game-advertising-report-with-twitch-ads-anzu-and-activision-blizzard/ (16 Jan 2024): URL and page carry 'About Twitch Ads Twitch Ads is a global creative powerhouse ...' next to 'About Anzu'. GAM-V-H-S05 https://about.ads.microsoft.com/en/blog/post/october-2024/harnessing-gaming-s-power-for-deeper-brand-connections: '("Play Attention: Calling Focus to Gaming," Dentsu Media in partnership with Lumen Research, Twitch Ads, Activision Blizzard Media, and Anzu, US, January 2024)'. | supported as stated | 09-outcomes: the Twitch sentence now says the report does not describe Twitch's role, but Lumen's release carries a Twitch Ads description and Microsoft credits Twitch Ads as a partner. GAM-C-S32 and GAM-V-H-S05 added to the citations. | Confirmed. Lumen's release (GAM-C-S32) carries an 'About Twitch Ads' description beside 'About Anzu'. Microsoft's October 2024 post (GAM-V-H-S05) cites '"Play Attention: Calling Focus to Gaming," Dentsu Media in partnership with Lumen Research, Twitch Ads, Activision Blizzard Media, and Anzu'. The ledger's statement that the report itself says nothing of Twitch's role is kept. | chapters/09-outcomes.md |
| IR1-g5-07 | Its April 2025 claim that intrinsic ads beat advertisers' cost-per-acquisition (CPA) goals by 21% on average mixes performance and brand campaigns; user-acquisition economics belong to the published mobile app growth paper (GAM-V-H-S03). | Internal review round 1 (major, fact): The release does not say which campaigns the 21% covers, so 'mixes performance and brand campaigns' is an unsupported statement of fact about the source. It is a reviewer's reading, not labelled as inference. The release gives no campaign count, objective mix, period or method. | GAM-V-H-S03 https://www.anzu.io/news/trends-report (30 April 2025): 'Brands tapping into intrinsic in-game advertising are achieving exceptional performance, surpassing advertiser CPA (Cost Per Acquisition) goals by an average of 21%. This finding comes from Anzu ... in its new research report'. There is no breakdown of campaigns by objective anywhere on the page. | supported as stated | 09-outcomes: the CPA sentence now says the release gives no campaign count, objective mix, period or method. | Confirmed on GAM-V-H-S03. The release states the 21% figure and gives no campaign count, objective mix, period or method; the full report is form-gated. The statement that the figure 'mixes performance and brand campaigns' had no support. | chapters/09-outcomes.md |
| IR1-g5-13 | A buyer who wanted to buy in-game inventory through an agent at the cutoff would find a gaming category in AdCP's schema and no in-game seller documented in AdCP's public registry or named by IAB Tech Lab's seller agent | Internal review round 1 (major, consistency): The closing sentence is a wider absence claim than the evidence allows, and it leaves out routes the chapter or the paper records. The paragraph above records Zynga's announced route to buy its inventory through PubMatic's AgenticOS. The public registry lists InMobi and LoopMe sales agents, and Roblox's app-ads.txt authorises InMobi (DIRECT) and LoopMe (RESELLER). IAB Tech Lab's seller agent connects to PubMatic, Magnite and Index Exchange, which Roblox authorises as direct sellers. No source sh | GAM-I-S36 https://agenticadvertising.org/api/registry/agents (fetched 29 Sep 2026, count 34): includes 'InMobi Sales Agent \| https://adcp-prod.exchange.inmobi.com/mcp' and 'LoopMe Sales Agent \| https://platform.loopme.ai/mcp/seller'. GAM-ROBLOX-S39 https://www.roblox.com/app-ads.txt (29 Sep 2026): 'inmobi.com, 351176ac251d40529aec1a0204f17303, DIRECT', 'loopme.com, 5679, RESELLER', 'pubmatic.com, 164509, DIRECT', 'rubiconproject.com, 19318, DIRECT', 'indexexchange.com, 214070, DIRECT'. GAM-V-R3-S03 https://www.pocketgamer.biz/pubmatic-and-zynga-partner-to-expand-advertiser-access-to-premium-mo | supported with qualification | 10-structure: the closing sentence now says no seller from the paper's register is in the registry or the seller agent, and adds the announced Zynga/PubMatic AgenticOS route. A new sentence notes that InMobi, LoopMe, PubMatic and Index Exchange, which Roblox's app-ads.txt authorises, reach the registry or the seller agent, and that no source shows game inventory offered through them. | Confirmed. The GAM-I-S36 registry (34 entries) lists 'InMobi Sales Agent' and 'LoopMe Sales Agent'. Roblox app-ads.txt (GAM-ROBLOX-S39) lists inmobi.com DIRECT, loopme.com RESELLER, pubmatic.com DIRECT and indexexchange.com DIRECT. The seller agent README (GAM-I-S40) connects to PubMatic, Index Exchange and Magnite. The Zynga/PubMatic AgenticOS route is in GAM-V-R3-S03. None of these intermediarie | chapters/10-structure.md |
| IR1-g5-14 | It sold them directly at launch, and by 15 September 2026 Digiday reported programmatic buying across EA's mobile games, with no DSP or deal type named | Internal review round 1 (major, fact): The sentence reads as though the ad units launched in June moved from direct sale to programmatic within three months. The launch release does not state a buying route for the ad units (ledger GAM-G-F21). Digiday says EA 'has already offered' programmatic buying across its mobile games, which is a different product line from the console EA SPORTS ad units. Digiday also says EA is only now extending more dynamic capabilities to console and PC. On 3 July EA said it was not yet working with DSPs su | GAM-G-S25 https://news.ea.com/press-releases/... (15 June 2026): 'EA Advertising works directly with brands and partners to create bespoke integrations ... with expanded buying capabilities planned'. GAM-R-S322 https://digiday.com/marketing/electronic-arts-is-betting-that-in-game-ads-can-out-earn-ctv/ (3 July 2026): '"So we're not working with those partners yet," said Dao' (re The Trade Desk and Criteo). GAM-V-R3-S01 https://digiday.com/marketing/electronic-arts-is-bundling-its-games-to-make-a-bigger-pitch-to-advertisers/ (15 Sep 2026): 'EA has already offered programmatic buying across its m | supported as stated | 10-structure: the EA sentence now separates the launch statement, the 3 July Digiday quote and the 15 September Digiday report on mobile programmatic buying and console/PC extension. | Confirmed. The EA release describes bespoke direct work 'with expanded buying capabilities planned'. Digiday on 3 July quotes '"So we're not working with those partners yet," said Dao' about The Trade Desk and Criteo. Digiday on 15 September says 'EA has already offered programmatic buying across its mobile games and is now extending more dynamic capabilities to its console and PC titles'. The old | chapters/10-structure.md |
| IR1-g5-15 | figure fig-events | Internal review round 1 (major, rule): The events figure classifies events differently from the chapter and breaks shared rules 8 and 11. It codes Admix as 'rebrand' where the text says merger and pivot. It codes the Bidstack plc administration and the Bidstack Ltd creditors' voluntary liquidation as 'retired', although insolvency is not product retirement. It shows 'Oracle Moat ends' as retired, although the chapter says 'this review opened no record of the retirement itself'. It codes Roblox's 1 April 2025 routes as a 'launch', alt | data/figure-data.json events: {'label':'Admix into LandVault','kind':'rebrand'}, {'label':'Bidstack plc administration','kind':'retire'}, {'label':'Bidstack Ltd liquidation','kind':'retire'}, {'label':'Oracle Moat ends','kind':'retire'}, {'label':'Roblox adds Google routes','kind':'launch','date':'2025-04-01'}. figures/fig-events.svg legend: 'announced\|completed\|rebrand\|launch\|retired'. GAM-C-S15 https://about.roblox.com/newsroom/2025/04/roblox-scales-video-ads-partners-with-google: 'In the coming weeks, brands can easily expand their video ad buys to purchase our new Rewarded Video ad format | contradicted | No change. | The finding is valid. The events are defined in scripts/derivations.mjs and generated into data/figure-data.json: Admix 'rebrand', both Bidstack events 'retire', 'Oracle Moat ends' 'retire', and 'Roblox adds Google routes' 'launch', although GAM-C-S15 says 'In the coming weeks'. Chapter 05 line 65 also says 'Oracle ended Moat'. All three files are outside my edit scope. The chapter 10 text classif | chapters/10-structure.md |
| IR1-g5-16 | Its public agent registry listed 34 agents on 28 and 29 September 2026, none operated by Roblox | Internal review round 1 (major, rule): Two of the 34 registry entries carry the name of No Fluff Advisory, this paper's publisher. The chapter uses the registry count as evidence without saying so. The shared rules require naming an interested party behind a number, and readers may reasonably want this interest disclosed where the paper assesses AdCP adoption. | GAM-I-S36 https://agenticadvertising.org/api/registry/agents (fetched 29 Sep 2026; count 34): entries 'No Fluff Advisory - Evgeny (Ev) Popov \| https://adcp.signal-stack.io/mcp' and 'No Fluff Advisory Sales Agent \| https://sell.nofluffadvisory.com/mcp'. | supported as stated | 10-structure: added 'Two of the 34 entries are listed under the name of No Fluff Advisory, this paper's publisher (GAM-I-S36).' | Confirmed: the registry fetched on 29 Sep 2026 includes 'No Fluff Advisory - Evgeny (Ev) Popov' (signals) and 'No Fluff Advisory Sales Agent' (sales). The rules require naming an interested party behind a number, and the abstract already discloses the author's AgenticAdvertising.org working-group role. The disclosure names no client and does not mention consulting. Chapter 03 line 81 cites the sam | chapters/10-structure.md |
| IR1-g5-minor | Minor wording, label and voice issues in 10-structure.md, 09-outcomes.md | Internal review round 1, minor findings | See review record | supported with qualification | Applied or rejected per finding: g5-08 (applied); g5-09 (applied_modified); g5-10 (applied); g5-11 (applied); g5-12 (rejected); g5-17 (applied); g5-18 (applied_modified); g5-19 (applied); g5-20 (applied); g5-21 (applied); g5-22 (applied); g5-23 (rejected) | Wording, labels, bounded absences and voice | chapters/10-structure.md, chapters/09-outcomes.md |
| IR1-G6-01 | Across the matrix stat cells_coverage_pct% of applicable cells carry a score. | Internal review round 1 (blocking, fact): The stat is scored cells divided by all cells, n/a included, not by applicable cells. Applicable cells are 266 (276 minus 10 n/a), so the share of applicable cells that carry a score is 94%, not 91%. The sentence states the wrong denominator. | scripts/stats.mjs line 68: cells_coverage_pct = scored / cells * 100. data/stats.json: cells 276, cells_scored 250, cells_ne 16, cells_na 10, cells_coverage_pct 91. 250/276 = 90.6%, 250/266 = 94.0%. | supported as stated | 11-matrix inshort: replaced the '% of applicable cells' sentence with 'Across the matrix stat cells_scored of the stat cells cells carry a score, stat cells_ne are n/e and stat cells_na are n/a.' | Confirmed: stats.json gives cells 276, cells_scored 250, cells_na 10. 250/276 = 91%, but the applicable base is 266 (94%), so the sentence used the wrong denominator. | chapters/11-matrix.md |
| IR1-G6-02 | Two non-US surveys put brand safety first, one inside sampling error | Internal review round 1 (blocking, fact): The ledger does not establish Advertiser Perceptions' sample as non-US. It says the geography is not stated. Calling both surveys non-US is a fact the evidence does not support. The row also says only one lead is inside sampling error, but GAM-S-F17 says IAB Australia's lead is not established as a distinguishable rank either. | claim-ledger GAM-S-F18 final_wording: 'Advertiser Perceptions 2024 (n=200, July 2024; geography not stated)'; 'neither sample is established as US'. GAM-S-F17: 'IAB Australia's 15-point lead ... is not established as a distinguishable rank either.' On 29 Sep 2026 Advertiser Perceptions' own summary page (https://www.advertiserperceptions.com/gaming-advertising-2024-top-takeaways-from-our-latest-report/, GAM-S-S03) gave no respondent geography. | supported as stated | 12 table H1 row: 'Two surveys not established as US put brand safety first, and neither lead is established as larger than sampling error; ...' | Confirmed. GAM-S-F18 says AP's geography is not stated, and AP's summary page (read 29 Sep) states no respondent geography. GAM-S-F17 says IAB Australia's lead is 'not established as a distinguishable rank either'. | chapters/12-implications.md |
| IR1-G6-03 | Four specialist exits since 2021 by primary record | Internal review round 1 (blocking, consistency): The fourth event, Misfits Ads, was a business line of an esports company, not a specialist company. The ledger, chapter 10 and chapter 13 all count three specialist companies plus one business line, so this row contradicts them. Plain.md and the abstract carry the same error. | claim-ledger GAM-U-F35 review_note: 'the threshold of three is met without Misfits Ads. Misfits Ads was a business line of Esports Now, LLC, an esports company, not a specialist company.' Chapter 10: 'Bloxbiz, Admix and Bidstack are companies, and Misfits Ads was an in-game ad business line that an esports company sold'. Super League 8-K (event date 30 Apr 2026, https://www.sec.gov/Archives/edgar/data/1621672/000143774926015035/slgg20260506_8k.htm, GAM-B-S44): 'Esports Now, LLC ("Misfits") ... agreed to sell certain assets strictly constituting the Misfits Ads Business'. | supported as stated | 12 table H5 row: 'At least three specialist companies sold, merged or put into insolvency since 2021, by primary record; comparable revenue series are not disclosed' | Confirmed by the GAM-U-F35 review_note: Misfits Ads was a business line of Esports Now, LLC, and the threshold of three is met without it. The abstract (00-abstract.md) is outside this fixer's files and may carry the same error. | chapters/12-implications.md |
| IR1-G6-04 | US buyers already investing in games rated it brand safe, and those who fold games into other channels cite scale first (GAM-S-S01) claim GAM-S-F09. | Internal review round 1 (blocking, rule): This is a commissioned study, sponsored by two products scored in this paper (Admazing and Venatus), but the sentence presents it with no sponsor (hard rule 5). The 'brand safe' result is in GAM-S-F08, not in the cited F09. F09's review note also says the four reasons sit in unnumbered boxes and that the report ties 'scale' to the advertiser's own spend, not to audience, so 'cite scale first' used as an audience argument is an ambiguous inference. | IAB PDF (https://www.iab.com/wp-content/uploads/2024/03/IAB_AP_Changing_The_Game_March_2024.pdf, GAM-S-S01), read 29 Sep 2026: cover 'SPONSORED BY:'; p.38 sponsor blocks for Admazing and Venatus; chart note 'IAB/Advertiser Perceptions 2024, among 300 U.S. advertisers'. GAM-S-F07: 'sponsored by Admazing and Venatus'. GAM-S-F08: '86% agreed gaming is brand safe'. GAM-S-F09 review_note: 'The boxes are unn[umbered]' and 'ties scale to the advertiser's own spend level ... not to a shortage of audience'. | supported with qualification | 12: sentence now names the March 2024 IAB survey of 300 US buyers, its sponsors Admazing and Venatus, and the 86% brand-safe result claim GAM-S-F07 claim GAM-S-F08. A separate sentence says 'not yet enough scale' comes first among four unnumbered reasons, which the report ties to buyers' own spend claim GAM-S-F09. | Confirmed in the IAB PDF: 'SPONSORED BY:' on the cover, Admazing and Venatus blocks on p.38, 'among 300 U.S. advertisers' and gaming 86% brand safe on p.13. p.20 shows four unnumbered reasons with the spend-tied quote. I split the fix into two sentences and folded in G6-05's 85% figure. | chapters/12-implications.md |
| IR1-G6-05 | The measurement half of that is well supported | Internal review round 1 (major, fact): The facts cited (no accreditation, verification that rests on announcements) show thin assurance on the supply side. They do not show that buyers cannot fit gaming into their measurement systems. The same IAB survey found 85% of current US buyers confident in measuring games advertising. That counterevidence is in the ledger and is left out. | claim-ledger GAM-A-F17 final_wording: '85% are confident in measuring it'. IAB PDF p.11 (GAM-S-S01), read 29 Sep 2026: '85% of advertisers are confident in measuring'. GAM-S-F08 review_note: 'The same report and release say 85% of these buyers are confident in their ability to measure games advertising'. | supported with qualification | 12: 'The supply-side facts behind the measurement half are well supported: ... They do not show that buyers cannot fit gaming into their measurement systems, nor that these gaps, rather than audience, hold spend back.' Added the 85% measurement-confidence result (GAM-S-S01, GAM-A-S21) claim GAM-A-F17. | Confirmed. The IAB release (GAM-A-S21) says '85% of those surveyed are confident in their ability to measure games advertising', and IAB PDF p.11 says the same. GAM-A-F17 carries it. The counterevidence had been left out. | chapters/12-implications.md |
| IR1-G6-06 | That turns out to be the only publisher with public US dollar figures, and its public statements disagree about when the series passes $10 billion (GAM-A-S13, GAM-A-S20, GAM-A-S15, GAM-A-S23) claim GAM-A-F21. | Internal review round 1 (major, rule): The absence claim is not bounded; the ledger, the abstract and chapter 2 all say 'the only publisher found' among those screened. The $10 billion disagreement is not in GAM-A-F21 or its sources. It is GAM-A-F06. | claim-ledger GAM-A-F21 final_wording: 'Among the publishers screened by 28 September 2026, EMARKETER is the only one found with public US dollar figures'. GAM-A-F06: 'its later public statements give no single revised crossing year ... the two statements conflict'. Chapter 0 cites the $10 billion point as (GAM-A-S02, GAM-A-S01) claim GAM-A-F06. | supported as stated | 12: 'Among the publishers screened by 28 September 2026, it is the only one found with public US dollar figures [[...]] claim GAM-A-F21, and its public statements disagree about when the series passes $10 billion (GAM-A-S02, GAM-A-S01) claim GAM-A-F06.' | Confirmed. GAM-A-F21 is bounded to 'publishers screened by 28 September 2026', and the $10 billion disagreement belongs to GAM-A-F06 (cited as (GAM-A-S02, GAM-A-S01) in chapter 0). | chapters/12-implications.md |
| IR1-G6-07 | Children are part of most game audiences. | Internal review round 1 (major, rule): This is an unsourced claim about the age of a population. The brief's landmine says never infer age and always cite the platform's own age data. The paper's only age data is Roblox's filed split. | RESEARCH-BRIEF landmine: 'Children's audiences: never infer age; cite the platform's own age data.' H4 rule_unresolved: 'Never infer a user's age; cite the platform's own age data.' Roblox Q2 2026 letter (https://www.sec.gov/Archives/edgar/data/1315098/000162828026051059/ex991-robloxq22026earnin.htm, GAM-E-S16), read 29 Sep 2026: 'As of the seven days ended June 30, 2026, on a global basis, 35% of users who have age-checked are under 13'; 'may not be representative of actual age demographics'. Ledger GAM-ROBLOX-F05. | supported with qualification | 12: 'Children can be a large share of a game audience: of Roblox users who had age-checked, 35% were under 13 globally in the seven days to 30 June 2026, and Roblox cautions that data drawn from age-checked users may not represent its whole user base (GAM-E-S16, GAM-E-S15) claim GAM-ROBLOX-F05.' | Confirmed as an unsourced age claim, which the brief's landmine forbids. The Roblox Q2 2026 letter (EDGAR, read 29 Sep) says '35% of users who have age-checked are under 13' and that data based on age-checked users 'may not be representative of actual age demographics on the platform'. I softened 'some game audiences' to 'a game audience', so the general claim rests only on the Roblox example. | chapters/12-implications.md |
| IR1-G6-08 | EA sells placements inside its own sports games directly, with programmatic buying only planned at its June 2026 launch (GAM-G-S25) claim GAM-G-F21 | Internal review round 1 (major, fact): EA's release says 'expanded buying capabilities planned'. It never says programmatic, and GAM-G-F21 says the release 'does not state the buying route for the ad units'. The sentence is also out of date and conflicts with chapter 10: by 15 September 2026, Digiday reported programmatic buying across EA's mobile games, with console and PC capabilities being extended. | EA release, 15 Jun 2026 (https://news.ea.com/press-releases/press-releases-details/2026/Electronic-Arts-Introduces-EA-Advertising-Launching-Brands-Directly-Into-Gameplay-and-Live-Experiences/, GAM-G-S25), read 29 Sep 2026: 'expanded buying capabilities planned as the platform continues to scale'. GAM-R-F329: 'By 15 September 2026 Digiday reported that EA already offers programmatic buying across its mobile games and is extending more dynamic capabilities to its console and PC titles. No DSP or deal type is named.' Chapter 10 cites (GAM-G-S25, GAM-R-S322, GAM-V-R3-S01) claim GAM-R-F329. | supported with qualification | 12: 'EA sold its in-game ad units directly at its June 2026 launch, and by 15 September 2026 Digiday reported programmatic buying across EA's mobile games while saying only that more dynamic capabilities were being extended to console and PC, with no DSP or deal type named (GAM-G-S25, GAM-R-S322, GAM-V-R3-S01) claim GAM-R-F329' | Confirmed. The EA release says only 'expanded buying capabilities planned', and Digiday on 15 Sep 2026 (GAM-V-R3-S01) says 'EA has already offered programmatic buying across its mobile games and is now extending more dynamic capabilities to its console and PC titles'. I did not claim that console programmatic buying exists, because Digiday says only 'more dynamic capabilities'. | chapters/12-implications.md |
| IR1-G6-09 | Trivver, which is not a live brand-selling network in the period (GAM-C-S09) claim GAM-C-F09 | Internal review round 1 (major, rule): GAM-C-F09 covers only the MRC listing. 'Not a live brand-selling network' is the review's own inference (GAM-R-F220), drawn from the absence of a named live title (GAM-R-F218). Here it is stated flatly, with no label and no citation. Chapter 3 bounds it correctly ('this review found no live title for it'). | GAM-R-F220 (evidence_class inference): 'Trivver in-game ads does not meet the register's condition for scoring (a live brand-selling network documented in the scoring period)'. GAM-R-F218: 'This review found no named game title, no named brand campaign and no published sellers.json ... in the sources screened by 29 September 2026.' MRC Services Under Review (https://mediaratingcouncil.org/accreditation/services-under-review/digital), read 29 Sep 2026: 'Trivver (Pre-audit reviewed 7/2022, Audit in process as of 4/2026)'. | supported as stated | 12: 'Trivver, which this review lists rather than scores because it found no named live title or campaign for it in the period (GAM-C-S09, GAM-R-S228) claim GAM-C-F09 claim GAM-R-F218' | Confirmed. GAM-C-F09 covers only the MRC listing, which I rechecked on 29 Sep: 'Trivver (Pre-audit reviewed 7/2022, Audit in process as of 4/2026)'. The 'not live network' point is GAM-R-F220's labelled inference, built on GAM-R-F218's bounded absence. | chapters/12-implications.md |
| IR1-G6-10 | Independent outcome studies for in-game formats: none found that meets the test above. | Internal review round 1 (major, consistency): GAM-H-F01 says studies with documented independent funding do exist: laboratory experiments. What was not found is a randomised, holdout or matched-market test of a live brand campaign funded by a party without a financial interest. As written, the sentence denies that any independent outcome study exists. | claim-ledger GAM-H-F01 final_wording: 'no public randomised, holdout or matched-market test of an in-game or game-adjacent brand campaign whose funding is documented as coming from a party without a financial interest ... The only studies with documented independent funding are laboratory experiments.' GAM-H-S30 is a peer-reviewed eye-tracking study (Eskisehir Technical University). | supported with qualification | 12: 'Independent outcome studies for in-game formats: none found for a live brand campaign that meets the test above, and the only study with documented independent funding is a laboratory experiment (GAM-H-S30, GAM-H-S01, GAM-H-S11) claim GAM-H-F01.' | Partly right. 'the test above' already bounds the sentence, but GAM-H-F01 says independently funded laboratory experiments exist, so the bare sentence understated the record. I added the qualifier and the citation. | chapters/12-implications.md |
| IR1-G6-11 | would move format execution, the weakest dimension in cohort A | Internal review round 1 (major, consistency): By the matrix data, coverage is cohort A's weakest dimension (mean 1.50 across its 6 counted cells), below format execution (1.56 across 9). | data/capability-matrix.json, cohort network, mean of counted cells: coverage 1.50 (Adverty 1, Anzu 3, Frameplay 1, Gadsme 2, IION 1, Kidoz 1), format_execution 1.56, creative_insertion 1.89, independent_assurance 2.00. | supported with qualification | 12: 'would move format execution, one of the two weakest dimensions in cohort A along with coverage.' | Confirmed from capability-matrix.json. Cohort A means of counted cells are coverage 1.50 (n6) and format_execution 1.56 (n9), so coverage is the lowest. | chapters/12-implications.md |
| IR1-G6-12 | which date Roblox applies for the developer revenue share, 30 September 2026 or 1 January 2027 (GAM-E-S10, GAM-E-S11) | Internal review round 1 (major, rule): This is listed as a signal to watch after the cutoff, but 30 September 2026 comes before the 2 October 2026 freeze. The shared rules require platform terms to be rechecked at the freeze, so the outcome for that date falls inside the evidence window. | SHARED-RESEARCH-RULES: 'Recheck status-sensitive facts (ownership, insolvency, accreditation, standards versions, platform terms) at the freeze'. Roblox Advertising Integrations Terms (help-centre API for article 47656162239124, GAM-E-S10), read 29 Sep 2026: 'Beginning September 30, 2026, You agree to pay Roblox a revenue share'. Creator Hub (https://create.roblox.com/docs/production/promotion/ad-integrations, GAM-E-S11): 'Starting January 1, 2027, ad integration campaigns will be subject to a revenue share'. | supported with qualification | 12 signals list: 'whether Roblox applies the developer revenue share from 30 September 2026, as its Terms say, or from 1 January 2027, as its Creator Hub says, the first being a date inside the evidence window to be rechecked at the 2 October freeze (GAM-E-S10, GAM-E-S11)' | Confirmed. The Roblox Terms (help-centre API, read 29 Sep) say 'Beginning September 30, 2026, You agree to pay Roblox a revenue share'. The Creator Hub says 'Starting January 1, 2027'. 30 September falls before the 2 October freeze, and the shared rules require platform terms to be rechecked at the freeze. | chapters/12-implications.md |
| IR1-G6-13 | The inventory leg holds, but the rule needs both. | Internal review round 1 (major, rule): The paragraph says the survey leg 'is met only on point estimates' and the inventory leg 'holds', which reads as both legs being met. It never names the registered condition that makes H1 unresolved: once the inference about IAB Australia's reach option is set aside, only one survey qualifies. The registered rule does not make US geography or sampling error a condition, so they cannot carry the coding. | hypotheses.json H1 rule_unresolved: 'Fewer than two usable surveys with base and period, or usable surveys disagree'. GAM-S-F18: 'The survey leg is therefore met on the rule's wording only on point estimates and on the inference that the reach option was offered in 2023; without that inference only one survey qualifies and the leg is unresolved.' | supported as stated | 13 H1 paragraph: last sentence now reads 'Without that inference only one usable survey remains, which the registered rule codes as unresolved; the inventory leg holds, but the rule needs both legs (GAM-S-S02, GAM-S-S06, GAM-S-S04) claim GAM-S-F18.' | Confirmed. H1 rule_unresolved reads 'Fewer than two usable surveys'. GAM-S-F18: 'without that inference only one survey qualifies and the leg is unresolved'. The paragraph never named the condition that decides the coding. | chapters/13-method.md |
| IR1-G6-14 | the first lead is inside sampling error | Internal review round 1 (major, fact): Saying only the first lead is inside sampling error implies the second is outside it. The ledger says IAB Australia's lead is not established as distinguishable either. | claim-ledger GAM-S-F17 final_wording: 'Brand safety's 3-point lead over lack of scale (29% vs 26%) is very unlikely to be a distinguishable rank. IAB Australia's 15-point lead ... has a 95% half-width of about ±7 to ±17 points ... so it is not established as a distinguishable rank either.' | supported as stated | 13: '...the first lead is inside sampling error, the second is not established as larger than it, and counting the second rests on the inference...' | Confirmed. GAM-S-F17 says IAB Australia's 15-point lead is 'not established as a distinguishable rank either'. | chapters/13-method.md |
| IR1-G6-15 | The verifier sample covers the material claims per file, not every claim; the rest are labelled unresolved rather than assumed. | Internal review round 1 (major, consistency): This caption on the pipeline figure, which renders in this chapter through figure fig-pipeline, contradicts the chapter's own count: every one of the 807 ledger claims was checked, and only 1 is unresolved. | figures/fig-pipeline.json caption and scripts/make-figures.mjs line 59. data/stats.json: claims 807, claims_checked 807, claims_unresolved 1. Chapter 13 standfirst: 'checked stat claims_checked of its stat claims claims'. | contradicted | No change. | The finding is valid. The caption contradicts stats.json (claims_checked 807 of 807, claims_unresolved 1). But the caption text is in figures/fig-pipeline.json and is generated by scripts/make-figures.mjs lines 55 and 59, including a second line: 'Claims outside the verifier sample are marked unresolved'. Neither file is one this fixer may edit, and the chapter holds only the figure fig-pipeline | chapters/13-method.md |
| IR1-G6-16 | the registers and scoring rules (fixed before scoring) in `data/` | Internal review round 1 (major, consistency): Chapter 13 says the product register was amended during research and that Medal was added and scored after the main pass. Only the hypotheses and scoring rules were fixed before scoring, so calling the registers 'fixed before scoring' contradicts chapter 13 and overstates the preregistration. The abstract repeats it ('the product register and the scoring rules were fixed before scoring'). | Chapter 13: 'The product register was written the same day and amended as research found products ... Medal's game-adjacent advertising, came from the second outside review and was scored after the main pass'. Git: scoring commit 6d50bfd0 (29 Sep 18:35) precedes the Medal commit 5b57d858 (19:03). Corrections ledger C05, C06, R2-18 amend the register. | supported as stated | 14 Rebuilding: 'the hypotheses and scoring rules (fixed before scoring) and the product register (amended during research, with the changes logged in the corrections ledger) in `data/`' | Confirmed. Chapter 13 says the product register was 'amended as research found products' and that Medal was scored after the main pass. The corrections ledger holds the register changes C05–C09, R2-17 and R2-18. The abstract's matching sentence is outside this fixer's files. | chapters/14-appendix.md |
| IR1-G6-17 | which puts it at the foot of Mid | Internal review round 1 (major, fact): Medal is not at the foot of Mid. Two Mid rows sit below it, and the chapter itself says it reports bands, not ranks. | data/capability-matrix.json, cohort adjacent, Mid offsets: Medal −5.63 (q 42.7), Enthusiast Gaming −7.12 (q 41.2), ESL FACEIT Group −7.35 (q 41.0). | supported with qualification | 11: '...documents formats only by name and targeting only through personal data, and sits in Mid (GAM-R-S701, GAM-R-S708).' | Confirmed from the offsets: Medal −5.63, Enthusiast Gaming −7.12, ESL FACEIT −7.35. Medal is not the lowest Mid row, and the chapter reports bands, not ranks. | chapters/11-matrix.md |
| IR1-G6-18 | its filed audience, combined with its published age rules, puts its ad-eligible base above the supply threshold | Internal review round 1 (major, rule): The eligible base is this review's own derivation, which the verifier accepted only 'as labelled inference'. It mixes a quarterly average DAU (daily active users) with end-of-quarter age data and counts accounts, not people. Here it is stated as fact without the inference label (hard rule 9). | capability-matrix Roblox supply cell: 'the share is this review's derivation from rules plus the age mix of age-checked users ... Score 4 is defensible as labelled inference'; 'The derivation mixes a quarterly average DAU with end-of-quarter age data and counts accounts, not people.' | supported as stated | 11: 'its filed audience and age mix, combined with its published age rules, put its ad-eligible base above the supply threshold by this review's derivation, an inference that counts accounts rather than people' | Confirmed. The Roblox supply cell rationale says the eligible share is 'this review's derivation', 'defensible as labelled inference', and 'counts accounts, not people'. | chapters/11-matrix.md |
| IR1-G6-19 | On Xbox and PlayStation, the ads we found sit outside the game. | Internal review round 1 (major, consistency): The same section then says EA sells ads in its own sports games and that some Microsoft ads could not be placed inside or outside the game. The paper records EA brand integrations inside three console titles (in-match LED boards). Saying all ads found sit outside the game contradicts both the paper and the next sentences. | claim-ledger GAM-G-F25: 'EA's direct-sold program names brand integrations in three console titles: EA SPORTS FC 26 (in-match LED ad-boards in the US ...)'; Microsoft's documents 'do not establish whether these render inside gameplay'. | supported as stated | plain.md: 'On Xbox and PlayStation, most ads we found sit outside the game.' | Confirmed. GAM-G-F25 records EA brand integrations inside console titles (FC 26 in-match LED boards in the US), and the next sentences in plain.md already mention EA in-game ads and Microsoft's unplaceable integrations. | chapters/plain.md |
