CTV, Streaming & Live Event Advertising.
The streaming execution layer: the CTV Ad Format Portfolio, the Live Event Ad Playbook and its Concurrent Streams API, SSAI delivery, the VAST CTV Addendum, OM SDK for CTV, device attestation, household measurement — and what all of it means when agents do the buying.
This is a deep dive inside the six-layer standards system, not a seventh pillar. CTV transacts on the same rails as the rest of programmatic, but almost every web assumption breaks at the television: the unit is an app, delivery is a stitched stream, the audience is a household, the device can be spoofed, and live events compress the whole problem into a single ad break. This page maps what has shipped, what is in public comment, and what is still roadmap — precisely.
CTV is not just digital video on a bigger screen. It is an app, device, stream, household, SSAI, verification, and measurement problem.
Fast read
- What it is
- A deep-dive guide to the standards and guidance that make CTV, streaming, and live-event advertising executable — and to where the formats, the live-scale signaling, the stitched delivery path, and the verification layer actually stand.
- What it covers
- The IAB Tech Lab CTV Ad Format Portfolio (six named formats, still in public comment as of this validation), the Guide to Programmatic CTV, the Live Event Ad Playbook (LEAP) with the Concurrent Streams API v1.0, SSAI and the VAST CTV Addendum 2024, OM SDK for CTV, device attestation, and household-level measurement physics.
- What it is not
- Not a seventh standards pillar — a deep dive inside the six-layer system. And not a claim that the format layer is finished: the portfolio guidelines and their signaling guidance were in public comment as of this validation, and only some LEAP components are final.
- Why it matters
- Agents buying CTV inherit physics the web never had — apps instead of pages, stitched streams instead of browser pixels, households instead of profiles, and live moments where the whole audience hits the ad break at once.
- Best for
- Streamers and publishers, DSPs and SSPs, SSAI and ad-serving vendors, measurement and verification vendors, agencies, and brands building or evaluating CTV and live-event execution in agentic workflows.
- Best next read
- Video & Mobile Ad Delivery, Measurement & Media Quality, and Event & Conversion APIs.
Where CTV fits in the standards stack.
CTV is a child of the six-layer system with television physics. Its delivery standards live in the Video & Mobile layer — VAST, the 2024 CTV Addendum, SSAI guidance, SIMID for interactivity; its transaction support lives on the Core AdTech rails, where the December 2025 Guide to Programmatic CTV details how portfolio formats are transacted over OpenRTB; its trust tooling — OM SDK for CTV, device attestation, MRC viewability definitions — belongs to the measurement layer; and its privacy perimeter runs through household and platform data rather than a browser consent prompt. The deep dives show where those standards become real in specific operating environments — this page is the close-up of the streaming one.
Why CTV is different.
Every CTV-specific standard on this page exists because a web assumption broke at the television. Six realities define the channel.
- 01
The unit is an app
CTV ads run inside apps on TV platforms, and every app is its own execution environment. There is no shared page, no common cookie, and no global capability set — format support, measurement SDKs, and signaling differ per app and per platform.
- 02
Delivery is stitched
Server-side ad insertion assembles ads into the stream before it reaches the device, so client-side verification assumptions break. The SSAI VAST Macros Guidance v1.0 (November 2020) exists precisely because verification partners and buyers still need basic user agent, environment, and content details.
- 03
The screen is shared
A household watches a TV, not a browser profile. Person-level reach, frequency, and co-viewing figures on CTV are typically modeled from household-level signals — methodology-bearing estimates, not observations.
- 04
The device can lie
Device spoofing is real enough that IAB Tech Lab launched device attestation support in the Open Measurement SDK (October 23, 2025), adopting the IETF Privacy Pass protocol for the ads-verification use case.
- 05
VAST predates the TV
Rather than mint a new VAST version, the VAST CTV Addendum 2024 (final July 2024) standardizes CTV-critical features across all VAST versions through backward-compatible extensions — retrofitting the installed base instead of replacing it.
- 06
Live compresses everything
Live viewership shifts dynamically mid-event, and the ad break arrives for everyone at once. Ad systems scaled for average load miss breaks — the problem the Concurrent Streams API v1.0 was standardized to address.
