Incremental today published a white paper quantifying what the ad industry has long suspected but struggled to measure: retail media advertising frequently drives sales at retailers other than the one where the ad ran. The study, titled "The Cross-Retailer Halo Effect: What Siloed Attribution Misses," draws on data from more than 150,000 campaigns and $350 million in ad spend over a 12-month period, and concludes that siloed last-touch attribution systematically understates the return on investment for upper-funnel retail media channels, in some cases by a factor of three to five.

The findings arrive at a moment when commerce media is absorbing a growing share of national brand budgets. According to Incremental, U.S. retail media ad spend is projected to reach approximately $70 billion in 2026, growing at roughly double the pace of the overall advertising market. Amazon alone accounts for more than three-quarters of that total. The sector's rapid expansion has brought brand dollars into retail media networks (RMNs) that were originally built to serve trade and shopper budgets - and the measurement infrastructure has not kept pace.

The structural problem with closed-loop measurement

Retail media's core measurement promise has always been the closed loop: a shopper sees an ad on a retailer's platform and buys the product at that same retailer, generating a traceable, deterministic attribution signal. According to Incremental, that model works reasonably well when the objective is to drive sales at a specific retailer. It breaks down when the objective is something broader - sustaining brand equity, generating demand across all sales channels, or capturing the full incremental lift that a media investment produces.

The problem is not with closed-loop measurement as a technical approach. Deterministic linkages across first-party data remain one of the most reliable attribution tools available to marketers. The problem, according to Incremental, is the lack of interoperability between retailers. Retailers are understandably reluctant to share purchase data with direct competitors, which means each RMN operates as a self-contained attribution system. Any sales impact that crosses from one retailer's ecosystem into another is, by definition, invisible to the network where the ad ran.

Shoppers, however, do not behave this way. Cross-shopping across multiple retailers is a normal pattern. A consumer exposed to a DSP campaign on one retailer's platform may complete the purchase at a competing retailer days later, or in a physical store entirely separate from the advertising platform. According to Incremental, that cross-retailer impact is being missed at scale, and the distortion is growing as retail media moves up the funnel toward brand-level objectives.

The measurement challenge in retail media has been a persistent theme across industry panels and research in 2025, with omnichannel attribution complexity and standardization gaps repeatedly identified as barriers. The IAB and IAB Europe released guidelines for incremental measurement in commerce media on November 3, 2025, establishing a shared definition of incrementality as the causal impact of marketing compared to what would have occurred without any campaign activity. That framework set conceptual groundwork that the Incremental study now attempts to quantify empirically at campaign scale.

What the data shows: search versus DSP

The Incremental analysis reviewed performance across four large retail media networks with both search and DSP(demand-side platform) capabilities. The dataset covered $565 million in gross merchandise value across nearly 350,000 unique products. Two of the four retailers operated physical stores alongside their e-commerce channels; the other two were online-only. For the brands analyzed, these four retailers represented between 40% and 80% of total sales - meaning the halo effects measured still likely understate the full cross-retailer impact.

Results were expressed as an indexed incremental return on investment (iROI), benchmarked against the average on-retailer impact across all media channels.

Search behaved close to expectations. According to Incremental, 93% of the measured impact from retailer search campaigns occurred within the same retailer's environment. Only 6% of the incremental effect was observed at other retailers. The explanation is straightforward: search captures high-intent shoppers who are already close to a purchase decision and tend to convert immediately. The short half-life of that intent means most of the effect is captured before the shopper has time to switch channels.

Even so, the modest cross-retailer signal from search carries an important implication. According to Incremental, consumers appear to be using retailer search not only for immediate purchase but also for product research, building consideration that sometimes converts at a different retailer. For omni-channel retailers specifically, search campaigns showed 11% to 17% of their impact occurring as in-store purchases - a finding that challenges the assumption that retailer search is purely an online channel.

DSP campaigns told a significantly different story. On-site DSP ads - display and video inventory served within the retailer's own digital environment - showed 29% of their incremental impact occurring at other retailers. That figure rises sharply for off-site DSP, where the ad inventory runs on third-party digital properties using the retailer's first-party data for targeting. According to Incremental, 52% of off-site DSP impact occurred outside the advertising retailer's own sales channels.

Breaking down DSP further by format reveals an even starker divide. Off-site display campaigns sent 41% of their impact off-retailer. Off-site video was more extreme: 62% of the measured incremental effect occurred at retailers other than the one whose data was used for targeting.

