Retail / Commerce Media Measurement.
The commerce execution layer: the IAB/MRC measurement guidelines, the in-store standards and their impression ladder, closed-loop attribution and its limits, incrementality, clean-room collaboration, and event signals — and what all of it means when agents do the optimizing.
Retail media measurement is the discipline of counting, attributing, and causally testing commerce media — onsite, offsite, and in-store. Five governing documents across IAB, MRC, and IAB Europe now define it, from the January 2024 IAB/MRC guidelines to the November 2025 incrementality rules. This page is the vendor-neutral synthesis.
Last validated July 12, 2026 · Maintained by Evgeny Popov
Retail media is not just an ad channel. It is a measurement negotiation between media, trade, commerce, and finance.
The negotiation
Three budget owners — media, trade, commerce — and a finance team read the same retail media report and see different proof. The standards on this page exist so all four can at least argue about the same numbers.
Fast read
- What it is
- A deep-dive guide to retail and commerce media measurement standards — the IAB/MRC measurement guidelines, the IAB and IAB Europe in-store standards, and the incrementality and data-collaboration guidance that decide what a retail media number actually means.
- What it covers
- The five governing documents: the IAB/MRC Retail Media Measurement Guidelines (January 2024), IAB Europe’s Commerce Media Measurement Standards V2 (final January 2026), the In-Store Retail Media definitions and measurement standards (December 2024) with their impression ladder — Ad Play, Gross Impression, Opportunity To See, LTS — and five store zones, the in-store maturity framework (December 2025), and the Guidelines for Incremental Measurement in Commerce Media (November 2025). Plus closed-loop attribution and its limits, clean-room-mediated measurement, and standardized event and conversion signals.
- What it is not
- Not a seventh standards pillar — a deep dive inside the six-layer system. And not a compliance scorecard: the guidelines themselves acknowledge that the majority of retail media organizations may not reflect viewability in attribution today.
- Why it matters
- Retail media sits where media, trade, commerce, and finance budgets meet — and every network measures itself. The standards on this page are the shared vocabulary that makes the numbers comparable, auditable, and worth optimizing against.
- Best for
- Brand, agency, retail media network, commerce platform, AdTech, and measurement leaders building or evaluating retail and commerce media measurement — especially where agents will do the optimizing.
- Best next read
- Measurement & Media Quality, DOOH & Place-Based Media, and the Retail Media Network ecosystem surface.
Where retail media fits in the standards stack.
The deep dives show where those standards become real in specific operating environments — and retail media’s operating environment is a site, a store, and a P&L at once. Its delivery and viewability metrics come from the measurement layer (begin-to-render impressions, MRC viewability thresholds, IVT filtration); its audience methods map to MRC digital audience standards; its consent metadata rides the privacy layer (GPP); its offsite execution runs on the core transaction rails. The January 2024 IAB/MRC guidelines function largely as a mapping of existing MRC standards onto retail media — which is precisely their value: retail media did not get a parallel measurement universe, it got connected to the existing one.
The standards concordance: five documents, one reference.
Retail media measurement is not governed by one standard. It is governed by five documents across two bodies — IAB (with MRC) and IAB Europe — published between January 2024 and January 2026, each deciding a different piece of what a number means. No standards body has unified them; this table does. Each row links its primary document. Statuses verified against the official pages, July 2026.
| Document | Body & status | Scope | What it governs |
|---|---|---|---|
| IAB/MRC Retail Media Measurement Guidelines | IAB with MRC · Final, January 2024 | Onsite, offsite, and in-store (digital place-based) retail media | The foundation: begin-to-render ad impressions, MRC viewability thresholds (required for attribution of outcomes), IVT filtration, audience methods, the attribution-vs-incrementality distinction, and the optional Chapter 5.3 outcome metrics (ROAS, New-to-Brand, Halo). Maps retail media onto existing MRC standards — and onto MRC audit practice. |
| In-Store Retail Media: Definitions and Measurement Standards | IAB × IAB Europe (MRC a collaborator) · Final, December 3, 2024 | In-store — digital screens, audio, and connected shopping formats | The store’s impression definitions: five store zones and the four-term impression ladder — Ad Play, Gross Impression, Opportunity To See, LTS — plus traffic and exposure metrics and store-level sales-lift measurement with purchase-cycle timeframes. |
| Guidelines for Incremental Measurement in Commerce Media | IAB × IAB Europe (Commerce Board + Incrementality Task Force) · Final, November 3, 2025 | Incrementality across commerce media — onsite, offsite, in-store | The causal layer: four methodology families — experiments, model-based counterfactuals, econometric models, hybrid proxies — grounded in three principles: credible counterfactuals, control of bias, separation of signal from noise. |
| A Viable Framework for Maturing In-Store Media Measurement | IAB · Published December 9, 2025 | In-store — maturity roadmap, including emerging formats (QR-enabled displays, digital endcaps, smart screens, addressable audio) | The adoption path: a phased progression from a measurement baseline retailers and vendors can adopt today toward comparable cross-network in-store measurement. A framework, not an IAB/MRC standard. |
| IAB Europe Commerce (incl. Retail) Media Measurement Standards V2 | IAB Europe · Final, January 22, 2026 (comment closed November 2025; V1-or-V2 grace period through July 31, 2026) | European commerce media — onsite, offsite, in-store; quick commerce added | The European reporting rules: a 30-day default attribution lookback, standardized gross and net sales definitions, a formalized incrementality definition with approved methodologies, five generic New-to-Brand / New-to-Category timeframes, and flexi ad sizes. |
Reading the ledger by topic
The topics buyers actually ask about each live in a specific document — and sometimes in two.
