{
 "published_before_scoring": "2026-11-03",
 "purpose": "Rules fixed before any product was scored. They define the cohorts and their applicability masks, what each dimension means, what a score of 0 to 5 requires, when a cell is n/e or n/a, which cells count, how availability, evidence grade and verification status are recorded, and how the Receipt Quotient summarises them without a league table.",
 "selection_rules": {
  "universe": "A reasoned sample, not a census, assembled in one bounded sweep (workstream U) from the brief's seed list, the round-1b reviewer's candidates (every line a lead until opened) and discovery, name by name: status, ownership, cohort by the boundary rule and the row-granularity rule, inclusion reason, US availability, the qualifying dimension-8 scale value with its perimeter code and the candidate H4a series. Excluded on purpose, as a labelled editorial exclusion: Google (Shopping ads, Performance Max, Merchant Center), Meta (Advantage+ shopping, Shops) and TikTok Shop ads, platforms whose primary business is media; TikTok Shop is the edge case revisited at U with its reason logged. Products found after the freeze enter only through a versioned register amendment that states its effect on every denominator and granularity; the original frozen register is kept for a sensitivity run; inclusion criteria never change after a score has been seen.",
  "scored_set": "Rows are named purchasable services for a defined buying task and placement scope (data/product-register.json lists every one): the purchasable service is registered first; placements, report tiers, access tiers and configurations a buyer cannot buy separately are child records, never rows; a separate scored row exists only for a separately purchasable task whose default method, analytical access or assurance scope differs; a rename alone never creates a row; evidence, parameters and accreditation never transfer between rows; no pooled parent RQ. Admission needs at least inclusion_min_scored counted numeric cells; n/a, n/e and unverified cells do not count. Live and beta services may be scored under the caps; announced and retired services are listed, not scored. A service whose task fits no ladder (an aggregate geo-lift or marketing-mix service with no attribution window and no household panel) is kept as a descriptive profile, never scored with vendor-specific n/a or substitute units.",
  "deep_profiles": "Six to eight networks and providers chosen by market weight and depth of public evidence, confirmed after workstream U: the brief names Amazon Ads, Walmart Connect, Instacart Ads, Kroger Precision Marketing, Target Roundel or Albertsons Media Collective, Criteo, one cohort B network and one cohort E provider. A network profile describes the portfolio and shows its service rows together without a pooled score."
 },
 "scale": {
  "0": "the record documents that the capability is absent or has failed the criterion",
  "1": "adjacent or incidental",
  "2": "partial or indirect, for example via a partner",
  "3": "real but thinly evidenced, narrow in scope, or beta",
  "4": "documented and specific but from vendor documentation only, or independently observed but narrow",
  "5": "the capability is documented in operational detail from an official document, filing or independent observation"
 },
 "cell_rules": {
  "scale": "0 to 5 integer, evidence-bounded, not absolute; every value is defined in \"scale\" above, and every applicable dimension carries its own anchors for at least 5, 3, 1 and 0 per cohort (anchors_by_cohort where the cohorts differ). Dimension anchors take precedence over the generic scale; the evidence cap is applied after scoring against the anchors.",
  "zero": "0 is used only when a source documents absence or failure of the criterion as the dimension's own 0 anchor defines it (for dimension 12, an affirmative, current, product-specific failure). Missing evidence is never scored 0.",
  "ne": "n/e (no evidence) when the opened record gives no basis to score either way, when a bounded disclosure screen (dimension 11) was not completed, when no qualifying scale figure exists (dimension 8), or when an assurance scope is unmapped (dimension 10). Distinct from 0. A partner's name alone is n/e; only documented, integrated functionality counts.",
  "na": "n/a comes only from the cohort's applicability mask (cohorts[].applicable), never vendor by vendor. n/a never enters k, a group mean, the coverage numerator or any imputation.",
  "availability": "Every scored cell carries one label: live (generally available at the cutoff), beta (limited access or beta), announced (not yet observable), retired, unconfirmed (claimed but no observable evidence).",
  "evidence_grade": "A = filing, official transaction document, regulator, court, accreditor or standards-body record; B = independent measurement or documentation by a party without financial interest; C = vendor technical documentation, developer documentation or help centre; D = vendor marketing, press release or case study; E = none. A third party paid by the vendor is not independent; if funding is unknown, independence is unknown.",
  "date": "The date of the most recent source the cell rests on (YYYY-MM or YYYY-MM-DD).",
  "verification_status": "Every scored cell carries one: verified (a verifier re-checked the cell's sources), qualified (re-checked; scored against the qualified claim) or unverified. Only verified and qualified cells count; an unverified cell is shown but treated as n/e for admission, k, the quotient and band eligibility.",
  "product": "The scoring unit is a named product that serves the cohort's buying task, one row per product, each mapped to one record in data/product-register.json. Evidence is not inherited between rows, and an alias or bundle never creates a second row.",
  "same_standard": "The same evidentiary standard applies to every vendor. Documentation volume is not product quality: a large help centre earns grade C, not a higher score.",
  "grade_caps": "Vendor documentation (grade C) caps a score at 4 and vendor marketing (grade D) at 3; a grade can cap a score but never raise it. scripts/check.mjs enforces the caps in grade_caps.",
  "maturity": "Live and beta products may be scored under the caps; announced and retired products are listed in the product register, not scored (check.mjs fails a scored announced or retired cell).",
  "us_first": "The US availability dimension scores documented US availability and coverage only. International breadth is recorded as an attribute and earns no points.",
  "interest": "A declared interest (the author's employer, the AdCP working-group role, a past employer) is marked on the row and disclosed; interest marks never alter a score, and a sensitivity run reports whether any conclusion changes when evidence from conflicted sources is excluded."
