Retail Media Attribution

Retail Media Attribution is the set of methods used to assign credit for—or measure the causal impact of—retail media advertising on purchase outcomes. It differs from general Marketing Attribution in one structural way: retail media networks (RMNs) possess deterministic first-party purchase data, enabling Closed-Loop Attribution — the direct matching of ad exposures to transaction records — without cookie or probabilistic inference. (IAB Europe, Commerce Media Measurement Standards V2.1, 2026)

Despite this structural advantage, retail media attribution is in a state of active measurement tension in 2026: only 15% of marketers describe their organisation as "very or extremely effective" at measuring retail media performance overall, and 75% cite Incrementality as their biggest unresolved measurement challenge. (Skai, 2026 State of Retail Media Measurement and Incrementality, 2026-02-04)

The attribution-incrementality distinction

IAB Europe and IAB US jointly define the core distinction: attribution assigns credit across touchpoints; Incrementality measures causal impact — the additional outcomes that would not have occurred without the ad. "Attribution alone was sufficient in the past" but "no single tool has all the answers — a combined approach using each tool's strengths is needed." (Think with Google, Modern Measurement Playbook, undated)

Andrew Covato (Growth by Science, 2026) frames it practically: "Everything we do is in service of estimating an 'i' something. Whether it's iROAS or incremental conversions or incremental whatever. That little 'i' means these are conversions you would not have gotten otherwise." (Kevel / Unlocking Retail Media, 2026-02-18)

Jason Wescott (Global Head of Commerce Solutions, WPP Media) states: "The overreliance on ROAS as the benchmark of value is over. Independent, transparent measurement is the baseline." (Skai, 2026 State of Retail Media Measurement, 2026-02-04)

IAB Europe V2.1 Standards (2026)

IAB Europe's Commerce (incl. Retail) Media Measurement Standards V2.1 — published January 2026, updated May 2026 — are the primary pan-European framework for retail media attribution. They replace V1 (April 2024). (IAB Europe, 2026-01-22)

IAB Europe ran a six-month V1→V2 transition grace period (January–July 2026); after July 2026, V1 certification is no longer available. (as-of 2026-07; [IAB Europe V2.1 PDF, 2026-05](https://iabeurope.eu/wp-content/uploads/IAB-Europes-Commerce-Media-Measurement-Standards-V2.1-May-2026.pdf))

Future editions are planned to extend to travel and finance media networks within 2026. (IAB Europe V2.1 PDF, 2026-05)

Attribution windows (V2.1)

Format	Default lookback	Notes
Endemic on-site / off-site retail media	30 days (post-view and post-click)	Brands must be offered flexible windows (e.g. 7 days for FMCG, 30 days for CE) on request
Quick Commerce (Uber Eats, DoorDash etc.)	7 days	Reflects immediate purchase nature

(IAB Europe V2.1 PDF, 2026-05)

Sale definitions and the gross/net gap

V2.1 distinguishes gross sales (recognised at payment/order confirmation, regardless of return status) from net sales (gross minus returns, cancellations, refunds). Anecdotal evidence cited in the standard suggests an average 20% delta between gross and net sales, meaning ROAS reported on gross sales materially overstates net performance. When a CMN does not specify gross or net, gross sale is assumed — search, social, and programmatic media also default to gross. (IAB Europe V2.1 PDF, 2026-05)

ROAS reporting requirements (V2.1)

IAB Europe V2.1 defines ROAS as outcomes attributable to a click (to a PDP or Add to Cart) or a viewable ad, typically based on a last-click or last-view model. CMNs must report click-based ROAS and view-based ROAS separately. (IAB Europe V2.1 PDF, 2026-05)

SKU-level and halo attribution (V2.1)

V2.1 mandates Same SKU (parent SKU) attribution for all sponsored product ads, including all size and colour variants of the clicked product. Halo attribution is defined as "Same Brand, Same Category" (category as defined by CMN catalogue); halo attribution is optional (not mandatory) for all ad formats. (IAB Europe V2.1 PDF, 2026-05)