The CTV Ad Format Portfolio.
In December 2025, IAB Tech Lab announced the CTV Ad Portfolio alongside a major update to the Guide to Programmatic CTV — six core formats distilled from more than 100 submissions through the “Ad Format Hero” initiative, released for public comment through January 31, 2026. A follow-on signaling guidance — updates to AdCOM, VAST, and Native Ads for four of the formats — was published May 6, 2026 with comment open to June 5, 2026. As of this validation, none of it is final: these are the exact official names, with their verified statuses attached.
- Prioritized
Pause Ad
One of the two formats (with Menu) the working group prioritized: the December 2025 Guide to Programmatic CTV adds OpenRTB support for transacting it, and it is covered by the May 2026 signaling guidance. Public documentation does not pin a specific OpenRTB version to that support.
- Prioritized
Menu Ad
The second prioritized format. Covered by the OpenRTB transaction support in the updated programmatic guide and by the May 2026 signaling guidance — the pair positioned to transact programmatically first. Validate current status before building.
- Signaling tranche
Overlay Ads
Included in the CTV Ad Portfolio Signaling Guidance published May 6, 2026 (AdCOM, VAST, and Native Ads updates; public comment to June 5, 2026). Sub-variation names circulate in summaries — cite the official documents for definitions.
- Signaling tranche
Squeeze Back Ads
Also in the May 2026 signaling guidance. The Tech Lab’s stated rationale for the signaling work: transacting these formats “has remained largely manual due to the lack of standardized definitions.”
- Named, not yet signaled
Screensaver Ad
One of the six core portfolio formats, but not among the four covered by the first signaling-guidance tranche. Treat it as a named format whose transaction mechanics are still ahead — validate current status.
- Named, not yet signaled
In Scene Insertion Ad
The sixth core format, also outside the first signaling tranche. Like the rest of the portfolio, its guidelines were derived from more than 100 submissions through the Ad Format Hero initiative.
| Document | Verified status | What to know |
|---|---|---|
| Ad Format Guidelines for Digital Video and CTV | Released for public comment December 2025; comment closed January 31, 2026 | No finalized release verified as of June 2026 — do not cite the portfolio as final or versioned |
| Guide to Programmatic CTV | Major update December 2025, released alongside the portfolio | Listed as “CTV Programmatic Guide” on the standards index — the same document; adds OpenRTB support for Pause and Menu |
| CTV Ad Portfolio Signaling Guidance (AdCOM, VAST, Native Ads) | Published May 6, 2026; public comment to June 5, 2026 | Covers Pause, Menu, Overlay, and Squeeze Back Ads; the signaling layer was still in public comment as of this validation |
The discipline
Public-comment drafts are not contracts. Until the guidelines and signaling finalize, “Pause Ad” means what a specific document version says it means — pin the version, or two systems will transact two different formats under one name.
Live-event ad serving and the Concurrent Streams API.
Live is where streaming physics get unforgiving: viewership shifts dynamically mid-event, the break arrives for everyone at once, and ad systems scaled for average load miss it. The Live Event Ad Playbook (LEAP) — launched August 5, 2025 and developed through the Live Event Ad Serving Working Group, with Amazon, NBCUniversal, FreeWheel, and Index Exchange collaborating on development — carries IAB Tech Lab’s technical recommendations and protocol enhancements for scalable, low-latency, reliable live ad delivery, integrating with the Deals API and OpenRTB. One component is final; the rest is in flight.
| LEAP component | Verified status | What it does |
|---|---|---|
| Concurrent Streams API | v1.0 — final (public comment closed July 28, 2025) | Standardizes how subscribers query the number of viewers in a live stream in real time; spec lives in the IABTechLab GitHub org, repo Live-Event-Ad-Protocols |
| Forecasting API | Phase Two — public comment February 19 to March 20, 2026 | Content owners share event schedules, expected ad breaks, and viewership forecasts with exchanges and DSPs in a schema aligned with OpenRTB and AdCOM; no finalization verified |
| Buyer Instructions API | Expected 2026 | Listed on the official LEAP page; not a published spec |
| Standardized Ad Pre-fetching | TBD | Listed on the official LEAP page; not a published spec |
| Creative Readiness | Expected 2026 | Listed on the official LEAP page; not a published spec |
Say it precisely
The Concurrent Streams API is a query and subscription interface for real-time concurrent viewership — it informs scaling and faster decisions; it does not push data into OpenRTB bid requests. And LEAP expands to Live Event Ad Playbook — Live-Event-Ad-Protocols is the GitHub repository’s name, not the standard’s.