 },
 "grade_caps": {
  "C": 4,
  "D": 3
 },
 "inclusion_min_scored": 4,
 "dimensions": [
  {
   "key": "default_method",
   "label": "Default outcome method and controlled-lift availability",
   "short": "Method",
   "definition": "The default method behind the service's outcome reporting and the availability of controlled lift, coded on the measurement record (assignment and its unit, counterfactual construction, estimation method, documented checks). A method label never settles assignment; allocation stays unknown unless the design discloses it. Cohort C: the control designs the platform documents as supported for its clients' campaigns. Cohort D: the incrementality testing the platform itself offers across networks. Cohort E: the control design of its own measurement service, read as its standard design.",
   "anchors": {
    "5": "random assignment is the default, with minimum samples, significance rule and uncertainty documented by an official record or independent observation",
    "4": "random assignment is the default without all of minimum samples, significance rule and uncertainty documented; or random-assignment lift available and documented but not the default, with the default's limits stated",
    "3": "a default lift method with non-random or unknown assignment and a documented counterfactual construction (any construction in the measurement record other than none; a pre/post comparison alone scores no higher than 2), with method, balance and uncertainty documented",
    "2": "such a counterfactual asserted without those checks (a pre/post comparison alone scores here at most)",
    "1": "documented exposure-to-purchase attribution with no campaign counterfactual, whatever the vendor calls it",
    "0": "a source documents that lift reported as causal was not; a row whose default is not documented is n/e, never 0"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "random assignment supported for clients' campaigns, with the method, minimum samples and significance rule documented",
     "4": "random assignment supported and documented without all of samples, significance rule and uncertainty",
     "3": "lift designs with non-random or unknown assignment supported, with the method documented",
     "2": "such a counterfactual asserted without method documentation",
     "1": "attribution only",
     "0": "a source documents that lift reported as causal by the platform's module was not; an undocumented module is n/e"
    },
    "optimisation": {
     "5": "random assignment supported for clients' campaigns, with the method, minimum samples and significance rule documented",
     "4": "random assignment supported and documented without all of samples, significance rule and uncertainty",
     "3": "lift designs with non-random or unknown assignment supported, with the method documented",
     "2": "such a counterfactual asserted without method documentation",
     "1": "attribution only",
     "0": "a source documents that lift reported as causal by the platform's module was not; an undocumented module is n/e"
    },
    "measurement": {
     "5": "random assignment is the default, with minimum samples, significance rule and uncertainty documented by an official record or independent observation",
     "4": "random assignment is the default without all of minimum samples, significance rule and uncertainty documented; or random-assignment lift available and documented but not the default, with the default's limits stated",
     "3": "a default lift method with non-random or unknown assignment and a documented counterfactual construction (any construction in the measurement record other than none; a pre/post comparison alone scores no higher than 2), with method, balance and uncertainty documented",
     "2": "such a counterfactual asserted without those checks (a pre/post comparison alone scores here at most)",
     "1": "documented exposure-to-purchase attribution with no campaign counterfactual, whatever the vendor calls it",
     "0": "a source documents that lift reported as causal was not; a row whose default is not documented is n/e, never 0"
    }
   }
  },
  {
   "key": "attribution_disclosure",
   "label": "Attribution definition disclosure",
   "short": "Attribution",
   "definition": "The parameters of the attributed metrics, per placement class: window (click and view, in days), touch model, match basis, in-store inclusion, new-to-brand lookback, sales basis, deduplication and halo. Cohort C: the parameters the platform documents as configurable and their defaults. Cohort D: the parameters it discloses for each network feed it reports (a normalised metric with undisclosed parameters scores 1; feed origin and limits are dimension 3). Cohort E: the parameters of its own metrics.",
   "anchors": {
    "5": "all eight parameters documented with values per placement class: window (click and view, days), touch model, match basis, in-store inclusion, new-to-brand lookback, sales basis, deduplication, halo",
    "4": "six or seven of the eight documented with values",
    "3": "window and touch model stated, the rest missing",
    "2": "one of window or touch model stated",
    "1": "\"closed loop\" or \"attributed sales\" asserted with no parameter",
    "0": "a documented inconsistency between stated and applied parameters"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "all eight parameters documented as configurable with their defaults stated, per placement class",
     "4": "six or seven of the eight documented with values",
     "3": "window and touch model stated, the rest missing",
     "2": "one of window or touch model stated",
     "1": "attribution offered with no parameter or default disclosed",
     "0": "a documented inconsistency between stated and applied parameters"
    },
    "optimisation": {
     "5": "all eight parameters disclosed for every network feed it reports",
     "4": "six or seven of the eight documented with values",
     "3": "window and touch model disclosed per feed, the rest missing",
     "2": "one of window or touch model stated",
     "1": "a normalised metric with undisclosed parameters",
     "0": "a documented inconsistency between stated and applied parameters"
    },
    "measurement": {
     "5": "all eight parameters documented with values per placement class: window (click and view, days), touch model, match basis, in-store inclusion, new-to-brand lookback, sales basis, deduplication, halo",
     "4": "six or seven of the eight documented with values",