- Attribution windows: the IAB/MRC guidelines define attribution and require disclosure of window logic; IAB Europe V2 sets the 30-day default lookback for European reporting.
- Viewability: the IAB/MRC guidelines carry the MRC thresholds and the requirement that viewable impressions underpin outcome attribution; the in-store analogue is Opportunity To See in the December 2024 standards.
- Impression definitions: begin-to-render (IAB/MRC) online; the Ad Play → Gross Impression → OTS → LTS ladder (in-store standards) inside the store — four terms, never summable as one.
- Incrementality: defined in the IAB/MRC guidelines, given its own methodology document in November 2025, and formalized for the European track in V2.
- Audit requirements: the IAB/MRC guidelines map retail media onto MRC standards and audit practice; the maturity framework stages what in-store measurement can credibly adopt today.
Why retail media measurement is different.
Every chapter in the guidelines exists because a generic media assumption broke at the shelf — digital or physical. Eight realities define the channel.
- 01
The seller grades its own homework
In retail media the media owner, the measurement provider, and the owner of the sales data are often the same company. That is the structural reason the January 2024 IAB/MRC guidelines exist: shared definitions, methodology disclosure, and independent audit are the counterweight to self-measurement.
- 02
The currency is the sale
Display trades on impressions; retail media is bought and defended on outcomes — ROAS, units, new buyers. The guidelines still require the media basics underneath: begin-to-render ad impressions, MRC viewability, IVT filtration. Outcome metrics without those basics are unanchored.
- 03
Attribution is not incrementality
The guidelines define attribution as “assigning credit for consumer actions, like sales or visits, to specific marketing efforts” and incrementality as “the potential causal impact of marketing.” Two chapters, two different questions — and most reporting answers only the first.
- 04
Viewability arrived late
Per the guidelines, MRC requires viewable impressions for attribution of outcomes to ad exposures — and the same document acknowledges that the majority of retail media organizations may not reflect viewability in attribution today. The gap between requirement and practice is measurable work.
- 05
Every network is a ruler
More than 70 companies participated in the working group precisely because every network reported differently. Even now, the Chapter 5.3 metric definitions — ROAS, New-to-Brand, Halo — are explicitly not mandatory for reporting. Comparing networks means comparing definitions first.
- 06
The store became media
In-store screens, audio, and connected shopping formats received their own standards in December 2024 (IAB and IAB Europe), with five store zones and a four-term impression ladder. The physical store now carries a defined impression vocabulary — and it is not the web’s.
- 07
Trade money meets media money
Retail media budgets arrive from media, trade, shopper marketing, and commerce lines — each with its own finance scrutiny and its own definition of proof. Measurement is the table where those owners negotiate what counts, which is why definitions carry commercial weight here.
- 08
Offsite stretches the loop
The working-group scope covers ads onsite, offsite, online, and in-store. The moment the ad runs off the retailer’s property, exposure and transaction live in different companies — and measurement becomes a data-collaboration problem, not a database query.
Core measurement dimensions.
