Retail Media's Measurement Crisis:
Why ANA's New Framework Still Can't Answer the One Question That Matters

Retail media grew up fast. Its measurement didn't keep pace.

Retail media networks have become one of the fastest-growing lines on the media plan. eMarketer projects the category will grow at a 17% compound annual rate through 2028, reaching roughly $90 billion in the US alone. Every major retailer now runs its own ad business, and every one of those ad businesses reports its own numbers, in its own dashboard, using its own definitions.

On August 18, 2026, the Association of National Advertisers (ANA) published a new report, "Retail Media Measurement Standardization," built with a working group that included PepsiCo, Hershey's, Clorox, Kimberly-Clark, Mondelez, Bayer and Intel. The headline finding: 55% of advertisers say the lack of consistent measurement standards is the single biggest barrier to retail media success - bigger than cost, bigger than fragmentation, bigger than anything else on the list.

That's a striking number for a channel advertisers are still pouring money into. It's also a useful moment to separate two problems that get conflated constantly in retail media conversations: the problem of inconsistent reporting, and the much bigger problem of unproven causality. ANA's framework is a real, needed step toward fixing the first one. It does almost nothing for the second. And the second is the one that actually determines whether retail media spend is working.

What the ANA framework actually recommends

The report is specific about where retail media measurement breaks down today. Networks use different attribution lookback windows, different definitions of a "new-to-brand" customer, different scopes for identity resolution, and different deduplication logic - which means a $1 million campaign on one network and a $1 million campaign on another can produce ROAS numbers that simply aren't comparable, even when the underlying sales impact is identical.

To fix that, the ANA is asking advertisers to push retail media networks toward a common baseline: standardized core metrics (impressions, viewability, clicks, invalid traffic), a standardized 14-day attribution lookback window, transparency into methodology and definitions, and - notably - greater use of independent, third-party measurement and validation rather than relying solely on the network's own self-reported numbers.

That last point is the closest the report gets to the real issue. It's a tacit admission that a retail media network marking its own homework is a structural problem, not just an inconvenience.

Standardized reporting still isn't incrementality

Here's the gap the framework doesn't close: even a perfectly standardized, perfectly transparent, third-party-audited retail media metric is still, at best, an attributed or exposure-based number. It tells you what happened to shoppers who saw an ad. It does not tell you what would have happened to those same shoppers if the ad hadn't run - which is the only question that actually determines incremental value.

Video advertising offers a clean, recent illustration of exactly this gap outside retail media. Measured's Q1 2026 analysis of 148 video advertisers found that CTV delivered the highest median incremental ROAS of any video channel, at $1.38, ahead of YouTube ($1.26) and linear TV ($0.75) - but with an enormous spread: the middle 50% of advertisers ranged from $0.50 to $3.07 in incremental ROAS on the exact same channel. In one geo holdout test on streaming TV that Measured ran, roughly half of the conversions the ad platform had claimed turned out to be incremental once tested properly - the channel still performed well, delivering a $2.46 incremental ROAS at 90% statistical significance, but at half the volume the platform reported.

That's the pattern retail media is walking into. Standardizing the lookback window and the "new-to-brand" definition across networks will make the numbers more comparable to each other. It will not tell you whether the sales attributed to a sponsored product placement would have happened anyway, because the shopper already had the product in their cart, already searched the brand name directly, or was already going to buy from that category that week regardless of the ad.

This is exactly why 75% of US buy-side leaders now say their core ad-measurement methods - attribution, incrementality testing and MMM alike - are underperforming, according to eMarketer's 2026 reporting on the state of measurement. The industry isn't short on metrics. It's short on causal proof.

Retail media doesn't run in a vacuum - and can't be measured like it does

There's a second, quieter flaw in evaluating retail media network by network: none of it happens in isolation. A sponsored product placement on a retailer's site interacts with the TV flight that built awareness for the product two weeks earlier, the paid social prospecting campaign running in parallel, the price promotion the retailer is also running, and the print circular that landed in mailboxes that same week. A retail media network's own dashboard has no visibility into any of that - it can only tell you what happened inside its walls, which means it will systematically take credit for demand that other channels, online and offline, actually created.

Isolating retail media's true contribution requires putting it inside a model that also accounts for everything else influencing sales at the same time - TV, out-of-home, radio, print, paid search, paid social, price, seasonality, competitive activity. That's the core function of marketing mix modeling, and it's why MMM has surged back into relevance industry-wide as identity-based attribution keeps losing ground to privacy constraints and walled-garden opacity. The IAB made this argument formally in its December 2025 "Modernizing MMM: Best Practices for Marketers" guide, which calls specifically for integrating MMM with attribution and experimentation data across paid, owned and earned media - retail media included - rather than evaluating any single channel's reported numbers on their own.

Where causal, unified measurement picks up where ANA's framework leaves off

ANA's framework is a genuinely useful step: standardized reporting makes retail media networks easier to compare and audit, and pushing for independent third-party validation is the right instinct. But comparability between networks was never the hard problem. The hard problem is knowing what retail media - and every other channel around it - actually caused, net of everything else happening in the market at the same time.

That's what a causal, always-on marketing mix model is built to answer, and it's why treating online and offline media as one connected system, rather than a stack of separately-reported channels, matters more than ever as retail media's share of the budget keeps growing. Animo builds retail media, paid search, paid social, TV, out-of-home, radio and print into a single always-on causal model, so a brand can see not just what a retail media network reported, but what it actually contributed once every other channel's influence - online and offline - is properly accounted for.

Standardized dashboards will keep multiplying as more networks launch ad businesses. The only way to cut through them is to stop asking each one what it did, and start asking the model what would have happened without it.

Sources

• ANA, "Retail Media Measurement Standardization" report, Aug 18, 2026: ana.net
• MediaPost, "ANA Urges Retail Media Framework, Measurement Consistency," Aug 18, 2026: mediapost.com
• Digiday, "ANA updates efforts to standardize retail media network measurement": digiday.com
• Marketing Dive, "ANA cautions marketers against overreliance on retail media data": marketingdive.com
• Path to Purchase Institute, "ANA Releases Framework for Retail Media Measurement Standardization": p2pi.com
• Measured, "Measured Data Reveals CTV Delivers Strong Incremental ROAS Despite Minimal Spend by Enterprise Brands" (Q1 2026, 148 video advertisers): measured.com
• eMarketer, "75% of marketers say measurement is broken - AI becomes the rebuild strategy": emarketer.com
• IAB, "Modernizing MMM: Best Practices for Marketers," Dec 2025: iab.com (PDF)