Closed-Loop Attribution in Retail Media: What It Is, How It Works and How to Offer It

Last updated: October 2026. Reviewed by Kunal Damgude, Growth and Product Marketing Manager.

Closed-loop attribution, also called closed-loop measurement or closed-loop reporting, is the practice of matching an ad impression or click to a purchase recorded in the retailer's own first-party transaction data, so the sales a campaign drove are counted from real orders. Industry standards call it event-level attribution, and in retail media and the wider commerce media category it is the evidence a retailer gives brands that its ads moved product.

Osmos is a white-label retail media operating system for retailers and marketplaces. Its modular apps for onsite, offsite and in-store ads sit over your existing stack, so you can break your monetisation ceiling faster without rebuilding it.

Osmos reports impressions, clicks and attributed sales to advertisers from the retailer's own orders, and its Bring Your Own Traffic app, now in Early Access, extends closed-loop attribution to every offsite campaign with tracker links that follow a click from an external campaign to the retailer's revenue. For how closed-loop attribution sits next to multi-touch models, media mix modeling and the rest of the measurement stack, see our guide to retail media attribution and measurement.

What Is Closed-Loop Attribution in Retail and Commerce Media?

Closed-loop attribution is a measurement method that links an ad exposure to a purchase inside the same retailer's data, at the level of an individual shopper or household. The "loop" is the path from the ad back to the till: the retailer served the ad, the retailer recorded the order, and the retailer can join the two. The IAB and Media Rating Council (MRC) Retail Media Measurement Guidelines (January 2024, section 3.6) put it this way:

"Many retailers report on media performance through event-level attribution (also sometimes referred to as closed-loop attribution). Closed-loop attribution often involves models that involve the use of extrapolation techniques to show the full impact of a campaign on brand sales."

The same guidelines define attribution itself as "the practice of assigning credit for consumer actions, like sales or visits, to specific marketing efforts," adding that when it is done at an individual or household level, "it aims to directly link ad exposure to outcomes." Closed-loop attribution carries that direct link from the ad exposure through to the final order.

The terminology covers four distinct parts of the practice:

• Closed-loop attribution is the credit rule: which campaign gets credit for which purchase.
• Closed-loop measurement is the wider practice of measuring media against transaction data, which includes attribution plus lift and incrementality reads.
• Closed-loop reporting is what the advertiser receives: the sales, units and return on ad spend (ROAS) tied to a campaign, usually by product.
• Closed-loop marketing attribution is the older B2B marketing use of the idea, where revenue from a customer relationship management (CRM) system is passed back to the team that ran the campaign. Retail media applies the same principle with the store's order data.

Commerce media is the umbrella term for ads that run on platforms where people buy, from grocers and marketplaces to delivery apps and payment networks, and closed-loop attribution works the same way there as it does in retail media.

How Does Closed-Loop Attribution Work?

Closed-loop attribution works by logging every ad event, identifying the shopper, matching that shopper's later purchases in the retailer's order data within a disclosed window, and reporting the matched sales back to the advertiser. The process follows seven steps based on the IAB/MRC guidelines:

1. Log the exposure. The ad server records each impression and click with a timestamp, the placement and the campaign. The IAB/MRC guidelines (section 3.5) note that "the MRC requires viewable impressions for attribution of outcomes to ad exposures," so the impression record needs a viewability signal.
2. Identify the shopper. A logged-in account, a loyalty ID or a payment card links the exposure to a known shopper or household.
3. Record the purchase. Online orders, app checkouts and, where loyalty or payment capture allows, store purchases land in the retailer's transaction data with the same identity.
4. Match exposure to purchase. The retailer joins ad events to orders. This is the step that closes the loop back to the ad platform.
5. Apply the attribution window. Only purchases inside the agreed window count. The IAB/MRC guidelines say "typical attribution windows for digital measurement are often 3, 7, 14, 28, or 30 days."
6. Model the unidentified share, and say so. Some buyers cannot be traced, for example cash payers or shoppers whose loyalty card was not scanned, so retailers extrapolate. The guidelines expect that modeling to be disclosed.
7. Report back by product. The advertiser sees attributed sales, units and ROAS per campaign, ideally per SKU, with an incrementality read where the retailer offers one.

Onsite retail media is the cleanest match path, because the ad and the order sit on the retailer's own site or app. Campaigns on offsite retail media channels need a link back to the retailer's revenue, which is the job of tracker links and audience matching. In-store retail media closes the loop through point-of-sale and loyalty data.

Why Is Closed-Loop Attribution Important for Retailers?

Closed-loop attribution matters because brands renew retail media budgets on sales they can see, and the retailer is the only party that holds the transaction record that proves those sales. With closed-loop attribution in place, the retailer can show advertisers the orders its media drove.

Skai's 2026 State of Retail Media survey of brands, a vendor survey tracked in Osmos's measurement guides, found that 52% of advertisers are moving display spend from open-web demand-side platforms (DSPs) to retail media DSPs, and closed-loop attribution was the driver cited by 57% of those reallocating. For a retailer, that means a share of the budget now moving into retail media follows the measurement it offers. The same report found that "only 15% of brands report strong confidence in measurement," which gives the retailer whose reports brands can check a clear lead.