The January 2024 IAB/MRC guidelines organize retail media measurement into chapters — data collection and quality control, audience, ad delivery and viewability, incrementality, reporting and transparency, and in-store digital place-based environments. Cutting across them, ten dimensions carry the practical weight. Each answers one question; each has a watch-out that decides whether the answer means anything.
| Dimension | What it answers | Watch-out |
|---|---|---|
| Ad impressions | Was the ad delivered? | Counted on begin-to-render per the guidelines. Served is not seen, and rendered is not noticed. |
| Viewable impressions | Did it have a chance to be seen? | MRC thresholds: 50% of pixels for 1 continuous second (display) or 2 seconds (video). Required for attribution of outcomes — and most networks do not yet apply it there. |
| Audience | Who, deduplicated, was reached? | Deterministic, probabilistic, or hybrid methods per MRC digital audience standards — ask which method, and how cross-device is handled. |
| Invalid traffic | Are the counts clean? | General and sophisticated IVT detection and filtration per MRC guidance. Unfiltered counts inflate every metric downstream of them. |
| Attribution | Which exposure gets credit for the sale? | Defined as assigning credit — a rule, not a proof. Windows, logic, and lookbacks vary by network and silently change the number. |
| Incrementality | Did the marketing cause sales that would not have happened anyway? | Defined as the potential causal impact of marketing. RCTs, synthetic controls, matched-market tests, and ML models differ sharply in evidential strength. |
| ROAS | How much attributed revenue per ad dollar? | An optional Chapter 5.3 definition, explicitly not mandatory — and attributed ROAS rewards harvesting demand that already existed. |
| New-to-Brand / New-to-Category | Is this buyer actually new? | The lookback window defines “new.” The in-store standards specify purchase-cycle timeframes — align definitions before comparing rates. |
| Halo / assisted outcomes | Did the rest of the brand benefit? | Optional definitions. The in-store standards scope Halo as same brand, same category — broader claims need broader evidence. |
| In-store exposure | What counts as an impression inside a store? | Four distinct official terms — Ad Play, Gross Impression, Opportunity To See, LTS — that are not interchangeable and must not be summed as one number. |
The optional tier
The guidelines’ Chapter 5.3 metric definitions — ROAS, New-to-Brand, New-to-Category, Halo and assisted outcomes, share of new buyers — are explicitly not mandatory for reporting. They are shared definitions for networks that choose to report them, which is exactly why asking for the definition is never rude.
How each network actually answers these dimensions — attribution model, windows, sales-metric definition, clean room, incrementality offering, and MRC accreditation status — varies enough that cross-network comparison without normalization is fiction. The maintained six-network matrix, sourced per cell, covers Amazon, Walmart Connect, Criteo, Instacart, Roundel, and Albertsons Media Collective.
In-store media measurement.
In-Store Retail Media: Definitions and Measurement Standards — published jointly by IAB and IAB Europe, final December 2024, with MRC acknowledged as a collaborator and MRC guidelines used as the basis — gave the physical store a defined measurement vocabulary for the first time. The scope, for now, is digital screens, audio, and connected shopping formats. The core of the document is an impression ladder: four terms making four progressively stronger claims about exposure, none of them interchangeable.
| Official term | What it counts | What it does not claim |
|---|---|---|
| Ad Play | The number of times an ad is displayed or rendered on a particular format — a playout count from the device or player. | That anyone was present. It counts plays, not people. |
| Gross Impression | Individuals present in the defined Display Exposure Zone while the display is functional — with ad impressions derived as audience × ad play. | Viewing. Presence in the zone is not attention to the screen. |
| Opportunity To See (OTS) | A single opportunity to view the advert — the viewable-impression analogue. The standards position it as the best available proxy for a viewable ad impression for in-store digital activations. | That the person actually looked — opportunity is a qualified chance, not a gaze. |
| LTS Impressions (Likelihood-To-See) | A further refinement of viewable impressions, adjusted for the likelihood individuals noticed the content — determined using sensor and analytic technology. | Sensor-free certainty. The standards attach a privacy caution to the sensing layer itself. |
- Five store zones: Zone 1 — Exterior to Store; Zone 2 — Entrance and Out of Category; Zone 3 — Check out; Zone 4 — In Aisles; Zone 5 — Other & Connected Store.
- Media vocabulary: Ad Loop Duration, Ad Segment, Ad Unit, Ad Unit Length, and Loop — the scheduling grammar of in-store screens and audio.
- Traffic vocabulary: In-Store Traffic, Dwell Time, and Display and Audio Exposure Zones — the denominators behind Gross Impressions.
- Sales vocabulary: Sales Lift, Brand Sales Lift (Halo — same brand, same category), Incremental Sales measured test versus control, New to Brand, and New to Category — with purchase-cycle timeframes defined for what “new” means.
- Reporting baseline: the standards state that today RMNs should report on Ad Play and Gross Impressions, with Opportunity To See as the forward proxy for a viewable in-store impression on digital activations.
- Scope: digital screens, audio, and connected shopping formats — the standards defer other in-store formats to future iterations.