| IR1-G6-20 | We found four since 2021. | Internal review round 1 (major, consistency): The paper counts at least three specialist companies. The fourth event is a business line sold by an esports company, not a small firm that sold ads in games (see GAM-U-F35 and chapter 10). | claim-ledger GAM-U-F35 review_note: 'Misfits Ads was a business line of Esports Now, LLC, an esports company, not a specialist company.' Chapter 10: 'at least three specialist companies sold, merged or put into insolvency, and one specialist business line sold'. | supported as stated | plain.md: 'We found at least three since 2021.' | Confirmed by the GAM-U-F35 review_note: three specialist companies, and Misfits Ads was a business line of an esports company. | chapters/plain.md |
| IR1-G6-21 | One person set the scores. | Internal review round 1 (major, consistency): The paper never says a person set the scores. It says scoring was one pass by the author's AI-assisted research process, with no second scorer. The plain edition says something the paper does not. | Chapter 13: 'one pass scored the products with no second scorer'; 'Scoring was one pass on 29 September 2026, after the verifier records were read'. SHARED-RESEARCH-RULES: 'Scoring is one pass by the author's research process after verification; no independent second scorer exists'. | supported as stated | plain.md: 'The scores were set in one pass, and no second scorer checked them.' | Confirmed. Chapter 13 says 'one pass scored the products with no second scorer' and never says a person scored them. | chapters/plain.md |
| IR1-g6-minor | Minor wording, label and voice issues in 11-matrix.md, 12-implications.md, 13-method.md, 14-appendix.md, plain.md | Internal review round 1, minor findings | See review record | supported with qualification | Applied or rejected per finding: G6-22 (applied_modified); G6-23 (applied_modified); G6-24 (applied); G6-25 (applied); G6-26 (applied); G6-27 (applied); G6-28 (applied); G6-29 (applied_modified); G6-30 (applied); G6-31 (applied_modified); G6-32 (applied); G6-33 (rejected); G6-34 (applied); G6-35 (applied); G6-36 (applied); G6-37 (applied_modified); G6-38 (applied); G6-39 (applied); G6-40 (applied); G6-41 (applied); G6-42 (applied); G6-43 (applied); G6-44 (applied); G6-45 (applied) | Wording, labels, bounded absences and voice | chapters/11-matrix.md, chapters/12-implications.md, chapters/13-method.md, chapters/14-appendix.md, chapters/plain.md |
| IR1-SCORING | Cells as scored on 29 September 2026 | Internal review round 1, scoring audit of 30 sampled cells against anchors, caps and sources | R1-01: admazing-mobile-gaming-ads-for-brands independent_assurance; R1-02: ign-advertising supply; R1-03: playwire-gaming-media suitability_age_data; R1-04: xbox suitability_age_data; R1-05: roblox supply; R1-07: admazing-mobile-gaming-ads-for-brands targeting_context; R1-08: superleague supply; R1-11: ea-advertising programmatic_access; R1-12: enthusiast-gaming-media format_execution; R1-13: iion-in-game-brand-network suitability_age_data; R1-14: superleague measurement_function; R1-15: anzu independent_assurance [396, 253, 43] | supported with qualification | Admazing independent_assurance 1 to n/e (a partner name alone); IGN supply 4 to 3 (filed audience not ad-eligible, no period); Playwire suitability 4 to 2 (property designation is not user age data); XBOX suitability 4 to 2 (brand-suitability page only in a January 2026 archive); Roblox supply 4 to 3 (ad-eligible share is this review's derivation, not vendor-documented); dates, sources and grades corrected for Super League measurement, Anzu assurance, EA programmatic, Enthusiast formats and IION suitability; verifier-corrected cells now open their rationale with the final score. | Anchors are applied literally; missing evidence is n/e, never a score | data/scores.json, data/matrix-extra.json, chapters/11-matrix.md |
| IR1-consistency | Statements that disagreed between chapters, figures, site texts and the ledger | Internal review round 1, cross-document consistency lens (25 findings) | See review record; each finding cites the registered rule, ledger record or source it relies on [107, 57, 190, 436, 337] | supported with qualification | All 25 applied. Chapter 2 now names the registered clause that codes H1 unresolved; chapter 7 states that the registered H4 rule names the Roblox revenue-share conflict as an example and why the review sets it aside; the pipeline, sellers, assurance, quotient and console figures match the text; the Twitch TAG cell cites TAG's own record; Maryland's data act is dated and cited in chapter 7; the AdCP interest note names AgenticAdvertising.org as the body that governs the protocol; the GAM-S-F09 review note records that the verifier reversed the two p. 19 labels. | The same fact must read the same way everywhere it appears | chapters, plain.md, paper.config.json, data/figure-data.json, data/claims-overlay.json |
| IR1-disclosure | Site texts and summaries that said more than the body | Internal review round 1, disclosure lens (24 findings) | GAM-A-F21 (only publisher found); GAM-C-F25 (MRC record, 29 September 2026); chapter 9 standfirst and GAM-H-F01 (parties with a stake); 13-method.md (reviews run with another company's AI model) [74, 443, 15] | supported with qualification | All 24 applied. Absences are bounded to what this review found; the MRC status carries its date; outcome evidence is attributed to parties with a stake, not only sellers; the footer names the two outside reviews as AI-model reviews the author ran; the scope line says products, bands only where a cohort is large enough, and claims checked by an AI verifier; the Frameplay quote is dated to its October 2024 launch; the Roblox MDL cites the JPML transfer order. | A summary may not claim more than the chapter it summarises | paper.config.json, chapters/00-abstract.md, chapters/02-market.md, chapters/03-networks.md, chapters/05-adjacent.md, chapters/06-measurement.md, chapters/07-safety.md, chapters/10-structure.md, chapters/12-implications.md, chapters/13-method.md, plain.md |
| IR2-g1 | Findings in 02-market.md, 01-defining.md, figures/fig-estimates.json (rendered in 02-market.md via figure fig-estimates), 00-abstract.md | Internal review round 2, chapter reviewer g1 in refute mode (23 findings) | Each finding cites its source passage or data value; the fixer re-checked each against the ledger and, for facts, the primary source | supported with qualification | Applied or rejected per finding: R2G1-01 (applied_modified); R2G1-02 (applied_modified); R2G1-03 (applied_modified); R2G1-04 (applied); R2G1-05 (applied_modified); R2G1-06 (applied); R2G1-07 (applied); R2G1-08 (applied_modified); R2G1-09 (applied); R2G1-10 (applied_modified); R2G1-11 (applied); R2G1-12 (applied); R2G1-13 (applied); R2G1-14 (applied); R2G1-15 (rejected); R2G1-16 (applied_modified); R2G1-17 (applied_modified); R2G1-18 (applied); R2G1-19 (applied); R2G1-20 (applied_modified); R2G1-21 (applied); R2G1-22 (applied_modified); R2G1-23 (applied_modified). Rejected findings were handled as follows: R2G1-15 applied in the figure script (the estimates title is now bounded); g3r2-04 needs no change (the DoubleVerify sentence remains accurate against the DEFM14A of 29 September 2026); g3r2-19 superseded by R3-05; g4-23 applied as a restatement of GAM-H-F01. | Facts, rules, consistency and voice | 02-market.md, 01-defining.md, figures/fig-estimates.json (rendered in 02-market.md via figure fig-estimates), 00-abstract.md |
| IR2-g2 | Findings in 04-platforms.md, 03-networks.md, 05-adjacent.md | Internal review round 2, chapter reviewer g2 in refute mode (13 findings) | Each finding cites its source passage or data value; the fixer re-checked each against the ledger and, for facts, the primary source | supported with qualification | Applied or rejected per finding: g2r2-01 (applied); g2r2-02 (applied_modified); g2r2-03 (applied); g2r2-04 (applied_modified); g2r2-05 (applied_modified); g2r2-06 (applied_modified); g2r2-07 (applied); g2r2-08 (applied); g2r2-09 (applied_modified); g2r2-10 (applied); g2r2-11 (applied); g2r2-12 (applied_modified); g2r2-13 (applied). Rejected findings were handled as follows: R2G1-15 applied in the figure script (the estimates title is now bounded); g3r2-04 needs no change (the DoubleVerify sentence remains accurate against the DEFM14A of 29 September 2026); g3r2-19 superseded by R3-05; g4-23 applied as a restatement of GAM-H-F01. | Facts, rules, consistency and voice | 04-platforms.md, 03-networks.md, 05-adjacent.md |
| IR2-g3 | Findings in 07-safety.md, 06-measurement.md, 09-outcomes.md, 08-console.md | Internal review round 2, chapter reviewer g3 in refute mode (19 findings) | Each finding cites its source passage or data value; the fixer re-checked each against the ledger and, for facts, the primary source | supported with qualification | Applied or rejected per finding: g3r2-01 (applied_modified); g3r2-02 (applied); g3r2-03 (applied_modified); g3r2-04 (rejected); g3r2-05 (applied); g3r2-06 (applied_modified); g3r2-07 (applied_modified); g3r2-08 (applied); g3r2-09 (applied); g3r2-10 (applied); g3r2-11 (applied_modified); g3r2-12 (applied); g3r2-13 (applied_modified); g3r2-14 (applied_modified); g3r2-15 (applied_modified); g3r2-16 (applied); g3r2-17 (applied); g3r2-18 (applied); g3r2-19 (rejected). Rejected findings were handled as follows: R2G1-15 applied in the figure script (the estimates title is now bounded); g3r2-04 needs no change (the DoubleVerify sentence remains accurate against the DEFM14A of 29 September 2026); g3r2-19 superseded by R3-05; g4-23 applied as a restatement of GAM-H-F01. | Facts, rules, consistency and voice | 07-safety.md, 06-measurement.md, 09-outcomes.md, 08-console.md |
| IR2-g4 | Findings in 13-method.md, 12-implications.md, 10-structure.md, plain.md, 11-matrix.md | Internal review round 2, chapter reviewer g4 in refute mode (24 findings) | Each finding cites its source passage or data value; the fixer re-checked each against the ledger and, for facts, the primary source | supported with qualification | Applied or rejected per finding: g4-01 (applied_modified); g4-02 (applied_modified); g4-03 (applied_modified); g4-04 (applied); g4-05 (applied); g4-06 (applied_modified); g4-07 (applied); g4-08 (applied_modified); g4-09 (applied); g4-10 (applied); g4-11 (applied); g4-12 (applied); g4-13 (applied); g4-14 (applied); g4-15 (applied); g4-16 (applied); g4-17 (applied); g4-18 (applied); g4-19 (applied); g4-20 (applied); g4-21 (applied); g4-22 (applied); g4-23 (rejected); g4-24 (applied). Rejected findings were handled as follows: R2G1-15 applied in the figure script (the estimates title is now bounded); g3r2-04 needs no change (the DoubleVerify sentence remains accurate against the DEFM14A of 29 September 2026); g3r2-19 superseded by R3-05; g4-23 applied as a restatement of GAM-H-F01. | Facts, rules, consistency and voice | 13-method.md, 12-implications.md, 10-structure.md, plain.md, 11-matrix.md |
| IR2-LOUISIANA | Louisiana's HB 570 took effect on 1 July 2026 (chapters 7 and 13, the rules table, the rules figure, GAM-E-F27) | Internal review round 2, blocking finding g4-01 | Enrolled Act No. 185 (2026 Regular Session, HB 977), read 30 September 2026: 'Section 1. Act No. 481 of the 2025 Regular Session shall not become effective'; Section 6: Sections 2 to 4 take effect on 1 July 2027 (GAM-NFA-S10) [434, 429] | contradicted | The 2025 act never took effect; chapter 7, the rules table, the rules figure and GAM-E-F27 now say so, and the Apple row no longer lists Louisiana as live. H4 does not count Louisiana and is unchanged. | A status-sensitive legal fact rechecked at the freeze | chapters/07-safety.md, chapters/13-method.md, data/rules-overlay.json, scripts/derivations.mjs, data/claims-overlay.json, data/sources-extra.json |
| IR2-AGENTIC | The agentic map said AdCP v3.1.24 contains no gaming terms and gave 'none documented' for schema representation | Internal review round 2, blocking finding R2CD-01 | GAM-I-F26: AdCP's channels enum lists 'gaming' ('In-game advertising across platforms', present since v3.0.0); chapter 10 already said so [185, 184] | contradicted | The AdCP rows now record the channel-level 'gaming' value, the deferral of in-game creative formats and v3.1.24 as the stable release; the other protocol states stay none documented for any in-game seller | The data table contradicted the chapter and misstated a standard | data/agentic-map-overlay.json |
| IR2-SCORING | Cells held to a looser standard than comparable rows | Internal review round 2, scoring lens (16 findings; 25 cells sampled that round 1 did not change) | Each cell compared with the same anchor reading on other rows (Gadsme and Kidoz suitability and format rows; Kidoz marketing cap; the Kidoz 20-F grading note) [73, 125, 59, 44, 48] | supported with qualification | IION US availability 2 to 1 (an announced US route is not availability); Adverty suitability 4 to 3 (a COPPA flag with no stated effect is not an age control, as for Gadsme); Gadsme format 2 to 1 (playback rules and a counting definition are not specifications, as for Super League and Kidoz); Overwolf transparency and programmatic access 4 to 3 (the advertiser FAQ is sales marketing, grade D); company self-descriptions in filings regraded A to C for six Super League cells, Adverty US availability and Enthusiast targeting (scores unchanged); Discord programmatic and Venatus format dates and sources corrected. Bands unchanged. | The same anchor reading for every row | data/scores.json, data/matrix-extra.json, chapters/11-matrix.md |
| IR2-lenses | Summaries and data rows stronger than, or different from, their chapters | Internal review round 2, consistency and disclosure lens (22 findings) and scoring-text findings | See review record [233, 439] | supported with qualification | Applied: R2CD-02 to R2CD-22 and R2S-10 to R2S-16. 'Almost all outcome evidence' became 'every published field result found for a live brand campaign' in the abstract, chapter 9, the site texts and the studies figure; the method and plain edition describe the internal AI review rounds; the Xbox quotation is attributed to a sponsored article's narration; Misfits Ads is marked recorded and not counted; Apple's Texas status is current; the pipeline figure counts scoring ledger entries, not corrections. | A summary may not say more than its chapter | chapters, plain.md, paper.config.json, data/rules-overlay.json, data/claims-overlay.json, scripts/paper-figures.mjs, scripts/make-figures.mjs |
| IR2-INTERESTS | A fixer added to the method's interests statement that two agents in AdCP's public registry are listed under No Fluff Advisory's name | Internal review round 2, finding g4-22 | The registered interests statement names only Samba, the Signals & Measurement co-lead role and past employers, and must match the abstract [121] | supported with qualification | The addition was reverted in the interests statement; the registry listing stays disclosed where the registry is discussed, in chapter 10 | The interests statement is fixed wording, identical in the abstract and the method | chapters/13-method.md |
| IR2-FREEZE | Status-sensitive facts as of 28 and 29 September 2026 | Internal review round 2, freeze recheck on 30 September 2026 (AdInMo at Companies House, MRC lists and Trivver, Roblox revenue-share documents, Xbox cloud ad test, EA and Roblox programmatic routes, Zynga Ads+, Discord S-1, new events) | Companies House SC648449 (active, proposal to strike off, suspended); MRC Digital lists unchanged; Roblox Terms and Creator Hub still conflict; Xbox test unchanged; no EA or Roblox DSP added; no Zynga Ads+ launch; no Discord S-1; Discord release of 29 September 2026 (GAM-NFA-S11) [248, 48, 15, 16, 57, 190] | supported with qualification | Every rechecked status held. Discord's Ads Manager and new Video Quests, announced on 29 September 2026, were added to chapters 5, 11 and 12 and to GAM-F-F01; the Discord programmatic cell stays 0 because Ads Manager is Discord's own auction with no DSP or deal type. The research cutoff became 30 September 2026, the date of the recheck and of publication; the registered scoring period, to 2 October 2026, is unchanged and its last two days are not covered. | Status-sensitive facts are rechecked at the freeze | chapters/05-adjacent.md, chapters/11-matrix.md, chapters/12-implications.md, chapters/13-method.md, paper.config.json, data/claims-overlay.json, data/sources-extra.json |
| R3-01 | The headline says no accredited in-game metric, which treats every ad in a game as an intrinsic placement | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | The IIG guidelines (GAM-C-S01, section 1) exclude interstitial, banner and in-stream or outstream video ads, which fall under the general desktop, mobile web and mobile in-app standards; DoubleVerify's accreditation covers mobile in-app traffic measured through the OM SDK (GAM-C-S10) [17, 18, 19] | supported as stated | Every statement of the MRC result now says 'intrinsic' in-game metric, and the abstract, chapter 6, the H2 row, the assurance figure and the plain edition say that other ad formats in game apps fall under the general mobile in-app standards. | Checked against the data and primary sources before any change | chapters, paper.config.json, scripts/paper-figures.mjs, plain.md |
| R3-02 | The PQ builds in preferences (personal-data targeting, AI creative tooling, direct-only buying) without a buyer rationale | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | Anchors checked in data/scoring-rules.json: personal-data-only targeting scores at most 1 (not 0, as the review said); creative 5 needs AI-assisted tooling with controls; direct-only buying is 0 on programmatic access | supported with qualification | The anchors were registered before scoring and are unchanged. Chapter 11 now says whose buying they describe, that a managed sponsorship or contextual specialist can be the right buy and still score low, and that buyers with such a brief should read the dimension columns and the buyer questions, since reweighting cannot undo a preference built into an anchor. | Checked against the data and primary sources before any change | chapters/11-matrix.md |
| R3-03 | Old evidence is labelled live, and the labels are inconsistent | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | Cell records: LIVAD targeting, optimisation, transparency and format rest on 2022 to 2024 material; Xbox targeting rests on an archived page; verifier notes recommended unconfirmed | supported as stated | One rule: live needs a source dated inside the scoring period or a vendor page still served; otherwise unconfirmed. Five cells relabelled. Unconfirmed cells still count under the registered rules; the matrix build now recomputes the PQ with them as n/e as a scenario, and chapter 11 reports it (LIVAD falls to a thin record; no other band changes). | Checked against the data and primary sources before any change | data/scores.json, data/matrix-extra.json, build/build-matrix.mjs, scripts/stats.mjs, chapters/11-matrix.md |
| R3-04 | Weight stability is presented as if it were evidence confidence | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | stats.json: q_sensitive_missing rows change band in the missing-evidence scenarios; cells_grade_d_pct of scored cells rest on vendor marketing; no test varies grades or the scorer | supported as stated | Chapter 11 names the band-sensitive rows as provisional and says the tests do not vary evidence grades or the scorer, and that a blind second scoring pass, not run, would test more than further reweighting. | Checked against the data and primary sources before any change | chapters/11-matrix.md |
| R3-05 | The study table repeats three Frameplay studies and omits the Ipsos GameTrack analysis; funding and design quality are mixed | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | data/study-evidence.json: Frameplay profile rows duplicated GAM-H-S08, S09 and S10; the Ipsos analysis (GAM-R-S328) was discussed in chapter 9 but not in the table [477, 475, 478, 479] | supported as stated | One record per study, with each publication's source attached; the Ipsos analysis added; a 'What it can support' column (causal for the tested stimulus, association, descriptive, not assessable) recorded apart from sponsor; the studies figure now counts the same records as the table. | Checked against the data and primary sources before any change | data/study-evidence-overlay.json, build/instruments.mjs, scripts/derivations.mjs, chapters/09-outcomes.md |
| R3-06 | The findings are not turned into buying decisions by type of buy, and the supply overlap stops at do not sum | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | Chapters 3 to 6 and 12 hold the routes, reporting and limits for each type of buy [119, 187, 1, 2, 280] | supported with qualification | Chapter 12 adds a table of four common buys (intrinsic placements, rewarded and other app formats, a branded world, streams and communities) with the buying route and reporting documented and what to ask for, each from sources already cited; the 'usually bought for' column is labelled as this review's reading. It adds what to ask two networks that claim the same titles. | Checked against the data and primary sources before any change | chapters/12-implications.md |
| R3-07 | The paper spends too much of its reader-facing space on its own process; the abstract runs to about 1,150 words | OpenAI round 3, a critical review of the draft run by the author with another company's AI model | Word count of the abstract before the change: 1,125 | supported as stated | The abstract was rewritten to lead with the three buying tasks and what can be bought and checked, with the process cut to one paragraph; it is about 870 words. The review history sits in the method chapter and the ledger. | Checked against the data and primary sources before any change | chapters/00-abstract.md, chapters/13-method.md |
| IR3-BANDS | Band names Front rank, Strong, Mid and Narrow | Internal review round 3 (series alignment with the audio paper): band names that describe documentation, as in the audio paper | data/scoring-rules.json composite_index.bands; the thresholds (+12, +1, -8) are unchanged | supported as stated | Bands renamed Most documented, Well documented, Partly documented and Least documented; bands_rule records the old names and that the thresholds did not change; chapters, plain edition and stat tokens follow. | A band describes how much is documented, not a rank | data/scoring-rules.json, chapters/00-abstract.md, chapters/11-matrix.md, chapters/13-method.md, plain.md |
| IR3-SECOND-PASS | One scoring pass; no inter-rater statistic | Internal review round 3 (series alignment with the audio paper): blind second scoring pass on the rows nearest a band line, then adjudication | data/second-pass.json and data/second-pass.csv: 10 rows, 115 cells; 98 exact (85%), 106 of 106 within one point, quadratic-weighted kappa 0.90, 3 cells n/e in one pass only; Kidoz and Venatus would move from Partly to Well documented on the second pass [69, 30, 400, 670, 686, 84] | supported with qualification | The 5 material disagreements were adjudicated against the deciding sources: none showed a factual, anchor-application or convention error in the first pass, so no score changed. Kidoz and Venatus are marked scorer-sensitive (†); chapters 11 and 13 and the plain edition report the pass. | A blind second pass tests the scorer, which the reweighting tests cannot | data/second-pass.json, build/build-matrix.mjs, chapters/11-matrix.md, chapters/13-method.md, plain.md |
| IR3-FLAGS | Band-sensitive rows marked only for missing evidence | Internal review round 3 (series alignment with the audio paper): provisional flags beside each band | capability-matrix.json band_sensitive_missing and band_sensitive_scorer | supported as stated | Matrix, quotient figure, CSV and Markdown export mark a provisional band ± (moves when missing evidence is filled) or † (a blind second scorer would move it); stats carry q_sensitive_scorer and q_provisional; the second pass is published as data/second-pass.csv and data/second-pass.json. | Readers should see which bands are provisional and why | build/build-matrix.mjs, build/instruments.mjs, scripts/make-figures.mjs, build/make-csv.mjs, build/datasets.mjs, scripts/stats.mjs |
| IR3-TEXT | Headlines, abstract and method notes | Internal review round 3 (series alignment with the audio paper): critiques likely to recur | Audio paper review rounds 4 and 5 | supported as stated | Abstract opens with a brief for buyers and says the paper audits the public proof, not whether games advertising works (also in the hero summary); chapter 12 scopes the H2 and H4 verdicts to their registered measures; the method says grade caps fold evidence type into the score and that the 20-million-monthly to 5-million-daily equivalence is an assumption; the interests line reads 'No vendor funded, reviewed or saw the paper before publication'. | Bound every headline to what was tested | chapters/00-abstract.md, chapters/11-matrix.md, chapters/12-implications.md, chapters/13-method.md, paper.config.json |
| IR3-PRINT | Matrix band label in upper case at .58rem; print matrix subtitles at 6.5pt; evidence panel led with the first wording of a claim | Internal review round 3 (series alignment with the audio paper): print fixes and the claim-display fix shared by the attention and audio sessions | build/paper.template.html | supported as stated | Band label not upper-cased, .6rem, letter-spacing .01em; print .mx-sg at 7pt; the evidence panel, source backlinks and search index show the final wording, with the first wording as As first written when it differs. | Long band names and the 7pt print floor; corrected wording must lead | build/paper.template.html |
| IR3-CUTOFF | Research cutoff 30 September 2026 | Internal review round 3 (series alignment with the audio paper): the cutoff is the publication date, with a freeze recheck that day | Freeze recheck on 1 October 2026: AdInMo Ltd unchanged (strike-off suspended 17 September 2026); MRC lists unchanged, Trivver still the only gaming entry; the Roblox Terms and Creator Hub still give 30 September 2026 and 1 January 2027; Xbox cloud ad test still an Insider test; no DSP named for EA or added on Roblox's brands page; Zynga Ads+ not live; no Discord S-1 and Discord's ad pages unchanged; no new market event, MRC accreditation or IAB Tech Lab in-game specification [15, 16, 57, 190, 48] | supported as stated | Research cutoff and publication date set to 1 October 2026; the registered scoring period (to 2 October 2026) is unchanged and the method chapter says its last day is not covered. | The cutoff must not run past the date the evidence was last checked | paper.config.json, chapters/00-abstract.md, chapters/06-measurement.md, chapters/13-method.md, plain.md |
| R4-01 | Profiles showed six Super League capabilities as grade A and Xbox targeting as live after the matrix corrected them | OpenAI round 4, a final critical review of the draft run by the author with another company's AI model | data/profiles.json against data/capability-matrix.json: 7 capability records disagreed on grade or availability | supported as stated | Profiles now take each scored cell's grade and availability from the matrix, so the two cannot disagree; the 7 records now match. | Checked against the build and sources before any change | build/build-matrix.mjs |
| R4-02 | Roblox opportunity: a published EDO result would fill the gap left by no independent outcome study | OpenAI round 4, a final critical review of the draft run by the author with another company's AI model | EDO is a commercial measurement partner (GAM-ROBLOX-S22); the paper's independence test needs documented disinterested funding and a stated design [486, 488] | supported as stated | The line now says a published EDO result could add public outcome evidence, meeting the independence test only with disclosed funding and a stated design. | Checked against the build and sources before any change | evidence/swot/roblox.json |
| R4-03 | Roblox threat: part of its programmatic supply runs through Google | OpenAI round 4, a final critical review of the draft run by the author with another company's AI model | Magnite and Index Exchange sellers.json list a supply connection (GAM-ROBLOX-S42, S43); no delivery was observed [539, 573] | supported as stated | The line now calls it a declared supply connection through Google, with no delivery observed through that route. | Checked against the build and sources before any change | evidence/swot/roblox.json |
| R4-04 | Chapter 12 left the Roblox revenue-share dates to be rechecked at the 2 October freeze | OpenAI round 4, a final critical review of the draft run by the author with another company's AI model | Freeze recheck of 1 October 2026: the Terms still say 30 September 2026 and the Creator Hub 1 January 2027 (GAM-E-S10, GAM-E-S11) [57, 190] | supported as stated | Replaced with a dated statement: on 1 October 2026 the two documents still conflicted and no source showed which date applies. | Checked against the build and sources before any change | chapters/12-implications.md |
| R4-05 | Chapter 13 and the pipeline and quotient figures still said single scorer or no second scorer | OpenAI round 4, a final critical review of the draft run by the author with another company's AI model | data/second-pass.json: 10 rows re-scored blind | supported as stated | Now one full scoring pass followed by blind AI re-scoring of the rows nearest a band line, kept distinct from independent human rating. | Checked against the build and sources before any change | chapters/13-method.md, scripts/make-figures.mjs |
| R4-06 | Chapter links in the Markdown reading edition pointed to anchors its headings did not carry | OpenAI round 4, a final critical review of the draft run by the author with another company's AI model | dist/gaming-advertising-2026.md: links to #implications, #matrix and 11 other chapter ids with no matching anchor | supported as stated | Each chapter heading in the Markdown edition now carries an explicit anchor with the chapter id. | Checked against the build and sources before any change | build/render.mjs |
### Data downloads
- `gaming-advertising-2026.pdf` — the print edition: A4, with page numbers and chapter bookmarks
- `gaming-advertising-2026.md` — the reading edition in Markdown, with numbered references
- `data/source-register.csv` — every source with publication, event and access dates, type, geography, funding, access limits and reliability notes
- `data/claim-ledger.csv` — every material claim with sources, passage locator, scope, evidence class, review outcome, confidence and final wording
- `data/capability-matrix-wide.csv` — vendors × dimensions, scores only (0–5, n/e, n/a)
- `data/capability-matrix-long.csv` — one row per cell with availability, evidence grade, date, rationale and sources
- `data/capability-dimensions.csv` — the dimension definitions and anchors
- `data/pq.csv` — Play Quotient per vendor: group means, coverage, band, the reweightings and every dimension score
- `data/pq-method.csv` — PQ groups, weights, shrinkage, field mean, band thresholds, and the reweighting, shrinkage and missing-evidence tests
- `data/pq-missing-evidence.csv` — every scored cell turned into n/e one at a time, with the PQ before and after
- `data/swot.csv` — every SWOT line for the profiled vendors with its basis (record, vendor documentation, vendor claim, inference) and sources
- `data/deep-profiles.csv` — the deep-profile summaries: what is sold, buyers, earnings, dependencies, fit, limits
- `data/deep-profile-records.csv` — every profiled capability, performance claim, financial figure and timeline event with sources
- `data/study-evidence.csv` — studies with design, sample, sponsor and limits
- `data/corrections-ledger.csv` — every change made after verification and challenge passes, with implementation and verification
- `data/vendor-universe.csv` — the reasoned vendor sample with scope group, owner, formats, routes and status
- `data/agentic-map.csv` — AI and agentic capabilities with status labels and observability
- `data/rules.csv` — rules and governance instruments by jurisdiction
- `data/estimates.csv` — market-size estimates with inclusions, exclusions and combinability
- `data/decision-tool.csv` — the decision tool rows
- `data/product-register.csv` — the product register: every scored or listed product with its purchased service, operating entity, owner, cohort, inclusion reason, aliases and US availability
- `data/hypotheses.csv` — the hypotheses registered before extraction, with the rules for supported, contradicted and unresolved
- `data/figure-data.json` — the canonical dataset behind every data figure
- `data/derived-checks.json` — every derived number with its formula and inputs
- `data/stats.json` — every count the prose and figures quote about the research itself, computed from the ledgers
- `data/second-pass.csv` — the blind second scoring pass: one row per re-scored cell with both passes’ scores, grades and reasons, and the adjudication of each material disagreement
- `data/second-pass.json` — the same cells with both passes’ sources, each re-scored row’s band under both passes, and the agreement statistics’ inputs
- `data/scoring-rules.json` — dimension definitions, anchors, cell rules and the PQ definition, fixed before scoring
### Rebuilding the paper
The package holds everything needed to rebuild the paper. `paper.md` lists the chapters; every citation and claim marker must resolve in `data/`, or the build fails. The research records are in `evidence/`; the hypotheses, scoring rules, product register and scores are in `data/`. `scripts/check.mjs` checks references, cells and counts; `README.md` is the full guide.