SSAI, VAST, OM SDK, and device attestation.
Underneath the formats sits the stack that decides whether a CTV impression is deliverable, measurable, and trustworthy. Most of it has shipped — the 2024 VAST CTV Addendum, the OM SDK CTV builds, device attestation — and knowing exactly what each piece is (a final addendum, a 2020 macro guidance, an SDK with a specific platform list) is the difference between an executable plan and a hopeful one.
- Shipped 2024
VAST CTV Addendum 2024
Final, published July 18, 2024. Standardizes CTV-critical features across all VAST versions through backward-compatible extensions, not a new VAST version: UniversalAdId recognition retrofit into VAST 2.0–3.x (a designated field in 4.x), Open Measurement, interactive video, high-resolution creative, and DSA Article 26 ad-transparency icons via the Icon element.
- Guidance, not a spec
SSAI
IAB Tech Lab publishes no standalone “SSAI spec.” The standalone document is the SSAI VAST Macros Guidance v1.0 (November 2020) — a curated VAST macro subset for sharing basic user agent, environment, and content details critical to verification partners and buyers — plus SSAI-supporting features inside VAST 4.x and SSAI-relevant extensions in the CTV Addendum.
- Post-VPAID
Interactivity: OMID + SIMID
VPAID is officially deprecated, replaced by OMID for measurement and verification and SIMID for interactivity. SIMID separates the interactive layer from the media asset — both delivered by VAST 4.x — which is what makes interactivity workable over SSAI and across OTT platforms.
- Shipped 2022–2024
OM SDK for CTV
Announced June 8, 2022 for tvOS and Android TV, adding CTV-specific signals such as TV-off and last-user-activity-in-app. OM SDK 1.5 extended support to Samsung and LG TVs (May 30, 2024) via the HTML5/Web Video SDK. Roku, Vizio, and Fire TV are not listed as OM SDK CTV measurement platforms on the official standards pages.
- Shipped Oct 2025
Device attestation
Launched October 23, 2025 — shipped, not proposed — adopting the IETF Privacy Pass protocol for the ads-verification use case. The launch release says it is “currently supported on Apple devices and FireTV”; a 2026 Tech Lab blog separately says the latest release brings attestation across Samsung, LG, AndroidTV, and tvOS. Both are official — attribute each list to its source rather than merging them. No OM SDK version number is stated.
- Who stands behind it
The verification bench
OM SDK work runs under the active Open Measurement Working Group; the October 2025 launch release names an Open Measurement Commit Group of AdsWizz, Amazon Ads, DoubleVerify, Google, HUMAN, and Integral Ad Science, with DoubleVerify, HUMAN, and Integral Ad Science quoted supporting attestation at launch. Support statements, not adoption guarantees.
Households, apps, and frequency.
CTV measurement runs on a different unit than the web: the household. Between co-viewing, per-app state silos, stitched delivery, and live-event volatility, most person-level numbers on CTV are models wearing metric names. That does not make the channel unmeasurable — it makes precision about what each number proves non-negotiable.
- 01
The household is the unit
A TV screen is shared. Person-level metrics on CTV are typically modeled from household-level signals, and co-viewing multiplies the gap between devices and people. Treat audience figures as methodology-bearing estimates and ask vendors how the model works.
- 02
Every app is a silo
The frequency a buyer sees is computed across apps, devices, and SSAI paths that do not naturally share state. Where supported, identifiers differ by platform and by delivery path — validate what each path actually carries before trusting a cross-app frequency number.
- 03
Creative identity has a handle
Repetition, misdelivery, competitive separation, and brand-safety tracking all hang on creative identity. In VAST 4.x the UniversalAdId element carries a unique creative identifier across platforms — and the 2024 CTV Addendum retrofits its recognition into VAST 2.0–3.x.