     "3": "window and touch model stated, the rest missing",
     "2": "one of window or touch model stated",
     "1": "\"closed loop\" or \"attributed sales\" asserted with no parameter",
     "0": "a documented inconsistency between stated and applied parameters"
    }
   }
  },
  {
   "key": "buyer_analysis",
   "label": "Buyer-controlled outcome analysis",
   "short": "Access",
   "definition": "Whether a buyer, or an independent party on the buyer's behalf, can run its own outcome analysis with documented lineage and constraints. Recorded beside the cell as separate fields, never pooled into it: row-level portability (event-level export) and delivery verification tags allowed on-site; neither alone earns 5. Cohort C: the access routes the platform documents as supported for its retailers' buyers. Cohort D: outcome-feed provenance, the denominator being every network outcome feed the service reports at the freeze. Cohort E: the access it gives buyers to its own data and method.",
   "anchors": {
    "5": "buyer-run outcome analysis documented (queries in a controlled environment or analysis on permitted exports), with output rules, constraints and lineage documented, and independent replication by a named third party documented",
    "4": "buyer-run outcome analysis documented with output rules and constraints",
    "3": "aggregates delivered through an API or clean room on the operator's queries only",
    "2": "dashboard reports plus documented access for a named third party to the operator's outcome data on the operator's terms",
    "1": "dashboard reports only",
    "0": "terms documented as prohibiting any third-party analysis of the inventory's outcomes"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "buyer-run outcome analysis documented (queries in a controlled environment or analysis on permitted exports), with output rules, constraints and lineage documented, and independent replication by a named third party documented",
     "4": "buyer-run outcome analysis documented with output rules and constraints",
     "3": "aggregates delivered through an API or clean room on the operator's queries only",
     "2": "dashboard reports plus documented access for a named third party to the operator's outcome data on the operator's terms",
     "1": "dashboard reports only",
     "0": "terms documented as prohibiting any third-party analysis of the inventory's outcomes"
    },
    "optimisation": {
     "5": "origin, granularity, access or transformation constraints and known limitations documented for every network outcome feed the service reports at the freeze",
     "4": "those four documented for more than half but not all of the feeds",
     "3": "those four documented for at least half of the feeds",
     "2": "origin documented for at least half of the feeds, the other fields not",
     "1": "normalised outcome reporting with provenance undisclosed",
     "0": "a source documents a materially false feed or provenance claim; an incomplete search is n/e"
    },
    "measurement": {
     "5": "buyer-run outcome analysis documented (queries in a controlled environment or analysis on permitted exports), with output rules, constraints and lineage documented, and independent replication by a named third party documented",
     "4": "buyer-run outcome analysis documented with output rules and constraints",
     "3": "aggregates delivered through an API or clean room on the operator's queries only",
     "2": "dashboard reports plus documented access for a named third party to the operator's outcome data on the operator's terms",
     "1": "dashboard reports only",
     "0": "terms documented as prohibiting any third-party analysis of the inventory's outcomes"
    }
   }
  },
  {
   "key": "identity_basis",
   "label": "Identity and data-basis disclosure",
   "short": "Identity",
   "definition": "What the attributed outcome is matched to and how completely that basis is disclosed (identity basis; the channels matched and how; the share of transactions matchable), scored on completeness of disclosure, never on how many channels a service matches. Not applicable to cohort D (outside its mask). Cohort E: the panel or feed basis of its own measurement.",
   "anchors": {
    "5": "identity basis, the method for every matched channel and a numeric matchable share documented",
    "4": "identity basis and the method for every matched channel documented, share stated qualitatively only",
    "3": "identity basis and matched channels documented, share not stated",
    "2": "identity basis documented, matched channels not",
    "1": "the basis asserted only",
    "0": "a source documents that attributed sales include modelled or unmatched sales without labelling"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "identity basis, the method for each ingested channel and coverage reporting documented for the platform's clients",
     "4": "identity basis and the method for each ingested channel documented, coverage reporting partial",
     "3": "identity basis and ingested channels documented, coverage reporting not",
     "2": "identity basis documented, ingested channels not",
     "1": "the basis asserted only",
     "0": "a source documents that attributed sales include modelled or unmatched sales without labelling; the breadth of ingestible channels is published in the profile, unscored"
    },
    "measurement": {
     "5": "panel or feed basis with size, US coverage and projection method documented",
     "4": "basis, size and US coverage documented, projection method partly",
     "3": "basis and size documented, projection method not",
     "2": "basis documented, size not",
     "1": "basis asserted only",
     "0": "a source documents that reported outcomes include modelled or unmatched components without labelling"
    }
   }
  },
  {
   "key": "validation",
   "label": "Validation",
   "short": "Validation",
   "definition": "Independent or out-of-sample validation of the service's outcome numbers (attributed against experimental results, panel against point-of-sale, lift method against replication), coded with the workstream B fields. The vendor-evidence cap applies after the anchor; independence is a per-claim field, not a rung. Cohort C: its measurement module; cohort D: its normalisation and optimisation claims; cohort E: its service.",
   "anchors": {
    "5": "independent validation with published design, replicated",
    "4": "independent validation on one dataset, or vendor validation with random assignment and a published method",