## References
The 753 sources cited in this paper, numbered in order of first citation. The full register of 982 sources opened during the research, with access dates, access limits and reliability notes, is in data/source-register.csv.
[1] Roblox Advertising Terms (Managed Campaigns). Roblox Advertising Services, LLC (Roblox Support), 2025-01-08. https://en.help.roblox.com/hc/en-us/articles/33703196637204-Roblox-Advertising-Terms-Managed-Campaigns Accessed 2026-09-28.
[2] Twitch ads - Discover marketing solutions for brands. Amazon Ads, undated (live page). https://advertising.amazon.com/channels/twitch Accessed 2026-09-28.
[3] 2026 Essential Facts About the U.S. Video Game Industry. Entertainment Software Association (ESA); survey by YouGov; spending data from Circana, 2026-06. https://www.theesa.com/wp-content/uploads/2026/06/2026-Essential-Facts-Booklet-05-27-26.pdf Accessed 2026-09-28.
[4] IAB/PwC Internet Advertising Revenue Report: Full-year 2025 results. IAB and PwC, 2026-04. https://iabhongkong.com/sites/default/files/2026-04/IAB_PwC_Internet_Ad_Revenue_Report_Full_Year_2025_April_2026.pdf Accessed 2026-09-28.
[5] Guide to video game advertising: Trends, tactics, and opportunities for marketers. EMARKETER (Caroline Shalabi), 2024-07-26. https://www.emarketer.com/learningcenter/guides/us-gaming-industry-ecosystem/ Accessed 2026-09-28.
[6] US game ad revenues growth slows. EMARKETER (Sara Lebow), 2024-06-10. https://www.emarketer.com/content/us-game-ad-revenues-growth-slows Accessed 2026-09-28.
[7] US In-Game Ad Revenues Forecast 2024: Mobile Still Dominates as New Ad Formats Attract Brands. EMARKETER (Yoram Wurmser), 2024-05-30. https://www.emarketer.com/content/us-in-game-ad-revenues-forecast-2024 Accessed 2026-09-28. Access: partial extract.
[8] In-Game Advertising 2023: Brands Eye Opportunities as Mobile Rebounds and New Platforms Open Up. EMARKETER (Yoram Wurmser), 2023-05-31. https://www.emarketer.com/content/in-game-advertising-2023 Accessed 2026-09-28. Access: partial extract.
[9] Anzu sellers.json. Anzu (Anzu Virtual Reality Ltd.). https://anzu.io/sellers.json Accessed 2026-09-28.
[10] Frameplay sellers.json. Frameplay, 2025-08-19. https://frameplay.gg/sellers.json Accessed 2026-09-28.
[11] IION sellers.json. iion, 2026-09-28. https://iion.io/sellers.json Accessed 2026-09-28.
[12] Gadsme sellers.json. Gadsme. https://www.gadsme.com/sellers.json Accessed 2026-09-28.
[13] Adverty sellers.json. Adverty AB (publ), 2026-09-27. https://adverty.com/sellers.json Accessed 2026-09-28.
[14] Gameloft sellers.json. Gameloft. https://gameloft.com/sellers.json Accessed 2026-09-28.
[15] Accredited Services: Digital. Media Rating Council. https://mediaratingcouncil.org/accreditation/digital Accessed 2026-09-28.
[16] Services Under Review: Digital. Media Rating Council. https://mediaratingcouncil.org/accreditation/services-under-review/digital Accessed 2026-09-28.
[17] Intrinsic In-Game Advertising Measurement Guidelines 2.0 (Initially released September 2009, Updated August 2022). IAB, IAB Tech Lab and Media Rating Council (IAB Intrinsic In-Game Advertising Measurement Task Force), 2022-08. https://mediaratingcouncil.org/sites/default/files/Standards/In-Game%20Advertising%20Measurement%20Guidelines%2008-31-22.pdf Accessed 2026-09-28.
[18] MRC letter to DoubleVerify: continuing accreditation of DoubleVerify's Pinnacle platform (GI041326). Media Rating Council (George W. Ivie), 2026-04-13. https://mediaratingcouncil.org/sites/default/files/Accreditation%20Letters/DoubleVerify/GI041326_DoubleVerify%20Accreditation%20Letter.pdf Accessed 2026-09-28.
[19] MRC letter to Integral Ad Science: continuing accreditation (GI070926). Media Rating Council (George W. Ivie), 2026-07-09. https://mediaratingcouncil.org/sites/default/files/Accreditation%20Letters/IAS/GI070926_Integral%20Ad%20Science%20Recurring_Google%20_Meta%20Accreditation%20Letter.pdf Accessed 2026-09-28.
[20] Advertising Standards. Roblox Corporation, 2026-06-05. https://en.help.roblox.com/hc/en-us/articles/13722260778260-Advertising-Standards Accessed 2026-09-28.
[21] Microsoft Privacy Statement. Microsoft Corporation, 2026-09. https://www.microsoft.com/en-us/privacy/privacystatement Accessed 2026-09-28.
[22] How to Confirm Your Age Group on Discord. Discord (Discord Support). https://support.discord.com/hc/en-us/articles/30326565624343-How-to-Confirm-Your-Age-Group-on-Discord Accessed 2026-09-28.
[23] Children's Online Privacy Protection Rule; Final Rule (90 FR 16918, FR Doc. 2025-05904). Federal Trade Commission (Federal Register), 2025-04-22. https://www.federalregister.gov/documents/2025/04/22/2025-05904/childrens-online-privacy-protection-rule Accessed 2026-09-28.
[24] Custom brand integrations in gaming. Microsoft Advertising (XBOX Media Solutions), undated. https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/custom-integrations Accessed 2026-09-28.
[25] Electronic Arts Introduces EA Advertising, Launching Brands Directly Into Gameplay and Live Experiences. Electronic Arts Inc. (Business Wire), 2026-06-15. https://news.ea.com/press-releases/press-releases-details/2026/Electronic-Arts-Introduces-EA-Advertising-Launching-Brands-Directly-Into-Gameplay-and-Live-Experiences/ Accessed 2026-09-28.
[26] Where Vision Meets Memory: An Eye-Tracking Study of In-App Ads in Mobile Sports Games with Mixed Visual-Quantitative Analytics. Büyükakgül, Ü. C.; Yüce, A.; Katırcı, H. (Eskisehir Technical University), Journal of Eye Movement Research 18(6):74, 2025-12-10. https://pmc.ncbi.nlm.nih.gov/articles/PMC12733859/ Accessed 2026-09-28.
[27] Play Attention: Calling Focus to Gaming. dentsu Media US (with Lumen Research), 2024-01. https://assets-eu-01.kc-usercontent.com:443/27bd3334-62dd-01a3-d049-720ae980f906/16471205-9146-4b6a-9709-7c5ae7883a9c/Play%20Attention_dentsu%20Media%20US_January%20202436.pdf Accessed 2026-09-28.
[28] New Study Points to Roblox's Positive Impact for Brands. Roblox (Allison McDuffee, Global Head of Brand Insights & Measurement), 2024-05-01. https://about.roblox.com/newsroom/2024/05/new-study-points-to-robloxs-positive-impact-for-brands Accessed 2026-09-28.
[29] Super League Gaming, Inc. Form 8-K (event date 4 October 2021) and Exhibit 99.1 'Super League Gaming Acquires Metaverse Ad Platform Bloxbiz'. Super League Gaming, Inc. (now Super League Enterprise, Inc.), 2021-10-07 (8-K filing date). https://www.sec.gov/Archives/edgar/data/1621672/000165495421010877/slgg_ex991.htm Accessed 2026-09-28.
[30] Super League Enterprise, Inc. Form 10-Q for the quarter ended 30 June 2026. Super League Enterprise, Inc. (SEC EDGAR), 2026-08-14. https://www.sec.gov/Archives/edgar/data/1621672/000143774926028066/slgg20260630_10q.htm Accessed 2026-09-28.
[31] Bidstack Group PLC (in administration): Joint Administrators' statement of proposals (form AM03, filed 10 April 2024). Alvarez & Marsal Europe LLP as Joint Administrators, filed at UK Companies House, 2024-04-10. https://find-and-update.company-information.service.gov.uk/company/04466195/filing-history/MzQxNjUwOTI5MGFkaXF6a2N4/document?format=pdf&download=0 Accessed 2026-09-28.
[32] BIDSTACK LTD (09835625): insolvency. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/09835625/insolvency Accessed 2026-09-28.
[33] Ad Gamesmanship: Advertisers Are Moving Dollars Into This Growing Category (Publishing Insider column). MediaPost (Ray Schultz, columnist), 2024-10-21. https://www.mediapost.com/publications/article/400445/ad-gamesmanship-advertisers-are-moving-dollars-in.html Accessed 2026-09-28.
[34] Game Advertising State of the Nation Survey, Wave 3, July 2023 (report PDF). IAB Australia (Game Advertising Working Group), 2023-07-27. https://assets-oce.mkt.dynamics.com/57dabe3f-beff-ef11-b015-002248942ac9/digitalassets/docs/57d0cf4a-1e7e-f011-b4cc-000d3a6a161d?ts=638913297110000000 Accessed 2026-09-28.
[35] AdPROs Network (panel description). Advertiser Perceptions, 2026. https://www.advertiserperceptions.com/join-ap-ad-pros/ Accessed 2026-09-28.
[36] (Author's work) Gaming Playbook: In-Game Advertising, Brand Safety & Americas GTM. No Fluff Advisory (Evgeny Popov), 2026. https://nofluffadvisory.com/services/playbooks/gaming/ Accessed 2026-09-28.
[37] US Mobile Game Ad Spending Forecasts, Estimates, & Historical Data (Source & Notes). EMARKETER, 2026. https://www.emarketer.com/forecasts/584b26021403070290f93a56/5c76cdc069f80503c46bb0d7/ Accessed 2026-09-28. Access: partial extract.
[38] WPP Media Global Advertising Forecast: Global ad revenue set to reach $1.14 trillion (This Year Next Year, 2025 Global End-of-Year). WPP Media (Kate Scott-Dawkins), 2025-12-10. https://www.wppmedia.com/news/campaigns-report-this-year-next-year-december-2025 Accessed 2026-09-28. Access: partial extract.
[39] State of Gaming 2026: Mobile, PC, & Console Trends (blog). Sensor Tower (Bryan Isagholian), 2026-02. https://sensortower.com/blog/state-of-gaming-2026 Accessed 2026-09-28. Access: partial extract.
[40] Why programmatic in-console game advertising is the next big opportunity for 2026. adWMG (The Drum Network member content), 2025-12-02. https://www.thedrum.com/industry-insight/why-programmatic-in-console-game-advertising-is-the-next-big-opportunity-for-2026 Accessed 2026-09-28.
[41] adWMG Partners with Gadsme to Unlock Mobile In-Game Advertising for Xbox & PlayStation Campaigns. ExchangeWire (adWMG announcement), 2025-05-23. https://www.exchangewire.com/blog/2025/05/23/adwmg-partners-with-gadsme-to-unlock-mobile-in-game-advertising-for-xbox-playstation-campaigns/ Accessed 2026-09-29.
[42] Immersive Ads on Roblox (brands advertising page). Roblox Corporation. https://brands.roblox.com/advertising Accessed 2026-09-28.
[43] Roblox Q2 2026 shareholder letter (Exhibit 99.1 to Form 8-K). Roblox Corporation (SEC EDGAR), 2026-07-30. https://www.sec.gov/Archives/edgar/data/1315098/000162828026051059/ex991-robloxq22026earnin.htm Accessed 2026-09-28.
[44] Advertising FAQ - Learn About Ads on Overwolf. Overwolf, undated (live page). https://www.overwolf.com/ads/faq/ Accessed 2026-09-28.
[45] Ad planning guidelines (Overwolf Native). Overwolf Developers, undated (live docs). https://dev.overwolf.com/ow-native/monetization/advertising/standard-ads/working-with-ads/ Accessed 2026-09-29.
[46] Overwolf Ads Launches Gamer Grid. ExchangeWire (Overwolf Ads announcement), 2026-04-27. https://www.exchangewire.com/blog/2026/04/27/overwolf-ads-launches-gamer-grid/ Accessed 2026-09-29.
[47] Twitch ads ad specifications | Premium Video. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/premium-video Accessed 2026-09-28.
[48] Discord Quests - Advertise Where the World Plays. Discord, undated (live page). https://discord.com/ads/quests Accessed 2026-09-28.
[49] Discord Announces First Mobile Ad Format, Broadening Advertising Opportunities. Discord, 2025-03-20. https://discord.com/press-releases/announcing-quests-advertising-on-mobile Accessed 2026-09-28.
[50] IAB Releases Intrinsic In Game Measurement IIG Guidelines For Public Comment. IAB, 2022-06-14. https://www.iab.com/news/iab-releases-intrinsic-in-game-measurement-iig-guidelines-for-public-comment/ Accessed 2026-09-28.
[51] IAB Releases Intrinsic In-Game (IIG) Measurement Guidelines 2022. IAB, 2022-08-31. https://www.iab.com/insights/iab-releases-intrinsic-in-game-iig-measurement-guidelines/ Accessed 2026-09-28.
[52] Standards & Guidelines (MRC standards library listing). Media Rating Council. https://mediaratingcouncil.org/standards-and-guidelines Accessed 2026-09-28.
[53] King app-ads.txt. King (Activision Blizzard, Microsoft). https://www.king.com/app-ads.txt Accessed 2026-09-28.
[54] Rovio app-ads.txt. Rovio. https://www.rovio.com/app-ads.txt Accessed 2026-09-28.
[55] In Game Advertising, Gameloft for Brands (COMBO!). Gameloft for brands. https://www.gameloft.com/for-brands/in-game-advertising Accessed 2026-09-28.
[56] Ads Manager UI & Targeting Enhancements. Roblox (Developer Forum announcement), 2026-02-04. https://devforum.roblox.com/t/ads-manager-ui-targeting-enhancements/4333400 Accessed 2026-09-29.
[57] Advertising Integrations Terms. Roblox Corporation (Roblox Support), 2026-03-27. https://en.help.roblox.com/hc/en-us/articles/47656162239124-Advertising-Integrations-Terms Accessed 2026-09-28.
[58] Medal Quests FAQ. Medal Support (help centre). https://support.medal.tv/support/solutions/articles/48001265371-medal-quests-faq Accessed 2026-09-29.
[59] Publisher's General Terms and Conditions - Gadsme. Gadsme, 2026-05-19. https://www.gadsme.com/legal/publisher-general-terms-and-conditions Accessed 2026-09-29.
[60] Unity Software Inc. Form 10-K for the fiscal year ended December 31, 2025. Unity Software Inc., 2026-02-11. https://www.sec.gov/Archives/edgar/data/1810806/000181080626000011/unity-20251231.htm Accessed 2026-09-28.
[61] AppLovin Corporation Form 10-K for the fiscal year ended December 31, 2025. AppLovin Corporation, 2026-02-19. https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/app-20251231.htm Accessed 2026-09-28.
[62] Discord Quests Frequently Asked Questions (advertiser FAQ). Discord, undated (live page). https://discord.com/ads/quests-faq Accessed 2026-09-29.
[63] Discord adds lower funnel with outcomes push. Digiday (Krystal Scanlon), 2026-08-21. https://digiday.com/marketing/discord-adds-lower-funnel-with-outcomes-push/ Accessed 2026-09-29.
[64] VIRTUAL SPORT TECHNOLOGY LIMITED (15133297): insolvency and persons with significant control. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/15133297/insolvency Accessed 2026-09-28.
[65] Kidoz Inc. Annual Report on Form 20-F for the fiscal year ended 31 December 2025. Kidoz Inc. (SEC EDGAR, CIK 1318482), 2026-04-29. https://www.sec.gov/Archives/edgar/data/1318482/000149315226019626/form20-f.htm Accessed 2026-09-29.
[66] Kidoz sellers.json. Kidoz (contact address Netanya, Israel), undated (no Last-Modified header). https://kidoz.net/sellers.json Accessed 2026-09-29.
[67] Breaking Through the Noise: A New Era for Kids App Monetization (Kidoz blog). Kidoz (blog feed feed.kidoz.net), 2026-05-25. https://feed.kidoz.net/kidoz/rss/ (item dated 2026-05-25; site path /blog/breaking-through-the-noise-a-new-era-for-kids-app-monetization) Accessed 2026-09-29.
[68] Kidoz home page (Where The World's Leading Brands Reach Kids and Teens; /brands renders the same content). Kidoz Inc., undated (live page). https://www.kidoz.net/ Accessed 2026-09-29.
[69] Kidoz Shield - Brand-safe by design, before the auction. Kidoz Inc., undated (live page). https://www.kidoz.net/privacy-shield Accessed 2026-09-29.
[70] Kidoz Inc. Investor Fact Sheet (June 12, 2026). Kidoz Inc., 2026-06-12. https://investor.kidoz.net/__l5e/assets-v1/6aae055d-f335-40eb-a7e7-a49bc0fc399d/kidoz-investor-fact-sheet-june-2026.pdf Accessed 2026-09-29.
[71] Kidoz Mobile SDK (GitHub: README, Direct/README, Direct/Android/README). Kidoz (Kidoz-SDK organisation), undated (repository pushed 2026-09-28 per GitHub API). https://github.com/Kidoz-SDK/kidoz-mobile-sdk/ Accessed 2026-09-29.
[72] Kidoz For Publishers and Kidoz Network (SDK reach) pages. Kidoz Inc., undated (live pages). https://www.kidoz.net/publishers ; https://www.kidoz.net/sdk-reach Accessed 2026-09-29.
[73] Frameplay and iion Form Strategic Partnership to Create Global In-Game Advertising Powerhouse. Frameplay; iion (PR Newswire), 2025-06-16. https://www.prnewswire.com/news-releases/frameplay-and-iion-form-strategic-partnership-to-create-global-in-game-advertising-powerhouse-302481336.html Accessed 2026-09-28.
[74] In-game ad firm Frameplay teams with others, including competitors, to launch its own exchange. Alexander Lee, Digiday, 2024-10-10. https://digiday.com/marketing/in-game-ad-firm-frameplay-teams-with-others-including-competitors-to-launch-its-own-exchange/ Accessed 2026-09-28.
[75] Miniclip app-ads.txt. Miniclip. https://miniclip.com/app-ads.txt Accessed 2026-09-28.
[76] Outfit7 app-ads.txt. Outfit7. https://outfit7.com/app-ads.txt Accessed 2026-09-28.
[77] Zynga app-ads.txt. Zynga (Take-Two Interactive). https://www.zynga.com/app-ads.txt Accessed 2026-09-28.
[78] Gameloft app-ads.txt. Gameloft. https://gameloft.com/app-ads.txt Accessed 2026-09-28.
[79] Adverty AB (publ) Delårsrapport januari-juni 2026 and release 'Adverty levererar stark tillväxt, högre marginaler och första positiva EBT-kvartalet'. Adverty AB (publ), 2026-07-28. https://storage.mfn.se/f5cff0f3-abca-466c-8fe8-375d2dba9f5d/adverty-q2-2026-sve.pdf Accessed 2026-09-28.
[80] Adverty SDK Change Log. Adverty AB (publ), 2026-09-09 (latest entry). https://adverty.com/sdk-change-log Accessed 2026-09-29.
[81] Brand Safety Policy. Adverty AB (publ), undated. https://adverty.com/brand-safety-policy/ Accessed 2026-09-28.
[82] Lihuhu x Adverty: How Revenue Increased by 15–30% with Seamless In-Play Advertising. Adverty AB (publ), 2026-01-14. https://adverty.com/news/lihuhu-x-adverty Accessed 2026-09-29.
[83] Gaming & Entertainment Advertising Growth Engine (Venatus home page). Venatus, undated (live page). https://www.venatus.com/ Accessed 2026-09-28.
[84] Away-from-the-game advertising for brands (Venatus). Venatus, undated (live page). https://www.venatus.com/advertisers/ad-placements/away-from-the-game Accessed 2026-09-28.
[85] US Total Game Ad Spending (forecast series page, Source & Notes). EMARKETER. https://www.emarketer.com/forecasts/5c76d87669f80503c46bb0e2/654404f0d05bf2db4beccd01/ Accessed 2026-09-29. Access: partial extract.
[86] Electronic Arts is betting that in-game ads can out-earn CTV. Digiday (Krystal Scanlon), 2026-07-03. https://digiday.com/marketing/electronic-arts-is-betting-that-in-game-ads-can-out-earn-ctv/ Accessed 2026-09-29.
[87] US Esports Ad Spending Forecasts, Estimates, & Historical Data (Source & Notes). EMARKETER, 2026. https://www.emarketer.com/forecasts/584b26021403070290f93a56/5c76d38069f80503c46bb0e0/ Accessed 2026-09-28. Access: partial extract.
[88] US Gaming Video Content Ad Spending Forecasts, Estimates, & Historical Data (Source & Notes). EMARKETER, 2026. https://www.emarketer.com/forecasts/584b26021403070290f93a56/5c76cf417ddc4308487caebd/ Accessed 2026-09-28. Access: partial extract.
[89] This Year Next Year 2026 Midyear Forecast: Global ad revenue to grow 8.9% to $1.3 trillion. WPP Media, 2026-06-16. https://www.wppmedia.com/news/report-this-year-next-year-midyear-2026 Accessed 2026-09-28. Access: partial extract.
[90] This Year Next Year Midyear June 2026 (report page). WPP Media (Kate Scott-Dawkins), 2026-06-16. https://www.wppmedia.com/thought-leadership/this-year-next-year/midyear-2026 Accessed 2026-09-28. Access: partial extract.
[91] Game Advertising State of the Nation 2023 (resource page). IAB Australia (Game Advertising Working Group), 2023-07-27. https://iabaustralia.com.au/resource/game-advertising-state-of-the-nation-2023/ Accessed 2026-09-28.
[92] FAQ on video game advertising: Market size, format analysis, and how to find players in 2026. EMARKETER (Peter Clark, EMARKETER Editors), 2026-01-10. https://www.emarketer.com/content/faq-on-video-game-advertising--market-size--ad-formats--how-find-players-2026 Accessed 2026-09-28.
[93] Video Game Industry Draws Less Than 5% of Global Ad Spend Despite 3.4 Billion Players, Dentsu Report Finds (EXCLUSIVE). Variety (Jennifer Maas), 2025-07-01. https://variety.com/2025/gaming/news/video-game-industry-gaming-ad-marketing-spend-report-1236444014/ Accessed 2026-09-28.
[94] Dentsu 2025 Gaming Trends Report (landing page). dentsu, 2025-07. https://www.dentsu.com/reports/dentsu_2025_gaming_trends_report Accessed 2026-09-28. Access: partial extract.
[95] Programmatic In-Game Advertising | Gadsme. Gadsme, undated. https://www.gadsme.com/programmatic Accessed 2026-09-29.
[96] Helium9 Games app-ads.txt. Helium9 Games, 2026-08-10. https://helium9games.com/app-ads.txt Accessed 2026-09-28.
[97] Gadsme app-ads.txt. Gadsme. https://www.gadsme.com/app-ads.txt Accessed 2026-09-29.
[98] 2026 Outlook: A Snapshot of U.S. Ad Spend, Opportunities, and Strategies for Growth. IAB, 2026-01-28. https://www.iab.com/insights/2026-outlook/ Accessed 2026-09-28. Access: partial extract.
[99] Game ad revenues grow, but spending falls behind consumer attention. EMARKETER (Marisa Jones), 2025-12-26. https://www.emarketer.com/content/game-ad-revenues-grow--spending-falls-behind-consumer-attention Accessed 2026-09-28. Access: access error; text or metadata via another route.
[100] US Digital Gamers and Penetration Forecasts, Estimates, & Historical Data (Source & Notes). EMARKETER, 2026. https://www.emarketer.com/forecasts/584b26021403070290f93a37/5989ed80aaf61c0ed0302dc9/ Accessed 2026-09-28. Access: partial extract.
[101] Gaming Advertising 2024: Top Takeaways from our Latest Report. Advertiser Perceptions (John Bishop, VP Business Intelligence), 2024-09-09. https://www.advertiserperceptions.com/gaming-advertising-2024-top-takeaways-from-our-latest-report/ Accessed 2026-09-28.
[102] Total global ad revenue grew to $1.14 trillion in 2025: WPP Media. Marketing Brew (Morning Brew Inc.), 2025-12-09. https://www.marketingbrew.com/stories/2025/12/09/total-global-ad-revenue-grew-nearly-9-in-2025-wpp-media Accessed 2026-09-28.
[103] Global entertainment and media industry revenues to hit US$3.5 trillion by 2029, driven by advertising, live events, and video games: PwC Global Entertainment & Media Outlook. PwC (release distributed via PR Newswire), 2025-07-24. https://www.prnewswire.com/in/news-releases/global-entertainment-and-media-industry-revenues-to-hit-us3-5-trillion-by-2029--driven-by-advertising-live-events-and-video-games-pwc-global-entertainment--media-outlook-302511891.html Accessed 2026-09-28.
[104] Global entertainment and media advertising revenues to hit US$1.4 trillion in 2030 - as global box office continues recovery (PwC release, as reproduced). PwC (reproduced by BizTechReports), 2026-07-09. https://www.biztechreports.com/news-archive/2026/7/1/global-entertainment-and-media-advertising-revenues-to-hit-us14-trillion-in-2030-as-global-box-office-continues-recovery-pwc-july-04-2026 Accessed 2026-09-28.
[105] 2026 global games market key numbers (Newzoo article). Newzoo, 2026-08-25. https://newzoo.com/articles/2026-global-games-market-key-numbers Accessed 2026-09-28. Access: partial extract.
[106] Game Advertising State of the Nation 2022 (report PDF, wave 2). IAB Australia (Game Advertising Working Group), 2022-07-05. https://assets-oce.mkt.dynamics.com/57dabe3f-beff-ef11-b015-002248942ac9/digitalassets/docs/4a2400ac-3a7e-f011-b4cc-000d3a6a161d?ts=638913419020000000 Accessed 2026-09-29.
[107] Changing the Game: How Games Advertising Powers Performance (report PDF). IAB and Advertiser Perceptions; sponsored by Admazing and Venatus, 2024-03. https://www.iab.com/wp-content/uploads/2024/03/IAB_AP_Changing_The_Game_March_2024.pdf Accessed 2026-09-28.
[108] New IAB Study Reveals Games Advertising Among Top Three Investment Growth Areas Alongside Digital Video. IAB, 2024-03-26. https://www.iab.com/news/new-iab-study-reveals-games-advertising-among-top-three-investment-growth-areas/ Accessed 2026-09-28.