Sales extrapolation disclosure

CMNs using sales extrapolation (estimating impact on non-identified users by projecting from identified users) must disclose the methodology used and which metrics the extrapolation is applied to. (IAB Europe V2.1 PDF, 2026-05)

Incrementality requirements (V2.1)

V2.1 defines incrementality as "the causal impact of marketing by identifying the additional business outcomes directly driven by a campaign or tactic, compared to what would have occurred in the absence of marketing activity." CMNs are not required to provide incrementality measurement to achieve certification — but if they do, they must use a specified method from the IAB/IAB Europe Guidelines for Incremental Measurement in Commerce Media (November 2025). (IAB Europe V2.1 PDF, 2026-05)

Measurement delivery infrastructure (V2.1)

V2.1 lists three measurement delivery platforms CMNs may offer: (1) CMN-provided reporting, (2) open data/API feeds analysed by the brand, (3) data clean rooms. CMNs must disclose which they support. Data clean rooms offer the most flexibility for incrementality methods but are expensive, have limited interoperability between platforms, and carry data leakage and IT security review costs. (IAB Europe V2.1 PDF, 2026-05)

Walled garden fragmentation and cross-retailer deduplication

Skai's 2026 State of Retail Media report (survey n=166 advertisers, conducted with Stratably) found advertisers work with an average of six retail media networks currently and expect that figure to grow to 11 by end-2026 (as-of 2026-02-12). Cross-retailer attribution is structurally blocked: no shared identity graph exists between walled gardens — Amazon Marketing Cloud, Walmart Connect, Instacart Data Hub, and Google Ads Data Hub all return aggregate outputs that cannot be trivially joined. (Skai, 2026-02-12)

Salim Bachatene (SVP Global Sales, NIQ): "What this survey really shows is that most retail media challenges are not media problems: they are measurement and integration problems." (Skai, 2026-02-12)

Enrico Babucci (Chief Strategy Officer, OmniShopper): European brands face additional measurement friction because of retailer-by-retailer differences in data quality and uneven standards adoption across markets. (Skai, 2026-02-12)

Andrew Covato (2026): geo-exclusion is the only viable cross-platform incrementality method when no user-level tie exists across walled gardens — "You pick cities or states or DMAs where you're not going to show ads." (Kevel / Unlocking Retail Media, 2026-02-18)

Industry confidence and measurement maturity (2026)

Measurement maturity indicator	Finding	Source
Rate brands as very/extremely effective at measuring retail media	15%	Skai, n=166, 2026-02-04
Cite incrementality as biggest measurement challenge	75%	Skai, 2026-02-04
Cite cross-channel measurement as challenge	59%	Skai, 2026-02-04
Cite retailer data access / attribution as challenges	36% each	Skai, 2026-02-04
Measure incrementality at only a basic level	50%	Skai, 2026-02-04
Not measuring incrementality at all	14%	Skai (registry), 2026
Primary barrier: limited analytics/data science resources	56%	Skai, 2026-02-04
Brands measuring incrementality — reduced wasted spend	54%	Skai (registry), 2026
Brands measuring incrementality — increased new customer acquisition	49%	Skai (registry), 2026

(as-of 2026-02-04)

The biggest gaps between achieved and hoped-for outcomes from incrementality measurement are in profit margin improvement (24% achieved vs 68% hoped) and customer lifetime value (22% achieved vs 42% hoped), indicating most brands use incrementality as a short-term efficiency tool rather than a strategic lever. (Skai, 2026-02-04)

What practitioners report

Albertsons Media Collective (Liz Roche, VP Media and Measurement): Closed-loop attribution is a design mandate — "we design with the end in mind, which means we're closing the loop at the end of this thing." Albertsons launched a matched market incrementality framework in January 2026 using nearly 60 store-level variables to compare test vs. control stores, and validated incremental sales lift for a Mondelēz/Sargento Cheese Bakes in-store campaign. Roche cites clean rooms and API access as the top technology requests from brand clients. Brand clients are moving away from campaign-level ROAS toward LTV and longitudinal metrics. (OmniTalk/Albertsons press release, 2025-03-27; 2026-01-06)