Terminology discipline
The official impression terms are Ad Play, Gross Impression, Opportunity To See, and LTS Impressions. Looser industry phrasings — “verified impressions” among them — are not defined terms in the December 2024 standards. A newer IAB framework, “A Viable Framework for Maturing In-Store Media Measurement” (published December 9, 2025), describes a phased path from a baseline stores can adopt today toward comparable in-store measurement — a framework, not an IAB/MRC standard, so adopt vocabulary from the standards text, not from coverage of the framework.
Closed-loop measurement and its limits.
The closed loop is retail media’s founding claim: the retailer observes both the exposure and the transaction, so it can connect them without anyone else in the room. The claim is real — and it has edges the standards now name out loud.
- The claim
What the loop actually shows
Exposure and transaction in one dataset, joined through retailer identity — loyalty, login, checkout. That supports attribution at a fidelity most channels cannot match. But what the join shows is a join: credit assignment under a chosen rule, not a demonstration of cause.
- The edges
Where the loop ends
Identity match rates, offsite exposures, and in-store purchases without loyalty all fall outside the loop; window choices quietly change the number; and the guidelines note MRC requires viewable impressions for attribution of outcomes — a condition most closed loops do not yet enforce.
- The upgrade
The causal toolkit
The guidelines define incrementality as the potential causal impact of marketing and catalogue methods — randomized controlled trials, synthetic controls, matched-market tests, machine-learning models. The Guidelines for Incremental Measurement in Commerce Media (final November 3, 2025) now formalize the methodology rules — decoded in the next section — and IAB Europe’s Commerce Media Measurement Standards V2 (final January 22, 2026) formalize an incrementality definition for the European track.
The distinction
A closed loop proves the loop closed. Whether the ad caused the sale is a separate question with its own methods — and its own chapter in the guidelines.
Incrementality: the November 2025 guidelines.
The Guidelines for Incremental Measurement in Commerce Media — IAB and IAB Europe, developed by the IAB Commerce Board and Incrementality Task Force, final November 3, 2025 — are the newest canonical document in the corpus, and the first to govern the causal question on its own. They name four methodology families and ground all of them in three principles: credible counterfactuals, control of bias, and separation of signal from noise. The table gives each family and how it builds its counterfactual in practice.
| Methodology family | The counterfactual it builds | Watch-out |
|---|---|---|
| Experiments | A designed control: exposure withheld from a comparable group, so the no-ad world is observed rather than modeled. | The strongest evidence and the most operationally expensive. Holdout construction and contamination decide whether the read is real. |
| Model-based counterfactuals | A modeled no-ad world: synthetic controls or machine-learning predictions standing in for a holdout. | Only as good as its bias controls — the counterfactual is estimated, not observed. |
| Econometric models | Aggregate inference across markets and time — the marketing-mix-modeling family. | Answers portfolio-level questions; resolution at the SKU or single-campaign level is limited. |
| Hybrid proxies | Attribution-based signals adjusted toward causality where a true test is not feasible. | A proxy, not a proof — the three principles are the checklist for how far to trust it. |
Why methodology choice is the whole game
A March 2026 analysis — Retail Media iROAS Demystified, by agency Ovative Group and RMN Albertsons Media Collective with professors from Northwestern’s Kellogg School of Management; published alongside Albertsons’ launch of its own onsite incrementality product, so treat the numbers as vendor-sourced — found within-campaign iROAS varying 6.5x on average (median 2.5x) purely by matching-methodology choices within one network, with 83% of campaigns flipping from positive to negative. And a February 2026 Skai/Stratably survey (n=166, also vendor-sourced) found 75% of advertisers cite incrementality as their top measurement challenge while only 15% strongly trust their measurement. A causal claim that does not disclose its methodology family, its counterfactual construction, and its bias controls — the guidelines’ own three principles — is a number wearing a lab coat.
- The IAB incrementality guidelines, decoded — the four method families graded, and what the document requires vs. recommends vs. stays silent on
- iROAS is not a number, it’s a negotiation — the buyer’s eight-question audit protocol for network-reported incrementality
- Research & Measurement Science — the evidence methods behind the guidelines
Clean rooms and retail media.
Everything the closed loop cannot see — offsite exposure, cross-retailer behavior, brand-side CRM — lives in someone else’s dataset. Clean rooms are how retail media measurement crosses company lines without shipping raw data across them, and the standards under them are further along than most teams assume.
- Why here
Where the loop needs a bridge
Offsite campaigns, brand–retailer joins, and cross-party reads all require analysis across datasets no single company should ship to another in the clear. IAB’s guidance on retail media advanced measurement and data collaboration (October 2024) places clean rooms at the center of that work, alongside experimental and counterfactual methods.
