Retail Media Has A Measurement Problem

Retail media is getting bigger, louder, and harder to ignore. It is also getting easier to overbuy.

That is the useful tension behind last week’s first retail media network Showcase, where rival networks pitched advertisers in an upfront-style event. The Wall Street Journal reported that DoorDash, PayPal Ads, Chase Media Solutions, Macy’s Media Network, Albertsons Media Collective, and others used the stage to argue that their data can connect attention to sales.

That pitch is attractive for ecommerce brands. It is also exactly why marketers should slow down before moving more budget into every retailer, marketplace, delivery app, and payment network that now sells ads.

The Channel Is Maturing Faster Than The Measurement

Retail media used to be easy to describe: sponsored products on Amazon, Walmart, Target, or another retailer’s site. Now the category stretches across onsite ads, offsite media, connected TV, digital out-of-home, in-store screens, loyalty data, and payment data.

The budget is following. EMARKETER’s May 2026 retail media forecast expects Amazon’s retail media revenue to exceed $75 billion by 2028, more than $65 billion ahead of the next-largest retail media network. That kind of growth will pull more ecommerce brands into the channel, even when their measurement systems are not ready for it.

That growth does not make the channel automatically efficient. It makes the measurement problem more expensive.

Skai and Stratably’s 2026 State of Retail Media research says only 15% of brands report strong confidence in retail media measurement. A companion Skai analysis found advertisers work with an average of six retail media networks today and expect that number to grow to 11 by the end of 2026.

That is a lot of dashboards, attribution windows, audience definitions, product feeds, and platform incentives for one marketing team to reconcile.

ROAS Is Too Easy To Misread

Retail media is appealing because it sits close to the purchase. A shopper is already browsing products, comparing prices, or completing an order. That proximity can make reports look cleaner than they really are.

The risk is that advertisers confuse platform-attributed revenue with incremental revenue. If a loyal customer was already going to buy the product, the ad should not get full strategic credit. If a retailer gives a campaign credit for sales influenced by pricing, inventory, brand demand, email, search, or social, the reported return can look stronger than the real lift.

IAB’s commerce media incrementality guidelines make the core point clear: marketers have to measure the business impact caused by commerce media investment, not just count sales that happened near an impression or click.

That sounds like an analytics detail. It is a budget decision.

If retail media gets treated as a bottomless performance channel, brands will keep adding networks until the operating cost, feed work, creative demand, and reporting confusion outweigh the upside. Better paid media optimization starts by asking which networks create sales the brand would not have captured otherwise.

Smaller Brands Should Be More Selective

Large CPG teams can afford dedicated retail media specialists, clean-room partners, data science support, and separate budgets for testing. Most ecommerce brands cannot.

That does not mean they should avoid retail media. It means they should enter with a sharper scorecard.

Before adding another network, ask four questions:

1. Does this retailer or marketplace already have meaningful demand for the product category?
2. Can the team see performance by SKU, margin, new customer, and repeat customer?
3. Can the platform support a clean incrementality test, holdout, matched-market test, or other causal read?
4. Does the campaign feed a broader business goal, or is it only buying visibility inside someone else’s store?

If the answer is vague, the brand is probably buying access, not performance.

That distinction matters because retail media can pull money away from the channels that created the demand in the first place. Search, social, email, creator content, and site improvements may still be doing the hard work of education and trust. Retail media may simply be catching the buyer at the register.

At Emarketed, we have seen how important that full-system view is for ecommerce growth. Sector 9 Skateboards grew sales by 165% and increased marketing-attributed revenue by 254%, while Nutcase Helmets grew sales by 306% and sales attributed to marketing by 774%. The lesson from those campaigns is not that one channel wins forever. It is that paid media, merchandising, conversion, and measurement have to be judged together. Our post on what ecommerce brands can learn from Nutcase and Sector 9 breaks down that pattern.

What To Do Before The Next Budget Shift

Treat retail media like a serious channel, not a budget spillover.

Start with the networks closest to actual product demand. Build a simple measurement table before launch: spend, attributed revenue, margin, new-customer rate, repeat purchase behavior, assisted search lift, and inventory constraints. Then separate two questions that often get mixed together:

1. Did the campaign produce reported sales?
2. Did the campaign create sales we were unlikely to win without it?

Those are different answers. The first keeps a dashboard happy. The second protects the business.

Retail media will keep growing because retailers, marketplaces, delivery apps, payment networks, and financial platforms all want ad dollars tied to their customer data. Marketers should not ignore that shift. They should demand cleaner proof before they treat every new network like a must-buy.

The move this week is simple: audit current retail media spend against incrementality, not just ROAS. If the channel cannot prove lift, cap the test, fix the measurement, and put the next dollar where the business can actually see it working.