    "3": "vendor validation with a controlled design of non-random or unknown assignment and documented identification and uncertainty checks",
    "2": "vendor validation with a published method and no counterfactual",
    "1": "accuracy claims only",
    "0": "an independent study finds the service's numbers systematically wrong beyond its stated error"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "independent validation with published design, replicated",
     "4": "independent validation on one dataset, or vendor validation with random assignment and a published method",
     "3": "vendor validation with a controlled design of non-random or unknown assignment and documented identification and uncertainty checks",
     "2": "vendor validation with a published method and no counterfactual",
     "1": "accuracy claims only",
     "0": "an independent study finds the service's numbers systematically wrong beyond its stated error"
    },
    "optimisation": {
     "5": "independent validation with published design, replicated",
     "4": "independent validation on one dataset, or vendor validation with random assignment and a published method",
     "3": "vendor validation with a controlled design of non-random or unknown assignment and documented identification and uncertainty checks",
     "2": "vendor validation with a published method and no counterfactual",
     "1": "accuracy claims only",
     "0": "an independent study finds the service's numbers systematically wrong beyond its stated error"
    },
    "measurement": {
     "5": "independent validation with published design, replicated",
     "4": "independent validation on one dataset, or vendor validation with random assignment and a published method",
     "3": "vendor validation with a controlled design of non-random or unknown assignment and documented identification and uncertainty checks",
     "2": "vendor validation with a published method and no counterfactual",
     "1": "accuracy claims only",
     "0": "an independent study finds the service's numbers systematically wrong beyond its stated error"
    }
   }
  },
  {
   "key": "placement_coverage",
   "label": "Placement measurement coverage",
   "short": "Placements",
   "definition": "For each placement class the service serves (on-site, off-site, in-store), whether measurement is documented per class. Placement breadth is published in the profile and is not scored, so a specialist is never marked down for scope. Cohort D: networks reachable through documented API integrations. Cohort E: networks or retailers covered by its measurement.",
   "anchors": {
    "5": "every served placement class has per-class measurement documented with parameters (exposure capture, match, reporting); a service that serves one class and documents its measurement scores 5",
    "4": "per-class measurement documented for every served class, parameters incomplete for at least one",
    "3": "documented for the primary class only",
    "2": "documented for the primary class with parameters incomplete",
    "1": "classes served with measurement pooled or undocumented",
    "0": "a source documents that reporting pools classes it claims to separate"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "every served placement class has per-class measurement documented with parameters (exposure capture, match, reporting); a service that serves one class and documents its measurement scores 5",
     "4": "per-class measurement documented for every served class, parameters incomplete for at least one",
     "3": "documented for the primary class only",
     "2": "documented for the primary class with parameters incomplete",
     "1": "classes served with measurement pooled or undocumented",
     "0": "a source documents that reporting pools classes it claims to separate"
    },
    "optimisation": {
     "5": "ten or more networks reachable through documented API integrations",
     "4": "seven to nine",
     "3": "three to nine (three to six where 4 applies)",
     "2": "two",
     "1": "one or two (one where 2 applies)",
     "0": "a documented integration withdrawn without replacement, leaving none"
    },
    "measurement": {
     "5": "ten or more networks or retailers covered by its measurement",
     "4": "seven to nine",
     "3": "three to nine (three to six where 4 applies)",
     "2": "two",
     "1": "one or two (one where 2 applies)",
     "0": "a documented coverage withdrawn without replacement, leaving none"
    }
   }
  },
  {
   "key": "routes",
   "label": "Activation and integration routes",
   "short": "Routes",
   "definition": "Documented route families by which a buyer activates or receives the service. Cohort C: the route families the platform documents (DSP and SSP connections, APIs, console). Cohort D: the families by which it connects to networks and delivers reporting to buyers (network APIs, console, agency and DSP integrations). Cohort E: the families by which its measurement reaches buyers and networks.",
   "anchors": {
    "5": "three or more route families including an API, documented (families: API; self-serve console; owned DSP; partner DSPs; agency or optimisation-platform integrations; clean-room integrations; managed service; several connectors of one kind are one family)",
    "4": "three or more families without an API",
    "3": "one or two families",
    "2": "one family documented thinly (beta or limited access)",
    "1": "managed service only",
    "0": "a documented route withdrawn without replacement"
   },
   "anchors_by_cohort": {
    "enablement": {
     "5": "three or more route families including an API, documented (families: API; self-serve console; owned DSP; partner DSPs; agency or optimisation-platform integrations; clean-room integrations; managed service; several connectors of one kind are one family)",
     "4": "three or more families without an API",
     "3": "one or two families",
     "2": "one family documented thinly (beta or limited access)",
     "1": "managed service only",
     "0": "a documented route withdrawn without replacement"
    },
    "optimisation": {
     "5": "three or more route families including an API, documented (families: API; self-serve console; owned DSP; partner DSPs; agency or optimisation-platform integrations; clean-room integrations; managed service; several connectors of one kind are one family)",
     "4": "three or more families without an API",
     "3": "one or two families",