The in-store dimension

One finding that may carry particular weight for CPG advertisers is the scale of in-store impact from digital campaigns. According to Incremental, on-site DSP campaigns showed 38% of their incremental effect occurring as in-store purchases. Off-site video campaigns showed 42% in-store impact. These figures are substantially higher than the 7% in-store share recorded for on-site search.

The structural reason is the lag effect. Display and video ads often build brand impressions over time rather than triggering immediate clicks. Consumers exposed to these formats may form purchase intent that takes days or weeks to convert - and when that conversion happens in a physical store, it is completely invisible to the retailer's attribution system. The ad ran online. The purchase happened in-store at that same retailer, or at a different one entirely. Neither outcome is captured by the standard closed-loop report.

The modality breakdown from Incremental's hierarchical measurement model makes this concrete. For on-site search: 66% online, 27% online pickup and delivery (OPD), 7% in-store. For on-site DSP: 24% online, 38% OPD, 38% in-store. For off-site video: 29% online, 29% OPD, 42% in-store.

An April 2026 report from In-Store Marketplace and Catalyst Media Consulting reached a parallel conclusion from a different direction, arguing that the retail industry's core in-store measurement problem is not a lack of technical capability but a fundamental misalignment between the scorecards that brands, agencies, and retailers each use to define success. The Incremental data suggests the misalignment is even deeper: upper-funnel digital campaigns are generating in-store outcomes that no current scorecard captures.

Sizing the total miss

Incremental's analysis of just four retailers still represents only 40% to 80% of the total sales universe for each brand studied. Extending the observed halo pattern to the remaining retailers - those not included in the dataset - produces an estimate of the total gap between reported and actual impact.

According to Incremental, factoring in the portion of the sales universe not captured by the four-retailer analysis suggests that 36% to 53% of total retail media impact is being missed by siloed attribution today. For off-site video specifically, the estimated miss ranges from 67% to 80% of actual impact.

The indexed iROI figures illustrate the magnitude. Search produces an on-retailer iROI of $1.24 and an off-retailer iROI of only $0.08, for a total of $1.31 with 6% off-retailer. On-site DSP produces $1.00 on-retailer, $0.40 off-retailer, total $1.40 at 29% off-retailer. Off-site DSP in total shows $0.77 on-retailer, $0.87 off-retailer, total $1.64, with 52% of impact occurring outside the advertising retailer. Off-site video is the most extreme: $0.74 on-retailer, $1.18 off-retailer, total $1.91, with 62% of impact invisible to the network's own attribution.

When accounting for the remaining universe of retailers beyond the four analyzed, those off-site video figures stretch to an estimated off-retailer halo of between 67% and 80%, and a total iROI range of $2.21 to $3.68 compared to the baseline.

The IAB's April 2026 white paper, which PPC Land covered at the time, argued directly that standard marketing mix modeling is structurally misaligned with retail media measurement, causing systematic undervaluation of commerce channels. The Incremental data provides campaign-level empirical weight to that argument.

Why this matters for brand budgets

The implications are not purely technical. According to Incremental, the gap between actual impact and reported impact creates a feedback loop that can damage both retailers and advertisers over time. When off-site DSP and video campaigns appear to underperform relative to search on a single-retailer ROAS basis, brands shift budget toward lower-funnel performance tactics. That reallocation reduces upper-funnel advertising pressure, which eventually depresses awareness and demand, making lower-funnel investments less efficient and raising customer acquisition costs. Collapsing margins then constrain the ability to re-invest at the top of the funnel.

Retail media's growth has already shifted substantial brand budget away from traditional national media channels into RMN environments. According to Incremental, 70% of retail media spending is now incremental to brands' annual trade budgets, with the majority funded by brand and media budgets rather than traditional trade or shopper funds. That migration brings with it a different set of measurement expectations. Trade and shopper dollars were evaluated against retailer-specific outcomes from the start. Brand dollars were not - and measuring them with the same retailer-specific framework creates a systematic distortion in how their performance is assessed.

Commerce media's growth in European markets has reflected similar dynamics. European retail media spending reached 13.7 billion euros in 2024, representing 21.1% growth compared to 6.1% for the broader advertising market. European advertisers working with four to six retail media networks simultaneously doubled from 10% to 24% in 2025, compounding both the measurement complexity and the stakes attached to accurate cross-network attribution.

A separate data point from Amazon's DSP measurement rollout in May 2026 - which opened self-service access to more than 50 third-party measurement products for DSP advertisers across 18 countries - signals growing industry acknowledgment that the standard closed-loop report is insufficient for evaluating full-funnel campaigns. Amazon's move to allow independent, incremental measurement within DSP workflows is consistent with the direction Incremental's findings point toward.