[109] AP3: Gaming Advertising - The Next Level (podcast page). Advertiser Perceptions, 2025-07-30. https://www.advertiserperceptions.com/ap3-gaming-advertising-the-next-level/ Accessed 2026-09-28. Access: partial extract.
[110] The 2026 Industry Pulse Report (web summary). Integral Ad Science (IAS), 2025-12-08. https://integralads.com/insider/the-2026-industry-pulse-report/ Accessed 2026-09-28. Access: partial extract.
[111] The Business Case for Gaming: A Guide to Play for CMOs and Their Teams. IAB, 2026-08-05. https://www.iab.com/guidelines/the-business-case-for-gaming/ Accessed 2026-09-28. Access: partial extract.
[112] DS01(ef) Application to strike off, ADINMO LTD. ADINMO LTD (filed with Companies House), 2026-09-07. https://find-and-update.company-information.service.gov.uk/company/SC648449/filing-history/MzU0Mjc4Njk2MWFkaXF6a2N4/document?format=pdf&download=0 Accessed 2026-09-29.
[113] AdInMo website and ad-serving hostname availability check. Author's access record (curl with browser user agent; Windows nslookup), 2026-09-29 (check date). https://www.adinmo.com/ Accessed 2026-09-29. Access: partial extract.
[114] News & Press. Trivver, Inc., undated (latest item dated 05.12.2025). https://trivver.com/news-press.html Accessed 2026-09-29.
[115] iion home page. iion, 2025. https://www.iion.io/ Accessed 2026-09-28.
[116] iion advertisers page. iion, 2025. https://www.iion.io/advertisers Accessed 2026-09-28.
[117] COMBO! 2026: Reaching 1.3 Billion Gamers Worldwide. Gameloft for brands, 2026-03. https://www.gameloft.com/for-brands/news/combo-2026-reaching-13-billion-gamers-worldwide Accessed 2026-09-28.
[118] In-Game Advertising for Advertisers (Gadsme). Gadsme. https://www.gadsme.com/advertisers Accessed 2026-09-28.
[119] Exchange (enum), Display & Video 360 API v4. Google for Developers, 2026-06-10. https://developers.google.com/display-video/api/reference/rest/v4/Exchange Accessed 2026-09-28.
[120] In-Game Advertising Platform (StackAdapt). StackAdapt, 2022. https://www.stackadapt.com/in-game-advertising Accessed 2026-09-28. Access: partial extract.
[121] AgenticAdvertising.org public agent registry (GET /api/registry/agents). AgenticAdvertising.org, 2026-09-28. https://agenticadvertising.org/api/registry/agents Accessed 2026-09-28. Access: partial extract.
[122] Anzu Unlocks Performance-Driven Intrinsic In-Game Advertising for Programmatic DSPs. Anzu (PRNewswire-PRWeb), 2026-02-09. https://www.prweb.com/releases/anzu-unlocks-performance-driven-intrinsic-in-game-advertising-for-programmatic-dsps-302682050.html Accessed 2026-09-28.
[123] Anzu advertisers page. Anzu. https://www.anzu.io/advertisers Accessed 2026-09-28. Access: partial extract.
[124] Ad Spaces | Frameplay SDK (Dev Guidelines). Frameplay, 2024-07-18. https://docs.frameplay.gg/guide/ad-spaces/ Accessed 2026-09-29.
[125] Adverty SDK documentation, current version 5.3. Adverty AB (publ), 2026 (version 5.3; changelog 5.3.0 dated 2026-09-09). https://adverty.com/5.3/documentation Accessed 2026-09-29.
[126] General Terms and Conditions (Gameloft advertising; linked as "Advertising Term & Conditions"). Gameloft SE and affiliates, 2021-05-31 (from the file name; linked from the live site footer). https://forbrands.assets.gameloft.com/assets/documents/Terms%20and%20Conditions%20Glad%20New%20Version_31052021_vGameloft.pdf Accessed 2026-09-29.
[127] immersiion game advertising platform. iion, undated. https://www.iion.io/immersiion-advertising-platform Accessed 2026-09-28.
[128] Super League Enterprise, Inc. Form 10-K for the fiscal year ended December 31, 2025. Super League Enterprise, Inc., 2026-03-31. https://www.sec.gov/Archives/edgar/data/1621672/000143774926010734/slgg20251231_10k.htm Accessed 2026-09-28.
[129] Admazing | Powerful Mobile Gaming Ads for Brands (home page). Admazing Co., undated (footer "© 2026 Admazing Co."). https://admazing.com/ Accessed 2026-09-29.
[130] Admazing Company (About Us) page. Admazing Co., undated (timeline to 2023). https://admazing.com/company Accessed 2026-09-29.
[131] Unlock Precise Audience Targeting and Lower-Funnel Measurement for Your In-Game Ad Campaigns. Anzu (Simon Whittick, Director of Product Marketing), 2025-09-09. https://www.anzu.io/blog/q3-product-updates Accessed 2026-09-29.
[132] Anzu Brings Intrinsic In-Game Ads To AAA Football Hit UFL. Anzu, 2026-05-13. https://www.anzu.io/news/anzu-brings-intrinsic-in-game-ads-to-aaa-football-hit-ufl Accessed 2026-09-28.
[133] Playgama and Anzu Bring Real-World Ads to Web Games. Anzu, 2025-11-19. https://www.anzu.io/news/playgama-and-anzu-bring-real-world-ads-to-web-games Accessed 2026-09-28.
[134] Gameloft Boosts Realism And Revenue, Expanding Intrinsic Ads To Asphalt 8: Airborne And Dragon Mania Legends. Anzu, 2025-11-26. https://www.anzu.io/case-studies/gameloft-case-study Accessed 2026-09-29.
[135] Non-Disruptive Doesn't Mean Non-Interactive. Anzu, 2026-08-11. https://www.anzu.io/blog/non-disruptive-doesnt-mean-non-interactive Accessed 2026-09-29.
[136] NexTide Media Partners with Frameplay to Deliver Scalable, Brand-Safe Livestream Advertising. NexTide Media (PR Newswire), 2026-06-09. https://www.prnewswire.com/news-releases/nextide-media-partners-with-frameplay-to-deliver-scalable-brand-safe-livestream-advertising-302794787.html Accessed 2026-09-29.
[137] Frameplay advertisers page. Frameplay. https://frameplay.com/advertisers/ Accessed 2026-09-28. Access: partial extract.
[138] How Voodoo Achieved 7X Revenue Growth Through Direct Integration (case study). iion, 2025-11-20. https://www.iion.io/case-studies/how-voodoo-achieved-7x-revenue-growth-through-direct-integration Accessed 2026-09-29.
[139] Gadsme teams up with Zynga to bring immersive in-game ads to select titles. Isa Muhammad, PocketGamer.biz, 2025-10-23. https://www.pocketgamer.biz/gadsme-teams-up-with-zynga-to-bring-immersive-in-game-ads-to-select-titles/ Accessed 2026-09-28. Access: partial extract.
[140] Sports Interactive ink three-year exclusive partnership with Gadsme. Sports Interactive (SEGA), 2025-09-29. https://www.sports-interactive.com/news/sports-interactive-ink-three-year-exclusive-partnership-gadsme Accessed 2026-09-28.
[141] Football Manager 26 on Steam. Valve (Steam store listing), 2025-11-04 (release date shown). https://store.steampowered.com/app/3551340/ Accessed 2026-09-29.
[142] In-game advertising : L'Agence Bradford et Gadsme lancent une campagne programmatique inedite pour Easy Cash. Minted (article 'realise en collaboration avec L'Agence Bradford et Gadsme'), 2025-12-11. https://www.mntd.fr/in-game-advertising-lagence-bradford-et-gadsme-lancent-une-campagne-programmatique-inedite-pour-easy-cash/ Accessed 2026-09-29.
[143] Adverty AB (publ) Årsredovisning 2025. Adverty AB (publ), org. nr. 559100-5250, 2026-04. https://storage.mfn.se/bfe211b8-610a-4bf8-82d3-011282e32513/adverty-arsredovisning-2025.pdf Accessed 2026-09-28.
[144] Super League Signs Definitive Agreement to Acquire Misfits Ads Division, Expanding Revenue and Accelerating Path to Positive Adjusted EBITDA. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-03-18. https://ir.superleague.com/news-events/press-releases/detail/291/super-league-signs-definitive-agreement-to-acquire-misfits Accessed 2026-09-29.
[145] Supply-path file checks: sellers.json, ads.txt and app-ads.txt at superleague.com and superbiz.gg. This review (direct observation), 2026-09-29. https://www.superleague.com/sellers.json Accessed 2026-09-29.
[146] Super League Acquires Stake in Top Roblox Experience, My Avatar!. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-04-09. https://ir.superleague.com/news-events/press-releases/detail/293/super-league-acquires-stake-in-top-roblox-experience-my Accessed 2026-09-29.
[147] Admazing supply-path files (admazing.com and admazing.co sellers.json, ads.txt, app-ads.txt; exchange.admazing.co/sellers.json). Admazing Co. (observed HTTP responses). https://admazing.com/sellers.json Accessed 2026-09-29. Access: partial extract.
[148] AdInMo launches player-first hybrid monetisation tech as a new SaaS platform for mobile developers. PocketGamer.biz, 2025-10-24. https://www.pocketgamer.biz/adinmo-launches-player-first-hybrid-monetisation-tech-as-a-new-saas-platform-for-mobile-developers/ Accessed 2026-09-29.
[149] Companies Act 2006, section 1004: Circumstances in which application not to be made: activities of company. UK Government (legislation.gov.uk), in force 2009-10-01. https://www.legislation.gov.uk/ukpga/2006/46/section/1004 Accessed 2026-09-29.
[150] ADINMO LTD (SC648449): company overview, filing history and officers. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/SC648449/filing-history Accessed 2026-09-29.
[151] Striking off action suspended (SOAS(A)), SC648449 ADINMO LTD. UK Companies House (Registrar), 2026-09-17. https://find-and-update.company-information.service.gov.uk/company/SC648449/filing-history/MzU0NTE4Nzg0NmFkaXF6a2N4/document?format=pdf&download=0 Accessed 2026-09-29.
[152] Internet Archive CDX capture index for www.adinmo.com and adinmo.com subdomains. Internet Archive, live index. https://web.archive.org/cdx/search/cdx?url=www.adinmo.com&matchType=host&from=20260601&to=20260930&fl=timestamp,original,statuscode Accessed 2026-09-29.
[153] Strike off your limited company from the Companies Register. GOV.UK (Companies House guidance), live page. https://www.gov.uk/strike-off-your-company-from-companies-register Accessed 2026-09-29.
[154] Services Under Review: Digital (Internet Archive capture of 11 April 2026). Media Rating Council (captured by the Internet Archive), 2026-04-11 (capture). http://web.archive.org/web/20260411120728/https://www.mediaratingcouncil.org/accreditation/services-under-review/digital Accessed 2026-09-29.
[155] Trivver Launches In-Game Advertising Platform on Google Cloud Marketplace. Trivver (GlobeNewswire release), 2026-06-11. https://www.globenewswire.com/news-release/2026/06/11/3310681/0/en/Trivver-Launches-In-Game-Advertising-Platform-on-Google-Cloud-Marketplace.html Accessed 2026-09-29.
[156] In-game Advertising (use case page). Trivver, Inc., undated (live). https://trivver.com/use-case/in-game-advertising.html Accessed 2026-09-29.
[157] Trivver supply-path files (trivver.com/sellers.json, /ads.txt, /app-ads.txt). Trivver, Inc. (observed HTTP responses). https://trivver.com/sellers.json Accessed 2026-09-29. Access: partial extract.
[158] The Greatest Mobile Game Publishers Team Up to Create COMBO! The Gaming Media Network. Gameloft, 2022-09-20. https://www.gameloft.com/newsroom/combo-gaming-media-network Accessed 2026-09-28.
[159] In-game advertising firm Gadsme opens first US office. Isa Muhammad, PocketGamer.biz, 2024-12-04. https://www.pocketgamer.biz/in-game-advertising-firm-gadsme-opens-first-us-office/ Accessed 2026-09-29.
[160] ARROVA、Gadsmeのインゲーム広告商品の取り扱いを開始. ARROVA Inc. (Hakuhodo DY ONE subsidiary), 2026-05-13. https://www.arrova.co.jp/2026/05/13/news/gadsme-ingame/ Accessed 2026-09-29.
[161] Super League Completes Acquisition of Misfits Ads Division, Strengthening Revenue Foundation and Advancing Path to Profitability. Super League Enterprise, Inc. (GlobeNewswire; ir.superleague.com), 2026-05-06. https://ir.superleague.com/news-events/press-releases/detail/297/super-league-completes-acquisition-of-misfits-ads-division Accessed 2026-09-28. Access: partial extract.
[162] Lihuhu app-ads.txt. Lihuhu Pte. Ltd.. https://lihuhugames.com/app-ads.txt Accessed 2026-09-29.
[163] How SUPERPRETZEL Baked Up a 29-Point Brand Lift with Intrinsic In-Game Ads. Anzu (Nick Woodford), 2025-03-25. https://www.anzu.io/case-studies/superpretzel-case-study Accessed 2026-09-29.
[164] NBCU Adds In-Game Ads To Its Sales Roster Through Partnership With Anzu. AdExchanger (Anthony Vargas), 2022-03-02. https://www.adexchanger.com/gaming/nbcu-adds-in-game-ads-to-its-sales-roster-through-partnership-with-anzu/ Accessed 2026-09-28.
[165] Gameloft for brands at POSSIBLE Miami 2026. Gameloft for brands, 2026-04/05 (no visible date; embedded timestamps 2026-04-23 to 2026-05-07). https://www.gameloft.com/for-brands/news/gameloft-for-brands-at-possible-miami-2026 Accessed 2026-09-29.
[166] About iion. iion, undated. https://www.iion.io/about Accessed 2026-09-29.
[167] Voodoo app-ads.txt. Voodoo. https://voodoo.io/app-ads.txt Accessed 2026-09-29.
[168] Frameplay, Adverty, and AdInMo Unveil Groundbreaking In-Game Advertising Partnership. Frameplay Corp, 2024-10-09. https://frameplay.com/resource/frameplay-adverty-and-adinmo-unveil-groundbreaking-in-game-advertising-partnership/ Accessed 2026-09-29.
[169] Frameplay resources feed (WordPress REST API, resources_post, 30 newest), including the post "Level Up: Pioneering Ad-Tech Solutions Provider Frameplay Announces Executive Leadership Changes...". Frameplay, 2025-06-04. https://frameplay.com/wp-json/wp/v2/resources_post?per_page=30 Accessed 2026-09-29.
[170] In-Game Advertising Platform (Gadsme home page). Gadsme. https://www.gadsme.com/ Accessed 2026-09-28.
[171] Anzu Marketplace Terms of Use (PDF embedded in anzu.io/anzu-marketplace-terms-of-use). Anzu Virtual Reality Ltd., 2026-07-12. https://www.anzu.io/hubfs/Anzu%20Market%20Place%20Template-%20July%2012%202026%20(1).pdf Accessed 2026-09-29.
[172] Exhibit 10.1: Exclusive Brand Partnership Agreement between Esports Now, LLC and Super League Enterprise, Inc., dated May 1, 2026. Super League Enterprise, Inc. and Esports Now, LLC (SEC EDGAR exhibit), 2026-05-06. https://www.sec.gov/Archives/edgar/data/1621672/000143774926015035/ex_957440.htm Accessed 2026-09-29.
[173] Ad Spaces | Frameplay SDK (Unity). Frameplay, 2024-07-18. https://docs.frameplay.gg/unity/ad-spaces/ Accessed 2026-09-29.
[174] GumGum sellers.json. GumGum, 2026-09-28. https://gumgum.com/sellers.json Accessed 2026-09-29.
[175] Smadex and Anzu bring performance and clickability to intrinsic in-game advertising (Partner Post - Smadex). Smadex, published on Business of Apps, 2026-02-05. https://www.businessofapps.com/news/smadex-and-anzu-bring-performance-and-clickability-to-intrinsic-in-game-advertising/ Accessed 2026-09-29.
[176] Eskimi Teams Up With Gadsme To Offer Clickable In-Game Ads. Eskimi (Philip Mahler), 2025-08-06. https://www.eskimi.com/blog/eskimi-gadsme-partnership Accessed 2026-09-28.
[177] In-Game Ad Platform Frameplay Raises $8 Million With An Eye On Growth. Allison Schiff, AdExchanger, 2021-12-08. https://www.adexchanger.com/gaming/in-game-ad-platform-frameplay-raises-8-million-with-an-eye-on-growth/ Accessed 2026-09-28. Access: partial extract.
[178] iion Expands Its Premium Game Inventory with Top Mobile Publisher Voodoo. iion (Bryan Bartlett), 2025-02-26. https://www.iion.io/blog/iion-expands-its-premium-game-inventory-with-top-mobile-publisher-voodoo Accessed 2026-09-29.
[179] iion Taps Azur Games' 9.5B+ Installs to Expand Its Premium Game Inventory. iion (Bryan Bartlett), 2025-03-31. https://www.iion.io/blog/iion-taps-azur-games-9-5b-installs-to-expand-its-premium-game-inventory Accessed 2026-09-29.
[180] Adverty selects PubMatic as global preferred PMP partner. Adverty AB (publ) (MFN), 2023-12-18. https://mfn.se/a/adverty/adverty-selects-pubmatic-as-global-preferred-pmp-partner-elevating-in-game-advertising-ecosystem Accessed 2026-09-29.
[181] Programmatic Advertising (Kidoz Full Stack Programmatic Ecosystem). Kidoz Inc., undated (live page). https://www.kidoz.net/programmatic Accessed 2026-09-29.
[182] Programmatic Sales Executive at Gameloft (job listing). GameJobs.co (listing of a Gameloft vacancy), about 2026-03 (listing says "6 months ago" on 2026-09-29; inference). https://gamejobs.co/Programmatic-Sales-Executive-at-Gameloft Accessed 2026-09-29.
[183] Gaming Wants To Prove It's Just Like Other Media Channels - While Also Owning How It's Different. AdExchanger (Anthony Vargas), 2026-09-18. https://www.adexchanger.com/gaming/gaming-wants-to-prove-its-just-like-other-media-channels-while-also-owning-how-its-different/ Accessed 2026-09-29.
[184] adcontextprotocol/adcp releases (Ad Context Protocol). AgenticAdvertising.org (AdCP maintainers), 2026-09-27. https://github.com/adcontextprotocol/adcp/releases Accessed 2026-09-28.
[185] AdCP documentation, version 3.1.24 (dist/docs/3.1.24). AgenticAdvertising.org (AdCP maintainers), 2026-09-23. https://github.com/adcontextprotocol/adcp/tree/main/dist/docs/3.1.24 Accessed 2026-09-28.
[186] French company register search API: GADSME (SIREN 884197799). French government (annuaire des entreprises, INSEE/INPI data), live register. https://recherche-entreprises.api.gouv.fr/search?q=gadsme&per_page=5 Accessed 2026-09-29.
[187] Roblox Launches Rewarded Video Ads, Partners With Google to Scale Immersive Advertising. Roblox Corporation, 2025-04-01. https://about.roblox.com/newsroom/2025/04/roblox-scales-video-ads-partners-with-google Accessed 2026-09-28.
[188] Roblox Expands Advertising Platform as Essential Channel for the Next Generations. Roblox Corporation (newsroom; Louqman Parampath), 2026-01-06. https://about.roblox.com/newsroom/2026/01/roblox-expands-advertising-platform Accessed 2026-09-28.
[189] Roblox Taps PubMatic to Offer Programmatic Immersive Video Ads on Its Platform. PubMatic, 2024-04-10. https://pubmatic.com/news/roblox-pubmatic-programmatic-immersive-video-ads/ Accessed 2026-09-29.
[190] Ad integrations (Roblox Creator Hub documentation). Roblox Corporation, 2026-04-15. https://create.roblox.com/docs/production/promotion/ad-integrations Accessed 2026-09-28.
[191] XBOX Media Solutions: Powering brand connections in gaming (Microsoft Advertising gaming solutions). Microsoft (Microsoft Advertising; XBOX Media Solutions), 2026-09-23. https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions Accessed 2026-09-28.
[192] Microsoft MCP Server (Microsoft Advertising). Microsoft (Microsoft Advertising). https://about.ads.microsoft.com/en/solutions/technology/mcp-server Accessed 2026-09-28.
[193] Electronic Arts is bundling its games to make a bigger pitch to advertisers. Digiday (Krystal Scanlon), 2026-09-15. https://digiday.com/marketing/electronic-arts-is-bundling-its-games-to-make-a-bigger-pitch-to-advertisers/ Accessed 2026-09-29.
[194] Outwit, Outplay, Out-Engage: How Zynga Turned Survivor Fans Into Players (Zynga Ads, Visionary Voices). Zynga, 2026-03-12. https://www.zyngaads.com/visionaryvoices/zynga-alexa-grandolfo-survivor-mobile-gaming-partnership Accessed 2026-09-29.
[195] The most efficient buy (Harry Wilkins on banner ads; Zynga Ads, Visionary Voices). Zynga, 2026-03-24. https://www.zyngaads.com/visionaryvoices/zynga-harry-wilkins-mobile-gaming-banner-ads Accessed 2026-09-29.
[196] PubMatic and Zynga partner to expand advertiser access to premium mobile games audiences. PocketGamer.biz (Isa Muhammad), 2026-07-10. https://www.pocketgamer.biz/pubmatic-and-zynga-partner-to-expand-advertiser-access-to-premium-mobile-games-audiences/ Accessed 2026-09-29.
[197] Roblox Corporation Form 10-K for the fiscal year ended 31 December 2025. Roblox Corporation (SEC EDGAR, accession 0001315098-26-000024), 2026-02-11. https://www.sec.gov/Archives/edgar/data/1315098/000131509826000024/rblx-20251231.htm Accessed 2026-09-28.
[198] Microsoft Corporation Form 10-K for the fiscal year ended June 30, 2026. Microsoft Corporation, 2026-07-29. https://www.sec.gov/Archives/edgar/data/789019/000119312526323660/msft-20260630.htm Accessed 2026-09-28.
[199] Amazon.com, Inc. Form 10-K for fiscal year 2025. Amazon.com, Inc., 2026-02-06. https://www.sec.gov/Archives/edgar/data/1018724/000101872426000004/amzn-20251231.htm Accessed 2026-09-28.
[200] Take-Two Interactive Software, Inc. Form 10-K for the fiscal year ended March 31, 2026. Take-Two Interactive Software, Inc., 2026-05-22. https://www.sec.gov/Archives/edgar/data/946581/000162828026037434/ttwo-20260331.htm Accessed 2026-09-28.
[201] Roblox Q4 and full-year 2025 shareholder letter (Form 8-K, Exhibit 99.1). Roblox Corporation, 2026-02-05. https://www.sec.gov/Archives/edgar/data/1315098/000131509826000009/ex991-q42025shareholder.htm Accessed 2026-09-28.
[202] Beginning Ad-supported Game Streaming Test & Learn. Team XBOX (Microsoft), Xbox Wire, 2026-07-23. https://news.xbox.com/en-us/2026/07/23/game-streaming-ad-supported-xbox-insiders/ Accessed 2026-09-28.
[203] Unlocking the Power of Mobile Gaming: Attention, Audiences & Impact with Zynga Ads+ (IAB Brand Academy event listing). IAB (event presented with Zynga Ads), undated listing (event 2026-10-21). https://www.iab.com/events/unlocking-the-power-of-mobile-gaming/ Accessed 2026-09-29.
[204] Venatus Partners with Roblox as Official Reseller for Rewarded Video Across the UK, France and Germany. Venatus, 2026-05. https://www.venatus.com/blog/venatus-partners-with-roblox-for-rewarded-video Accessed 2026-09-28.
[205] Gaming Gets Serious: Playwire Partners with Roblox to Deliver Premium Gaming Inventory. Playwire (Playwire Strategy Team), 2025-10-14. https://www.playwire.com/blog/gaming-gets-serious-playwire-partners-with-roblox-to-deliver-premium-gaming-inventory Accessed 2026-09-29.
[206] SuperAwesome: Roblox's Sole Partner for Gen Alpha Advertising (Roblox Ads product page). SuperAwesome, 2026-05-18 (page metadata). https://www.superawesome.com/awesome-gaming/roblox-u13/ Accessed 2026-09-28.
[207] Rewarded video ads (Creator Hub documentation source). Roblox Corporation (Roblox/creator-docs). https://github.com/Roblox/creator-docs/blob/main/content/en-us/production/promotion/rewarded-video-ads.md Accessed 2026-09-28.
[208] Immersive ads (Creator Hub documentation source). Roblox Corporation (Roblox/creator-docs), 2026-07-30. https://github.com/Roblox/creator-docs/blob/main/content/en-us/production/monetization/immersive-ads.md Accessed 2026-09-28.
[209] Boost your game earnings with Rewarded Video enhancements (Developer Forum announcement). Roblox Corporation (DevForum, staff account superburr0), 2026-05-20. https://devforum.roblox.com/t/boost-your-game-earnings-with-rewarded-video-enhancements/4645335 Accessed 2026-09-28.
[210] Roblox Q1 2026 shareholder letter (Form 8-K, Exhibit 99.1). Roblox Corporation, 2026-04-30. https://www.sec.gov/Archives/edgar/data/1315098/000162828026028882/ex991-q12026earningsshar.htm Accessed 2026-09-28.
[211] Magnite Becomes a Global Partner to Power Rewarded Video on Roblox. Magnite (blog), 2026-01-06. https://www.magnite.com/blog/magnite-becomes-a-global-partner-to-power-rewarded-video-on-roblox/ Accessed 2026-09-29.
[212] Google expands Immersive ads and Roblox partnership (Ads & Commerce blog). Google, 2025-04-01. https://blog.google/products/ads-commerce/immersive-ads-roblox/ Accessed 2026-09-29. Access: partial extract.
[213] Ads Preferences. Roblox Corporation (Roblox Support), 2024-07-30. https://en.help.roblox.com/hc/en-us/articles/28943243301780-Ads-Preferences Accessed 2026-09-28.
[214] Ads on Roblox. Roblox Corporation (Roblox Support), 2026-04-21. https://en.help.roblox.com/hc/en-us/articles/48454017762580-Ads-on-Roblox Accessed 2026-09-28.
[215] New Advertising Policies & Standards (Developer Forum announcement, with 17 June 2026 update). Roblox Corporation (DevForum, staff account superburr0), 2026-03-20. https://devforum.roblox.com/t/new-advertising-policies-standards/4527365 Accessed 2026-09-29.
[216] Roblox Opens Up Advertising To Kids Under 13. Anthony Vargas, AdExchanger, 2026-06-04. https://www.adexchanger.com/gaming/roblox-opens-up-advertising-to-kids-under-13/ Accessed 2026-09-28.
[217] Brand Safety on Roblox. Roblox Corporation (Roblox for Brands), undated (footer 2026). https://brands.roblox.com/brand-safety Accessed 2026-09-28.
[218] Roblox to overhaul Ad policies, introduce Revenue sharing in 2027 (Reuters report, syndicated). Reuters, syndicated by TechDailyPost, 2026-03-27. https://techdailypost.co.za/2026/03/27/roblox-to-overhaul-ad-policies-introduce-revenue-sharing-in-2027/ Accessed 2026-09-29.
[219] creator-docs: content/en-us/production/promotion/ad-integrations.md and its commit history. Roblox Corporation (Roblox/creator-docs GitHub repository), 2026-09-16. https://github.com/Roblox/creator-docs/blob/main/content/en-us/production/promotion/ad-integrations.md Accessed 2026-09-28.
[220] Fortnite Developer Rules. Epic Games, 2026-04-29. https://legal.epicgames.com/fortnite/developer-rules Accessed 2026-09-28.
[221] Play Age Signals overview (including Terms of service). Google (Android Developers), 2026-07-20. https://developer.android.com/google/play/age-signals/overview Accessed 2026-09-28.
[222] Activision Blizzard Media is Now Xbox Media Solutions: What This Means for Marketers. XBOX Media Solutions (LinkedIn company page article), 2026-02-04. https://www.linkedin.com/pulse/activision-blizzard-media-now-xbox-solutions-what-d6fye Accessed 2026-09-29.
[223] Who we are in the world of gaming. Microsoft Advertising (XBOX Media Solutions), undated. https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/about-us Accessed 2026-09-28.
[224] In-game video ads that drive powerful engagement (Rewarded Video). Microsoft Advertising (XBOX Media Solutions), undated. https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/rewarded-video Accessed 2026-09-28.
[225] Unlock deeper engagement with Rewarded Portals. Microsoft Advertising (Caroline Bauer, Senior Product Marketing Manager, XBOX Media Solutions), 2026-05-12. https://about.ads.microsoft.com/en/blog/post/may-2026/unlock-deeper-engagement-with-rewarded-portals Accessed 2026-09-29.
[226] The sweet spot: Where Hershey's becomes part of play (case study). Microsoft Advertising (Xbox Media Solutions), 2026-05-26. https://about.ads.microsoft.com/en/resources/discover/case-studies/hersheys-holiday-bells-success-story Accessed 2026-09-29.
[227] How KitKat brought "Have a Break" into Candy Crush Saga (case study). Microsoft Advertising (XBOX Media Solutions), 2026-08-18. https://about.ads.microsoft.com/en/resources/discover/case-studies/kitkat-and-candy-crush-saga-success-story Accessed 2026-09-29.
[228] Microsoft Solitaire Collection - Free download and play on Windows | Microsoft Store. Microsoft Store (publisher listed: Xbox Game Studios), undated. https://apps.microsoft.com/detail/9wzdncrfhwd2 Accessed 2026-09-29.
[229] Activision Blizzard Media grows revenue with a balanced monetization model. Meta Audience Network (success story), undated (cites MOAT King Performance Q4 2020). https://www.facebook.com/audiencenetwork/resources/success-stories/activision Accessed 2026-09-29.