Closed-loop attribution is also what makes ROAS a number a brand can act on. ROAS in retail and commerce media is attributed sales divided by ad spend: a campaign that spends $10,000 and is matched to $45,000 of orders inside its window reports a ROAS of 4.5x. The figure depends on the window, the match rules, and the platform and ad format behind it. Attributed ROAS also differs from incremental ROAS (iROAS), which counts only the sales that would not have happened without the ad.

In one Osmos customer example, a premium beauty marketplace replaced an in-house ad server that offered fixed banners, priced per day, with no performance reporting. After moving to Osmos, which added Product Ads on search and category pages and impression, click and attributed-sales reporting on display, the marketplace grew attributed GMV about 63% and ad revenue about 90% comparing June to July 2026, its flagship sale period.

What Should You Look For When Evaluating Closed-Loop Attribution?

A retailer's measurement should meet the standards a brand's agency checks before it trusts the report. Most of the eight below come from the IAB/MRC guidelines:

1. A window set and disclosed before the campaign. The guidelines say windows "must be disclosed before campaign execution and measurement and as part of disclosures corresponding to reported results," and encourage day-level detail so results from different windows can be reconciled.
2. A transaction-level match. The report should say whether it matches individuals or households, and which identifiers (account, loyalty, payment) it uses.
3. Disclosed extrapolation. The modeled share of sales should be visible, along with the reason (cash, card or loyalty capture limits).
4. SKU-level reporting. The guidelines state: "At minimum, SKU-level reporting is recommended where relevant unless otherwise unavailable (where disclosed and supported), given most investment is to promote specific SKUs."
5. Viewable impressions. Attribution should start from MRC-compliant viewable impressions, with accreditation as the long-term goal.
6. Independent validation. The guidelines add that if outcomes "are reported for retail media organization activity and used for pricing or optimization considerations, they are encouraged to be subject to third-party audit and validation."
7. Incrementality next to attribution. Incrementality tests show how many of the attributed sales the ad caused. Method choices change the answer: research by Albertsons Media Collective, Ovative and Northwestern University's Kellogg School of Management found that "iROAS can vary by 6.5x and flip results in 83% of campaigns based on methodology alone" (April 2026). Osmos's Incrementality app, currently in Early Access, uses a configurable holdout to measure lift at view, add-to-cart and purchase for onsite sponsored display, with an 8-week minimum window.
8. Comparable definitions across partners. To compare closed-loop results across advertising partners or streaming vendors, line up the window, the match basis, the share that is modeled and whether the figure is attributed or incremental, because ROAS figures built on different windows cannot be compared directly.

Liz Roche, VP of Media and Measurement at Albertsons Media Collective, described why incrementality carries weight with brands: "Incrementality is a critical metric because it helps brands understand whether their media investment is creating new demand or simply capturing existing sales."

The strongest closed-loop attribution tool for a retailer is the one that meets these eight points on its own transaction data and reports them in a form brands can audit.

Which Platforms Offer Closed-Loop Attribution?

The large retail media networks offer closed-loop attribution on their own shopping data, each with its own identity, clean room and reporting setup, and retailers building their own network can offer it through a platform such as Osmos:

• Walmart Connect: Walmart describes closed-loop measurement powered by its first-party data that ties ad exposure across Walmart's digital properties to results advertisers read in its self-serve Walmart Connect Ad Center.
• Amazon Ads: Amazon measures campaigns against purchases on its own store, and Amazon Marketing Cloud gives advertisers a clean room for custom analysis of that data.
• Instacart Ads: Instacart says its ads give "closed-loop visibility from impression to purchase across environments," backed by MRC-accredited measurement, and reports that brand partners see a 25% sales lift on average (company-reported, 2026).
• Target Roundel: Roundel says it is "built on closed-loop measurement" that connects ad exposure to real guest behaviors, using first-party data from in-store, in-app and online shopping.
• Criteo: Criteo says its closed-loop measurement and real-time reporting reveal sales impact down to the SKU level.
• Osmos: Osmos, a white-label retail media operating system for retailers and marketplaces, reports attributed sales from the retailer's own orders, and a premium beauty marketplace on Osmos grew attributed GMV about 63% and ad revenue about 90% in its flagship sale; its Bring Your Own Traffic app for closed-loop attribution on every offsite campaign, and its Incrementality, Brand Lift Study and Survey Ads apps, are in Early Access.

A retailer without Amazon's or Walmart's scale can still offer closed-loop attribution tied to real transactions, because what it needs is its own order data, a shopper identifier and an ad server that can join the two. Osmos builds attribution on the retailer's own order data for this reason, so the retailer keeps its transactions as the ground truth that every advertiser report starts from.

Does Closed-Loop Attribution Work for TV and Streaming?