- 04
Viewability reads differently
MRC viewability thresholds — 50% of pixels for one continuous second for display, two for video — set the baseline definitions. On a full-screen TV rendered through SSAI, the harder questions are render confirmation and device truth, which is where OM SDK CTV and attestation earn their place.
- 05
Counting moves server-side
With SSAI, impression events can originate from servers rather than the device. The SSAI macro guidance exists so verification partners get basic environment details — treat counts without independent verification coverage as seller-attested, and price that into the readout.
- 06
Live measurement is a forecast
For live events, planning data is a projection by definition. The LEAP Forecasting API — public comment closed March 2026, no finalization verified — would let content owners publish schedules, expected breaks, and viewership forecasts in a standard schema. Until then, validate what each publisher actually shares.
Agentic implications.
An agent buying CTV does not buy a pageview — it reasons about apps, platforms, stitched streams, households, and live moments. The standards on this page are the constraint set that reasoning has to satisfy. Six implications follow.
- 01
Capability discovery is per platform
An agent cannot assume a format, an SDK, or an attestation check exists on a given TV platform. Pause and Menu transaction support, OM SDK presence, and attestation coverage all differ by app and platform — capability discovery is a per-target check, not a channel-level fact.
- 02
Formats are not yet a stable contract
The portfolio and its signaling guidance were still in public comment as of this validation. An agent transacting a “Pause Ad” needs the definition pinned to a document version — name-matching against a moving draft is how mismatched buys happen.
- 03
Live buying needs scale-awareness
Agents buying live inventory should consume concurrency signals where publishers support the Concurrent Streams API, and tolerate forecast error everywhere else. A bidder tuned for steady-state load is mis-engineered for a stream where the audience arrives at once.
- 04
Weight spend by verification
Device spoofing motivated attestation; stitched delivery complicates counting. Agentic spend allocation should weight inventory by verification and attestation coverage — attributed per official source, since the platform lists differ between them.
- 05
Frequency goals need household logic
An agent optimizing person-level frequency on a shared screen is optimizing a model. Encode the household assumption explicitly — including how co-viewing and cross-app state gaps are handled — or the optimizer will treat model artifacts as audience behavior.
- 06
Reconciliation is server-shaped
SSAI delivery, pod-level dynamics, and server-originated events mean reconciliation logic built for browser pixels does not transfer. Agents need stitched-delivery reconciliation: what the ad server says, what verification observed, and what the contract says counts.
Implementation lens.
The same standards land differently depending on where you sit in the chain. Select your company type for the version of this page that applies to you.
Require format definitions pinned to document versions (the portfolio was still in public comment as of this validation), verification coverage stated per platform, attestation claims attributed to their official source, and household-modeled frequency labeled as such. Keep outcome claims inside what stitched, modeled delivery can support.
Maintain a capability matrix per streamer and platform: which portfolio formats are supported, what signaling each path carries, whether live concurrency data is available, and which measurement vendors cover which apps. Build live-event flighting around forecast error, not around the average minute.
Parse and honor the VAST CTV Addendum extensions, carry the SSAI VAST macros for verification partners, ship OM SDK where your platforms are supported, and evaluate the Concurrent Streams API for live inventory. Documented transparency about delivery and verification is a sales asset.
The Addendum names SSAI vendors among its implementers: parse UniversalAdId across VAST versions, pass the curated macro set faithfully, and expose pod and break structure to buyers. For live, engineer for concurrency spikes — the Concurrent Streams API exists because ad systems scaled for average load miss breaks.
Implement portfolio-format transaction support as the signaling finalizes (Pause and Menu first, per the programmatic guide), subscribe to concurrency data where offered, and weight buying by verification and attestation coverage. Remove web assumptions deliberately: no cookies, no client-side scripts, no person-level certainty.
State OM SDK CTV platform coverage precisely — AndroidTV and tvOS native, Samsung and LG via the web SDK — and attribute attestation support to the right source. Publish SSAI-aware methodology, household-model assumptions, and live-event measurement limits rather than letting buyers infer them.
The format surfaces — pause, menu, screensaver — are operating-system surfaces, and attestation depends on device-level participation. Where supported, aligning platform capabilities with the portfolio’s signaling work and the OM SDK builds is what makes the inventory tradable rather than bespoke.