[230] Press Start: Engage Players and Boost Your Brand with Microsoft Advertising. Microsoft Advertising, 2024-09-26. https://about.ads.microsoft.com/en/blog/post/september-2024/press-start-engage-players-and-boost-your-brand-with-microsoft-advertising Accessed 2026-09-29.
[231] Microsoft is reportedly looking to bring ads to free-to-play Xbox games. Aisha Malik, TechCrunch (relaying Business Insider), 2022-04-18. https://techcrunch.com/2022/04/18/microsoft-ads-free-to-play-xbox-games/ Accessed 2026-09-28.
[232] Anzu PC & console game developers page. Anzu. https://www.anzu.io/pc-console-game-developers Accessed 2026-09-28.
[233] When everybody's a player, 'it makes no sense' for brands to still be on the sidelines. The Drum, 2026-01-16. https://www.thedrum.com/news/when-everybody-s-a-player-it-makes-no-sense-for-brands-to-still-be-on-the-sidelines Accessed 2026-09-29.
[234] King Privacy Policy. King (Activision Blizzard, part of Microsoft), 2026-02-26. https://www.king.com/privacyPolicy Accessed 2026-09-28.
[235] Xfinity and Peacock Team Up with EA Sports FC 26 to Bring Fans Exclusive In-Game Offers and Kits. Comcast Corporation, 2025-11-17. https://corporate.comcast.com/press/releases/xfinity-and-peacock-team-up-with-ea-sports-fc-26-to-bring-fans-exclusive-in-game-offers-and-kits Accessed 2026-09-29.
[236] EA Advertising Could Bring In-Game Ads to Future EA FC and Madden NFL Titles. Operation Sports, 2026-06-15. https://www.operationsports.com/ea-advertising-could-bring-in-game-ads-to-future-ea-fc-and-madden-nfl-titles/ Accessed 2026-09-29.
[237] EA Advertising - Advertise with Electronic Arts. Electronic Arts Inc., undated (live page). https://www.ea.com/ea-advertising (reached from https://www.ea.com/brand-partnerships via redirect) Accessed 2026-09-29.
[238] Lowe's - Madden 26 Ultimate Team Program. MUT.GG (Stormstrike Inc., fan database), undated. https://www.mut.gg/programs/lowes/ Accessed 2026-09-29.
[239] Mountain Dew Joins EA SPORTS 'College Football 26' with Dew University, Limited Merch Drops Ahead. The Source, 2025-07-15. https://thesource.com/2025/07/15/mountain-dew-college-football-26-dew-university/ Accessed 2026-09-29.
[240] EA sellers.json and app-ads.txt (HTTP checks). Electronic Arts Inc., undated. https://www.ea.com/sellers.json ; https://www.ea.com/app-ads.txt Accessed 2026-09-29. Access: partial extract.
[241] EA Announces Completion of Acquisition by PIF, Silver Lake, and Affinity Partners (Form 8-K, Exhibit 99.1). Electronic Arts Inc. (SEC EDGAR, CIK 712515), 2026-08-04. https://www.sec.gov/Archives/edgar/data/712515/000114036126031157/ef20079099_ex99-1.htm Accessed 2026-09-29.
[242] How Hyundai Drove Its Football Passion Into a Mobile Game (Zynga Ads, Visionary Voices). Zynga, 2026-06-15. https://www.zyngaads.com/visionaryvoices/hyundai-top-eleven-non-gaming-brands Accessed 2026-09-29.
[243] Electronic Arts Inc. Annual Report on Form 10-K for the fiscal year ended 31 March 2026. Electronic Arts Inc. (SEC EDGAR), 2026-05-11. https://www.sec.gov/Archives/edgar/data/712515/000162828026033617/ea-20260331.htm Accessed 2026-09-29.
[244] Ad Solutions | Zynga Ads. Zynga, undated (live page). https://www.zyngaads.com/ad-solutions Accessed 2026-09-29.
[245] Resources | Zynga Ads (Media Kit - 2026 link). Zynga, undated (live page). https://www.zyngaads.com/resources Accessed 2026-09-29. Access: partial extract.
[246] Roblox Corporation Form 10-Q for the quarter ended June 30, 2026. Roblox Corporation (SEC EDGAR), 2026-07-30. https://www.sec.gov/Archives/edgar/data/1315098/000162828026051082/rblx-20260630.htm Accessed 2026-09-28.
[247] Microsoft Form 8-K (Item 7.01), Exhibit 99.1: FY27 Segments and Investor Metrics. Microsoft Corporation (SEC EDGAR), 2026-09-02. https://www.sec.gov/Archives/edgar/data/0000789019/000119312526380280/d291965dex991.htm Accessed 2026-09-29.
[248] Discord Opens Advertising to Brands of All Sizes With Ads Manager and Video Quests. Discord, 2026-09-29. https://discord.com/press-releases/discord-opens-advertising-to-brands-of-all-sizes-with-ads-manager-and-video-quests Accessed 2026-09-30.
[249] PMP for gaming publishers (Venatus). Venatus Media Limited, undated (live page). https://www.venatus.com/platform/tech/pmp Accessed 2026-09-29.
[250] Enthusiast Gaming Holdings Inc. Annual Information Form for the year ended 31 December 2025 (EGLX-2026-AIF.pdf). Enthusiast Gaming Holdings Inc., 2026-04 (uploaded April 2026; share data 'As of April 14, 2026'). https://www.enthusiastgaming.com/wp-content/uploads/2026/04/EGLX-2026-AIF.pdf Accessed 2026-09-29.
[251] faceit.com/ads.txt (Publift Fuse, 'Updated: May 2026'). FACEIT / Publift, 2026-05 (file header). https://www.faceit.com/ads.txt Accessed 2026-09-29.
[252] Grow on Medal - Reach Millions of Gamers. Medal.tv. https://medal.tv/grow Accessed 2026-09-29.
[253] Ziff Davis, Inc. Form 10-K for the fiscal year ended 31 December 2025. Ziff Davis, Inc. (SEC filing), 2026-02-24. https://www.sec.gov/Archives/edgar/data/1084048/000108404826000005/zd-20251231.htm Accessed 2026-09-29.
[254] How brands can get the most out of YouTube Shorts ads. Google (The Keyword blog), 2024-04-17. https://blog.google/products/ads-commerce/youtube-shorts-ads-select-lineups-abcds/ Accessed 2026-09-28.
[255] About reservation: reserved media placements on YouTube (Google Ads Help). Google, undated (live help article). https://support.google.com/google-ads/answer/6030919?hl=en Accessed 2026-09-28.
[256] In-the-game advertising (Venatus). Venatus, undated (live page). https://www.venatus.com/advertisers/ad-placements/in-the-game Accessed 2026-09-28.
[257] Twitch Terms of Service. Twitch Interactive, Inc., 2026-08-12 (Last modified on 08/12/2026). https://legal.twitch.com/legal/terms-of-service/ Accessed 2026-09-29.
[258] Twitch display inventory now available in Amazon DSP. Amazon Ads (launch announcement), 2022-12-14. https://advertising.amazon.com/resources/whats-new/twitch-display-inventory-amazon-dsp Accessed 2026-09-29.
[259] Create deals with Twitch premium preferred deals now generally available globally in Amazon DSP. Amazon Ads (launch announcement), 2023-10-25. https://advertising.amazon.com/resources/whats-new/enhance-twitch-programmatic-video-buying-experience Accessed 2026-09-29.
[260] Streaming TV ads: Advertise on shows, movies, live sports. Amazon Ads, undated (live page). https://advertising.amazon.com/solutions/products/streaming-tv-ads Accessed 2026-09-29.
[261] Bounty Board ad specifications and guidelines. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/bounty-board Accessed 2026-09-28.
[262] Channel Takeover, a creator-centric media solution. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/channel-takeover Accessed 2026-09-28.
[263] Twitch ads ad specifications | Streamables. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/streamables Accessed 2026-09-28.
[264] Twitch ads - How to start advertising with Twitch. Amazon Ads (guide), 2021-10-27 (page metadata datePublished). https://advertising.amazon.com/library/guides/twitch-ads Accessed 2026-09-29.
[265] Introducing Content Classification Labels. Twitch (blog), 2023-06-20. https://blog.twitch.tv/en/2023/06/20/introducing-content-classification-labels/ Accessed 2026-09-28.
[266] Twitch's approach to brand safety, monetization, and brand suitability controls. Amazon Ads (guide), 2026-02-22 (page metadata). https://advertising.amazon.com/library/guides/twitch-brand-safety Accessed 2026-09-28.
[267] Content Classification Labels. Twitch Help. https://help.twitch.tv/s/article/content-classification-labels?language=en_US Accessed 2026-09-29.
[268] Twitch API Reference: Get Content Classification Labels. Twitch Developers. https://dev.twitch.tv/docs/api/reference/#get-content-classification-labels Accessed 2026-09-29.
[269] New Content Controls for Twitch Premium Video now available on Amazon DSP. Amazon Ads (launch announcement), 2023-11-06. https://advertising.amazon.com/resources/whats-new/new-content-controls-for-twitch-premium-video Accessed 2026-09-28.
[270] LIVAD For Brands - Leverage the power of smart live stream advertising. LIVAD Technologies, undated (live page). https://www.livad.stream/brands Accessed 2026-09-29.
[271] Capture the Moments That Matter: Introducing Live Event Detection. LIVAD (post by Emre Eskihoran), 2023-04-12. https://www.livadstream.com/blog/live-event-detection Accessed 2026-09-29.
[272] Livad Technologies launches virtual product placements for livestreaming creators | exclusive. GamesBeat, 2026-09-17. https://gamesbeat.com/livad-technologies-launches-virtual-product-placements-for-livestreaming-creators-exclusive/ Accessed 2026-09-29.
[273] Monetize your live streams with LIVAD's non-intrusive video ads (home page). LIVAD Technologies Inc., 2026 (copyright line). https://www.livad.stream/ Accessed 2026-09-29.
[274] What if livestreamers could do product placement without real products?. Tubefilter (James Hale), 2026-09-18. https://www.tubefilter.com/2026/09/18/livad-technologies-livestream-product-placements/ Accessed 2026-09-29.
[275] Terms and Conditions (LIVAD Contracted Streamer terms). LIVAD Medya Teknoloji A.S. (LIVAD), undated (live page). https://www.livad.stream/terms-and-conditions Accessed 2026-09-29.
[276] ads.txt for www.livad.stream. LIVAD Technologies, 2026-08-03 (server Last-Modified). https://www.livad.stream/ads.txt Accessed 2026-09-29.
[277] Adform sellers.json. Adform A/S, live file (no date field). https://adform.com/sellers.json Accessed 2026-09-29.
[278] Improve Digital sellers.json. Improve Digital, live file. https://sellers.improvedigital.com/sellers.json Accessed 2026-09-29.
[279] OpenCorporates company search: LIVAD TECHNOLOGIES. OpenCorporates (aggregating the Delaware Division of Corporations record), live search. https://opencorporates.com/companies?q=LIVAD+TECHNOLOGIES Accessed 2026-09-29. Access: partial extract.
[280] Discord Advertising IAB Addendum. Discord Inc., 2026-05-12 (updated). https://support.discord.com/hc/en-us/articles/28455100031511-Discord-Advertising-IAB-Addendum Accessed 2026-09-29.
[281] Discord Launches Newest Ad Format and Partners with AppsFlyer, Gamesight for Ads Measurement. Discord, 2025-10-02. https://discord.com/press-releases/discord-launches-newest-ad-format-and-partners-with-appsflyer-gamesight-for-ads-measurement Accessed 2026-09-28.
[282] Discord Deepens Its Ability to Drive Growth for Games. Discord, 2026-03-09. https://discord.com/press-releases/discord-deepens-its-ability-to-drive-growth-for-games Accessed 2026-09-29.
[283] Discord and Zach Cregger's Resident Evil Bring the First-Ever Quest Activation to Life. Discord, 2026-07-22. https://discord.com/press-releases/discord-zach-cregger-resident-evil-missions-quest-activation Accessed 2026-09-29.
[284] Introducing New Tools to Power Game Discovery and Social Play. Discord, 2026-08-20. https://discord.com/press-releases/introducing-new-tools-to-power-game-discovery-and-social-play Accessed 2026-09-29.
[285] Discord Quests FAQ. Discord (Discord Support). https://support.discord.com/hc/en-us/articles/22225719947543-Discord-Quests-FAQ Accessed 2026-09-28.
[286] Discord Ads - Gaming (Get players in your game with Quests). Discord, undated (live page). https://discord.com/ads/gaming Accessed 2026-09-29.
[287] Data Used for Sponsored Content on Discord. Discord (Discord Support), 2025-09-18 (edited_at). https://support.discord.com/hc/en-us/articles/22225542459415 Accessed 2026-09-29.
[288] Ads Policy. Discord, 2026-09-09. https://support.discord.com/hc/en-us/articles/25516720403223-Ads-Policy Accessed 2026-09-28.
[289] Safer for Teens, Same Discord for Adults. Discord (post signed Stan), 2026-09-22. https://discord.com/blog/safer-for-teens-same-discord-for-adults Accessed 2026-09-28.
[290] medal.tv ads.txt. Medal.tv. https://medal.tv/ads.txt Accessed 2026-09-29.
[291] Advertise with Overwolf - Reach Gamers in the World's Top Titles. Overwolf, undated (live page). https://www.overwolf.com/ads/ Accessed 2026-09-28.
[292] Custom Activations (Overwolf Ads). Overwolf, undated (live page). https://www.overwolf.com/ads/ad-solutions/custom-activations/ Accessed 2026-09-28.
[293] Overwolf Acquires Gaming Ad Tech Leader NitroPay. Overwolf, 2024-05-08. https://blog.overwolf.com/overwolf-acquires-gaming-ad-tech-leader-nitropay/ Accessed 2026-09-28.
[294] Advertising service providers (Overwolf legal). Overwolf, Ltd., 2025-12-29 (Last updated). https://legal.overwolf.com/docs/overwolf/advertising-service-providers Accessed 2026-09-29.
[295] Ziff Davis, Inc. 10-K Exhibit 21.1 (subsidiaries). Ziff Davis, Inc. (SEC filing), 2026-02-24. https://www.sec.gov/Archives/edgar/data/1084048/000108404826000005/zd20251231ex-211.htm Accessed 2026-09-29.
[296] IGN Entertainment - Contact Us (Advertise). IGN Entertainment, undated (live page). https://corp.ign.com/advertise Accessed 2026-09-29.
[297] Ad Specs & Guidelines - IGN Entertainment (home). IGN Entertainment, Inc. (a Ziff Davis company), 2026 (copyright line). https://specs.ign.com/ Accessed 2026-09-29.
[298] Ad Specs - Pre-Roll. IGN Entertainment, undated (live page). https://specs.ign.com/ad-unit/pre-roll Accessed 2026-09-29.
[299] Advertising Policies (IGN ad specs site). IGN Entertainment, undated (live page). https://specs.ign.com/advertising-policies Accessed 2026-09-29.
[300] IGN Entertainment Launches 'IMAGINE': A Cultural & Cognitive AI Platform for Audience Intelligence. IGN Entertainment, Inc. (via PR Newswire), 2025-09-24. https://www.prnewswire.com/news-releases/ign-entertainment-launches-imagine-a-cultural--cognitive-ai-platform-for-audience-intelligence-302565292.html Accessed 2026-09-29.
[301] IMAGINE (imagine.ign.com). IGN Entertainment, n/a. https://imagine.ign.com/ Accessed 2026-09-29. Access: partial extract.
[302] PROJECT VENUS TOPCO LIMITED (13588236): persons with significant control. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/13588236/persons-with-significant-control Accessed 2026-09-28.
[303] VENATUS MEDIA LIMITED (07183957): persons with significant control. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/07183957/persons-with-significant-control Accessed 2026-09-28.
[304] MEDIA SCALE CAPITAL LIMITED (17252129): persons with significant control. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/17252129/persons-with-significant-control Accessed 2026-09-28.
[305] Venatus Enters New Chapter with Acquisition and Appointment of New CEO & CFO. Venatus, 2026-07-14 (page datePublished 'Jul 14'; displayed 'July 2026'). https://www.venatus.com/blog/venatus-enters-new-chapter-with-acquisition-by-media-scale-capital-limited Accessed 2026-09-28.
[306] High-Impact Formats (Venatus, publisher page including direct sales). Venatus Media Limited, undated (live page). https://www.venatus.com/platform/tech/high-impact-formats Accessed 2026-09-29.
[307] OP.GG ads.txt. OP.GG (file managed by RevIQ per managerdomain), 2026-09-28 ('Last Updated: September 28, 2026'). https://op.gg/ads.txt Accessed 2026-09-29.
[308] Venatus sellers.json. Venatus Media Limited, live file. https://www.venatus.com/sellers.json Accessed 2026-09-29.
[309] Playwire sellers.json. Playwire, LLC, live file. https://www.playwire.com/sellers.json Accessed 2026-09-29.
[310] Playwire | Brand Advertising. Playwire, LLC, undated (live page). https://www.playwire.com/brand-advertising Accessed 2026-09-29.
[311] Playwire advertisers. Playwire, undated (live page). https://www.playwire.com/advertisers Accessed 2026-09-28.
[312] Playwire Advertising Terms. Playwire, LLC, 2025-11-21 (Publication Date). https://www.playwire.com/advertising-terms Accessed 2026-09-29.
[313] Playwire | Gaming Advertising. Playwire, undated (live page). https://www.playwire.com/game-advertising Accessed 2026-09-28.
[314] Playwire | In-Game Advertising Services. Playwire, undated (live page). https://www.playwire.com/in-game-advertising-services Accessed 2026-09-28.
[315] Unlocking the Power of Minecraft Advertising: Insights from Vortex Games. Playwire, 2025-02-17. https://www.playwire.com/blog/unlocking-the-power-of-minecraft-advertising Accessed 2026-09-28.
[316] How Vortex Games Maximizes Player Value Through Strategic Ad Monetization. Playwire, undated. https://www.playwire.com/vortex-games-case-study Accessed 2026-09-28.
[317] Enthusiast Gaming Completes Direct Sales Business Divestment. Enthusiast Gaming Holdings Inc., 2025-09-02. https://www.enthusiastgaming.com/enthusiast-gaming-completes-direct-sales-business-divestment/ Accessed 2026-09-28.
[318] Enthusiast Gaming Holdings Inc. Management Discussion and Analysis for the three and six months ended 30 June 2026. Enthusiast Gaming Holdings Inc., 2026-08-12. https://www.enthusiastgaming.com/wp-content/uploads/2026/08/EGHI-MDA-Jun-30-2026-FINAL-1.pdf Accessed 2026-09-29.
[319] u.gg/ads.txt (Freestar-hosted, 'Last updated 09-25-2026'). Enthusiast Gaming / Freestar (file served from a.pub.network), 2026-09-25 (file header). https://u.gg/ads.txt Accessed 2026-09-29.
[320] ads.txt of icy-veins.com, thesimsresource.com, pocketgamer.com and addictinggames.com (Freestar-hosted). Enthusiast Gaming / Freestar (files served from a.pub.network), 2026-09-25 (Icy Veins header) and 2026-09-28 (The Sims Resource, Pocket Gamer headers). https://www.icy-veins.com/ads.txt ; https://www.thesimsresource.com/ads.txt ; https://www.pocketgamer.com/ads.txt ; https://www.addictinggames.com/ads.txt Accessed 2026-09-29.
[321] ESL and FACEIT merge to create the world's leading competitive gaming platform. ESL FACEIT Group, 2022-01-24. https://eslfaceitgroup.com/press/esl-and-faceit-merge-to-create-the-worlds-leading-competitive-gaming-platform/ Accessed 2026-09-28.
[322] MTG completes divestment of ESL Gaming and becomes a pure-play gaming group. Modern Times Group MTG AB, 2022-04-21. https://news.cision.com/mtg/r/mtg-completes-divestment-of-esl-gaming-and-becomes-a-pure-play-gaming-group,c3549678 Accessed 2026-09-28.
[323] Ad Solutions (ESL FACEIT Group). ESL FACEIT Group, undated (live page). https://eslfaceitgroup.com/advertising/ Accessed 2026-09-28.
[324] Publift sellers.json. Publift, live file. https://www.publift.com/sellers.json Accessed 2026-09-29.
[325] mobalytics.gg ads.txt ('Updated: 02 September 2026'). Mobalytics, 2026-09-02 (file header). https://mobalytics.gg/ads.txt Accessed 2026-09-29.
[326] NitroPay sellers.json. Nitro (Overwolf division; GG Software LP), live file. https://nitropay.com/sellers.json Accessed 2026-09-29.
[327] EXCLUSIVE: ESL FACEIT Group acquires analytics platform Mobalytics. Esports Insider, 2025-03-17. https://esportsinsider.com/2025/03/esl-faceit-group-mobalytics-acquisition Accessed 2026-09-29.
[328] EFG Social Media Channels - ESL FACEIT Group (Wayback Machine capture of 4 December 2025). ESL FACEIT Group (archived by the Internet Archive), 2025-12-04 (capture). http://web.archive.org/web/20251204042006id_/https://eslfaceitgroup.com/advertising/streaming-and-social-media-channels/ Accessed 2026-09-29.
[329] FACEIT Mission Privacy Notice. ESL FACEIT Group Limited (FACEIT Help Center), 2025-08-08. https://support.faceit.com/hc/en-us/articles/21634454927132-FACEIT-Mission-Privacy-Notice Accessed 2026-09-29.
[330] FACEIT Content Standards. FACEIT Help Center, 2025-07-22. https://support.faceit.com/hc/en-us/articles/21295054606620-FACEIT-Content-Standards Accessed 2026-09-29.
[331] ESL FACEIT Group imprint. ESL FACEIT Group GER GmbH, 2022-09-02 (datePublished). https://eslfaceitgroup.com/imprint/ Accessed 2026-09-29.
[332] MRC letter to Integral Ad Science: accreditation of IAS server-to-server integration with Amazon DSP (GI110725). Media Rating Council, 2025-11-07. https://mediaratingcouncil.org/sites/default/files/Accreditation%20Letters/IAS/GI110725_IAS%203P%20Amazon%20DSP%20Accreditation%20Letter%20.pdf Accessed 2026-09-29.
[333] TAG Registry (public registry search, company records). Trustworthy Accountability Group (TAG). https://www.tagtoday.net/registry-search Accessed 2026-09-28.
[334] TAG Registry search (public endpoint behind tagtoday.net/registry-search). Trustworthy Accountability Group. https://hubgloo.opxar.com/search?q=Medal Accessed 2026-09-29.
[335] DoubleVerify Expands Global Measurement Capabilities To Include Amazon Custom Audiences. DoubleVerify, 2023-10-31. https://doubleverify.com/company/newsroom/doubleverify-expands-global-measurement-capabilities-to-include-amazon-custom-audiences Accessed 2026-09-29.
[336] IAS Expands Global Measurement of Amazon Properties. Integral Ad Science, 2024-05-29. https://integralads.com/news/ias-expands-global-measurement-amazon-properties-twitch/ Accessed 2026-09-29.
[337] TAG Registry: company record for Twitch. Trustworthy Accountability Group (TAG), live registry. https://hubgloo.opxar.com/search?q=Twitch Accessed 2026-09-29.
[338] TAG Registry record: Venatus. Trustworthy Accountability Group (registry search backend hubgloo.opxar.com, called by tagtoday.net/registry-search), live registry. https://hubgloo.opxar.com/search?q=Venatus Accessed 2026-09-29.
[339] Oracle Advertising EOL Frequently Asked Questions, Version 3.0. Oracle, 2024-08-12. https://www.oracle.com/a/ocom/docs/oracle-advertising-end-of-life-faqs.pdf Accessed 2026-09-28.
[340] Discord for Business Vol. 2: Cannes-worthy ad product updates. Discord, 2025-07-16. https://discord.com/blog/discord-for-business-cannes-worthy-ad-product-updates Accessed 2026-09-28.
[341] DOOM: The Dark Ages Video Quest Success Story. Discord, undated (live page). https://discord.com/quests-success-stories/doom-the-dark-ages Accessed 2026-09-28.
[342] The world's leading esports company - ESL FACEIT Group (home page). ESL FACEIT Group, undated (live page, copyright 2026). https://eslfaceitgroup.com/ Accessed 2026-09-28.
[343] Audience - Reach the Most Engaged Gaming Community (Overwolf). Overwolf, undated (live page). https://www.overwolf.com/ads/audience/ Accessed 2026-09-28.
[344] Enthusiast Gaming Launches New Subscription Offering, Provides Operational and Corporate Updates. Enthusiast Gaming Holdings Inc., 2026-05-26. https://www.enthusiastgaming.com/enthusiast-gaming-launches-new-subscription-offering-provides-operational-and-corporate-updates/ Accessed 2026-09-29.
[345] LIVAD For Streamers. LIVAD Technologies, 2025 (copyright line). https://www.livad.stream/streamers Accessed 2026-09-29.
[346] 'Our Audience in Turkey' slide (image livad-audience-demography.webp). DigitalsCom (slide footer: Twitch Advertising, Confidential), 2024-11-12 (server Last-Modified). https://www.digitalscom.com/asset/images/livad/livad-audience-demography.webp Accessed 2026-09-29.
[347] IAB Gaming Measurement Framework (PDF, June 2025). IAB, 2025-06-26. https://www.iab.com/wp-content/uploads/2025/06/IAB_Gaming_Measurement_Framework_June_2025.pdf Accessed 2026-09-28.
[348] IAB Launches Gaming Measurement Framework to Standardize and Elevate Advertising in Gaming. IAB, 2025-06-26. https://www.iab.com/news/iab-launches-gaming-measurement-framework-to-standardize-and-elevate-advertising-in-gaming/ Accessed 2026-09-28.
[349] Media Rating Council Accreditation Decisions, January 2026 through March 2026 (Accreditation Status Quarterly Update Q1 2026). Media Rating Council (George W. Ivie), 2026-04-17. https://www.mediaratingcouncil.org/sites/default/files/News/MRC%20Accreditation%20Status%20Quarterly%20Update_Q1%202026.pdf Accessed 2026-09-29.
[350] IAS and Anzu Partner to Provide Media Quality Measurement for In-Game Advertising Environments. Integral Ad Science, 2022-07-28. https://integralads.com/news/ias-anzu-partner-for-in-game-advertising-environments/ Accessed 2026-09-28.
[351] DoubleVerify Extends Gaming Measurement Suite Through Strategic Partnership with Roblox. DoubleVerify, 2025-04-01. https://doubleverify.com/blog/social/verify/doubleverify-extends-gaming-measurement-suite-through-strategic-partnership-with-roblox Accessed 2026-09-28.
[352] Gold Standard Certified and Registered. IAB UK. https://www.iabuk.com/goldstandard/certified-and-registered Accessed 2026-09-28.
[353] Adverty home page. Adverty AB (publ). https://www.adverty.com/ Accessed 2026-09-28. Access: partial extract.
[354] IAB Gaming Measurement Framework (guideline page). IAB, 2025-06-26. https://www.iab.com/guidelines/gaming-measurement-framework/ Accessed 2026-09-29. Access: partial extract.
[355] Understanding Revenue | Frameplay SDK. Frameplay, 2023-09-12. https://docs.frameplay.gg/guide/revenue/ Accessed 2026-09-29.
[356] Frameplay Announces New Attention Metric in Video Game Environments, Validated in Partnership with dentsu's Attention Economy Team by First-to-Market Studies. Frameplay Corp. (PR Newswire), 2022-04-05. https://www.prnewswire.com/news-releases/frameplay-announces-new-attention-metric-in-video-game-environments-validated-in-partnership-with-dentsus-attention-economy-team-by-first-to-market-studies-301518119.html Accessed 2026-09-28. Access: partial extract.
[357] About us | Frameplay. Frameplay, undated (timeline ends 2024). https://frameplay.com/about-us/ Accessed 2026-09-29.
[358] A Guide To Understanding Intrinsic In-Game Ad Placements. Anzu (Nick Woodford), 2024-12-04. https://www.anzu.io/blog/in-game-ad-placements-guide Accessed 2026-09-29.
[359] The Ultimate Guide to Measuring Intrinsic In-Game Ads. Anzu, 2024-08-22. https://www.anzu.io/blog/in-game-measurement-guide Accessed 2026-09-28.
[360] Real Attention, Guaranteed Outcomes. iion, undated (cites 2025 market figure). https://www.iion.io/guaranteed-outcomes Accessed 2026-09-29.
[361] In-Game advertising measurement & optimization (Microsoft Advertising gaming solutions). Microsoft Advertising (XBOX Media Solutions). https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/ad-performance-measurement Accessed 2026-09-28.
[362] Advertising in and around gaming (ad solutions). Microsoft Advertising (XBOX Media Solutions), undated. https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/ad-solutions Accessed 2026-09-29.
[363] Anzu Advertisers' Terms and Conditions (PDF embedded in anzu.io/advertisers-terms-and-conditions). Anzu Virtual Reality Ltd., 2025-09. https://www.anzu.io/hubfs/20250928%20Anzu%20AdvertisersTCs.pdf Accessed 2026-09-29.
[364] Measure, Optimize, and Protect Your Intrinsic In-Game Advertising Campaigns (Media Value page). Anzu Virtual Reality Ltd.. https://www.anzu.io/media-value Accessed 2026-09-28.
[365] Services Under Review: Digital (Internet Archive capture of 10 October 2025). Media Rating Council (captured by the Internet Archive). http://web.archive.org/web/20251010032302/https://www.mediaratingcouncil.org/accreditation/services-under-review/digital Accessed 2026-09-29.
[366] DoubleVerify to support in-game ads to win over brands' budgets for the channel. Digiday (Alexander Lee), 2023-07-27. https://digiday.com/marketing/doubleverify-to-support-in-game-ads-to-win-over-brands-budgets-for-the-channel/ Accessed 2026-09-28.
[367] IAS & Anzu Launch First-To-Market In-Game Measurement Solution. Integral Ad Science (IAS Insider), 2023-07-11. https://integralads.com/insider/ias-anzu-first-to-market-in-game-measurement/ Accessed 2026-09-28. Access: partial extract.