     "2": "one family documented thinly (beta or limited access)",
     "1": "managed service only",
     "0": "a documented route withdrawn without replacement"
    },
    "measurement": {
     "5": "three or more route families including an API, documented (families: API; self-serve console; owned DSP; partner DSPs; agency or optimisation-platform integrations; clean-room integrations; managed service; several connectors of one kind are one family)",
     "4": "three or more families without an API",
     "3": "one or two families",
     "2": "one family documented thinly (beta or limited access)",
     "1": "managed service only",
     "0": "a documented route withdrawn without replacement"
    }
   }
  },
  {
   "key": "scale",
   "label": "Scale",
   "short": "Scale",
   "definition": "The service's disclosed numeric scale in a stated period, in one unit per cohort fixed here before scoring, with thresholds set from the distribution workstream U reports and frozen before any cell is scored. The register records each figure's line definition and perimeter code.",
   "scale_rules": {
    "unit_by_cohort": {
     "retailer_network": "annual advertising revenue (US$, stated fiscal period)",
     "commerce_network": "annual advertising revenue (US$, stated fiscal period)",
     "enablement": "distinct named US retailer networks with dated production evidence in the screened record, deduplicated across aliases and service rows, published as a documented minimum footprint",
     "optimisation": "ad spend under management (US$, stated period)",
     "measurement": "US households in the panel or matched feed (stated period)"
    },
    "perimeter_codes": "Mutually exclusive, mapped by name before thresholds: service (scoped to the service) and network (scoped to the network) score in full; broader parent proxy (a line that spans other named networks or non-advertising revenue: Walmart's global advertising business \"including VIZIO\"; Instacart's \"advertising and other\"; Cardlytics' company-wide revenue; a company-wide retail media segment that includes other businesses) caps at 3; company-wide unallocated line (Amazon's \"advertising services\") is context only and never scores; third-party estimate scoped to the network, with publisher and vintage, caps at 3. Undisclosed service revenue is never prorated from a broader line.",
    "evidence_caps": "Applied after the perimeter cap: an audited filing amount scores in full; an unaudited operational disclosure in a filing or an earnings-release approximation caps at 4; a figure only in a press release or product page caps at 3; a figure whose measurement period (not its page date) ends before 2024 caps at 2. No qualifying figure, or no named deployment for cohort C, is n/e, never 0. Confidence is published beside the cell; no hypothesis derives a quartile from these scores.",
    "threshold_algorithm": "Frozen 2026-11-03 from the distribution workstream U reported (data table scale_values in evidence/research/U.json), before any cell was scored. Cohorts A and B (annual advertising revenue, US$): half-decade steps (a factor of about 3.16) anchored at US$100 million, spanning the observed service- and network-perimeter range (Ibotta \"Ad & other\" US$45.2 million to Walmart's \"nearly $6.4 billion\"): 5 = US$3 billion or more; 4 = US$1 billion to under 3 billion; 3 = US$300 million to under 1 billion; 2 = US$100 million to under 300 million; 1 = a qualifying figure under US$100 million. Cohort C (distinct named US retailer networks with dated production evidence, deduplicated): the observed footprints run from 1 to 3 (Moloco 1, Vantage 1, STRATACACHE 1, Epsilon 1 dated 2022, Pentaleap 2, Kevel 2, Swiftly 3), so the ladder places them on rungs 1 to 3 and reserves the upper rungs for footprints no platform documented by the freeze: 5 = 10 or more; 4 = 5 to 9; 3 = 3 to 4; 2 = 2; 1 = 1. Cohort D (ad spend under management, US$): half-decade steps anchored at US$300 million spanning Intentwise US$1.5 billion (Amazon only) to Pacvue \"over US$12 billion\": 5 = US$10 billion or more; 4 = US$3 billion to under 10 billion; 3 = US$1 billion to under 3 billion; 2 = US$300 million to under 1 billion; 1 = under US$300 million. Cohort E (US households in the panel or matched feed): half-decade steps anchored at 3 million spanning Numerator \"1 Million+\" to Catalina \"130 million households\": 5 = 100 million or more; 4 = 30 million to under 100 million; 3 = 10 million to under 30 million; 2 = 3 million to under 10 million; 1 = under 3 million. A figure in another unit (Attain's \"14M+ verified shoppers\", consumers not households) is recorded beside the cell and the cell is n/e. Perimeter caps then evidence caps apply after the rung."
   },
   "anchors": {
    "5": "annual advertising revenue of US$3 billion or more (service or network perimeter)",
    "4": "US$1 billion to under US$3 billion",
    "3": "US$300 million to under US$1 billion",
    "2": "US$100 million to under US$300 million",
    "1": "a qualifying figure under US$100 million",
    "0": "no 0 is possible on this dimension by design: no qualifying figure is n/e, never 0; 0 is reserved for a source documenting that the service earned no advertising revenue in the period"
   },
   "anchors_by_cohort": {
    "commerce_network": {
     "5": "annual advertising revenue of US$3 billion or more (service or network perimeter)",
     "4": "US$1 billion to under US$3 billion",
     "3": "US$300 million to under US$1 billion",
     "2": "US$100 million to under US$300 million",
     "1": "a qualifying figure under US$100 million",
     "0": "no 0 is possible on this dimension by design: no qualifying figure is n/e, never 0; 0 is reserved for a source documenting that the service earned no advertising revenue in the period"
    },
    "enablement": {
     "5": "ten or more distinct named US retailer networks with dated production evidence",
     "4": "five to nine",
     "3": "three to four",
     "2": "two",
     "1": "one",
     "0": "a source documents that the platform has no US retailer network in production in the period; no named deployment is n/e, never 0"
    },
    "optimisation": {
     "5": "ad spend under management of US$10 billion or more in a stated period",
     "4": "US$3 billion to under US$10 billion",
     "3": "US$1 billion to under US$3 billion",
     "2": "US$300 million to under US$1 billion",
     "1": "a qualifying figure under US$300 million",
