Amazon Ads announced on July 07, 2026 that Omnichannel Metrics, the measurement tool tracking how Amazon DSP campaigns influence sales beyond Amazon's own store, now breaks down offline purchases by the specific retailer where they happened. The update follows an April update that grouped results by product category rather than by name, and it marks the first time Amazon has surfaced retailer identity within its own attribution product, according to the launch announcement published on Amazon's advertising site.

What the update actually changes

Omnichannel Metrics, commonly abbreviated OCM inside Amazon's documentation, exists to answer a question that has nagged retail advertisers for years: when a shopper sees an Amazon DSP ad and later buys the product somewhere else, where exactly does that "somewhere else" sit? Until this update, the tool could tell an advertiser that a purchase happened off Amazon. It could not say whether that purchase happened at Target, at Walmart, at a regional grocery chain, or at a specialty retailer three states away.

According to the announcement, the new retailer-level breakouts surface both units and sales at two levels of granularity: the overall study level, which aggregates performance across a campaign's full flight, and the campaign level, which isolates individual campaigns within a broader account. Advertisers running Amazon DSP through either Self-Service or Managed Service can access the feature, and Amazon has limited availability to United States advertisers only, at least for now. There is no mention in the launch material of a planned international rollout, nor any stated timeline for one.

The mechanism behind the retailer identification draws on two data sources working together. First-party signals come from the Amazon Shopper Panel, a rewards-based consumer community that Amazon launched in 2020 and that pays participants for scanning receipts and completing short surveys about purchases made anywhere, not just on Amazon.com. Panelists who submit ten or more eligible receipts in a month receive a ten-dollar reward, according to Amazon's original description of the program; the payment structure has not changed materially since launch. Third-party signals supplement that panel data, filling gaps where panel participation alone would not produce a statistically reliable sample.

That combination matters because attribution built entirely on a voluntary panel runs into a familiar problem: panels are never large enough, on their own, to represent every retailer a shopper might visit. Layering third-party signals on top allows Amazon to extend coverage without waiting for panel enrollment to catch up, though the company has not disclosed what share of the underlying data comes from each source, nor has it named which third-party providers contribute the supplementary signals.

Exposure-based attribution, not last-click

The launch announcement describes the methodology as "exposure-based attribution," a term that recurs throughout Amazon's recent measurement releases. Rather than crediting a sale to the last ad a shopper clicked, exposure-based attribution asks whether a shopper who saw an Amazon DSP ad subsequently completed a purchase, regardless of whether they clicked, and regardless of how much time passed in between. According to the announcement, this design is meant to reflect where "ad-exposed shoppers actually purchased," which is a broader and arguably more honest standard than click-based tracking, since most retail purchases involve no click at all.

This distinction is not cosmetic. Click-based measurement systematically undercounts the influence of display, audio, and video formats, which build awareness over time rather than driving immediate clicks. A shopper who sees an Amazon DSP display ad for a blender on Monday and buys that blender at a home goods retailer the following weekend generates no click to attribute. Exposure-based attribution is the only way to capture that shopper's journey at all.

Building on April's category breakouts

This retailer-level release does not arrive in isolation. Amazon introduced product category-level breakouts to Omnichannel Metrics in April 2026, using the same exposure-based methodology to show advertisers which product categories, rather than which specific stores, their campaigns were driving purchases within. That April release gave large, multi-category advertisers a way to validate omnichannel strategy against purchasing behavior instead of broad market assumptions, according to Amazon's own framing of the update at the time.

The category-level work itself extended an even earlier expansion. On November 04, 2025, Amazon broadened Omnichannel Metrics to five product categories - Consumer Electronics, Fashion, Home Goods and Furniture, Home Improvement, and Toys and Games - after the tool had previously operated with narrower vertical coverage. That November update also automated the study creation process, removing a requirement that advertisers previously navigate manual setup through Amazon's managed service teams before measurement could begin. Fashion advertisers, in particular, benefited from that change, since fashion purchases often involve extended research periods spanning both online and in-store touchpoints, complicating any attempt to trace a single, tidy conversion path.

Read together, the three releases form a clear progression. November 2025 widened which product categories could be measured at all. April 2026 sharpened that measurement down to the category level. July 2026 sharpens it further still, down to the individual retailer. Each step has narrowed the unit of analysis without changing the underlying data sources or the exposure-based methodology that ties them together.

Why retailer-level detail matters for budget decisions

Amazon's own justification for the update centers on strategic specificity. According to the announcement, the retailer breakouts give advertisers "the retailer-level specificity needed to optimize omnichannel strategies," a claim that reflects a real and long-standing gap in retail measurement. Knowing that a campaign drove some unspecified volume of offline sales is directionally useful. Knowing that it drove sales specifically at one national chain rather than another is operationally useful, because it changes what an advertiser can actually do with the information.

Consider a brand that sells across dozens of retail partners nationally. Under the category-level view introduced in April, that brand could learn that its Amazon DSP campaigns were driving strong offline sales in, say, the home improvement category. Under the retailer-level view introduced this month, the same brand can learn whether those sales concentrated at one large home improvement chain or spread evenly across several competitors. That distinction bears directly on where a brand negotiates trade marketing dollars, where it prioritizes in-store merchandising support, and which retail partnerships it might deepen or reconsider.