[368] IAS Partners with Gadsme to Provide Greater Transparency for In-Game Advertising Performance. Integral Ad Science, 2022-12-15. https://integralads.com/news/ias-gadsme-partnership-in-game-advertising/ Accessed 2026-09-28.
[369] Frameplay and Integral Ad Science Partnership Brings Greater Transparency to Intrinsic In-Game Advertising. Frameplay Corp, 2023-10-05. https://frameplay.com/resource/frameplay-and-integral-ad-science-partnership-brings-greater-transparency-to-intrinsic-in-game-advertising/ Accessed 2026-09-28.
[370] IAS Announces First-to-Market Integration With Roblox to Provide 3D Immersive Measurement. Integral Ad Science, 2024-05-01. https://integralads.com/news/ias-integration-roblox-3d-immersive-measurement/ Accessed 2026-09-28.
[371] IAS Launches First-to-Market Integration with Roblox for 3D Immersive Measurement. Integral Ad Science (IAS Insider, UK site), 2024-09-12. https://integralads.com/uk/insider/first-to-market-roblox-measurement/ Accessed 2026-09-28. Access: partial extract.
[372] IAS Partners with Gadsme to Provide Greater Transparency for In-Game Advertising Performance. Integral Ad Science (via PR Newswire), 2022-12-15. https://www.prnewswire.com/news-releases/ias-partners-with-gadsme-to-provide-greater-transparency-for-in-game-advertising-performance-301703555.html Accessed 2026-09-29.
[373] IAS expands media quality solutions for in-game media buys (IAS Insider). Integral Ad Science, 2022-12-15. https://integralads.com/insider/gadsme-gaming-partnership/ Accessed 2026-09-29.
[374] DoubleVerify Expands Global Measurement Capabilities to Include Microsoft Advertising's Proprietary Inventory and Data. DoubleVerify (newsroom), 2025-10-15. https://doubleverify.com/company/newsroom/doubleverify-expands-global-measurement-capabilities-to-include-microsoft-advertisings-proprietary-inventory-and-data Accessed 2026-09-29.
[375] doubleverify.com sitemap.xml. DoubleVerify, live file. https://doubleverify.com/sitemap.xml Accessed 2026-09-29.
[376] Integral Ad Science sitemaps (post-sitemap 1 to 4, page-sitemap, portfolio-sitemap 1 and 2). Integral Ad Science, undated (sitemap index). https://integralads.com/sitemap.xml Accessed 2026-09-29.
[377] Meaningful Connections & Gaming Experiences | Gameloft for Brands (home page). Gameloft for brands, undated (cites 2025 studies). https://www.gameloft.com/for-brands Accessed 2026-09-29.
[378] Inside Zynga's Bid to Make Gaming Ads as Measurable as the Open Web (Zynga Ads, Visionary Voices). Zynga, 2026-07-13. https://www.zyngaads.com/visionaryvoices/zynga-corentin-leydis-liveramp-rampid-gaming-addressability Accessed 2026-09-29.
[379] TAG registry endpoint re-queries (Gameloft SE, TAG-ID 9df283b2158909d5; controls King, Zynga, Gadsme). Trustworthy Accountability Group, live registry. https://hubgloo.opxar.com/search?q=9df283b2158909d5 Accessed 2026-09-29.
[380] Nielsen to Acquire DoubleVerify, Creating a Leading, Independent Media Intelligence Platform. Nielsen Holdings and DoubleVerify (joint release on nielsen.com), 2026-08-06. https://www.nielsen.com/news-center/2026/nielsen-to-acquire-doubleverify-creating-a-leading-independent-media-intelligence/ Accessed 2026-09-29.
[381] DoubleVerify Holdings, Inc. Form 8-K (Items 1.01, 5.02, 9.01), event date 6 August 2026. DoubleVerify Holdings, Inc. (SEC filing), 2026-08-10. https://www.sec.gov/Archives/edgar/data/1819928/000110465926092934/tm2621904d5_8k.htm Accessed 2026-09-29.
[382] Anzu Awarded TAG Seals and IAB UK Gold Standard 2.0 Certification. Anzu (republishing an ExchangeWire release), 2021-08-12. https://www.anzu.io/news/anzu-awarded-tag-seals-and-iab-uk-gold-standard-2.0-certification Accessed 2026-09-28.
[383] TAG Registry search (registry-search page's GET endpoint). Trustworthy Accountability Group. https://hubgloo.opxar.com/search?q=Anzu Accessed 2026-09-29.
[384] Dentsu and Lumen Release Gaming & Advertising Attention Report 2024. Lumen Research (reproducing a dentsu release), 2024-01-16. https://lumen-research.com/blog/dentsu-and-lumen-release-in-game-advertising-report-with-twitch-ads-anzu-and-activision-blizzard/ Accessed 2026-09-28. Access: partial extract.
[385] TAG registry search results for 'iion', 'adverty' and 'adinmo'. Trustworthy Accountability Group (registry served at hubgloo.opxar.com), live registry. https://hubgloo.opxar.com/search?q=iion Accessed 2026-09-29.
[386] Companies passing Tech Lab Compliance Programs (Compliance Registry). IAB Tech Lab. https://iabtechlab.com/compliance-programs/compliant-companies/ Accessed 2026-09-28.
[387] TAG Registry search (public lookup endpoint used by tagtoday.net/registry-search). Trustworthy Accountability Group (TAG). https://hubgloo.opxar.com/search?q=Frameplay Accessed 2026-09-29.
[388] iion's game-changing platform is now IAB Tech Lab Open Measurement certified. iion (Sanjaya Molligoda), 2024-09-10. https://www.iion.io/blog/iions-game-changing-platform-is-now-iab-tech-lab-open-measurement-certified-making-it-easier-to-track-your-ad-campaigns--across-gaming-environments Accessed 2026-09-29.
[389] TAG Registry search (public GET endpoint behind tagtoday.net/registry-search). Trustworthy Accountability Group, live registry. https://hubgloo.opxar.com/search?q=admazing Accessed 2026-09-29.
[390] TAG Registry company record: Activision Blizzard Media. Trustworthy Accountability Group (registry endpoint), undated (live record). https://hubgloo.opxar.com/company/6951145640 Accessed 2026-09-29.
[391] TAG Registry company record: King.com Ltd.. Trustworthy Accountability Group (registry endpoint), undated (live record). https://hubgloo.opxar.com/company/52706912291 Accessed 2026-09-29.
[392] TAG Registry search: Gadsme. Trustworthy Accountability Group, live registry. https://hubgloo.opxar.com/search?q=Gadsme Accessed 2026-09-29.
[393] IAB Tech Lab Compliance Programs. IAB Tech Lab. https://iabtechlab.com/compliance-programs/ Accessed 2026-09-28.
[394] Claritas & Anzu Join Forces to Prove the ROI of Intrinsic In-Game Advertising Through Attribution. Anzu (Nick Woodford), 2025-03-27. https://www.anzu.io/news/claritas Accessed 2026-09-29.
[395] Venatus delivers 5x higher video ad attention by activating audiences In the Zone, Lumen Research finds. Venatus Media (release hosted by Lumen Research), 2026-03-18. https://lumen-research.com/press-releases/venatus-delivers-5x-higher-video-ad-attention-by-activating-audiences-in-the-zone-lumen-research-finds/ Accessed 2026-09-28.
[396] Unlocking the Gaming Mindset: How Admazing Helps Brands Connect. Admazing Co. (Eddy Prado), 2025-10-05 (date in blog index). https://admazing.com/blog/unlocking-the-gaming-mindset-how-admazing-helps-brands-connect Accessed 2026-09-29.
[397] Brand Safety Policy & Ethic Guidelines (PDF embedded in anzu.io/brand-safety-policy). Anzu Virtual Reality Ltd., 2023-04-30. https://www.anzu.io/hubfs/Anzu%20Brand%20Safety%20Policy%20-%20April%2030%202023.pdf Accessed 2026-09-28.
[398] Brand Safety & Takedown Policy. Gadsme, 2026-02. https://www.gadsme.com/brand-safety Accessed 2026-09-28.
[399] Inventory Partner General Terms & Conditions (including iion Inventory Guidelines and a data processing addendum). iion Pty Ltd, 2026-09-23 (sitemap lastmod). https://www.iion.io/inventory-partner-general-terms-conditions Accessed 2026-09-29.
[400] Super League and Common Sense Networks Join Forces To Launch Global Initiative For Safe, Age-Appropriate Advertising And Immersive Content for Marketers Looking To Reach Youth Audiences. Super League Enterprise, Inc. (GlobeNewswire release on the investor site), 2024-03-26. https://ir.superleague.com/news-events/press-releases/detail/217/super-league-and-common-sense-networks-join-forces-to Accessed 2026-09-28.
[401] Brand Safety Policy - AdInMo (Wayback Machine capture of 7 October 2025). AdInMo Ltd (archived by the Internet Archive), undated. https://web.archive.org/web/20251007180118id_/https://www.adinmo.com/brand-safety-policy/ Accessed 2026-09-29.
[402] WFA discontinues GARM. World Federation of Advertisers, 2024-08-09. https://wfanet.org/knowledge/item/2024/08/09/wfa-discontinues-garm Accessed 2026-09-28.
[403] Taxonomies: implementation.md (Content Taxonomy 3.0 Implementation Guidance). IAB Technology Laboratory (GitHub), 2025-08-05 (last commit). https://github.com/InteractiveAdvertisingBureau/Taxonomies/blob/develop/implementation.md Accessed 2026-09-29.
[404] Gameloft for brands Launches COMBO! Kids, a Media Offer Designed to Reach Generation Alpha & Parents within a Secure Gaming Environment. Gameloft for brands, 2023-07-11. https://www.gameloft.com/for-brands/news/gameloft-for-brands-launches-combo-kids-a-media-offer-designed-to-reach-generation-alpha-parents-within-a-secure-gaming-environment Accessed 2026-09-28.
[405] Brand Safety Policy (Gameloft for brands). Gameloft, undated (footer copyright 2022). https://www.gameloft.com/for-brands/policy Accessed 2026-09-29.
[406] Terms & policies (Terms of Use; Corporate and Website Privacy Policy; Technology Privacy Policy; Do Not Sell or Share; Cookie Policy). Frameplay Corporation, 2023-01-01. https://frameplay.com/terms/ Accessed 2026-09-29.
[407] Immersive ads (Creator Hub documentation). Roblox Corporation, undated. https://create.roblox.com/docs/production/monetization/immersive-ads Accessed 2026-09-28.
[408] Immersive Ads: Expanded Publisher Eligibility and Brand Suitability for Advertisers. Roblox Corporation (DevForum announcement), 2024-01-24. https://devforum.roblox.com/t/immersive-ads-expanded-publisher-eligibility-and-brand-suitability-for-advertisers/2807446 Accessed 2026-09-28.
[409] Activision Blizzard Media | Brand Safety (archived). Activision Blizzard Media; archived by the Internet Archive Wayback Machine, undated (footer copyright 2022). http://web.archive.org/web/20260121071220/https://www.activisionblizzardmedia.com/brand-safety Accessed 2026-09-29.
[410] EA Privacy and Cookie Policy. Electronic Arts Inc., 2025-06-23 (Last Updated). https://www.ea.com/legal/privacy-and-cookie-policy Accessed 2026-09-29.
[411] Take-Two Interactive Software Privacy Policy. Take-Two Interactive Software, Inc., 2026-02 (Last Update: February 2026). https://www.take2games.com/privacy/en-US/ (linked as "Privacy" from zyngaads.com) Accessed 2026-09-29.
[412] Content Maturity Labels. Roblox Corporation (Roblox Support), undated. https://en.help.roblox.com/hc/en-us/articles/8862768451604-Content-Maturity-Labels Accessed 2026-09-28.
[413] Roblox Partners With IARC to Enhance Global Age and Content Ratings. Roblox Corporation (Business Wire release on the investor relations site), 2025-09-03. https://ir.roblox.com/news/news-details/2025/Roblox-Partners-With-IARC-to-Enhance-Global-Age-and-Content-Ratings/default.aspx Accessed 2026-09-28.
[414] Content maturity and compliance (Creator Hub documentation source). Roblox Corporation (Roblox/creator-docs), 2026-09-26. https://github.com/Roblox/creator-docs/blob/main/content/en-us/production/promotion/content-maturity.md Accessed 2026-09-29.
[415] SuperAwesome Partners with Roblox to Bring Brand Safe Advertising to Gen Alpha. SuperAwesome press release, syndicated by Roastbrief, 2026-06-04. https://roastbrief.us/superawesome-partners-with-roblox-to-bring-brand-safe-advertising-to-gen-alpha/ Accessed 2026-09-28.
[416] Minecraft Usage Guidelines for Fans and Creators. Mojang AB / Microsoft, undated. https://www.minecraft.net/en-us/usage-guidelines Accessed 2026-09-28.
[417] As Election Day arrives, Twitch clarifies the types of "Sensitive Social Issues" that require content labels. Tubefilter, 2024-11-05. https://www.tubefilter.com/2024/11/05/twitch-content-classification-labels-politics-sensitive-issues/ Accessed 2026-09-29.
[418] WFA and X settle over GARM litigation. World Federation of Advertisers, 2026-07-29. https://wfanet.org/knowledge/item/2026/07/29/wfa-and-x-settle-over-garm-litigation Accessed 2026-09-28.
[419] Content Taxonomy 3.0 Implementation Guide. IAB Technology Laboratory, 2021-09. https://iabtechlab.com/wp-content/uploads/2021/09/Implementation-Guide-Content-Taxonomy-3-0-pc-Sept2021.pdf Accessed 2026-09-28.
[420] Verify intrinsic in-game buys in Frameplay's gaming ecosystem. Integral Ad Science, 2023-10-05. https://integralads.com/insider/frameplay-transparency-media-measurement/ Accessed 2026-09-29.
[421] FTC Finalizes Changes to Children's Privacy Rule Limiting Companies' Ability to Monetize Kids' Data. Federal Trade Commission, 2025-01-16. https://www.ftc.gov/news-events/news/press-releases/2025/01/ftc-finalizes-changes-childrens-privacy-rule-limiting-companies-ability-monetize-kids-data Accessed 2026-09-28.
[422] COPPA - Enforcement Policy Statement Promoting the Adoption of Age-Verification Technology. Federal Trade Commission, 2026-02-25. https://www.ftc.gov/system/files/ftc_gov/pdf/coppa-age-verification-policy-statement.pdf Accessed 2026-09-28.
[423] FTC Issues COPPA Policy Statement to Incentivize the Use of Age Verification Technologies to Protect Children Online. Federal Trade Commission, 2026-02-25. https://www.ftc.gov/news-events/news/press-releases/2026/02/ftc-issues-coppa-policy-statement-incentivize-use-age-verification-technologies-protect-children Accessed 2026-09-28.
[424] Fortnite Video Game Maker Epic Games to Pay More Than Half a Billion Dollars over FTC Allegations of Privacy Violations and Unwanted Charges. Federal Trade Commission, 2022-12-19. https://www.ftc.gov/news-events/news/press-releases/2022/12/fortnite-video-game-maker-epic-games-pay-more-half-billion-dollars-over-ftc-allegations Accessed 2026-09-28.
[425] FTC Will Require Microsoft to Pay $20 million over Charges it Illegally Collected Personal Information from Children without Their Parents' Consent. Federal Trade Commission, 2023-06-05. https://www.ftc.gov/news-events/news/press-releases/2023/06/ftc-will-require-microsoft-pay-20-million-over-charges-it-illegally-collected-personal-information Accessed 2026-09-28.
[426] Genshin Impact Game Developer Will be Banned from Selling Lootboxes to Teens Under 16 without Parental Consent, Pay a $20 Million Fine to Settle FTC Charges. Federal Trade Commission, 2025-01-17. https://www.ftc.gov/news-events/news/press-releases/2025/01/genshin-impact-game-developer-will-be-banned-selling-lootboxes-teens-under-16-without-parental Accessed 2026-09-28.
[427] COPPA Safe Harbor Program (approved programs, related cases and releases). Federal Trade Commission. https://www.ftc.gov/enforcement/coppa-safe-harbor-program Accessed 2026-09-28.
[428] Supreme Court allows Texas to enforce law requiring age verification and parental consent on apps. Amy Howe, SCOTUSblog, 2026-07-06. https://www.scotusblog.com/2026/07/supreme-court-allows-texas-to-enforce-law-requiring-age-verification-and-parental-consent-on-app/ Accessed 2026-09-28.
[429] House Bill No. 570 (Enrolled), 2025 Regular Session. Louisiana Legislature, 2025. https://www.legis.la.gov/legis/ViewDocument.aspx?d=1425304 Accessed 2026-09-28.
[430] AB-1043 Age verification signals: software applications and online services (Chapter 675, Statutes of 2025). California Legislature (leginfo), 2025-10-14. https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260AB1043 Accessed 2026-09-28.
[431] Update on age requirements for apps distributed in Texas. Apple Developer, 2025-12-23. https://developer.apple.com/news/?id=8jzbigf4 Accessed 2026-09-28.
[432] S.B. 142 App Store Accountability Act (Enrolled Copy), 2025 General Session. Utah State Legislature, 2025. https://le.utah.gov/Session/2025/bills/enrolled/SB0142.pdf Accessed 2026-09-28.
[433] H.B. 498 Utah App Store Accountability Act Amendments (Enrolled Copy), 2026 General Session. Utah State Legislature, 2026. https://le.utah.gov/Session/2026/bills/enrolled/HB0498.pdf Accessed 2026-09-28.
[434] Act No. 185 (2026 Regular Session), House Bill No. 977, enrolled. Louisiana Legislature, 2026-05-15. https://legis.la.gov/legis/ViewDocument.aspx?d=1475238 Accessed 2026-09-30.
[435] AB-1856 Age verification signals: software applications (Chapter 184, Statutes of 2026): text, status and history. California Legislature (leginfo), 2026-09-10. https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260AB1856 Accessed 2026-09-29.
[436] Maryland Online Data Privacy Act of 2024, Chapter 455 (Senate Bill 541). Maryland General Assembly, 2024-05-09. https://mgaleg.maryland.gov/2024rs/Chapters_noln/CH_455_sb0541e.pdf Accessed 2026-09-28.
[437] Declared Age Range (framework documentation). Apple Developer Documentation. https://developer.apple.com/documentation/declaredagerange Accessed 2026-09-28. Access: partial extract.
[438] Next steps for apps distributed in Texas. Apple Developer, 2025-11-04. https://developer.apple.com/news/?id=2ezb6jhj Accessed 2026-09-28.
[439] Update for Apps Distributed in Texas. Apple Developer, 2026-06-03. https://developer.apple.com/news/?id=sg176nne Accessed 2026-09-29.
[440] Apple Developer Program License Agreement, § 3.3.3(O) Declared Age Range API and Significant App Update Topic API. Apple Inc., 2026-08-18. https://developer.apple.com/support/terms/apple-developer-program-license-agreement/ Accessed 2026-09-29.
[441] Viewer Controls. Twitch Safety Center, undated. https://safety.twitch.tv/s/article/Viewer-Controls?language=en_US Accessed 2026-09-29.
[442] Introducing Parental Controls for Teen Accounts. Twitch Safety Center (Christine Weber), 2026-07-21. https://safety.twitch.tv/s/article/Introducing-Parental-Controls-for-Teen-Accounts?language=en_US Accessed 2026-09-29.
[443] Transfer Order, In re: Roblox Corporation Child Sexual Exploitation and Assault Litigation, MDL No. 3166. United States Judicial Panel on Multidistrict Litigation, 2025-12. https://www.jpml.uscourts.gov/sites/jpml/files/MDL-3166-Transfer_Order-12-25.pdf Accessed 2026-09-29.
[444] Gadsme Privacy Policy. Gadsme, 2026-02-24. https://www.gadsme.com/privacy-policy Accessed 2026-09-28.
[445] Anzu Privacy Policy (PDF embedded in anzu.io/privacy-policy). Anzu Virtual Reality Ltd., 2025-12. https://www.anzu.io/hubfs/Anzu%20-%20Privacy%20Policy%20-%20December%202025.pdf Accessed 2026-09-29.
[446] Live TV on PS5 launches today, bringing free ad-supported TV to PS5 players in the U.S.. Grace Chen, PlayStation.Blog (Sony Interactive Entertainment), 2026-08-31. https://blog.playstation.com/2026/08/31/live-tv-on-ps5-launches-today-bringing-free-ad-supported-tv-to-ps5-players-in-the-u-s/ Accessed 2026-09-28.
[447] Frequently Asked Questions (FAQs) for GeForce NOW. NVIDIA, undated. https://www.nvidia.com/en-us/geforce-now/faq/ Accessed 2026-09-28.
[448] Amazon Luna: Everything you need to know about Amazon's cloud gaming service. Android Authority, 2025-05-06. https://www.androidauthority.com/amazon-luna-1170676/ Accessed 2026-09-28.
[449] US 2026/0260258 A1: Contextually Aware Management of Interactive Software Application Access. United States Patent and Trademark Office; applicant Microsoft Technology Licensing, LLC, 2026-09-03. https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/20260260258 Accessed 2026-09-28.
[450] US 2026/0138036 A1: Delivery of Alternate Content During Video Game Wait Times. United States Patent and Trademark Office; applicant Sony Interactive Entertainment Inc., 2026-05-21. https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/20260138036 Accessed 2026-09-28.
[451] Sony is scaling up its Playstation advertising ambitions. Digiday, 2026-08-05. https://digiday.com/media-buying/sony-is-scaling-up-its-playstation-advertising-ambitions/ Accessed 2026-09-28.
[452] Nintendo Switch store blocks indie dev's release over 'in-game advertising' for including references to previous titles in series. AUTOMATON WEST, 2025-05-20. https://automaton-media.com/en/news/nintendo-switch-store-blocks-indie-devs-release-over-in-game-advertising-for-including-references-to-previous-titles-in-series/ Accessed 2026-09-28.
[453] Advertising on Steam (Steamworks Documentation). Valve, undated. https://partner.steamgames.com/doc/marketing/advertising Accessed 2026-09-28.
[454] Epic Games Store Requirements. Epic Games (Epic Developer Resources), undated. https://dev.epicgames.com/docs/epic-games-store/requirements-guidelines/distribution-requirements/requirements-overview Accessed 2026-09-28.
[455] Xbox Activities on landing experiences and other product updates. Microsoft Advertising (Caroline Bauer, XBOX Media Solutions), 2025-06-30. https://about.ads.microsoft.com/en/blog/post/june-2025/2025-q2-product-updates Accessed 2026-09-29.
[456] The Future Is In Play: Gaming as Advertising's Most Powerful Ecosystem. Microsoft Advertising (Claire Nance), 2026-04-06. https://about.ads.microsoft.com/en/blog/post/april-2026/the-future-is-in-play-gaming-as-advertisings-most-powerful-ecosystem Accessed 2026-09-28.
[457] Microsoft Invest - Programmatic buying on Xbox for marketers. Microsoft Learn (Microsoft Invest documentation), 2024-05-17 (last updated). https://learn.microsoft.com/en-us/xandr/invest/programmatic-buying-on-xbox-for-marketers-retired Accessed 2026-09-29.
[458] Xbox testing ad-supported streaming, Amazon brings Luna to Prime Video. CNBC, 2026-07-23. https://www.cnbc.com/2026/07/23/amazon-microsoft-cloud-gaming.html Accessed 2026-09-28.
[459] Xbox Is Asking Customers to Volunteer to Test Ad-Supported Game Streaming. Jennifer Maas, Variety, 2026-07-23. https://variety.com/2026/gaming/news/xbox-tests-ad-supported-game-streaming-1236819193/ Accessed 2026-09-28.
[460] Upcoming Changes to XBOX Cloud Gaming. Team XBOX (Microsoft), Xbox Wire, 2026-09-03. https://news.xbox.com/en-us/2026/09/03/xbox-cloud-gaming-changes/ Accessed 2026-09-28.
[461] Microsoft reportedly wants to show ads in free-to-play Xbox games. XDA Developers (relaying Business Insider), 2022-04-18. https://www.xda-developers.com/microsoft-wants-to-show-ads-xbox-games/ Accessed 2026-09-28.
[462] Anzu.io Expands Programmatic In-Game Advertising Offerings on Xbox. Anzu (The Drum profile; originally Gamasutra, 6 November 2019), 2020-07-06. https://www.thedrum.com/profile/anzu-io/news/anzu-io-expands-programmatic-in-game-advertising-offerings-on-xbox Accessed 2026-09-29.
[463] Anzu Platform (Microsoft Marketplace listing). Anzu Virtual Reality Ltd. (publisher listing on Microsoft Marketplace). https://marketplace.microsoft.com/en-us/product/web-apps/anzuvirtualrealityltd1620134365610.anzu_platform?tab=overview Accessed 2026-09-29.
[464] Frameplay | Maximize Revenue with Seamless Intrinsic In-Game Advertising (developers page). Frameplay, undated (footer 2026). https://frameplay.com/developers/ Accessed 2026-09-29.
[465] Changelog | Frameplay SDK (Unity). Frameplay, 2024-10-16. https://docs.frameplay.gg/unity/upgrading/changelog/ Accessed 2026-09-29.
[466] Unity Quick Start (SDK v3) | Frameplay SDK. Frameplay, 2025-10-08. https://docs.frameplay.gg/unity/ Accessed 2026-09-29.
[467] Gadsme in the Press (newsroom). Gadsme. https://www.gadsme.com/press Accessed 2026-09-28.
[468] Sony Aims to Connect Advertisers to PlayStation Users With Launch of 'Live TV on PS5' (EXCLUSIVE). Variety, 2026-08-31. https://variety.com/2026/tv/news/sony-playstation-advertisers-fast-channels-live-1236847258/ Accessed 2026-09-28.
[469] Anzu brings intrinsic in-game ads to AAA football hit UFL. Anzu (via Games Press), 2026-05-13. https://www.gamespress.com/ANZU-BRINGS-INTRINSIC-IN-GAME-ADS-TO-AAA-FOOTBALL-HIT-UFL Accessed 2026-09-28.
[470] UFL Launches on PC - Free to Play. Operation Sports, 2025-12-05. https://www.operationsports.com/ufl-launches-on-pc-free-to-play/ Accessed 2026-09-29.
[471] UFL on Steam (store page). Valve (Steam store listing). https://store.steampowered.com/app/3023930/UFL/ Accessed 2026-09-29.
[472] Make-A-Wish International and Gadsme bring 'Unwrap the Power of a Wish' campaign to millions of gamers worldwide. Mobidictum, 2025-12-11. https://mobidictum.com/make-a-wish-international-gadsme-partnership/ Accessed 2026-09-29.
[473] Game On: Lowe's to Appear in Multiple EA SPORTS Titles for the First Time. Lowe's Companies (PR Newswire), 2025-04-11. https://www.prnewswire.com/news-releases/game-on-lowes-to-appear-in-multiple-ea-sports-titles-for-the-first-time-302426343.html Accessed 2026-09-29.
[474] Intrinsic In-Game Ads Drive Stronger Purchase Power Than Display, Video, and CTV, New Research Reveals. Anzu, 2024-03-26. https://www.anzu.io/news/anzu-happydemics-full-funnel-research Accessed 2026-09-28.
[475] Frameplay and Happydemics Study: In-Game Ads Outperform Traditional Digital on Key Brand Metrics. Frameplay Corp. (PR Newswire), 2025-10-27. https://www.prnewswire.com/news-releases/frameplay-and-happydemics-study-in-game-ads-outperform-traditional-digital-on-key-brand-metrics-302595522.html Accessed 2026-09-28.
[476] The Black Phone 2 Quest Success Story. Discord, undated (live page). https://discord.com/quests-success-stories/black-phone-2 Accessed 2026-09-28.
[477] Ipsos Study of 37,000 Gamers Finds In-Game Advertising Significantly Improves Brand Affinity. Ipsos (UK), 2026-08-07. https://www.ipsos.com/en-uk/game-advertising-significantly-improves-brand-affinity Accessed 2026-09-29.
[478] Frameplay's Intrinsic In-Game Advertising Solution Outperforms Other Media Channels in Key Ad Recall and Creative Categories Per Happydemics' Meta Analysis. Frameplay Corp. (PR Newswire), 2024-03-21. https://www.prnewswire.com/news-releases/frameplays-intrinsic-in-game-advertising-solution-outperforms-other-media-channels-in-key-ad-recall-and-creative-categories-per-happydemics-meta-analysis-302095290.html Accessed 2026-09-28.
[479] Comscore and Frameplay Launch Brand Lift Survey Series to Better Understand Intrinsic In-Game Advertising Performance. Comscore (PR Newswire), 2021-01-12. https://www.prnewswire.com/news-releases/comscore-and-frameplay-launch-brand-lift-survey-series-to-better-understand-intrinsic-in-game-advertising-performance-301205748.html Accessed 2026-09-28.
[480] From Play to Purchase: Harnessing Gaming’s Power for Deeper Brand Connections. Microsoft Advertising (Activision Blizzard Media blog, redirected), 2024-10. https://about.ads.microsoft.com/en/blog/post/october-2024/harnessing-gaming-s-power-for-deeper-brand-connections Accessed 2026-09-29.
[481] Dentsu's New Study Unveils the Power of Intrinsic In-Game Ads in Capturing Attention. Anzu, 2024-01-26. https://www.anzu.io/blog/anzu-dentsu-lumen-attention-study Accessed 2026-09-28.
[482] New research shows in-game ad campaigns drive extremely high levels of attention (Anzu and Lumen). Anzu, 2022-10-06. https://www.anzu.io/news/lumen-study Accessed 2026-09-28.
[483] Anzu and Lumen two-year attention study (news release, Cannes Lions 2023). Anzu, 2023-06-21. https://www.anzu.io/news/attention-report-pr Accessed 2026-09-28.
[484] Roblox courts advertisers with fresh data on its ad effectiveness. Digiday, 2025-05-12. https://digiday.com/media/roblox-courts-advertisers-with-fresh-data-on-its-ad-effectiveness/ Accessed 2026-09-28.
[485] Roblox says people have better brand recall with its ads. Do companies care?. Tubefilter, 2025-05-12. https://www.tubefilter.com/2025/05/12/roblox-ad-effectiveness-study-brand-recall-eye-tracking-newfronts/ Accessed 2026-09-28.