- The stack
The standards under the room
IAB Tech Lab’s portfolio carries the interoperability layer: Data Clean Room Guidance v1.0 (the official page describes a July 2024 v1.0 release), PAIR v1.1 (finalized July 2025) for privacy-centric audience activation, and ADMaP v1.0 (finalized February 2025) for attribution data matching. Protocols and guidance — not magic.
- The honest claim
Output policy is the control
A clean room changes where analysis happens — not automatically what leaves it. The aggregates, thresholds, and recipients a room permits are the real privacy posture, and nothing in this chain is privacy-safe by default. Evaluate configuration and output policy, not the category label.
Event and outcome signals.
Retail media’s outcome metrics ride on event data — purchases, carts, leads — moving from commerce systems to advertising platforms. Until recently that movement ran entirely through per-platform conversion APIs, each with its own structure. In 2026 it acquired a shared standard.
- Released
ECAPI 1.0 is real
IAB Tech Lab’s Event & Conversion API went to public comment on January 20, 2026, closed comment on February 20, 2026, and was then finalized — Tech Lab’s blog describes ECAPI 1.0 as available for implementation, with the canonical specification on GitHub. It standardizes server-to-server marketing events from advertiser systems to platforms and partners.
- The schema
A commerce-shaped event language
A named event taxonomy — purchase, add_to_cart, begin_checkout, page_view, ad_impression, generate_lead, sign_up, subscribe, and more — with value and currency, transaction identifiers, user-match identifiers, and extension objects for platform-specific fields. One shared structure beneath many per-platform conversion APIs — a standardization layer, not a replacement for them.
- The constraint
Consent travels with the event
GPP consent metadata is built into the schema: gpp_string and gpp_sid, plus an mmt_only flag marking events for measurement only — excluded from ads-delivery optimization. The spec leaves legal compliance to implementers: a standardized schema is not a legal opinion, and platform adoption should be validated per partner.
Agentic implications.
An agent optimizing retail media is optimizing against numbers whose definitions vary by network, whose causality is usually unproven, and whose consent terms ride along as metadata. The standards on this page are the constraint set that keeps that optimization from becoming confident nonsense. Six implications follow.
- 01
Attributed ROAS is a trap objective
An agent optimizing attributed ROAS will migrate spend toward shoppers who were buying anyway — the metric rewards demand harvesting. Encode incrementality reads, where they exist, as the senior signal, and treat attributed outcomes as routing hints, not ground truth.
- 02
Every network is a different ruler
Cross-network optimization requires normalization first: attribution windows, New-to-Brand lookbacks, and optional metric definitions vary by network. An agent comparing unaligned ROAS figures is comparing definitions, not performance.
- 03
Impression terms are not summable
Ad Play, Gross Impression, Opportunity To See, and LTS are four different claims about in-store exposure. An agent that aggregates them as “impressions” miscounts by design — carry the term with the number, always.
- 04
Windows are parameters, not facts
A 14-day click window and a 30-day view window produce different truths from identical campaigns. Agents must persist window, method, and metric-version metadata with every outcome they ingest — or they will optimize an artifact of configuration.
- 05
Encode the proof ladder
Played, rendered, viewable, attributed, incremental — each rung answers a different question with different evidence. Agentic workflows should gate spend decisions on the rung the claim actually reached, not the rung the report implies.
- 06
Respect measurement-only signals
Standardized event schemas carry consent metadata — GPP strings and a measurement-only flag, where supported. An agent that feeds measurement-only events into bidding optimization is not being clever; it is violating the signal’s declared terms.
And when the shopper itself is an agent — no click, no viewable impression, no session — most of this stack's assumptions come up for review. Which metrics break, which survive, and what has to be built new: Measurement for Agentic Commerce.
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 the media basics under every outcome number — begin-to-render impressions, MRC viewability, IVT filtration — and the definition behind every metric: which attribution window, which New-to-Brand lookback, which incrementality method. Fund periodic causal reads on the spend that closed-loop ROAS calls best-performing; that is where the surprises live.
Maintain a per-network measurement matrix: impression definitions, attribution windows, optional Chapter 5.3 metrics in use, in-store terms reported, and incrementality methods offered. Normalize before you compare — cross-network league tables built on unaligned definitions are fiction with decimals.
Map reporting to the January 2024 guidelines chapter by chapter — data collection and QC, audience, delivery and viewability, incrementality, reporting transparency — and adopt the December 2024 in-store vocabulary as written: Ad Play and Gross Impressions today, Opportunity To See as the forward proxy. Disclosed methodology is a sales asset; independent audit raises the trust ceiling.