The broader measurement landscape gives this update additional weight. Research published by Incremental, a causal intelligence platform for retail media measurement, in June 2026 found that siloed, retailer-specific attribution misses between 36% and 53% of total retail media impact across the industry, with the miss rate climbing as high as 80% for off-site video formats specifically. That research examined a different measurement problem than the one Omnichannel Metrics addresses; Incremental's work concerns cross-retailer halo effects, where an ad on one retailer's platform drives sales at a competing retailer entirely outside the advertising retailer's own attribution system. Amazon's retailer-level OCM breakouts approach the same underlying tension from the opposite direction: rather than measuring how much impact escapes a single retailer's closed loop, they aim to show, within Amazon's own DSP data, precisely which outside retailers actually captured that escaped impact. Neither tool alone solves the full cross-retailer measurement problem the industry has been grappling with throughout 2025 and 2026, but the two approaches sit adjacent to one another conceptually.

Amazon's measurement stack has been expanding on parallel tracks throughout this period as well. The company opened self-service access to more than fifty third-party measurement products for Amazon DSP advertisers across eighteen countries in May 2026, a move that let advertisers layer independent verification atop Amazon's own first-party reporting without requiring managed-service coordination. Separately, Amazon extended Amazon Marketing Cloud's lookback window from thirteen months to twenty-five months in November 2025, giving advertisers the ability to compare performance across two full holiday seasons rather than one. None of these releases directly modifies Omnichannel Metrics, but together they describe a company steadily removing friction and adding granularity across its measurement products, of which the retailer-level OCM breakouts are simply the latest, and most narrowly targeted, installment.

Technical scope and current limitations

The July 07, 2026 announcement is explicit about scope, and the limitations are worth stating plainly rather than glossing over. The feature is available only in the United States; Amazon has not indicated any international expansion plans within the announcement itself. It applies only to Amazon DSP campaigns; Sponsored Products, Sponsored Brands, and Sponsored Display are not mentioned as part of this release. Both Self-Service and Managed Service advertisers can access the retailer breakouts, and the access point remains the Amazon DSP console itself, consistent with how OCM has always been surfaced to advertisers.

Amazon has not published which specific retailers appear as named breakouts, nor has it disclosed a minimum sample size threshold below which a retailer might be grouped into an "other" category rather than named individually. Given that the underlying methodology depends partly on Amazon Shopper Panel participation, it stands to reason that retailers with denser panel coverage, or those that draw more third-party signal, would produce more granular, more reliable breakouts than smaller or more regional chains. Amazon's announcement does not address this question directly, and advertisers evaluating the feature will likely need to test it against their own retail partner mix to understand where the granularity holds up and where it does not.

The announcement also does not disclose adoption figures, performance benchmarks, or any quantified outcome from advertisers who tested the retailer-level view ahead of general release. This is, in the strictest sense, a feature launch rather than a performance disclosure, and Amazon's own framing treats it as such throughout the announcement.

Context for the marketing community

Retail media measurement has spent much of 2025 and 2026 wrestling with a structural problem: attribution systems built by individual retailers naturally see only what happens inside their own walls. Amazon has been addressing that problem from several angles simultaneously, expanding coverage by category in 2025, sharpening it by category in April 2026, and now sharpening it further by named retailer in July 2026.

For advertisers managing budgets across a genuinely omnichannel retail footprint, the practical significance is easy to state plainly. A tool that once told an advertiser only that offline sales occurred now tells that advertiser where those sales occurred, by name. Whether that additional precision changes actual budget allocation decisions will depend on how consistently the retailer breakouts hold up across different product categories and different advertiser scales, questions this announcement does not yet answer but that advertisers testing the feature over the coming months will likely begin to.

Timeline

  • October 2020 - Amazon launches the Amazon Shopper Panel, a rewards-based mobile app collecting receipt data and survey responses on purchases made outside Amazon.com.
  • November 04, 2025 - Amazon expands Omnichannel Metrics to five product categories: Consumer Electronics, Fashion, Home Goods and Furniture, Home Improvement, and Toys and Games, and automates study creation.
  • April 14, 2026 - Amazon introduces product category-level breakouts to Omnichannel Metrics, using exposure-based attribution at the study and campaign level.
  • May 15, 2026 - Amazon opens self-service access to more than fifty third-party measurement products for Amazon DSP advertisers across eighteen countries.
  • June 15, 2026 - Incremental publishes research finding that siloed retail media attribution misses between 36% and 53% of true campaign impact industry-wide.
  • July 07, 2026 - Amazon announces retailer-level breakouts for Omnichannel Metrics, available to US Amazon DSP Self-Service and Managed Service advertisers.

Summary

Who: Amazon Ads, specifically the team behind Omnichannel Metrics within Amazon DSP, issued the update for US-based Amazon DSP Self-Service and Managed Service advertisers.

What: Omnichannel Metrics now breaks down offline sales and units by specific retailer name, at both the study and campaign level, using exposure-based attribution built on Amazon Shopper Panel and third-party signals.

When: Amazon published the announcement on July 07, 2026, building on category-level breakouts introduced in April 2026 and the five-category expansion from November 04, 2025.

Where: The feature is available exclusively in the United States, accessed through the Amazon DSP console.

Why: Advertisers with offline sales spanning multiple retail partners previously could see that Amazon DSP campaigns drove purchases beyond Amazon, without knowing which specific retailers captured those purchases. Retailer-level detail lets advertisers connect campaign performance to individual trade relationships rather than treating "offline" as a single undifferentiated category.