No Fluff POV.
CTV requires execution-aware standards — and it is finally getting them, in public, in real time, and mostly in public comment. The honest inventory: the delivery and verification layers have shipped (the 2024 CTV Addendum, OM SDK CTV, device attestation, the Concurrent Streams API v1.0); the format and signaling layers were still drafts as of this validation; and the rest of LEAP is roadmap. Teams that treat CTV as display with a bigger screen automate the wrong assumptions. Teams that respect its physics — apps, streams, households, live moments — get the channel the standards are actually being written for.
- Use the official format names — Pause Ad, Menu Ad, Screensaver Ad, In Scene Insertion Ad, Overlay Ads, Squeeze Back Ads — and pin definitions to document versions: the portfolio guidelines and signaling guidance were still in public comment as of this validation.
- Say LEAP precisely: the Live Event Ad Playbook. The Concurrent Streams API v1.0 is final; the Forecasting API closed public comment in March 2026 without a verified finalization; everything else listed is expected or TBD.
- Treat SSAI as a delivery reality, not a spec: the standalone document is macro guidance from 2020, and the rest lives inside VAST 4.x and the 2024 CTV Addendum.
- Attribute attestation platform claims to their sources — the October 2025 launch release and the 2026 Tech Lab blog list different platforms, and reconciling them is speculation, not citation.
- Label household-modeled numbers as household-modeled: frequency, reach, and co-viewing on CTV are estimates with assumptions, and agents will optimize whatever you feed them.
The point
CTV is not just digital video on a bigger screen. It is an app, device, stream, household, SSAI, verification, and measurement problem — and the standards that matter are the ones that treat it that way.
Primary sources to validate.
Standards, guidance, and implementation references last validated: June 2026. Specifications, public-comment status, frameworks, APIs, and implementation guidance change. Validate against official documentation before implementation.
Primary sources to validate 9 sources
- CTV Ad Portfolio (standards page) ↗ Official standards page
Canonical hub for the CTV Ad Portfolio (last updated May 6, 2026). Exact format names: Pause Ad, Screensaver Ad, Menu Ad, In Scene Insertion Ad, Overlay Ads, and Squeeze Back Ads. Tracks two documents: 'Ad Format Guidelines for Digital Video and CTV' (public comment ended Jan 31, 2026) and 'CTV Ad Portfolio Signaling Guidance' (AdCOM/VAST/Native Ads updates, public comment until June 5, 2026). STATUS CAUTION: no finalized portfolio release was verified as of June 12, 2026 — do not call the portfolio 'final' or give it a version number. Supports: Exact official CTV format names, Portfolio document inventory and statuses, Signaling guidance scope (AdCOM, VAST, Native Ads).
- IAB Tech Lab Announces CTV Ad Portfolio and Updated Guide to Programmatic CTV (Released for Public Comment) ↗ Official press release
December 2025 announcement of the CTV Ad Portfolio plus a major update to the Guide to Programmatic CTV; both released for public comment until January 31, 2026. Names six core formats (Pause, Menu, Screensaver, In Scene, Squeezebacks, Overlays) derived from 100+ submissions via the 'Ad Format Hero' initiative. The updated guide adds OpenRTB support for the two prioritized formats: Pause and Menu (no specific OpenRTB version number is claimed). 'Guide to Programmatic CTV' (press release) and 'CTV Programmatic Guide' (standards index) are the same December 2025 document. Supports: Portfolio launch and public-comment window, Ad Format Hero origin, Pause/Menu OpenRTB prioritization.
- CTV Ad Portfolio Standardized Signals Open for Public Comment ↗ Official press release
May 6, 2026: standardized signaling spec updates (AdCOM, VAST, Native Ads) for four formats — Pause Ads, Menu Ads, Overlay Ads, Squeeze Back Ads — open for public comment until June 5, 2026, developed through the Ad Format Hero Task Force. Rationale quoted: transacting these formats 'has remained largely manual due to the lack of standardized definitions.' The signaling layer was still in public comment as of early June 2026 — not final. Supports: Most recent (May–June 2026) CTV format signaling status, Which four formats got transaction signaling first.