What cross-retailer measurement requires

According to Incremental, correcting the current distortion requires two shifts. The first is moving from last-touch attribution to an incrementality-based performance measure. Last-touch attribution inherently under-credits upper-funnel channels because it assigns conversion credit to the most recent touchpoint, ignoring the awareness and consideration work that upper-funnel media does much earlier in the purchase path. Longer delays between ad exposure and purchase, which are typical for display and video formats, make this problem worse.

The second shift is moving from a measurement framework tied to a single retailer to one that captures cross-retailer impact. Marketing mix models provide one route, since they typically use total sales across all retailers as the dependent variable. That approach captures the halo effect in aggregate but cannot isolate retailer-specific contributions. Experimentation provides another route, provided the study design captures sales across all retailers rather than just the advertising retailer - but legal restrictions on commingling user-level data across retailers make this difficult to execute using first-party retailer data directly.

Incremental frames the requirement as a dual measurement structure: retailer-specific outcome measurement to support sales and shopper teams' objectives within a retailer, alongside broader cross-retailer measurement to ensure brand dollars are evaluated against their actual impact across all channels where consumers purchase.

The tension that emerges is organizational as much as technical. As retail media absorbs budget from trade, shopper, and brand pools simultaneously, questions arise about which team owns the planning process and how agencies managing separate budget lines coordinate with each other and with non-media functions managing inventory, assortment, and product content. These questions do not have straightforward answers and are not resolved by better measurement alone - but better measurement is a prerequisite for anything else.

Walmart's April 2026 launch of the Scintilla Media Data Feed - an API giving agency and technology partners structured access to approximately 500 Walmart operational and retail data elements, including omni sales figures covering both physical and online channels - reflects one direction retailers are moving to improve signal quality for media partners. Whether that kind of first-party data opening can be combined with cross-retailer views remains an open question.

Limitations and context

The Incremental study is vendor-supplied research, a factor the company acknowledges explicitly in its outreach materials. The four retailers analyzed are unnamed, and the brands and categories are not identified beyond references to CPG. The finding that these four retailers represent 40% to 80% of the sales universe is presented as a range across brands, which means the halo extension to unmeasured retailers involves assumptions that vary considerably by brand and category.

The paper also notes that the four-retailer sample included two omni-channel retailers and two e-commerce pure plays, which affects how in-store impact is distributed across the dataset. For categories where physical retail dominates purchase volume, the in-store figures may be more representative; for categories that skew heavily online, they may not.

Those caveats aside, the scale of the dataset - 150,000 campaigns and $350 million in spend over a defined 12-month window - is large enough to suggest the directional findings are robust. The IAB's November 2025 incrementality guidelines and the related incrementality framework published in September 2025 both establish the theoretical basis for exactly the kind of cross-retailer causal measurement Incremental's study attempts to quantify. The data does not prove the industry's attribution gap with precision, but it gives that gap a concrete numerical range for the first time.

Timeline

Summary

Who: Incremental, a causal intelligence platform for retail media measurement, published the research. The study analyzed campaigns across four unnamed large retail media networks, covering brands in multiple categories including CPG.

What: A meta-analysis titled "The Cross-Retailer Halo Effect: What Siloed Attribution Misses," quantifying how much of retail media's true incremental impact is missed by retailer-specific attribution systems. Key findings include: 52% of off-site DSP impact occurs at retailers other than the one that ran the ad; off-site video sends 62% of impact off-retailer; DSP campaigns drive 38% to 42% in-store purchase impact that closed-loop systems cannot capture; and the total attribution gap across the full retailer universe is estimated at 36% to 53% of actual impact, rising to 67% to 80% for off-site video.

When: The white paper was published June 15, 2026. The underlying dataset covers a 12-month period of campaign activity analyzed prior to publication.

Where: The analysis covers U.S. retail media networks, with four large retailers representing between 40% and 80% of each brand's total sales universe. Two retailers included physical stores; two were online-only.

Why: As retail media captures an increasing share of national brand budgets - with 70% of retail media spending now incremental to trade budgets and largely funded from brand and media pools - the mismatch between how brands set campaign objectives (full-funnel, cross-channel growth) and how RMNs report performance (retailer-specific, last-touch) creates systematic under-measurement of upper-funnel investment. The research argues that continued siloed attribution will produce a feedback cycle of under-reported brand media performance, reduced upper-funnel investment, and rising acquisition costs.