Sam's Club (Harvey Ma, VP & GM Retail Media): Sam's rebuilt its search infrastructure to break the highest-bidder auction model, running ads through a relevancy layer rather than pure revenue maximisation — "from a contextual and personalisation standpoint, members see what they want to see." Ma describes ROAS as a "drug" and is actively moving supplier KPIs toward "new member conversion" and "new basket growth." Sam's Club Scan & Go display ads achieved 71% incremental reach (as-of 2025-03-27). Ma on cross-format standardisation: "How do you standardise ROAS on search from a fitness club and a warehouse club? You can't." (OmniTalk, 2025-03-27)

Andrew Covato (Growth by Science, independent measurement consultant): Platform-reported attribution is inherently unreliable because "ad platforms are not in the business of optimising advertiser outcomes — they're in the business of optimising their own revenue." First-party RMN lift products should be treated as "a sales tool versus a measurement tool." MMM does not require user-level data — RMNs overengineer the data problem; aggregated time-series by geography is sufficient. Measurement capability gaps are causing deals to fail: "If I can't measure it the way that I want to measure it, I'm not going to spend, especially if it's a relatively new platform." (Kevel / Unlocking Retail Media, 2026-02-18)

In-store attribution: the "digital envy" problem

Internet Retailing (June 2026) reports that ISM/Catalyst research found the same in-store retail media activation is evaluated against four entirely different success definitions simultaneously: Media Mix Scorecard (agencies), Retail Sales Scorecard (merchants), Media Revenue Scorecard (RMNs), and Efficiency Scorecard (brands/shopper marketing). When three of the four are satisfied, the misaligned stakeholder can stall the next campaign buy. One RMN reported that five separate brand campaigns were blocked in a single week in 2025 due to measurement-alignment requirements, despite budget and inventory being in place. (Internet Retailing / ISM + Catalyst Media Consulting, 2026-06-08)

ISM/Catalyst identify "digital envy" — the assumption that in-store media must produce closed-loop, one-to-one attribution comparable to ecommerce — as the root cause, arguing it forces inappropriate frameworks on physical retail. Collin Colburn (VP, Commerce and Retail Media, IAB) and Paul Brenner (SVP, Retail Media, ISM) have proposed a "Shopper Purchase Rate" (SPR) framework as a fit-for-purpose alternative: it captures dollars spent, units purchased, and shopper behaviour across four segments (loyal, occasional, lapsed, new), uses matched-market tests and pre/post analyses, and produces a single number mapping to multiple stakeholder scorecards. SPR entered a validation phase with retailers and CPGs in H2 2026. (Internet Retailing, 2026-06-08)

A joint IAB and Instacart white paper found that Media Mix Modeling (MMM) systematically undercounts the impact of in-store and commerce channels due to the more nuanced consumer behaviour in those environments. (Internet Retailing, 2026-06-08)

Contradictions

Advertiser trust levels: 6% vs 15% A claim circulating in search results ("Six Percent. That Is How Many Advertisers Trust the Numbers That Retail Media Sends Them," attributed to Bidstream/Amit Goel) implies a much lower trust figure than Skai's 15% finding. Skai's figure is "very or extremely effective at measuring retail media performance overall" (one's own effectiveness rating). The Bidstream figure implies direct trust in network-reported numbers. Different question framings may explain the gap, but the Bidstream page returned empty and the claim cannot be verified from a readable source. (Skai, 2026-02-04) VS (unverified: Bidstream/Amit Goel — page returned empty; treat as hearsay until confirmed)