Build begin-to-render impression counting, MRC-threshold viewability measurement, and IVT filtration into onsite serving — and carry definition metadata (window, method, metric version) with every reported number so downstream systems can normalize instead of guess.
Treat each network’s metrics as schema-bearing inputs, not comparable scalars. Encode attribution and incrementality as separate signal classes, honor measurement-only flags on event data where supported, and optimize against causal reads where they exist — attributed ROAS alone steers agents toward demand harvesting.
Anchor retail media products in the published frameworks — the IAB/MRC guidelines, the in-store standards, the incrementality guidance — and publish methodology, coverage, and error honestly. The in-store LTS layer carries a privacy caution in the standards themselves: sensor-based adjustment needs governance, not just calibration.
Retail media’s offsite and cross-party measurement increasingly runs through clean rooms. Support the IAB Tech Lab guidance and protocol set where applicable — Data Clean Room Guidance, PAIR for activation, ADMaP for attribution — and make output policy a first-class, auditable control: what a room permits to leave is its actual privacy posture.
No Fluff POV.
Retail media has the raw material to be the most accountable channel in advertising — the transaction is right there. It is also the easiest channel in which to dress correlation as causation, because the party selling the media owns the data that grades it. The standards exist now; the work is declining numbers that ignore them.
- Say which question a number answers: attribution assigns credit, incrementality estimates causal impact — the guidelines define both, and they are not interchangeable.
- Demand the media basics beneath every outcome: begin-to-render impressions, MRC viewability, IVT filtration — viewable impressions are required for attribution of outcomes, and most networks are not there yet.
- Use the official in-store vocabulary — Ad Play, Gross Impression, Opportunity To See, LTS — and label every in-store number with the term it actually is; looser industry phrasings are not defined in the standards.
- Treat ROAS, New-to-Brand, and Halo as optional, definition-bearing metrics — the guidelines say so explicitly — and never compare them across networks without aligning definitions first.
- Plan offsite measurement as data collaboration: clean rooms, standardized event signals, consent metadata — the closed loop ends at the retailer’s data perimeter.
- Watch the moving parts: IAB Europe’s Commerce Media Measurement Standards V2 is final (January 22, 2026) with a V1-or-V2 grace period through July 31, 2026, and the in-store maturity framework (December 2025) stages what stores can adopt today — validate versions before you build.
The point
Retail media is not just an ad channel. It is a measurement negotiation between media, trade, commerce, and finance — and the standards are the negotiating table.
Frequently asked questions.
The definitional questions behind the queries that reach this page — each answered from the standards themselves, with the primary documents linked in the concordance above.
What do publishers need from commerce media measurement?
For the publisher side — retailers and commerce media networks selling inventory — the need is audited, comparable currency: ad impressions counted on begin-to-render with MRC viewability and IVT filtration underneath, methodology disclosed rather than asserted, and an independent-audit path per the IAB/MRC guidelines. Add standardized metric definitions (Chapter 5.3) so ROAS and New-to-Brand are tradable across buyers, and the November 2025 incrementality guidelines to defend outcomes beyond attributed revenue. Self-measurement without those counterweights is the structural problem the standards exist to fix.
What must advertisers look for in commerce media measurement?
Four questions before comparing networks: Is viewability actually applied to attribution (the guidelines require it; most networks do not yet do it)? What attribution window and logic produced the number — and would a different lookback quietly change it? Is attribution separated from incrementality, or is harvested demand being sold as caused demand? And do the network’s metric definitions match Chapter 5.3 — because those definitions are explicitly optional, two networks’ ROAS figures are only comparable after the definitions are.
How is commerce media improving transparency in measurement and reporting?
Through shared definitions rather than open data: a 70+-company working group produced the IAB/MRC guidelines precisely because every network reported differently. Transparency now arrives as required methodology disclosure, an independent audit path, an official in-store impression ladder (Ad Play, Gross Impression, OTS, LTS) replacing vendor-invented terms, and incrementality methodology families named in the November 2025 guidelines. The gap that remains is adoption — the guidelines themselves acknowledge most networks do not yet reflect viewability in attribution.
What are the IAB/MRC Retail Media Measurement Guidelines?
The foundational retail media measurement standard: published January 2024 by IAB with the Media Rating Council after public comment. Six chapters cover data collection and quality control, audience, ad delivery and viewability, incrementality, reporting and transparency, and in-store digital place-based measurement — largely mapping existing MRC standards onto retail media.