- Live Event Ad Playbook (LEAP) — standards page ↗ Official standards page
Canonical LEAP page (title also renders 'Live Events Ad Playbook' — expand LEAP as a Playbook, not 'Protocol'). LEAP is technical recommendations and protocol enhancements for scalable, low-latency, reliable ad delivery in live streaming, integrating with the Deals API and OpenRTB. Component table: Concurrent Streams API 1.0 FINAL (public comment closed July 28, 2025); Forecasting API (public comment closed March 20, 2026 — no finalization verified); Buyer Instructions API (expected 2026); Standardized Ad Pre-fetching (TBD); Creative Readiness (expected 2026). Cite components individually rather than a fixed phase count. Supports: LEAP official name, publisher, scope, Component list with versions/statuses, Concurrent Streams API 1.0 final status.
- IAB Tech Lab Releases Live Event Ad Playbook in Collaboration with Amazon, NBCUniversal, FreeWheel, and Index Exchange ↗ Official press release
August 5, 2025 launch of LEAP, introducing the (then-proposed) Concurrent Streams API: real-time concurrent-viewership data so publishers and ad systems make faster decisions and reduce missed ad breaks during high-concurrency moments such as live sports. Development led via the Live Event Ad Serving Working Group with Amazon, NBCUniversal, FreeWheel, and Index Exchange — named as development collaborators only; no production deployment or adoption is verified. Supports: LEAP launch date and named collaborators (development, not deployment), Live-event scale problem statement.
- IAB Tech Lab Opens Public Comment on LEAP Forecasting API ↗ Official press release
February 19, 2026: Phase Two of LEAP. The Forecasting API lets content owners share structured metadata — event schedules, expected ad breaks, viewership forecasts — with exchanges and DSPs using a schema aligned with OpenRTB and AdCOM. Public comment ran until March 20, 2026; no finalization announcement was verified as of June 2026 — do not cite the Forecasting API as final. Supports: Forecasting API scope and public-comment window, LEAP phase sequencing.
- Concurrent Streams API Specification (Live-Event-Ad-Protocols/Concurrent-Streams.md) ↗ Official GitHub
The Concurrent Streams API spec, v1.0 final (public comment closed July 28, 2025). Standardizes how subscribers query the number of viewers in a stream during a live event — a query/subscription API for concurrent viewership that DSPs, SSPs, and publisher ad servers use to scale ad-decisioning for concurrency spikes (it does not push data into OpenRTB bid requests). Note the repo lives under the IABTechLab GitHub org (not InteractiveAdvertisingBureau) and is named Live-Event-Ad-Protocols. Supports: Canonical spec text and version (1.0 final), Live-event scale signaling problem statement, Correct repo org/name.
- Advanced TV — Standards index ↗ Official standards page
Advanced TV Working Group document index listing 'CTV Ad Portfolio' and 'CTV Programmatic Guide', both dated December 2025 — confirming the programmatic guide exists as a distinct standing Tech Lab document. No version numbers or final/public-comment labels are shown on this index. Supports: Confirming the CTV Programmatic Guide as an official document, Document naming as listed in the standards index.
- Ad Format Guidelines for Digital Video and CTV — public-comment draft (December 2025 PDF) ↗ Official docs
The December 2025 public-comment draft PDF of the portfolio's format guidelines (internal title 'DRAFT - TV Ad Format Guidelines'). Text extraction failed during validation, so format definitions inside were not independently read — format names are confirmed from the CTV Ad Portfolio standards page instead. Do not treat sub-variation names (L-shaped, double-box, etc.) as load-bearing facts from this document. Supports: Canonical download link for the December 2025 public-comment draft.
Platform capabilities and naming change quickly. Last validated: June 13, 2026. Check current documentation before implementation.Standards, guidance, and implementation references last validated: June 2026. Specifications, public-comment status, frameworks, APIs, and implementation guidance change. Validate against official documentation before implementation.
Building CTV, streaming, or live-event workflows?
The operating work is to wire format definitions, live-scale signaling, SSAI-aware verification, and household-honest measurement into the buying path — before agents scale whatever the stream actually delivers.