Retail media is a $53.7 billion channel built on a promise: closed-loop attribution. Brands spend because retailers can prove what the money did.

But there is a structural problem with that proof. It only covers the fraction of sales that happen online.

Ecommerce represents roughly 17% of total U.S. retail sales. The other 83% happens in stores. For grocery and mass retail, the split is starker. Walmart does 91% of its sales in-store. Target sits at 92%. The revenue that matters most to these businesses is the revenue their media measurement cannot see.

That creates an uncomfortable dynamic for anyone selling retail media or buying it. The attribution that justifies the spend covers a fraction of the commerce it is supposed to explain. A retail media network pitching closed-loop measurement to a CPG brand is telling a true story about a small audience. The bigger audience, the one shopping in physical stores five or six days a week, remains unmeasured. The promise of closed-loop attribution is real. Its scope is the problem.

‍

How much retail media spend goes unmeasured

77% of retail media spend goes toward online inventory. In-store retail media accounts for less than 1% of the $71.67 billion U.S. market. Meanwhile, 80% of consumer spending happens in physical stores.

That is a lopsided equation. Retail media became the fastest-growing advertising channel because it connects ad spend to purchase data. But that connection only exists for the minor share of purchases that happen through ecommerce. The other 80% generates revenue without generating attribution data.

In-store media works, but the measurement has not caught up. 29% of retail media buyers avoid in-store investment entirely because they cannot prove outcomes. When a CPG brand allocates $10 million to a retail media network, the media team needs to show what that spend did. If the proof only covers online conversions, the brand is evaluating the campaign against a sliver of its actual impact.

Retailers, for their part, are rushing to offer in-store formats. The share not offering in-store media dropped from 28% to 15% in a single year. Digital screens in aisles, checkout displays, in-app promotions triggered by store entry. The formats are proliferating, and retailers see the opportunity, but buyers see the measurement gap.

The supply side is moving faster than the proof side, with retailers building inventory they cannot prove performs and buyers being asked to invest in formats they cannot measure.

And it is not a small problem. The in-store retail media market is growing at 11.62% CAGR compared to 6.35% for the overall market. The demand exists, but measurement remains the bottleneck.

‍

Loyalty programs and the in-store attribution gap

The default answer to in-store attribution is the loyalty program. Scan the card, tie the purchase to a profile, connect the profile to media exposure. In theory, the loop closes.

In practice, the loop is partial. 80% of Americans are enrolled in at least one loyalty program. Sounds comprehensive. But the annual activity rate is 59%. For every ten loyalty members, four made no purchase through the program in the past year.

Do the math. A retailer with 80% enrollment and 59% activity has loyalty-based visibility into roughly 47% of in-store transactions. More than half of purchases happen without the signal that attribution depends on. And that is the optimistic scenario. Retailers with weaker programs see much lower effective coverage.

And the shoppers who skip the loyalty scan are the ones brand advertisers most want to measure. Occasional shoppers. New movers. Price-sensitive customers who bounce between retailers. These are the incremental customers, the ones whose behavior advertising might actually change. Loyalty-based measurement misses them by design.

Kroger is the exception that proves the rule. Through 84.51, Kroger captures 96% of all transactions via loyalty card. At that penetration, loyalty-based attribution genuinely works because the data covers nearly every transaction and the measurement gap functionally disappears.

But Kroger's number is the product of decades of program investment in a grocery format where weekly shopping habits create constant card usage. The fuel points, the personalized coupons, the digital integration with the app, all of it drives consistent identification at checkout. Most retailers operate far below that ceiling. And replicating Kroger's loyalty penetration is a multi-year program investment, not a technology decision.

The industry does not publicly report what percentage of in-store transactions are tied to a loyalty scan. That silence tells you something. If the numbers were flattering, they would be marketed as a selling point to advertisers.

‍

Retail media attribution covers the online minority

Retail media grew 23% to $53.7 billion in 2024. The growth is driven by the perceived advantage of closed-loop attribution: connecting ad exposure to purchase data in ways open web advertising cannot replicate. The advantage is real, but it is narrower than the headline suggests.

If 83% of retail revenue is in-store and 77% of retail media spend measures ecommerce, the "closed loop" covers a minority of the commercial activity it is supposed to attribute.