What is iROAS (incremental ROAS)?
iROAS is return on ad spend computed on incremental sales only — revenue that would not have occurred without the ad — rather than on all attributed revenue. It is heavily methodology-dependent: a vendor-sourced March 2026 analysis by Ovative Group and Albertsons Media Collective, with Northwestern Kellogg professors, found within-campaign iROAS varying 6.5x on average purely by matching-methodology choices within one network, which is why the November 2025 IAB incrementality guidelines matter.
What is incrementality in retail media?
The IAB/MRC guidelines define incrementality as the potential causal impact of marketing — sales that would not have happened anyway. It is distinct from attribution, which assigns credit under a chosen rule. The Guidelines for Incremental Measurement in Commerce Media (final November 3, 2025) name four methodology families: experiments, model-based counterfactuals, econometric models, and hybrid proxies.
What is closed-loop attribution, and where does it break?
Closed-loop attribution joins ad exposure and transaction inside one retailer’s identity data — loyalty, login, checkout. It breaks at the loop’s edges: offsite exposures, unmatched shoppers, and in-store purchases without loyalty fall outside it — and it assigns credit under a chosen rule; it does not demonstrate cause.
What counts as an in-store retail media impression?
Four official terms from the December 2024 in-store standards, none interchangeable: Ad Play (a playout count), Gross Impression (people present in the exposure zone), Opportunity To See (the viewable-impression analogue), and LTS (likelihood-to-see, sensor-adjusted). The standards say networks should report Ad Play and Gross Impressions today, with OTS as the forward proxy.
What role do clean rooms play in retail media measurement?
Everything the closed loop cannot see — offsite exposure, brand-side CRM, cross-retailer behavior — lives in someone else’s dataset. Clean rooms let those datasets be analyzed together without shipping raw data across company lines, under IAB Tech Lab’s Data Clean Room Guidance, PAIR, and ADMaP. Output policy, not the category label, is the real privacy control.
Primary sources to validate.
Standards, guidance, and implementation references last validated: July 2026. Specifications, public-comment status, frameworks, APIs, and implementation guidance change. Validate against official documentation before implementation.
Primary sources to validate 11 sources
-
The final joint guidelines, dated January 2024 on the cover (no version number — do not attach one). Developed by the IAB Retail Media Measurement Working Group 'through a collaborative endeavor involving both the Interactive Advertising Bureau (IAB) and the Media Rating Council (MRC).' Six chapters: data collection/QC; audience measurement; ad delivery and viewability (begin-to-render ad impressions; MRC viewable thresholds — 50% of pixels for 1 continuous second display, 2 seconds video; viewable impressions required for attribution of outcomes); incrementality ('the potential causal impact of marketing'); reporting and transparency (ROAS, New-to-Brand, Halo et al. are optional metric definitions, explicitly 'not mandatory for reporting'); and in-store Digital Place-Based (DPB) environments measurement. The guidelines themselves acknowledge most retail media organizations do not yet incorporate viewability into attribution — do not overclaim adoption or compliance. Supports: Chapter structure and exact definitions (impressions, viewability, attribution, incrementality), MRC requirement that viewable impressions underpin outcome attribution, Optional (not mandatory) status of ROAS / New-to-Brand / Halo metric definitions.
-
The same January 2024 guidelines hosted in the MRC's own Standards directory on mediaratingcouncil.org — corroborates MRC's role as co-publisher/co-developer. Note MRC has not published a separate standalone 'MRC Retail Media Standard'; its retail media involvement is via this joint document. Supports: Evidence of MRC co-ownership of the guidelines, Citing MRC as joint publisher alongside IAB.
-
Companion 'Executive Playbook'/explainer published January 2024 alongside the final guidelines. The official PDF and explainer carry only 'January 2024' — do not assert a specific release day from secondary coverage. Supports: Practical companion to the guidelines, Dating the final release to January 2024.
- IAB and MRC Release Retail Media Measurement Guidelines (public comment announcement) ↗ Official press release
September 13, 2023 announcement of the draft guidelines for a 30-day public comment period (to October 13, 2023); final followed January 2024. Confirms stated coverage areas: audience measurement, in-store digital place-based measurement, ad delivery, viewability, incrementality, reporting, and transparency. Supports: Timeline: draft for public comment Sept 13, 2023; final January 2024, Stated coverage areas.