[486] Roblox Announces New Ad Research and Measurement Partnerships at Cannes Lions. Roblox, 2026-06-18. https://about.roblox.com/newsroom/2026/06/roblox-ipsos-edo-immersive-effective-measurement-partnerships-cannes-lions Accessed 2026-09-28.
[487] Roblox's Immersive is Effective (report download page). Roblox, 2026-06. https://brands.roblox.com/immersiveiseffective Accessed 2026-09-28. Access: partial extract.
[488] EDO and Roblox Bring Investment-Grade Outcomes Measurement to Gaming Advertising for the First Time. EDO, 2026-06-18. https://www.edo.com/resources/roblox-gaming-ad-outcomes Accessed 2026-09-28.
[489] Obsession Quest Success Story (Focus Features). Discord, undated (live page). https://discord.com/quests-success-stories/obsession Accessed 2026-09-28.
[490] World of Warcraft x Mountain Dew: Crown of the Violet Rose Campaign Case Study. Discord, undated (live page). https://discord.com/quests-success-stories/world-of-warcraft-mountain-dew Accessed 2026-09-28.
[491] Wendy's x Wednesday 'Meal of Misfortune' Video Quest success story. Discord, undated (live page). https://discord.com/quests-success-stories/wendys Accessed 2026-09-29.
[492] Science has spoken: gaming is a brand-building power play. The Drum (Drum Studios, sponsored by Gameloft for brands), 2026-06-01. https://www.thedrum.com/news/science-has-spoken-gaming-is-a-brand-building-power-play Accessed 2026-09-28.
[493] Attention & Emotion in Play: How Gaming is Redefining Advertising Effectiveness (report download page). Gameloft for brands, 2026. https://www.gameloft.com/for-brands/attention-and-emotion-in-play Accessed 2026-09-28. Access: partial extract.
[494] Digital Turbine Mobile Video Ads Grab 22 Seconds of User Attention, Driving Better Brand Recall - Amplified Intelligence Research Finds. Digital Turbine, Inc. (PR Newswire), 2023-04-03. https://www.prnewswire.com/news-releases/digital-turbine-mobile-video-ads-grab-22-seconds-of-user-attention-driving-better-brand-recall---amplified-intelligence-research-finds-301787771.html Accessed 2026-09-28.
[495] How Garnier Achieved 7x the Industry Benchmark in Video Game Ad Attention. iion, undated (live page). https://www.iion.io/case-studies/garnier-achieved-7x-the-industry-benchmark-in-video-game-ad-attention Accessed 2026-09-28.
[496] Super League | Case Studies. Super League Enterprise, Inc., undated (live page, copyright 2026). https://www.superleague.com/case-studies Accessed 2026-09-28.
[497] In-Game Ads Beat CPA Goals by 21%, Revealing Hidden Advertiser Opportunities, According to New Report from Anzu. Anzu (Nick Woodford), 2025-04-30. https://www.anzu.io/news/trends-report Accessed 2026-09-29.
[498] The Gamification of Branded Content: A Meta-Analysis of Advergame Effects. van Berlo, Z. M. C.; van Reijmersdal, E. A.; Eisend, M., Journal of Advertising, 2021-01-26. https://doi.org/10.1080/00913367.2020.1858462 Accessed 2026-09-28. Access: abstract only.
[499] Brand placements in video games: How local in-game experiences influence brand attitudes. Ingendahl, M.; Vogel, T.; Maedche, A.; Wänke, M., Psychology & Marketing, 2022-12-05. https://doi.org/10.1002/mar.21770 Accessed 2026-09-28. Access: abstract only.
[500] Beyond Affect Transfer: Attribute Associations in Video-Game Brand Placements and Their Impact on Brand Attitudes. Ingendahl, M.; Brückner, L.; Vogel, T., Journal of Advertising, 2024-04-26. https://doi.org/10.1080/00913367.2024.2337428 Accessed 2026-09-28. Access: abstract only.
[501] From intrusion to advantage: uncovering the mediating role of perceived benefit in consumers' responses to in-game advertising. Zagni, L. M.; Rizzo, C.; Filieri, R., International Journal of Advertising, 2026-06-18. https://doi.org/10.1080/02650487.2026.2689797 Accessed 2026-09-28. Access: abstract only.
[502] Tap Intention Toward Rewarded Mobile In-Game Banner Advertising: The Roles of Category-Based Contextual Fit and Reward Type. Hatakeyama, Y. (Rissho University), Journal of Current Issues & Research in Advertising, 2026-08-03. https://doi.org/10.1080/10641734.2026.2701036 Accessed 2026-09-28. Access: abstract only.
[503] The Promotional Effects of Live Streams by Twitch Influencers. Huang, Y.; Morozov, I., Marketing Science 44(4), 2025-02-20. https://doi.org/10.1287/mksc.2022.0400 Accessed 2026-09-28. Access: abstract only.
[504] The Promotional Effects of Live Streams by Twitch Influencers (IDEAS/RePEc record with full abstract). Huang, Y.; Morozov, I. (Marketing Science 44(4):916-932), 2025. https://ideas.repec.org/a/inm/ormksc/v44y2025i4p916-932.html Accessed 2026-09-29. Access: abstract only.
[505] Effect of in-game situations and advertisement animation in eSports on visual attention, memory, brand attitude and behavioral intentions. Min, B.; Kim, D.; Ko, Y. J., Internet Research, 2025-03-14. https://doi.org/10.1108/intr-03-2024-0491 Accessed 2026-09-28. Access: abstract only.
[506] Super League Enterprise, Inc. Form 8-K (event date 30 April 2026). Super League Enterprise, Inc. (SEC EDGAR), 2026-05-06. https://www.sec.gov/Archives/edgar/data/1621672/000143774926015035/slgg20260506_8k.htm Accessed 2026-09-28.
[507] Metaverse's Biggest Builder Merges With Game Tech Firm To Bring Brands Into Web3. Admix / LandVault (company-posted release on The Drum's company profile pages), 2022-06-21. https://www.thedrum.com/profile/admix/article/metaverses-biggest-builder-merges-with-game-tech-firm-to-bring-brands-into-web3-1 Accessed 2026-09-28.
[508] Roblox Eyes Expanded Placements For 'Rewarded' Video Ads. MediaPost (Colin Kirkland), 2026-09-15. https://www.mediapost.com/publications/article/417959/roblox-eyes-expanded-placements-for-rewarded-vid.html Accessed 2026-09-28. Access: partial extract.
[509] BIDSTACK GROUP PLC (04466195): filing history. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/04466195/filing-history Accessed 2026-09-28.
[510] Israel Registrar of Companies dataset (list of companies), records for 514656453 (Sayollo Media Ltd), 515647808 (Anzu Virtual Reality Ltd.), 514330331 (Overwolf Ltd) and 514191527 (Kidoz Ltd). Israel Corporations Authority, Registrar of Companies, via data.gov.il, 2026-09-28 (dataset last modified). https://data.gov.il/api/3/action/datastore_search?resource_id=f004176c-b85f-4542-8901-7b3176f9a054&q=514656453 Accessed 2026-09-28.
[511] Super Biz / Evolving. Super League Enterprise, Inc. (Super Biz), undated. https://superbiz.gg/ Accessed 2026-09-28.
[512] BIDSTACK GROUP PLC (04466195): insolvency. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/04466195/insolvency Accessed 2026-09-28.
[513] Super League Enterprise, Inc. Form 8-K (items 1.01, 3.02, 8.01, 9.01): Metaplanet and Evo subscription agreements. Super League Enterprise, Inc. (SEC EDGAR), 2026-08-18. https://www.sec.gov/Archives/edgar/data/1621672/000143774926028267/slgg20260818_8k.htm Accessed 2026-09-29.
[514] Super League Enterprise, Inc. definitive proxy statement (DEF 14A), 2026 annual meeting. Super League Enterprise, Inc. (SEC EDGAR), 2026-09-25. https://www.sec.gov/Archives/edgar/data/1621672/000143774926031210/slgg20260924_def14a.htm Accessed 2026-09-29.
[515] Infinite Reality Closes $350 Million Investment; Acquires Landvault in $450 Million Deal, Valuation Soars to $5.1 Billion (Exhibit 99.2 to Newbury Street Acquisition Corp Form 425). Infinite Reality, Inc. (release filed with the SEC by Newbury Street Acquisition Corp), 2024-07-09. https://www.sec.gov/Archives/edgar/data/1831978/000121390024059983/ea020912402ex99-2_newbury.htm Accessed 2026-09-28.
[516] Admix - Connecting Brands to Casual Gamers (admixplay.com home page and its JavaScript bundle). Site operator named on the site as W Media Ltd / W Media International Ltd, 2026-07-21 (embedded seller data last_updated). https://admixplay.com/ Accessed 2026-09-28.
[517] Companies House company search results for 'w media international'. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/search/companies?q=w+media+international Accessed 2026-09-28.
[518] Sayollo launches direct-to-consumer in-game marketplace gComm. PocketGamer.biz, 2021-12-01. https://www.pocketgamer.biz/news/77641/sayollo-direct-to-consumer-in-game-marketplace-gcomm/ Accessed 2026-09-29. Access: access error; text or metadata via another route.
[519] BIDSTACK LTD (09835625): persons with significant control. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/09835625/persons-with-significant-control Accessed 2026-09-28.
[520] Enthusiast Gaming press releases (listing). Enthusiast Gaming Holdings Inc., live page. https://www.enthusiastgaming.com/press-releases/ Accessed 2026-09-28. Access: partial extract.
[521] VENATUS MEDIA LIMITED (07183957): company overview. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/07183957 Accessed 2026-09-29.
[522] Microsoft Corporation Form 8-K dated 13 October 2023 (Item 2.01, completion of the Activision Blizzard acquisition). Microsoft Corporation, 2023-10-13. https://www.sec.gov/Archives/edgar/data/789019/000119312523255762/d537928d8k.htm Accessed 2026-09-28.
[523] Israel Registrar of Companies dataset, record 515647808 (Anzu Virtual Reality Ltd.). Israel Corporations Authority via data.gov.il. https://data.gov.il/api/3/action/datastore_search?resource_id=f004176c-b85f-4542-8901-7b3176f9a054&q=515647808 Accessed 2026-09-29.
[524] French company register search API (recherche-entreprises.api.gouv.fr): GADSME (SIREN 884197799) and GAMELOFT SE (SIREN 429338130). French government (annuaire des entreprises, INSEE/INPI data), live register. https://recherche-entreprises.api.gouv.fr/search?q=gadsme&per_page=3 Accessed 2026-09-28.
[525] ABN Lookup: current details for ABN 50 663 746 487 (IION PTY LTD). Australian Business Register, live register. https://abr.business.gov.au/Search/ResultsActive?SearchText=663746487 Accessed 2026-09-28.
[526] 2026 IAB PlayFronts (event page). IAB, 2026-09. https://www.iab.com/events/playfronts-2026/ Accessed 2026-09-29.
[527] Ads Manager (Creator Hub documentation source). Roblox Corporation (Roblox/creator-docs). https://github.com/Roblox/creator-docs/blob/main/content/en-us/production/promotion/ads-manager.md Accessed 2026-09-28.
[528] Super League home page. Super League Enterprise, Inc.. https://www.superleague.com/ Accessed 2026-09-28.
[529] Advertise on Roblox (Creator Hub documentation source). Roblox Corporation (Roblox/creator-docs). https://github.com/Roblox/creator-docs/blob/main/content/en-us/production/promotion/advertise-on-roblox.md Accessed 2026-09-28.
[530] RDC 2026: The World Needs More Play. Roblox Corporation (newsroom), 2026-09-11. https://about.roblox.com/newsroom/2026/09/rdc-2026-the-world-needs-more-play Accessed 2026-09-28.
[531] Game Monetization for Game Studios | Gadsme (publishers page). Gadsme, undated. https://www.gadsme.com/publishers Accessed 2026-09-29.
[532] Create assets for Twitch in seconds with Gen-AI. Amazon Ads (launch announcement), 2024-09-30. https://advertising.amazon.com/resources/whats-new/create-assets-for-twitch-in-seconds-with-gen-ai Accessed 2026-09-29.
[533] IABTechLab/agentic-real-time-framework release v1.0. IAB Tech Lab, 2026-07-20. https://github.com/IABTechLab/agentic-real-time-framework/releases/tag/v1.0 Accessed 2026-09-28.
[534] ARTF (Agentic Real Time Framework). IAB Tech Lab, 2025-11-13. https://iabtechlab.com/standards/artf/ Accessed 2026-09-28.
[535] AAMP 2.0 Release Brings Transaction-ready Buyer and Seller Agent SDKs. IAB Tech Lab, 2026-04-23. https://iabtechlab.com/aamp-2-0-release-brings-transaction-ready-buyer-and-seller-agent-sdks/ Accessed 2026-09-28.
[536] IABTechLab/seller-agent (README and releases). IAB Tech Lab, 2026-08-18. https://github.com/IABTechLab/seller-agent Accessed 2026-09-28.
[537] Agentic Mobile specification (docs/specification.md). IAB Tech Lab. https://github.com/IABTechLab/agentic-mobile/blob/main/docs/specification.md Accessed 2026-09-28.
[538] Roblox app-ads.txt. Roblox Corporation, 2026-09-29 (as served). https://www.roblox.com/app-ads.txt Accessed 2026-09-29.
[539] Magnite (Rubicon Project) sellers.json. Magnite, Inc.. https://sellers.rubiconproject.com/sellers.json Accessed 2026-09-29.
[540] Twitch ad specifications and guidelines. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch Accessed 2026-09-28.
[541] Medal Privacy Policy (global / GDPR version). Medal B.V., 2026-07-13. https://medal.tv/privacy Accessed 2026-09-29.
[542] EDGAR submissions record, Discord Inc. (CIK 0001763840). US Securities and Exchange Commission (EDGAR), live record. https://data.sec.gov/submissions/CIK0001763840.json Accessed 2026-09-28.
[543] Report: Discord confidentially files for IPO. Game Developer, 2026-01-07. https://www.gamedeveloper.com/business/report-discord-confidentially-files-for-ipo Accessed 2026-09-28.
[544] Mobile In-Game Advertising: More Game Developers Turn to Advertising. EMARKETER (Yoram Wurmser), 2019-10-17. https://www.emarketer.com/content/mobile-in-game-advertising Accessed 2026-09-28. Access: partial extract.
[545] 2025 Essential Facts About the U.S. Video Game Industry (web page). Entertainment Software Association (ESA), 2025. https://www.theesa.com/resources/essential-facts-about-the-us-video-game-industry/2025-data/ Accessed 2026-09-28.
[546] Campaign measurement page. Admazing Co., undated. https://admazing.com/campaign-measurement Accessed 2026-09-29.
[547] Admazing Privacy Policy. Admazing Co. LLC, 2023-07-02 ("Last updated: July 2, 2023."). https://admazing.com/privacy Accessed 2026-09-29.
[548] Admazing Co. Terms of Use. Admazing Co., 2023-07-02 ("Last updated: July 2, 2023."). https://admazing.com/terms Accessed 2026-09-29.
[549] Adverty and NumberEight: McDonald's Achieves a 58% Uplift in Brand Preference. Adverty AB (publ) (MFN), 2024-01-22. https://mfn.se/a/adverty/adverty-and-numbereight-mcdonalds-achieves-a-58-uplift-in-brand-preference-through-the-fusion-of-gaming-and-id-less-demographic-targeting Accessed 2026-09-29.
[550] Adverty Privacy Policy (footer link /privacy-policy-2021). Adverty AB (publ), 2021-01-12. https://adverty.com/privacy-policy-2021 Accessed 2026-09-29.
[551] Anzu and IAS Launch First-To-Market Measurement Solution To Validate 3D In-Game Media Quality. Anzu (Nick Woodford), 2023-07-11. https://www.anzu.io/news/ias-anzu-partnership Accessed 2026-09-28.
[552] Frameplay | The Future of Intrinsic In-Game Advertising (home page). Frameplay, undated (footer 2026). https://frameplay.com/ Accessed 2026-09-29.
[553] Frameplay Powers Intrinsic In-Game Ad Attribution with Kochava (case study). Kochava, undated (quotes Jonathon Troughton as CEO, a title he held until 2 October 2024; images uploaded 2022-03). https://www.kochava.com/case-studies/frameplay Accessed 2026-09-29.
[554] Data We Collect | Frameplay SDK. Frameplay, 2024-07-02. https://docs.frameplay.gg/guide/data-collected/ Accessed 2026-09-29.
[555] IAB Europe TCF Global Vendor List (vendor-list.json, GVL specification v3). IAB Europe, 2026-09-24. https://vendor-list.consensu.org/v3/vendor-list.json Accessed 2026-09-29.
[556] We raised $8 million in seed round (Gadsme blog). Gadsme, 2022-04-06. https://www.gadsme.com/blog/we-raised-8-million-in-seed-round Accessed 2026-09-29.
[557] iion Demand Policies. iion Pty Ltd, undated. https://www.iion.io/demand-policies Accessed 2026-09-29.
[558] IION Privacy Policy. Iion Pty Ltd and Iion Digital Pty Ltd, 2024-11-04. https://www.iion.io/privacy-policy Accessed 2026-09-29.
[559] McDonald's Reaches Young Audiences Through In-Game Advertising (case study). iion, 2025-06-20. https://www.iion.io/case-studies/mcdonalds-reaches-young-audiences-through-in-game-advertising Accessed 2026-09-29.
[560] Ad Units (High-Impact Playable Creatives) and Frequently Asked Questions. Kidoz Inc., undated (live pages). https://www.kidoz.net/ad-units ; https://www.kidoz.net/faq Accessed 2026-09-29.
[561] Kite IQ - Contextual AI Targeting for Kids' Brands. Kidoz Inc., undated (live page). https://www.kidoz.net/kite-iq Accessed 2026-09-29.
[562] PRIVO Case Study: Kidoz Privacy Shield - Creating a compliant contextual mobile advertising network. PRIVO (Privacy Vaults Online, Inc.), undated (HTTP Last-Modified 2026-09-16). https://www.privo.com/kidoz Accessed 2026-09-29.
[563] Kidoz Website and SDK Privacy Policy. Kidoz Inc., undated (live page). https://www.kidoz.net/privacy-policy Accessed 2026-09-29.
[564] Participate in the Families Self-Certified Ads SDK Program (Play Console Help). Google, undated (live page). https://support.google.com/googleplay/android-developer/answer/9283445?hl=en Accessed 2026-09-29.
[565] SYBO Games app-ads.txt. SYBO Games, undated. https://sybogames.com/app-ads.txt Accessed 2026-09-29.
[566] TAG Registry public search (registry-search GET endpoint). Trustworthy Accountability Group (TAG), live registry. https://hubgloo.opxar.com/search?q=Super%20League Accessed 2026-09-29.
[567] New Super League White Paper Sizes Up the Web’s Most Overlooked Ad-Friendly Audience: 100M+ Browser-Based Gamers. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-09-17. https://ir.superleague.com/news-events/press-releases/detail/305/new-super-league-white-paper-sizes-up-the-webs-most Accessed 2026-09-29.
[568] Super League Reports Second Quarter 2026 Financial Results, Highlighted by Improving Margins and Operating Performance. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-08-14. https://ir.superleague.com/news-events/press-releases/detail/303/super-league-reports-second-quarter-2026-financial-results Accessed 2026-09-29.
[569] Super League Enterprise Privacy Policy. Super League Enterprise, Inc., 2023-09-20. https://www.superleague.com/privacy-policy Accessed 2026-09-29.
[570] TAG registry company search for 'Roblox' (registry search backend). Trustworthy Accountability Group (registry served from hubgloo.opxar.com), 2026-09-29. https://hubgloo.opxar.com/search?q=Roblox Accessed 2026-09-29.
[571] Roblox ads.txt. Roblox Corporation, 2026-09-29 (as served). https://www.roblox.com/ads.txt Accessed 2026-09-29.
[572] PubMatic sellers.json. PubMatic, Inc., 2026-09-29. https://pubmatic.com/sellers.json Accessed 2026-09-29.
[573] Index Exchange sellers.json. Index Exchange, 2026-09-29 (as served). https://cdn.indexexchange.com/sellers.json Accessed 2026-09-29.
[574] Microsoft Casual Games app-ads.txt. Microsoft Casual Games, 2026-03-26 (date in file header). https://www.microsoftcasualgames.com/app-ads.txt Accessed 2026-09-29.
[575] Ninth Circuit Issues Mixed Ruling on California Age-Appropriate Design Code Act. Holland & Knight LLP, 2026-03. https://www.hklaw.com/en/insights/publications/2026/03/ninth-circuit-issues-mixed-ruling-on-california-age-appropriate-design Accessed 2026-09-28.
[576] Regulation (EU) 2022/2065 (Digital Services Act). European Parliament and Council of the European Union, 2022-10. http://publications.europa.eu/resource/cellar/3ff67256-55c4-11ed-92ed-01aa75ed71a1.0006.03/DOC_1 Accessed 2026-09-28.
[577] Age appropriate design: a code of practice for online services, standard 12 Profiling. Information Commissioner's Office (UK). https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/childrens-information/childrens-code-guidance-and-resources/age-appropriate-design-a-code-of-practice-for-online-services/12-profiling/ Accessed 2026-09-28.
[578] Age requirements for apps distributed in Brazil, Australia, Singapore, Utah, and Louisiana. Apple Developer, 2026-02-24. https://developer.apple.com/news/?id=f5zj08ey Accessed 2026-09-28.
[579] In-Game Ad Creative Best Practices Boost Brand Recommendation by 21% and Purchase Intent by 9%. Anzu, 2024-06-27. https://www.anzu.io/news/creative-best-practices Accessed 2026-09-28.
[580] The cross-platform advantage of reaching players everywhere. Activision Blizzard Media / Microsoft Advertising (Melinda Spence), 2025-02-18. https://www.activisionblizzardmedia.com/blog/gaming-industry/cross-platform-advantage-reaching-players-everywhere Accessed 2026-09-28.
[581] TAG Registry company record: DoubleVerify. Trustworthy Accountability Group (TAG), live registry. https://hubgloo.opxar.com/search?q=DoubleVerify Accessed 2026-09-29.
[582] DoubleVerify Holdings, Inc. Annual Report on Form 10-K for the year ended 31 December 2025. DoubleVerify Holdings, Inc. (SEC filing), 2026-02-26. https://www.sec.gov/Archives/edgar/data/1819928/000110465926020499/dv-20251231x10k.htm Accessed 2026-09-29.
[583] DoubleVerify Holdings, Inc. Quarterly Report on Form 10-Q for the quarter ended 30 June 2026. DoubleVerify Holdings, Inc. (SEC filing), 2026-08-07. https://www.sec.gov/Archives/edgar/data/1819928/000110465926092375/dv-20260630x10q.htm Accessed 2026-09-29.
[584] DoubleVerify Holdings, Inc. preliminary proxy statement (PREM14A) for the special meeting on the merger. DoubleVerify Holdings, Inc. (SEC filing), 2026-09-11. https://www.sec.gov/Archives/edgar/data/1819928/000110465926107126/tm2621904-10_prem14a.htm Accessed 2026-09-29.
[585] Wizard Games app-ads.txt. Wizard Games, 2026-03-02. https://wizardgames.ca/app-ads.txt Accessed 2026-09-29.
[586] FUN-GI Games app-ads.txt. FUN-GI GAMES LLC. https://fun-gi.com/app-ads.txt Accessed 2026-09-29.
[587] Castry app-ads.txt. Castry. https://castryve.com/app-ads.txt Accessed 2026-09-29.
[588] Leadership team | Frameplay. Frameplay, undated (live at access). https://frameplay.com/team/ Accessed 2026-09-29.
[589] Sports Interactive: Our Partners. Sports Interactive (SEGA), undated (live page). https://www.sports-interactive.com/partners Accessed 2026-09-29.
[590] US12211141B2: Method for calculating the visibility of objects within a 3D scene. Google Patents (USPTO data); assignee Gadsme, 2025-01-28. https://patents.google.com/patent/US12211141B2/en Accessed 2026-09-29.
[591] Pourquoi les marques vont enfin investir dans l'in-game advertising en 2026. Julia Luczak-Rougeaux, Minted, 2026-01-13. https://www.mntd.fr/pourquoi-les-marques-vont-enfin-investir-dans-lin-game-advertising-en-2026/ Accessed 2026-09-29.
[592] Gadsme advertisers page route script (_layout.advertisers). Gadsme, undated. https://www.gadsme.com/build/routes/_layout.advertisers-U5W5CFC2.js Accessed 2026-09-29.
[593] docs.gadsme.com (developer documentation link from the Gadsme site footer). Gadsme. https://docs.gadsme.com Accessed 2026-09-29. Access: partial extract.
[594] Terms of Use - Gadsme. Gadsme, 2023-04-11. https://www.gadsme.com/terms-of-use Accessed 2026-09-29.
[595] Football Manager 27 | Out November (official site home page). Sports Interactive (SEGA), undated (live page). https://www.footballmanager.com/ Accessed 2026-09-29.
[596] The In-Game Ad Market Is Expanding, One SDK At A Time. AdExchanger, 2024-10-21. https://www.adexchanger.com/gaming/the-in-game-ad-market-is-expanding-one-sdk-at-a-time/ Accessed 2026-09-29.
[597] EDGAR filing index for DoubleVerify Holdings, Inc. (submissions JSON, CIK 0001819928). US Securities and Exchange Commission, live endpoint. https://data.sec.gov/submissions/CIK0001819928.json Accessed 2026-09-29.
[598] Super League Enterprise, Inc. Second Quarter 2026 Earnings Call transcript (Exhibit 99.2 to Form 8-K filed 20 August 2026). Super League Enterprise, Inc., 2026-08-20. https://www.sec.gov/Archives/edgar/data/1621672/000143774926028647/ex_1006627.htm Accessed 2026-09-29.
[599] Super League Named U.S. Ad Sales Partner for Play.Works Connected Television Advergaming Inventory. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-06-09. https://ir.superleague.com/news-events/press-releases/detail/301/super-league-named-u-s-ad-sales-partner-for-play-works Accessed 2026-09-29.
[600] Roblox Help Center API records for Advertising Standards (13722260778260) and Advertising Integrations Terms (47656162239124). Roblox Corporation (Zendesk help-centre API), 2026-04-30. https://en.help.roblox.com/api/v2/help_center/en-us/articles/47656162239124.json Accessed 2026-09-29.
[601] Metaplanet to Invest 2,100 Bitcoin in Super League to Launch U.S. Bitcoin Treasury Platform, Superplanet. Super League Enterprise, Inc. and Metaplanet, Inc., 2026-08-18. https://ir.superleague.com/news-events/press-releases/detail/304/metaplanet-to-invest-2100-bitcoin-in-super-league-to Accessed 2026-09-29.
[602] Twitch Vertical Video Ad Specs. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/twitch-vertical-video-ads Accessed 2026-09-29.
[603] MRC letter to Amazon: accreditation of Amazon DSP (GI052926). Media Rating Council (George W. Ivie), 2026-05-29. https://mediaratingcouncil.org/sites/default/files/Accreditation%20Letters/Amazon/GI052926_Amazon%20DSP%20Accreditation%20Letter.pdf Accessed 2026-09-29.
[604] Twitch ads ad specifications | Super Leaderboard; Medium Rectangle. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/super-leaderboard Accessed 2026-09-29.
[605] ads.txt file for www.twitch.tv. Twitch (OWNERDOMAIN=amazon.com), undated (served 2026-09-29). https://www.twitch.tv/ads.txt Accessed 2026-09-29.
[606] Amazon Publisher Services sellers.json. Amazon Publisher Services, 2026-09-14 (HTTP Last-Modified). https://aps.amazon.com/sellers.json Accessed 2026-09-29.
[607] Sponsorships Are Now Available to Affiliates. Twitch (blog), 2026-08-20. https://blog.twitch.tv/en/2026/08/20/sponsorships-are-now-available-to-affiliates/ Accessed 2026-09-29.
[608] TwitchCon Rotterdam 2026 showed how livestreaming is rewriting what "attention" means. Amazon Ads (news), 2026-06-09. https://advertising.amazon.com/library/news/twitchcon-rotterdam-2026-recap Accessed 2026-09-29.
[609] Amazon Ads policies (redirect target of advertising.amazon.com/resources/ad-policy/twitch). Amazon Ads Support Center, 2026-09-02 (Updated on Sep 2, 2026). https://advertising.amazon.com/help/G29MAQ3YTK5LQ7VC Accessed 2026-09-29. Access: partial extract.
[610] Turkey fines Twitch 2 million lira for data breach. Turkish Minute (citing Anadolu), 2024-11-16. https://turkishminute.com/2024/11/16/turkey-twitch-2-million-lira-data-breach/ Accessed 2026-09-29.
[611] Twitch.tv Privacy Notice. Twitch Interactive, Inc., 2026-08-12 (Last modified on 08/12/2026). https://legal.twitch.com/legal/privacy-notice/ Accessed 2026-09-29.
[612] Interactive Ads (Playables and Interactive End Cards). Microsoft Advertising (XBOX Media Solutions), undated. https://about.ads.microsoft.com/en/solutions/technology/gaming-solutions/interactive-advertising Accessed 2026-09-29.
[613] Microsoft is actively working on adding in-game ads that will literally intrude on gameplay. ixbt.games, 2026-09-17. https://ixbt.games/en/news/2026/09/17/436487-microsoft-aktivno-truditsia-nad-tem-ctoby-dobavit-reklamu-v-igry-kotoraia-budet-bukvalno-vklinivatsia-v-geimplei.html Accessed 2026-09-29.
[614] Developers (format suite). Anzu. https://www.anzu.io/developers Accessed 2026-09-29.
[615] Unity: Step-by-step intrinsic ads integration guide for Unity (Anzu SDK). Playgama (wiki.playgama.com, GitBook). https://wiki.playgama.com/playgama/guides/monetization/intrinsic-ads/unity Accessed 2026-09-29.
[616] A Guide to Creating Intrinsic In-Game Ads. Anzu (Nick Woodford), 2024-11-11. https://www.anzu.io/blog/in-game-creative-guide Accessed 2026-09-29.