Closed-loop attribution works for connected TV (CTV) and streaming when the retailer matches streaming ad exposure to shoppers in its own sales data, and the incrementality side is still being proven. Digiday reported in April 2026 that Walmart's tie-up with Vizio promises marketers "the very thing they've wanted across the retail media landscape: closed-loop measurement," while early buyers were cautious. Laura Knebusch, senior vice president of marketing and customer experience at Georgia-Pacific, told Digiday: "Their measurement wasn't quite where we needed it to be to execute a test, to truly understand incrementality."

Retailers close the loop on CTV against their own order data. When a retailer extends its audiences to CTV through offsite retail media, the evaluation checklist applies unchanged, for Roku, any other streaming partner or a social platform alike: whose purchase data closes the loop, over which window, and with what lift test. A network that grows beyond its own site into CTV sets its credit rules before the new channel goes live, so the first CTV report already uses the agreed window and match basis.

How Can Retailers Improve Closed-Loop Attribution Quality?

Retailers improve closed-loop attribution by raising the share of purchases they can tie to a known shopper and by making every report something a brand can check.

1. Grow identified traffic. More logged-in sessions and loyalty scans mean more matched orders and a smaller modeled share.
2. Keep the product catalog and order feed in step with the ad server. The same SKU identifiers on both sides make product-level reporting possible.
3. Publish a window policy. State the default window by format before campaigns run, and keep it stable so quarter-on-quarter results compare.
4. Bring store transactions into the same data. Point-of-sale and loyalty data let in-store purchases count, which matters most for grocery and pharmacy brands.
5. Give brands self-serve reporting. Attributed sales, units, new-to-brand buyers and ROAS by campaign and SKU, without waiting on a manual report.
6. Run lift tests on a schedule. A regular incrementality read is the answer to the brand that asks how many of the attributed sales were caused by the ad.
7. Use one standard across banners. A group that runs several store banners can apply one window and one match rule everywhere, so brands read results from every banner the same way.

What Are the Main Challenges in Closed-Loop Attribution?

The IAB/MRC guidelines name the identity gap directly: modeled attribution is "typically needed because not every ad that is served can be traced back to an identified user/consumer," and they list cash transactions and card or loyalty capture rates among the reasons. The fix is disclosure, so a brand knows how much of the result is matched and how much is modeled.

Offline matching depends on loyalty and payment capture at the till, so retailers with strong loyalty programs report in-store results with more confidence.

Across networks, different windows make retail media attribution results hard to compare unless each retailer publishes its windows with day-level detail.

Retailers measure incrementality by running holdout tests alongside the attributed report.

Explore Osmos Measurement or book a demo.

Frequently Asked Questions

How is closed-loop attribution different from last-click attribution?

Last-click gives all the credit to the final click before a conversion, wherever the conversion is recorded. Closed-loop attribution credits views as well as clicks, and it counts only purchases found in the retailer's transaction data. The two can also be combined, since a closed-loop report can still apply a last-touch rule inside its window, one of several attribution models a retailer can choose.

Does closed-loop attribution need third-party cookies?

Closed-loop attribution runs on the retailer's first-party data: logged-in accounts, loyalty IDs and order records. Where a brand wants to combine its own customer data with the retailer's, a clean room lets both sides run the analysis without handing over raw customer records.

How long should a closed-loop attribution window be?

Match the window to how often shoppers buy the product. Frequently bought grocery items suit a short window, while considered purchases such as electronics need a longer one to capture the sale. Whatever the default, keep it consistent by format so brands can compare one campaign with the next.

How does Amazon Marketing Cloud support closed-loop attribution?

Amazon describes Amazon Marketing Cloud as a privacy-safe, cloud-based clean room in which advertisers run analytics and build audiences across pseudonymized signals, which is where custom closed-loop analysis of Amazon Ads campaigns happens.

Can in-store sales be part of closed-loop attribution?

Yes, when the shopper is identified at the till by a loyalty card, an app scan or a payment card. Store purchases without an identifier are the main reason retailers model part of the result, so loyalty capture rate is the number that decides how much of in-store attribution is matched to a real order.

Sources

1. IAB and Media Rating Council, Retail Media Measurement Guidelines, January 2024, sections 3.5 and 3.6.
2. Skai, "2026 State of Retail Media Report," 2026. skai.io/reports-and-whitepapers/2026-state-of-retail-media-report/
3. Albertsons Companies, Albertsons Media Collective Introduces Onsite Incrementality Measurement, April 16, 2026.
4. Digiday, Kimeko McCoy, Walmart-Vizio's CTV measurement faces incrementality hurdle, April 1, 2026.
5. Walmart Connect, "Measurement," checked October 12, 2026. walmartconnect.com/measurement
6. Amazon Ads, "Amazon Marketing Cloud is now available on AWS Clean Rooms." advertising.amazon.com/blog/amazon-marketing-cloud-available-on-aws-clean-rooms
7. Instacart, "One Campaign. 100,000 stores and beyond. The Instacart Ads Ecosystem Explained," 2026. company.instacart.com/instacart-ads/growth-incremental-sales-and-scale
8. Roundel, "Measurement," checked October 12, 2026. roundel.com/measurement
9. Criteo, "Retail media," checked October 12, 2026. criteo.com/retail-media/