- In-Store Retail Media: Definitions and Measurement Standards (FINAL, December 2024 PDF) ↗ Official docs
Final joint IAB/IAB Europe in-store standards (public comment Sept–Nov 2024; MRC credited in acknowledgements, and MRC DOOH/DPB guidelines used as the basis — but the document is branded IAB/IAB Europe, not IAB/MRC). Defines five store zones (Exterior to Store; Entrance and Out of Category; Check out; In Aisles; Other & Connected Store); the EXACT impression terms 'Ad Play', 'Gross Impression', 'Opportunity To See' (viewable), and 'LTS Impressions' (Likelihood-To-See); media terms (Ad Loop Duration, Ad Segment, Ad Unit, Loop); traffic terms (In-Store Traffic, Dwell Time, Display/Audio Exposure Zones); and sales metrics (Sales Lift, Brand Sales Lift/Halo, Incremental Sales via test-vs-control, New to Brand, New to Category). 'Verified impression' does NOT appear anywhere in this document (full-text search: zero occurrences) — never present it as an official term here. Scope is digital screens, audio, and connected shopping formats only; other formats deferred to future iterations. Supports: Exact official in-store impression, media, traffic, and sales terms, Five store-zone taxonomy, Refuting 'verified impression' as a defined term in these standards.
- In-Store Retail Media: Definitions and Measurement Standards (guidelines landing page) ↗ Official standards page
IAB landing page for the final in-store standards, dated December 3, 2024, noting they build on the IAB/MRC Retail Media Measurement Guidelines. Full document behind a free-account wall; the PDF is directly fetchable. Supports: Canonical landing URL and December 3, 2024 publication date, Confirming final status.
-
IAB blog announcing the final in-store standards (December 3, 2024) after the public feedback period; summarizes the impression metrics (ad play, gross impression, opportunity to see, likelihood-to-see), media metrics, sales lift metrics, and adherence to the IAB/MRC Retail Media Measurement Guidelines. Supports: Official summary of the in-store standards' contents, Relationship to the IAB/MRC guidelines.
- IAB Europe and IAB Release First-Ever In-Store Retail Media Standards for Public Comment ↗ Official press release
September 2024 announcement of the draft in-store standards for public comment. Historical context for the timeline (draft Sept 2024 → comment Sept–Nov 2024 → final Dec 3, 2024); the September 2024 PDF is the draft, not the final. 'First-ever' framing is IAB's own. Supports: Timeline of the in-store standards.
- A Viable Framework for Maturing In-Store Media Measurement ↗ Official standards page
IAB framework published December 9, 2025 (full document gated behind a free account). Accessible official text describes a phased approach offering 'a standard measurement baseline that retailers and vendors can adopt today' for QR-enabled screens, digital endcaps, smart displays, in-store audio, and other addressable formats. CAUTION: trade coverage attributes a 'verified impressions' baseline and 'Three Ps' (Play, Presence, Pairing) to this framework, but that wording was NOT confirmable from accessible official IAB text — attribute it as 'reported' if used at all, and do not call this an IAB/MRC document or a 'standard'. Supports: Existence, title, publisher, and Dec 9, 2025 date of the newest in-store measurement framework, Context for where 'verified impressions' terminology likely originates (reported, not verified official text).
- Guidelines for Incremental Measurement in Commerce Media ↗ Official standards page
Published November 3, 2025 (final). Establishes incrementality measurement frameworks for commerce media — broader than retail media — covering four methodology families: experiments, model-based counterfactuals, econometric models, and hybrid proxies; grounded in 'credible counterfactuals, control of bias, and separation of signal from noise.' Supports: IAB's commerce-media-scoped incrementality guidance, Commerce media as a distinct 2025 IAB workstream.
- IAB Europe Commerce (incl. Retail) Media Measurement Standards V2 — FINAL (January 22, 2026) ↗ Official docs
IAB Europe's V2 of its commerce (including retail) media measurement standards — FINAL as of January 22, 2026, after a public comment period that closed in November 2025; a grace period allows V1-or-V2 compliance through July 31, 2026. Key changes: 30-day default attribution lookback, standardized gross/net sales definitions, a formalized incrementality definition with approved methodologies, five generic new-to-brand/new-to-category timeframes, quick-commerce metrics, and flexi ad sizes. The European workstream that widened scope from 'retail media' to 'commerce media' (V1: IAB Europe Retail Media Measurement Standards 2024, a separate document from the IAB US/MRC guidelines). Supports: European standards track and commerce-media V2 expansion, Final status (Jan 22, 2026), 30-day default lookback, grace period to July 31, 2026.
Platform capabilities and naming change quickly. Last validated: July 12, 2026. Check current documentation before implementation.Standards, guidance, and implementation references last validated: July 2026. Specifications, public-comment status, frameworks, APIs, and implementation guidance change. Validate against official documentation before implementation.
Building or evaluating retail media measurement?
The operating work is to wire shared definitions, viewability-anchored delivery metrics, incrementality designs, and clean-room-mediated outcome data into the buying path — before agents scale whatever each network happens to report.