[617] A Complete Guide to Targeting in Intrinsic In-Game Ads. Anzu (Nick Woodford), 2024-10-18. https://www.anzu.io/blog/in-game-targeting-guide Accessed 2026-09-29.
[618] Unlocking Precision In-Game: Anzu Empowers Advertisers With Global, Privacy-Safe Audience Targeting. Anzu (Nick Woodford), 2025-09-03. https://www.anzu.io/news/unlocking-precision-in-game-anzu-empowers-advertisers-with-global-privacy-safe-audience-targeting Accessed 2026-09-29.
[619] Mobile Game Developers. Anzu. https://www.anzu.io/mobile-game-developers Accessed 2026-09-29.
[620] The Gamer Stereotype Is Dead: Today's US Players Are Educated, Family-Oriented, And Nearly Half Are Women. Anzu (Nick Woodford), 2025-10-23. https://www.anzu.io/news/the-gamer-stereotype-is-dead-todays-us-players-are-educated-family-oriented-and-nearly-half-are-women Accessed 2026-09-29.
[621] Contact (offices). Anzu. https://www.anzu.io/contact Accessed 2026-09-29.
[622] Solutions Privacy Notice (within DoubleVerify's Privacy Notice page). DoubleVerify, 2025-10-09 (effective date). https://doubleverify.com/privacy-notice Accessed 2026-09-29.
[623] Rewarded video ads (Roblox Creator Hub documentation). Roblox Corporation, undated (live documentation). https://create.roblox.com/docs/production/promotion/rewarded-video-ads Accessed 2026-09-29.
[624] DoubleVerify Expands Measurement Coverage on TikTok Pangle In the United States and International Markets. DoubleVerify (newsroom), 2026-08-13. https://doubleverify.com/company/newsroom/doubleverify-expands-measurement-coverage-on-tiktok-pangle-in-the-united-states-and-international-markets Accessed 2026-09-29.
[625] Brand Suitability (DV product page). DoubleVerify, undated (live page). https://doubleverify.com/products/advertisers/verify/brand-suitability Accessed 2026-09-29.
[626] FRAUD ALERT: Fake Gaming Apps May Be Gaming Your Ads. DoubleVerify (DV Fraud Lab blog), 2025-11-19. https://doubleverify.com/blog/web/prove/fraud-alert-fake-gaming-apps-may-be-gaming-your-ads Accessed 2026-09-29.
[627] Case studies | Frameplay. Frameplay, undated (case-study sitemap lastmod 2024-07-23). https://frameplay.com/case-studies/ Accessed 2026-09-29.
[628] Adverty accelerates faster than expected in the US through partnership with Frameplay. Adverty AB (publ) (MFN), 2024-12-09. https://mfn.se/a/adverty/adverty-accelerates-faster-than-expected-in-the-us-through-partnership-with-frameplay Accessed 2026-09-28.
[629] Roblox Expands Immersive Video Ads Access to All Advertisers, Adds New Measurement Partners. Roblox Corporation, 2024-05-01. https://about.roblox.com/newsroom/2024/05/roblox-expands-immersive-video-ads-access-to-all-advertisers-adds-new-measurement-partners Accessed 2026-09-28.
[630] Roblox for Brands (brands.roblox.com home page). Roblox Corporation, 2026 (copyright line). https://brands.roblox.com/ Accessed 2026-09-28.
[631] Super League Debuts AI-Driven Playable Media Product via Exclusive Partnership with AdArcade. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-01-28. https://ir.superleague.com/news-events/press-releases/detail/289/super-league-debuts-ai-driven-playable-media-product-via Accessed 2026-09-29.
[632] Super League Reports First Quarter 2026 Financial Results, Advancing Transition from Corporate Restructuring to Operational Execution. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-05-15. https://ir.superleague.com/news-events/press-releases/detail/298/super-league-reports-first-quarter-2026-financial-results Accessed 2026-09-29.
[633] Stream Display Ads | Twitch. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/stream-display-ads Accessed 2026-09-29.
[634] Twitch ads ad specifications | First Impression Takeover. Amazon Ads, undated (live page). https://advertising.amazon.com/resources/ad-specs/twitch/first-impression Accessed 2026-09-29.
[635] Amazon Brand Lift: Study the impact of your advertising. Amazon Ads, undated (live page). https://advertising.amazon.com/measurement-analytics/brand-lift Accessed 2026-09-29.
[636] app-ads.txt file for twitch. Twitch (OWNERDOMAIN=amazon.com), undated (served 2026-09-29). https://www.twitch.tv/app-ads.txt Accessed 2026-09-29.
[637] Activision Blizzard Media Rewarded Video Package (Lot 5521094). The Ad Club Auction, 2019-09-16 (auction start). https://www.adclubmedia.com/item/5521094 Accessed 2026-09-29.
[638] Maybelline sees 100,000 daily visitors and millions of plays (case study). Microsoft Advertising, 2024-04-16. https://about.ads.microsoft.com/en/resources/discover/case-studies/maybelline-case-study Accessed 2026-09-29.
[639] Arla Anchor spreads joy in gaming (case study). Microsoft Advertising, 2024-12-09. https://about.ads.microsoft.com/en/resources/discover/case-studies/arla-anchor-case-study Accessed 2026-09-29.
[640] 40,000 Prada samples requested in less than a day (case study). Microsoft Advertising, 2023-02-09. https://about.ads.microsoft.com/en/resources/discover/case-studies/prada-case-study Accessed 2026-09-29.
[641] Anzu privacy policy page (site footer and legal links). Anzu Virtual Reality Ltd., undated. https://www.anzu.io/privacy-policy Accessed 2026-09-28. Access: partial extract.
[642] Privacy Policy - AdInMo: InGamePlay brand advertising (Wayback Machine capture of 28 May 2025). AdInMo Ltd (archived by the Internet Archive), 2025-02-18 (page dateModified). https://web.archive.org/web/20250528172800/https://www.adinmo.com/privacy-policy/ Accessed 2026-09-28.
[643] Vivendi home page (list of assets). Vivendi SE, 2026-09-03 (dateModified). https://www.vivendi.com/en/ Accessed 2026-09-28.
[644] Admazing | Powerful Mobile Gaming Ads for Brands (home page). Admazing Co., 2026 (copyright line). https://www.admazing.co/ Accessed 2026-09-28.
[645] SuperAwesome Targets New Opportunities In Kids Advertising Beyond The Metaverse. AdExchanger (Anthony Vargas), 2024-04-02. https://www.adexchanger.com/gaming/superawesome-targets-new-opportunities-in-kids-advertising-beyond-the-metaverse/ Accessed 2026-09-28.
[646] Zynga Ads (home page). Zynga, undated (live page). https://www.zyngaads.com/ Accessed 2026-09-29.
[647] Discord Terms of Service. Discord Inc., undated (current version at access). https://discord.com/terms Accessed 2026-09-28.
[648] Overwolf Website Terms of Use (last updated 23 June 2024). Overwolf, Ltd., 2024-06-23. https://www.overwolf.com/legal/terms/ Accessed 2026-09-28.
[649] PROJECT VENUS BIDCO LIMITED (13587744): persons with significant control. UK Companies House, live register. https://find-and-update.company-information.service.gov.uk/company/13587744/persons-with-significant-control Accessed 2026-09-28.
[650] Playwire Terms of Service (last updated 11 October 2024). Playwire, LLC, 2024-10-11. https://www.playwire.com/terms-of-service Accessed 2026-09-28.
[651] Precisify | Contextual & Targeted YouTube Advertising (home page). Precisify, 2026 (copyright line). https://www.precisify.com Accessed 2026-09-28.
[652] Precisify Gaming | Precisify Your Media Buy. Precisify, undated. https://www.precisify.com/gaming Accessed 2026-09-29.
[653] Ad Ops Tech Startup Aditude Acquires CPMStar To Snag More Direct Gaming Deals. AdExchanger, 2024-03-12. https://www.adexchanger.com/gaming/ad-ops-tech-startup-aditude-acquires-cpmstar-to-snag-more-direct-gaming-deals/ Accessed 2026-09-28.
[654] GEEIQ | The Data Platform Behind Brands and Gaming (home page). GEEIQ, 2026 (copyright line). https://www.geeiq.com/ Accessed 2026-09-28.
[655] Medal Terms of Service. Medal B.V., 2026-07. https://medal.tv/terms Accessed 2026-09-29.
[656] Changing the Game: How Games Advertising Powers Performance. IAB with Advertiser Perceptions, 2024-03-26. https://www.iab.com/insights/changing-the-game-how-games-advertising-powers-performance/ Accessed 2026-09-28. Access: partial extract.
[657] iion Unlocks Access to 1.2m Daily Gamers Through Premium Arkadium Partnership. iion, 2025-10-23. https://www.iion.io/blog/iion-unlocks-access-to-1-2m-daily-gamers-through-premium-arkadium-partnership Accessed 2026-09-29.
[658] Rediscover COMBO! and connect with every player. Gameloft for brands, 2026-01-27 (date shown for this article in related-article lists). https://www.gameloft.com/for-brands/news/rediscover-combo-and-connect-with-every-player Accessed 2026-09-29.
[659] Our Audience | Gameloft for Brands. Gameloft for brands, undated (data line January 2026). https://www.gameloft.com/for-brands/our-audience Accessed 2026-09-29.
[660] Super League Enters Next Phase of Execution, Bringing the Power of Play to Modern Marketing. Super League Enterprise, Inc. (GlobeNewswire release on ir.superleague.com), 2026-01-21. https://ir.superleague.com/news-events/press-releases/detail/287/super-league-enters-next-phase-of-execution-bringing-the Accessed 2026-09-29.
[661] Services Under Review: Digital (Internet Archive capture of 5 February 2026). Media Rating Council (captured by the Internet Archive), 2026-02-05 (capture). http://web.archive.org/web/20260205111824/https://www.mediaratingcouncil.org/accreditation/services-under-review/digital Accessed 2026-09-29.
[662] COMBO!'s Strategic Partnerships Boost the Power of Its In-Game Advertising Network. Gameloft for brands. https://www.gameloft.com/for-brands/news/combos-strategic-partnerships-boost-the-power-of-its-in-game-advertising-network Accessed 2026-09-28.
[663] Discord Privacy Policy. Discord, 2025-08-29 (last updated). https://discord.com/privacy Accessed 2026-09-28.
[664] Standard Media - High-Impact Display & Video Ads (Overwolf). Overwolf, undated (live page). https://www.overwolf.com/ads/ad-solutions/standard-media/ Accessed 2026-09-28.
[665] Playwire kids advertisers. Playwire, undated (live page). https://www.playwire.com/kids-advertisers Accessed 2026-09-28.
[666] Discord Launches 'Arena Quests' Ad Product Tied to Length of Time Spent Playing Sponsored Games (EXCLUSIVE). Variety (Jennifer Maas), 2025-10-02. https://variety.com/2025/gaming/news/discord-arena-quests-ad-sponsored-games-1236536780/ Accessed 2026-09-29.
[667] DigitalsCom Programmatic Media - LIVAD. DigitalsCom (reseller agency), 2024 (copyright line). https://www.digitalscom.com/livad.html Accessed 2026-09-29.
[668] LIVAD Creative Guide / Spec sheet (PDF, livad-creavite-specs.pdf). LIVAD (hosted by DigitalsCom), 2024-11-12 (server Last-Modified). https://www.digitalscom.com/asset/images/livad/livad-creavite-specs.pdf Accessed 2026-09-29.
[669] Privacy Policy. LIVAD Medya Teknoloji A.S., undated (live page). https://www.livad.stream/privacy-policy Accessed 2026-09-29.
[670] LIVAD Brand Safety slide (image livad-brand-safety.webp). LIVAD / DigitalsCom, 2024-11-12 (server Last-Modified). https://www.digitalscom.com/asset/images/livad/livad-brand-safety.webp Accessed 2026-09-29.
[671] Ad Specs - Genie. IGN Entertainment, undated (live page). https://specs.ign.com/ad-unit/genie Accessed 2026-09-29.
[672] Ziff Davis International, Ltd. Brand Safety Guidelines (PDF). Ziff Davis International, Ltd., 2022-09-13 (server Last-Modified). https://globalstudios.ign.com/ads/docs/brand-safety-guidelines.pdf Accessed 2026-09-29.
[673] Contact us - Enthusiast Gaming. Enthusiast Gaming Holdings Inc., live page. https://www.enthusiastgaming.com/contact-us/ Accessed 2026-09-29.
[674] Enthusiast Gaming home page (By the numbers; navigation; latest news). Enthusiast Gaming Holdings Inc., live page (copyright 2026; news items to 12 August 2026). https://www.enthusiastgaming.com/ Accessed 2026-09-29.
[675] Wayback Machine captures of u.gg/ads.txt (2 December 2025, 1 February 2026, 2 May 2026). Internet Archive, 2026-05-02 (latest capture read). http://web.archive.org/web/20260502071543id_/https://u.gg/ads.txt Accessed 2026-09-29.
[676] Freestar sellers.json. Freestar (contact sellers.json@freestar.io, Scottsdale, Arizona), live file. https://freestar.com/sellers.json Accessed 2026-09-29.
[677] Reporting Ads on Medal. Medal Support (help centre), 2025-05-16. https://support.medal.tv/support/solutions/articles/48001271830-reporting-ads-on-medal Accessed 2026-09-29.
[678] Connecting Discovery to Purchase: Green Man Gaming Partners with Medal.tv. Green Man Gaming Limited, 2026-09-03. https://www.greenmangaming.com/company/news-category/latest-news/medaltv-and-greenmangaming-partnership/ Accessed 2026-09-29.
[679] Medal SVP on Gaming Clips, FAST Channels & Creator Monetization. MarTech Edge (interview with Yuriy Yarovoy, SVP Revenue, Medal), 2025-07-18. https://martechedge.com/interview/medal-svp-on-gaming-clips-fast-channels--creator-monetization Accessed 2026-09-29.
[680] Sponsored TV - Now part of Amazon Ads streaming tv ads offering. Amazon Ads, undated (live page). https://advertising.amazon.com/products/sponsored-tv Accessed 2026-09-29.
[681] Discord Cloud Play Quests. Discord (Discord Support), 2025-10-29 (edited_at). https://support.discord.com/hc/en-us/articles/35372187686295 Accessed 2026-09-29.
[682] Xbox Media Solutions (member directory). IAB UK, undated. https://www.iabuk.com/member-directory/xbox-media-solutions Accessed 2026-09-29.
[683] Content Taxonomy (standard page). IAB Technology Laboratory, 2026-02-11. https://iabtechlab.com/standards/content-taxonomy/ Accessed 2026-09-28.
[684] EDGAR full-text search: 'Discord Inc' in S-1, S-1/A, F-1 and DRS filings. US Securities and Exchange Commission (EDGAR full-text search), live search. https://efts.sec.gov/LATEST/search-index?q=%22Discord%20Inc%22&forms=S-1,S-1/A,F-1,DRS Accessed 2026-09-28.
[685] Trivver home page (Spatial AI Measurement). Trivver, Inc., undated (live). https://trivver.com/ Accessed 2026-09-29.
[686] Venatus Guide to Desktop Ad Formats (blog). Venatus Media Limited, 2024-05-14 (datePublished; dateModified reads 2022-07-11). https://www.venatus.com/blog/venatus-guide-to-online-advertising-formats Accessed 2026-09-29.
[687] Ad format gallery. Admazing Co., undated. https://admazing.com/ad-format-gallery Accessed 2026-09-29.
[688] Revolutionizing Fast Food Advertising in the Mobile Gaming Space (use case). Admazing Co., 2022-04-08. https://admazing.com/use-cases/revolutionizing-fast-food-advertising-in-the-mobile-gaming-space Accessed 2026-09-29.
[689] OpenCorporates company records: ADMAZING CO. LLC (Delaware 6164106; Florida branch M19000006420). OpenCorporates (aggregating state registry data), live aggregator records. https://opencorporates.com/companies/us_fl/M19000006420 Accessed 2026-09-29. Access: partial extract.
[690] Best practises for ad viewability. Adverty AB (publ), undated. https://adverty.com/best-practises-for-ad-viewability/ Accessed 2026-09-29.
[691] Adverty: Monetize with in-game advertising (Developers page). Adverty AB (publ), undated. https://adverty.com/developers/ Accessed 2026-09-29.
[692] Adverty Demand Terms and Conditions. Adverty AB (publ), undated. https://adverty.com/demand-terms-and-conditions/ Accessed 2026-09-29.
[693] Adverty: Reach global audience with in-game ads (Advertisers page). Adverty AB (publ), undated. https://adverty.com/advertisers/ Accessed 2026-09-29.
[694] LEGO CITY USA THINK FAST CHALLENGE (success story). Gameloft for brands, undated; campaign in 2025. https://www.gameloft.com/for-brands/success-stories/lego-city-usa-think-fast-challenge Accessed 2026-09-29.
[695] Zynga and Gameloft for brands Join Forces to Connect Brands with Gaming’s Largest and Most Diverse Audience. Gameloft for brands, 2025-02-11 (date shown for this article in related-article lists). https://www.gameloft.com/for-brands/news/zynga-and-gameloft-for-brands-join-forces-to-connect-brands-with-gamings-largest-and-most-diverse-audience Accessed 2026-09-29.
[696] The Gaming Advertising Advantage: Our Presence at IAB PlayFronts 2026. Gameloft for brands, 2026-09 (event 2026-09-17; embedded timestamps to 2026-09-23). https://www.gameloft.com/for-brands/news/the-gaming-advertising-advantage-our-presence-at-iab-playfronts-2026 Accessed 2026-09-29.
[697] How iion Helped Menulog Cook Up Almost x8 Engagement (and Plenty of Tasty Deliveries). iion, 2025-11-25. https://www.iion.io/blog/how-iion-helped-menulog-cook-up-almost-x8-ctr-and-plenty-of-tasty-deliveries Accessed 2026-09-29.
[698] Case Study - Chupa Chups (landing page). iion, undated. https://www.iion.io/lp/iion-case-study-chupa-chups Accessed 2026-09-29. Access: partial extract.
[699] Beyond the Noise: Why Kidoz's Focused Strategy Unlocks Premium Value (And Predictable Returns) (Kidoz blog). Kidoz (blog feed feed.kidoz.net), 2025-11-18. https://feed.kidoz.net/kidoz/rss/ (item dated 2025-11-18; site path /blog/beyond-the-noise-why-kidozs-focused-strategy-unlocks-premium-value-and-predictable-returns) Accessed 2026-09-29.
[700] Discord Quests - Premium Video Ads (Media & Entertainment). Discord, undated (live page). https://discord.com/ads/media-entertainment Accessed 2026-09-29.
[701] discord.com sellers.json. Discord Inc., live file (no date field). https://discord.com/sellers.json Accessed 2026-09-29.
[702] Advertise - Icy Veins. Icy Veins (Enthusiast Gaming property), 2026-03-30 (Last Updated). https://www.icy-veins.com/advertise Accessed 2026-09-29.
[703] Enthusiast Gaming Privacy Policy. Enthusiast Gaming Holdings Inc., undated (no effective date found on the page). https://www.enthusiastgaming.com/privacy-policy/ Accessed 2026-09-29.
[704] U.GG privacy policy (access error). U.GG (Enthusiast Gaming). https://u.gg/privacy-policy Accessed 2026-09-29. Access: partial extract.
[705] FACEIT Advertising - ESL FACEIT Group (Wayback Machine capture of 3 December 2025). ESL FACEIT Group (archived by the Internet Archive), 2025-12-03 (capture). http://web.archive.org/web/20251203072746id_/https://eslfaceitgroup.com/advertising/faceit/ Accessed 2026-09-29.
[706] Partners - ESL FACEIT Group. ESL FACEIT Group, 2021-11-18 (datePublished). https://eslfaceitgroup.com/partners/ Accessed 2026-09-29.
[707] Intel Extreme Masters Cologne 2026 Major Breaks Records as the Most-Watched Counter-Strike Tournament of All-Time. ESL FACEIT Group, 2026-06-21. https://eslfaceitgroup.com/press/intel-extreme-masters-cologne-2026-major-breaks-records-as-the-most-watched-counter-strike-tournament-of-all-time/ Accessed 2026-09-29.
[708] FACEIT Underage consent FAQ (access error). FACEIT Help Center. https://support.faceit.com/hc/en-us/articles/10533341902748-Underage-consent-FAQ Accessed 2026-09-29. Access: partial extract.
[709] Ad Specs - Review Module. IGN Entertainment, undated (live page). https://specs.ign.com/ad-unit/review-module Accessed 2026-09-29.
[710] Ad Specs - ThemeScape. IGN Entertainment, undated (live page). https://specs.ign.com/ad-unit/themescape Accessed 2026-09-29.
[711] IGN Entertainment corporate site (home). IGN Entertainment, undated (live page). https://corp.ign.com/ Accessed 2026-09-29.
[712] ads.txt for www.ign.com. IGN (OWNERDOMAIN ziffdavis.com), 2026-09-25 (server Last-Modified). https://www.ign.com/ads.txt Accessed 2026-09-29.
[713] ziffdavis.com sellers.json. Ziff Davis, live file. https://www.ziffdavis.com/sellers.json Accessed 2026-09-29.
[714] Privacy Policy - IGN. IGN Entertainment, 2025-01 (effective). https://www.ign.com/privacy-policy Accessed 2026-09-29.
[715] Vodafone and Netflix Achieve 76% Lift in Ad Recall with Collaborative Overlays on Live Streams powered by LIVAD Technologies. LIVAD (post by Arda Genc), 2022-03-15. https://www.livadstream.com/blog/vodafone-and-netflix-bls-case-study Accessed 2026-09-29.
[716] LIVAD Reporting slide (image livad-reporting.webp). LIVAD / DigitalsCom, 2024-11-12 (server Last-Modified). https://www.digitalscom.com/asset/images/livad/livad-reporting.webp Accessed 2026-09-29.
[717] Medal media deck (Google Slides link on medal.tv/grow). Medal. https://docs.google.com/presentation/d/1KA33JjcvVRexbC89O4mx_mlNugvqmnHZ7ReUlrqXGN0/edit Accessed 2026-09-29. Access: partial extract.
[718] "If players can clip it in 15 seconds, they’ll market it for you" | GamesBeat Summit recap. Medal (blog), 2026-06-23. https://medal.tv/blog/posts/gamesbeat-medal Accessed 2026-09-29.
[719] OpenX sellers.json. OpenX Technologies, Inc.. https://www.openx.com/sellers.json Accessed 2026-09-29.
[720] How Medal.tv saw a MoM lift of 19.8% in revenue with Aditude. Aditude Inc.. https://www.aditude.com/case-studies/medal.tv Accessed 2026-09-29.
[721] $100 Valorant Clip That! Contest — Valorant Points Giveaway (page and Official Rules). Medal / Ferox US, Inc., 2026-09-22. https://medal.tv/features/VALORANT Accessed 2026-09-29.
[722] Custom Media - Takeovers & High-Impact Creative Formats (Overwolf). Overwolf, undated (live page). https://www.overwolf.com/ads/ad-solutions/custom-media/ Accessed 2026-09-29.
[723] Gamer Grid (Overwolf Native monetization docs). Overwolf Developers, undated (live docs). https://dev.overwolf.com/ow-native/monetization/advertising/gamer-grid/ Accessed 2026-09-29.
[724] How JBL Quantum Made Competitive FPS Players Rethink Their Headset (Overwolf case study). Overwolf, undated. https://www.overwolf.com/ads/case-studies/jbl-quantum/ Accessed 2026-09-29.
[725] How MilkPEP drove players to register and tune into The Milk Cup (Overwolf case study). Overwolf, undated. https://www.overwolf.com/ads/case-studies/milkpep/ Accessed 2026-09-29.
[726] How Tropicana's Naked Smoothie owned the 78-minute gamer break (Overwolf case study). Overwolf, undated. https://www.overwolf.com/ads/case-studies/tropicana/ Accessed 2026-09-29.
[727] Advertising (Overwolf Native monetization overview). Overwolf Developers, undated (live docs; footer 2026). https://dev.overwolf.com/ow-native/monetization/advertising/overview/ Accessed 2026-09-29.
[728] Overwolf sellers.json. Overwolf, live file. https://www.overwolf.com/sellers.json Accessed 2026-09-29.
[729] Overwolf Ads Expands Global Brand Partnerships Team with Senior Hires Across EMEA & US. ExchangeWire (Overwolf Ads announcement), 2026-04-23. https://www.exchangewire.com/blog/2026/04/23/overwolf-ads-expands-global-brand-partnerships-team-with-senior-hires-across-emea-us/ Accessed 2026-09-29.
[730] Case Studies (Overwolf Ads). Overwolf, undated (live page). https://www.overwolf.com/ads/case-studies/ Accessed 2026-09-29.
[731] TAG Registry name searches with no match: Overwolf, Playwire, Playwire LLC, GG Software, Nitro, TAG-ID fcddfba7adc2d929. Trustworthy Accountability Group (registry backend hubgloo.opxar.com), live registry. https://hubgloo.opxar.com/search?q=Playwire Accessed 2026-09-29.
[732] Overwolf Platform Privacy Policy. Overwolf, Ltd., 2024-06-23 (policy 'Last updated'); page footer 'Last updated on Jul 21, 2025'. https://legal.overwolf.com/docs/overwolf/platform/platform-privacy-policy Accessed 2026-09-29.
[733] User Identity / Hashed Emails (Overwolf Native monetization docs). Overwolf Developers, undated (live docs). https://dev.overwolf.com/ow-native/monetization/advertising/user-identity Accessed 2026-09-29.
[734] The Playwire Flex Suite. Playwire, LLC, undated (live page). https://www.playwire.com/flex-suite Accessed 2026-09-29.
[735] Gaming | Data and Audience Solutions (Playwire). Playwire, undated (live page). https://www.playwire.com/data-and-audience-solutions-gaming Accessed 2026-09-28.
[736] kidSAFE Member List. kidSAFE Seal Program, live list. https://kidsafe.com/members Accessed 2026-09-29. Access: partial extract.
[737] Playwire Privacy Policy. Playwire, LLC, 2026-01-13 (Last Updated). https://www.playwire.com/playwire-privacy-policy Accessed 2026-09-29.
[738] Venatus Brand Safety Procedures. Venatus Media Limited, undated (live page). https://www.venatus.com/brand-safety Accessed 2026-09-29.
[739] Why focus is the brain games advantage (Venatus blog). Venatus Media Limited, 2026-06-24. https://www.venatus.com/blog/why-focus-is-the-brain-games-advantage Accessed 2026-09-29.
[740] Venatus Core Gamers (audience page). Venatus Media Limited, undated (live page). https://www.venatus.com/advertisers/audiences/core-gaming Accessed 2026-09-29.
[741] Venatus Young Gamers (Kids Gaming audience page). Venatus Media Limited, undated (live page). https://www.venatus.com/advertisers/audiences/kids-gaming Accessed 2026-09-29.
[742] For Advertisers | Contact Us | Venatus. Venatus Media Limited, undated (live page). https://www.venatus.com/contact/for-advertisers Accessed 2026-09-29.
[743] Venatus Privacy Notice. Venatus Media Limited, undated in text captured. https://www.venatus.com/privacy-policy Accessed 2026-09-29.
[744] Anzu Announces Investment From Amex Ventures To Accelerate Advertising In Premium Gaming. Anzu (Nick Woodford), 2025-06-12. https://www.anzu.io/news/anzu-announces-investment-from-amex-ventures-to-accelerate-advertising-in-premium-gaming Accessed 2026-09-29.
[745] Anzu News (newsroom listing). Anzu. https://www.anzu.io/news Accessed 2026-09-29.
[746] DoubleVerify Grows Q2 Revenue By Expanding Brand Safety To Retail Media, TikTok And Gaming. AdExchanger (Alyssa Boyle), 2022-08-03. https://www.adexchanger.com/online-advertising/doubleverify-grows-q2-revenue-by-expanding-brand-safety-to-retail-media-tiktok-and-gaming/ Accessed 2026-09-29.
[747] SEC EDGAR company search ("frameplay") and EDGAR full-text search ("Frameplay"). U.S. Securities and Exchange Commission. https://www.sec.gov/cgi-bin/browse-edgar?company=frameplay&type=&dateb=&owner=include&count=40&action=getcompany Accessed 2026-09-29.
[748] Gadsme partners with Konami to bring immersive ads to games and esports. Isa Muhammad, PocketGamer.biz, 2025-07-28. https://www.pocketgamer.biz/gadsme-partners-with-konami-to-bring-immersive-ads-to-games-and-esports/ Accessed 2026-09-29.
[749] Exhibit 2.1: Asset Purchase Agreement between Esports Now, LLC and Super League Enterprise, Inc., dated March 16, 2026 (with Exhibit F, form of Brand Partnership Agreement). Super League Enterprise, Inc. (SEC EDGAR exhibit to the 8-K filed 2026-03-20), 2026-03-20. https://www.sec.gov/Archives/edgar/data/1621672/000143774926009204/ex_934180.htm Accessed 2026-09-29.
[750] SEC EDGAR submissions index for CIK 0001621672 (Super League Enterprise, Inc.). US Securities and Exchange Commission, live index. https://data.sec.gov/submissions/CIK0001621672.json Accessed 2026-09-29.
[751] Amazon.com, Inc. Form 10-K for fiscal year 2014. Amazon.com, Inc., 2015-01-30. https://www.sec.gov/Archives/edgar/data/1018724/000101872415000006/amzn-20141231x10k.htm Accessed 2026-09-28.
[752] Introducing the Amazon Ads MCP Server, the foundation for the next generation of advertising workflows. Amazon Ads (Paula Despins), 2026-02-02. https://advertising.amazon.com/library/news/amazon-ads-mcp-server-open-beta Accessed 2026-09-29.
[753] Amazon.com, Inc. Form 10-Q for the quarter ended 30 June 2026. Amazon.com, Inc., 2026-07-31. https://www.sec.gov/Archives/edgar/data/1018724/000101872426000026/amzn-20260630.htm Accessed 2026-09-29.