     "0": "a source documents no spend under management in the period; no qualifying figure is n/e, never 0"
    },
    "measurement": {
     "5": "100 million or more US households in the panel or matched feed",
     "4": "30 million to under 100 million",
     "3": "10 million to under 30 million",
     "2": "3 million to under 10 million",
     "1": "under 3 million",
     "0": "a source documents no US panel or matched feed in the period; no qualifying figure is n/e, never 0"
    }
   }
  },
  {
   "key": "us_availability",
   "label": "US availability and coverage",
   "short": "US",
   "definition": "Documented US availability and production use (shared rules). International breadth earns no points. One ladder for every cohort.",
   "anchors": {
    "5": "documented US production use in the period (named US clients, campaigns or a filing) and documented availability to new US buyers",
    "4": "documented US production use in the period, availability to new US buyers not documented; or documented availability with production use documented only through a partner",
    "3": "US availability documented, production use not",
    "2": "available in the US only through a partner or reseller, with the arrangement documented",
    "1": "availability asserted only",
    "0": "documented as not available in the US"
   }
  },
  {
   "key": "assurance",
   "label": "Independent assurance",
   "short": "Assurance",
   "definition": "Recorded per instrument as provider, standard, stage (completed; audit in process; pre-audit, in process or reviewed; not begun), product, metric, environment and exclusions, and scored on completed assurance of the service's own claims, by scope rather than by provider, taking the highest applicable rung and recording the other instruments beside it. Each IAB Europe certificate is read for its compliance version, service, metrics, geography, environments and exclusions and placed on the same ladder (its compliance table tests attribution reporting and any sales-uplift A/B test; incrementality is optional), so the programme badge alone proves neither a US service nor certified lift; an unmapped scope is n/e. Delivery accreditation attests nothing about outcomes and the cell says so. One ladder for every cohort.",
   "anchors": {
    "5": "completed third-party assurance (MRC accreditation, AAM certification, IAB Europe certification or a named audit under a published standard) of the service's outcome metrics or outcome method, covering the environments sold",
    "4": "completed assurance of the service's delivery metrics in the environments sold, or outcome assurance covering a narrower scope than sold",
    "3": "completed assurance of a narrower delivery scope",
    "2": "any mapped service listed in an accreditor's process, including a reviewed pre-audit (a status, not assurance)",
    "1": "\"aligned with\" or \"built to\" the guidelines, self-attested",
    "0": "accreditation denied, withdrawn or lapsed, documented"
   }
  },
  {
   "key": "governance",
   "label": "Change and correction governance",
   "short": "Governance",
   "definition": "How the service governs its own numbers: metric versioning and change logs, restatement and correction policy, documented reporting revisions, challenge or dispute procedures, reproducibility records. Scored only after a completed, bounded disclosure screen of the service's documentation and terms; an incomplete screen is n/e. Attribution parameters live in dimension 2, control design in 1, identity and data basis in 4, placement coverage in 6 and validation in 5; this dimension never scores them again. One ladder for every cohort.",
   "anchors": {
    "5": "all five documented with examples: metric versioning and change logs, restatement and correction policy, documented reporting revisions, challenge or dispute procedures, reproducibility records (methodology documents with version history)",
    "4": "all five documented",
    "3": "three or four documented",
    "2": "one or two documented",
    "1": "none found in the screened record (a bounded absence after a completed disclosure screen, never proof that no controls exist)",
    "0": "a documented restatement or revision with no disclosure to buyers"
   }
  },
  {
   "key": "privacy",
   "label": "Privacy and consent posture",
   "short": "Privacy",
   "definition": "Current, product-specific controls against a fixed checklist: consent basis for using purchase, loyalty or payment data in advertising; consent or notice for matching with partners; treatment of sensitive purchase categories (health, pharmacy, alcohol, children's products); state-law handling (sale and sharing opt-outs, sensitive data); data-broker registration where it applies; independent assurance of the controls. Adverse events are recorded separately (event type, date, final or pending status, remediation, obligations); a pending complaint is recorded, not scored; the absence of a public remediation record is not proof of continuing failure. One ladder for every cohort.",
   "anchors": {
    "5": "consent basis and sensitive-category handling documented for the product, with named legal frameworks and an independent record (a named privacy certification or audit covering the product, or a regulator-approved assessment)",
    "4": "both documented with named legal frameworks",
    "3": "one of the two documented for the product",
    "2": "product-specific notice without named legal frameworks",
    "1": "generic company policy only",
    "0": "an affirmative, current, product-specific failure under this frozen criterion: a regulator order or court finding in force whose obligations the record shows unmet, or a documented breach of the product's shopper data within the three years before the cutoff that the record shows unremedied; a silent product record is n/e, not 1"
   }
  }
 ],
 "peer_groups": [
  {
   "key": "retailer",
   "label": "Retailers: store-based retailers and pharmacies selling media",
   "definition": "Operators whose media business sits on a physical and online retail estate (Walmart, Target, Kroger, Albertsons, Ahold Delhaize USA, Best Buy, The Home Depot, Lowe's, CVS, Walgreens, Macy's, Ulta, Kohl's, Dollar General, DICK'S, Petco, Staples, 7-Eleven, Hy-Vee)."
  },
  {
   "key": "marketplace",
   "label": "Marketplaces: online marketplaces, delivery and travel marketplaces",
   "definition": "Operators whose commerce surface is a marketplace or booking surface (Amazon, Instacart, DoorDash, Uber Eats, Chewy, Wayfair, eBay, Etsy, Gopuff, Expedia Group)."