Gross ROAS vs net ROAS — an embedded industry blind spot IAB Europe V2.1 explicitly acknowledges a ~20% average delta between gross and net sales, and notes all major channels (search, social, programmatic) default to gross. This means ROAS benchmarks routinely cited across the industry — including Skai's retail media 6.1× (existing ROAS vault page) — are gross-based and materially overstate net return. No source directly names and reconciles this gap; it is structurally embedded in current standards. (IAB Europe V2.1 PDF, 2026-05)

MMM role in retail media: structurally ill-suited vs indispensable IAB Europe published a white paper in April 2026 arguing MMM is structurally ill-suited to retail media and causes brands to undervalue the channel, urging closed-loop and incremental methods instead. (URL: iabeurope.eu — specific white paper URL not captured; sourced from search snippet only.) A joint IAB/Instacart white paper found MMM undercounts in-store/commerce channels. (Internet Retailing, 2026-06-08) VS Skai's fragmentation report and Andrew Covato (2026) include MMM as one of three components in a practical cross-retailer measurement triangulation stack — "periodic, diligent modelling can deliver a robust view without building a full internal lab." (Skai, 2026-02-04) These reflect unresolved debate on whether MMM has a role in retail media attribution.

Closed-loop attribution: growth driver vs in-store investment blocker Skai (2026): 52% of advertisers are reallocating display spend from open web DSPs to retail media DSPs — 57% cite closed-loop attribution as the primary driver. (Skai, 2026) VS ISM/Catalyst (June 2026): The same closed-loop attribution expectation — imported from digital and applied to in-store — is actively blocking campaign investment ("digital envy"), with five campaigns stalled in a single week at one RMN. (Internet Retailing, 2026-06-08) The same feature is simultaneously the reason brands are entering retail media, and the reason in-store retail media specifically is blocked from budget.

First-party RMN lift as measurement tool vs sales tool Albertsons Media Collective (Liz Roche, 2026-01-06) launched and promotes its matched market incrementality product as a credible causal measurement product for advertisers — "we are laser focused on closed loop everywhere." (OmniTalk, 2025-03-27) VS Andrew Covato (2026): "I would look at first-party lift as more of a sales tool versus a measurement tool... we understand that you may not look at this as purely objective because it's our own platform." (Kevel, 2026-02-18) Practitioner-insider vs. independent-consultant framing of the same capability.

ROAS departure pace: actively abandoning vs still using alongside incrementality Harvey Ma (Sam's Club) framed ROAS as a "drug" and has rebuilt search infrastructure to deprioritise revenue maximisation, actively moving supplier KPIs to new member conversion. (OmniTalk, 2025-03-27) VS Liz Roche (Albertsons): "we of course do ROAS campaign over campaign" — incrementality is described as a complement, not a replacement. (OmniTalk, 2025-03-27) Both practitioners are at the same conferences but at materially different points on the ROAS→incrementality transition.

Key terms

Term	Meaning
Closed-loop attribution	Directly connecting ad impressions to purchase transactions by matching customer IDs from ad exposures to purchase records, using deterministic first-party data
Gross ROAS	Revenue at point of order confirmation ÷ ad spend; does not subtract returns/refunds
Net ROAS	Revenue minus returns, cancellations, refunds ÷ ad spend; more conservative; ~20% lower than gross on average
Same SKU attribution	Credit only for the exact advertised product (parent SKU, including all variants)
Halo attribution	Credit for other products in the same brand and category purchased after ad exposure
Sales extrapolation	Estimating attributed sales for non-identified/unmatched users by projecting from identified users
Data clean room	Privacy-safe environment where brands and RMNs combine datasets for measurement without raw data exposure
Digital envy	The assumption that in-store media must deliver closed-loop, one-to-one attribution comparable to ecommerce — applied inappropriately to physical retail
Shopper Purchase Rate (SPR)	Proposed IAB/ISM metric combining dollars spent, units purchased, and shopper segment behaviour; in validation H2 2026
iROAS	Incremental revenue ÷ ad spend; requires experiment (test vs. control); see iROAS
CMN	Commerce Media Network — the term used in IAB Europe V2.1 for retail media networks