Think about what this looks like in practice. A CPG brand runs a campaign on a major retailer's media network. The performance report shows strong ecommerce ROAS, maybe 3x or 4x. The media team is satisfied. But ecommerce represents 9% of that retailer's sales. The campaign may have driven ten times more revenue in-store, and none of it appears in the attribution report.

The brand is making allocation decisions based on the visible fraction, not the actual impact. The retailer is pricing its media against the measurable fraction, not the total value delivered. Both sides are operating with an incomplete picture, and both know it.

This is the daily reality for media teams at every major retailer with significant physical store revenue. The performance report tells a clean story. The P&L tells a bigger one. The gap between those two stories is where retail media's credibility will be tested.

Advertisers know this. Nearly two-thirds increased their retail media investments in 2024 despite what Forrester described as "lackluster performance and measurement concerns." As analyst Nikhil Lai put it, there is a "divergence between retail media, sponsored content and its reality."

That divergence is sustainable only as long as the channel keeps growing. The moment growth slows, the measurement gap becomes a credibility gap. Advertisers who accepted partial attribution during a growth phase will demand full accountability in a flat one.

86% of retail media practitioners rate measurement and attribution as a high or critical priority. The urgency is there, but the infrastructure to act on it is not.

‍

Better analytics will not fix in-store measurement

The temptation is to frame this as a data science problem. Better models, smarter algorithms, more sophisticated match-back analysis. But no amount of analytical sophistication compensates for missing identity data.

If a retailer cannot connect a digital ad impression to the household that walked into a store and bought the product, no model bridges that gap cleanly. Match-back analysis against loyalty data tells you what happened among card-scanners. Geo-lift studies tell you what happened in aggregate at the store level. Neither provides the individual-level, cross-surface attribution that advertisers increasingly expect from retail media. Loyalty programs provide one signal, but as the enrollment-to-activity gap shows, it is an incomplete signal. Closing the loop requires identity resolution that works independently of whether the customer scanned a card: payment data, device signals, household matching, behavioral patterns.

The gap is in the infrastructure, not the analysis. Most retail media networks do not have the identity layer connecting digital exposure to physical-store transactions at the individual level. They have loyalty data covering a portion of transactions and modeled estimates covering the rest. The models can be sophisticated, but they are still models. And advertisers increasingly want proof, not projections.

The technology to do this exists. Walmart built it. The identity graph, the real-time resolution, the closed-loop measurement across digital and physical all work. But the gap between what the largest retailer in the world has invested and what the average retail media network operates with is enormous. Most retailers are not starting from Walmart's position. They need the same capability without the same resource commitment.

The industry is starting to respond. The IAB published a framework for in-store media measurement in December 2025. eMarketer declared CES 2026 the moment retail media entered its "accountability era." The frameworks are arriving, but the infrastructure to implement them is what most retailers still lack.

‍

Full-estate retail media measurement in practice

The best proof comes from the retailers who have actually built it.

At CES 2026, Albertsons Media Collective launched incrementality measurement for in-store campaigns. The methodology: test stores with media exposure matched against control stores without, across nearly 60 variables. The beta with Mondelez across 116 stores delivered $2.41 incremental ROAS and a 14% lift in in-store sales.

Those numbers represent what happens when a retailer can prove digital media drove physical store purchases. Not modeled. Measured. The 14% lift was real revenue that would have been invisible under ecommerce-only attribution. And it changes the conversation with the CPG brand from "here's your online ROAS" to "here's what your campaign actually did across our stores."

BCG estimates the U.S. in-store retail media opportunity at $20 to $30 billion. That is not a projection based on advertiser surveys. It is the scale of the opportunity when measurement catches up to what is already happening in physical stores. The market is not waiting on demand; it is waiting on the infrastructure to measure what is already happening.

When retailers can attribute digital spend to in-store outcomes at the individual level, the economics of retail media change. CPM premiums become justifiable because advertisers can see what their money did across the full estate. Audience scale extends beyond loyalty members to the full customer base. Retail media moves from a trade spend line item to a brand investment, the kind of spend that shows up in annual planning, not just quarterly negotiations.

Seven in ten advertisers now meet or exceed their retail media performance goals. The channel works for the metrics it measures. The question is what happens when those metrics expand to include the 80% of revenue sitting outside the measurement window.

That is the subject of the next piece in this series: how identity resolution connects digital media exposure to in-store transactions.