  },
  {
   "key": "commerce_data",
   "label": "Commerce-data operators: banks, payments, travel, rideshare and reward apps selling non-endemic media",
   "definition": "Operators selling advertising against transaction, loyalty, payment or travel data to outside brands (Chase, PayPal, Cardlytics, American Express, Capital One, Klarna, Affirm, Marriott, United, Uber, Lyft, Fetch, Ibotta, Zillow)."
  },
  {
   "key": "adtech",
   "label": "Ad tech: enablement and optimisation software",
   "definition": "Platforms retailers license to run a network (cohort C) and software brands and agencies use to buy across networks (cohort D)."
  },
  {
   "key": "measurer",
   "label": "Measurers: outcome measurement providers",
   "definition": "Services that measure retail or commerce media outcomes and sell the result (cohort E)."
  }
 ],
 "cohorts": [
  {
   "key": "retailer_network",
   "label": "A · Retailer and marketplace media networks",
   "definition": "Services selling advertising for goods or services purchasable through the operator's own store, marketplace or booking surface (endemic advertising), on its surfaces, with its data or in its stores. Bought by brands and their agencies. One row per purchasable service; a network's rows are shown together in its profile without a pooled score.",
   "applicable": [
    "default_method",
    "attribution_disclosure",
    "buyer_analysis",
    "identity_basis",
    "validation",
    "placement_coverage",
    "routes",
    "scale",
    "us_availability",
    "assurance",
    "governance",
    "privacy"
   ],
   "banded": true
  },
  {
   "key": "commerce_network",
   "label": "B · Commerce media networks",
   "definition": "Services selling non-endemic advertising against transaction, loyalty, payment or travel data (the advertised goods or services are not purchasable through the operator's surface). Bought by brands and their agencies. Scored only if workstream U admits enough rows; shown with quotients and no bands if fewer than six rows are band-eligible, or as a described table if fewer than four are admitted.",
   "applicable": [
    "default_method",
    "attribution_disclosure",
    "buyer_analysis",
    "identity_basis",
    "validation",
    "placement_coverage",
    "routes",
    "scale",
    "us_availability",
    "assurance",
    "governance",
    "privacy"
   ],
   "banded": true
  },
  {
   "key": "enablement",
   "label": "C · Enablement platforms",
   "definition": "Ad serving, sponsored-products engines, demand- and supply-side platforms, self-serve portals and measurement modules that retailers license to run a network. Bought by retailers and operators. Scored on the functions the platform documents as supported and configurable and on its constraints, never on what a particular client deployed.",
   "applicable": [
    "default_method",
    "attribution_disclosure",
    "buyer_analysis",
    "identity_basis",
    "validation",
    "placement_coverage",
    "routes",
    "scale",
    "us_availability",
    "assurance",
    "governance",
    "privacy"
   ],
   "banded": true
  },
  {
   "key": "optimisation",
   "label": "D · Cross-network optimisation platforms",
   "definition": "Software that plans, buys, bids and reports across several networks for brands and agencies. Scored on its own functions and on the provenance and limits of the outcome feeds it consumes; dimension 4 (identity and data basis) is outside its mask and dimension 3 is read as outcome-feed provenance.",
   "applicable": [
    "default_method",
    "attribution_disclosure",
    "buyer_analysis",
    "validation",
    "placement_coverage",
    "routes",
    "scale",
    "us_availability",
    "assurance",
    "governance",
    "privacy"
   ],
   "banded": true
  },
  {
   "key": "measurement",
   "label": "E · Outcome measurement providers",
   "definition": "Services that measure retail or commerce media outcomes and sell the result to brands, agencies or networks. Scored on the measurement service itself; independence is coded claim by claim from the reporting-party fields, never assumed from the cohort.",
   "applicable": [
    "default_method",
    "attribution_disclosure",
    "buyer_analysis",
    "identity_basis",
    "validation",
    "placement_coverage",
    "routes",
    "scale",
    "us_availability",
    "assurance",
    "governance",
    "privacy"
   ],
   "banded": true
  }
 ],
 "cohorts_note": "Each cohort lists its applicable dimensions (the mask); n/a comes only from the mask. Dimensions 1-8 carry per-cohort anchors (anchors_by_cohort); dimensions 9-12 apply one ladder to every cohort. Every applicable cohort x dimension has a complete ladder including 0, confirmed by scripts/check.mjs before the freeze. A service whose task fits no ladder is kept as a descriptive profile, never scored with vendor-specific n/a or substitute units.",
 "row_order": [],
 "matrix_title": "The capability matrix",
 "matrix_note": "One row per purchasable service, scored within its cohort on documented measurement capability, never on measurement quality or effectiveness. Click a cell for its anchor, rationale, sources and verification status; n/e means the opened record gave no basis to score, n/a that the dimension is outside the cohort's mask. A network's rows sit together; there is no pooled network score.",
 "matrix_foot": "Only verified or qualified cells count. Unverified cells are shown dashed and treated as n/e. Interest marks (the author's employer, the AdCP working-group role, past employers) are disclosed and never alter a score.",
 "composite_index": {
  "used": true,
  "name": "Receipt Quotient (RQ)",
  "added": "2026-11-03, with these rules and before any service was scored, as the quotient of the shared method (shared research rules v3) first used for the attention, audio and gaming papers and then the location paper; the dimension-8 thresholds are frozen after workstream U and before any cell is scored.",
  "groups": {
   "Depth": {
    "label": "Depth",
    "meaning": "What the number is: the default outcome method and controlled-lift availability, the attribution parameters disclosed, buyer-controlled outcome analysis, the identity and data basis disclosed, and validation.",
    "dims": [
     "default_method",
     "attribution_disclosure",
     "buyer_analysis",
     "identity_basis",
     "validation"
    ]
   },
   "Reach": {
    "label": "Reach",
    "meaning": "Where it works: per-class placement measurement coverage, activation and integration routes, disclosed scale and US availability.",
    "dims": [
     "placement_coverage",
     "routes",
     "scale",
     "us_availability"
    ]
   },
   "Standing": {
    "label": "Standing",
    "meaning": "Whether a buyer can trust and trade it: independent assurance, change and correction governance, and privacy and consent posture.",
    "dims": [
     "assurance",
     "governance",
     "privacy"
    ]
   }
  },
  "weights": {
   "Depth": 0.45,
   "Reach": 0.25,
   "Standing": 0.3
  },
  "field_mean_method": "raw_mean",
  "prior_strength": 4,
  "shrink_alternatives": [
   0,
   2,
   8
  ],
  "missing_scenario_values": [
   1,
   4
  ],
  "missing_envelope_values": [
   0,
   5
  ],
  "bands": [
   [
    "Most documented",
    12
   ],
   [
    "Well documented",
    1
   ],
   [
    "Partly documented",
    -8
   ],
   [
    "Least documented",
    -1000
   ]
  ],
  "thin_record_min": 6,
  "cohort_min_banded": 6,
  "method": "Within each cohort, over its admitted rows (thin rows included; listed and historical products never count): each group score is the mean of the row's counted numeric cells in the group (n/e, n/a and unverified cells skipped, never zero). Raw = 0.45 Depth + 0.25 Reach + 0.30 Standing, renormalised over the groups present. F = the arithmetic mean of the cohort's Raw scores (field_mean_method raw_mean). Shrunk = (k x Raw + K x F) / (k + K), k = the row's counted numeric cells, K = prior strength 4. Quotient = 100 x Shrunk / 5 at full precision; only the displayed value is rounded, halves up. An empty cohort has no quotient.",
  "reweightings": [
   {
    "name": "Equal thirds",
    "weights": {
     "Depth": 0.3333333333333333,
     "Reach": 0.3333333333333333,
     "Standing": 0.3333333333333333
    }
   },
   {
    "name": "Depth-led",
    "weights": {
     "Depth": 0.6,
     "Reach": 0.2,
     "Standing": 0.2
    }
   },
   {
    "name": "Reach-led",
    "weights": {
     "Depth": 0.3,
     "Reach": 0.45,
     "Standing": 0.25
    }
   },
   {
    "name": "Standing-led",
    "weights": {
     "Depth": 0.3,
     "Reach": 0.2,
     "Standing": 0.5
    }
   }
  ],
  "why_not_a_sum": "A sum counts an unscored dimension as zero and so penalises a thin public record rather than a weak service; a plain average of scored cells rewards missing evidence. Group means plus shrinkage limit both effects but do not remove the reward for omission: a row that leaves a documented weakness unrecorded can score higher than one that discloses it. The deletion test and the missing-evidence scenarios report that effect row by row. The quotient is conditional on the evidence observed.",
  "limits": "The index summarises publicly documented measurement capability within a cohort, not measurement quality, product performance, data accuracy or business quality, and inherits the single-scorer limitation. Bands are reported because exact positions depend on the weights. Rows are never compared across cohorts, and a network never receives a pooled score.",
  "bands_rule": "Bands come from the unrounded offset 20 x (Shrunk - F): Most documented at +12 or more, Well documented at +1 or more, Partly documented at -8 or more, Least documented below; equal values share a band. These are the Contextual Quotient's offsets under the names the attention and location papers gave them, which say what the bands measure: how much measurement capability is documented in public.",
  "thin_rule": "A row is band-eligible with k >= thin_record_min and at least one counted cell in every group that applies under the cohort's mask; a row missing a whole applicable group is shown with that group named and no band. Count the eligible rows first: a cohort with fewer than cohort_min_banded of them shows quotients and coverage without bands.",
  "shrink_rationale": "Prior strength K = 4 is an author-defined heuristic, not a statistical estimator: the dimensions are different capabilities, not repeated measurements of one quantity. The build reruns the index with K = 0 (Raw), 2 and 8 and reports how far positions and bands move.",
  "missing_rule": "Missing evidence is not neutral. Deletion: each counted cell turned into n/e in turn, recomputing the row's k and eligibility; quotient change, admission loss and eligibility loss are reported separately. Scenarios (not bounds): row by row, with F and every other row frozen, all of the row's applicable n/e cells filled with missing_scenario_values[0], then [1]; a banded row whose band changes under either is flagged band sensitive to missing evidence. Fills of missing_envelope_values are reported separately as a hypothetical full-scale envelope, never used for the flag. Imputed cells never create admission or band eligibility: the row keeps its observed k and eligibility.",
  "reweighting_rule": "Each reweighting recomputes Raw and F across the fixed cohort."
 },
 "scorer": {
  "independent_second_scorer": false,
  "disclosure": "One scoring pass by the author's research process, applied after the verifier records were read; facts were adversarially verified, scores were not independently re-scored. No inter-rater statistic exists."
 }
}
