Retail Media Has Reached the Audit Stage
The most useful finding in ISBA and mediasense’s new study of UK retail media is not that the channel has momentum. That much has been obvious for some time.
It is that, after a period in which retailers have built impressive first-party data propositions and advertisers have moved budget towards them, the market has reached a more exacting phase. Brands are asking a question that growth forecasts cannot answer: can we compare what we are buying?
The study, published on 8 September, audited more than 200 capability points across Boots, Lidl, Nectar 360, Ocado and Tesco. It found mature first-party data capabilities and improving closed-loop measurement, but substantial variation in definitions, reporting, attribution and service models. No participating network combined every leading practice.
That is not an indictment of retail media. It is a predictable consequence of a category developing quickly across businesses with different customer bases, operating models, data assets and commercial priorities. But it changes the task for advertisers. The question is no longer simply whether retail media deserves a place on the plan. It is whether the organisation has the measurement discipline to prevent a promising channel from becoming another opaque line in the budget.
The sale at the end of a campaign is valuable evidence. It is not, by itself, proof of advertising effect.
Closed-loop reporting is not the same as causality
Retail media has a compelling proposition because it appears to join what marketers have long wanted kept together: audience data, advertising exposure and a purchase outcome. For a consumer-goods brand, seeing product sales after an on-site search placement or retailer display campaign can be far more tangible than reading a conventional media dashboard.
Yet this apparent completeness can create a category error. A transaction recorded by the retailer is undoubtedly more proximate to commercial behaviour than an impression, click or completed video view. It does not automatically establish that the campaign caused an additional sale.
Some shoppers would have bought the product anyway. Others may have switched from a competing item, brought forward a planned purchase, or responded to a price promotion, improved availability or a stronger in-store display rather than the advertising itself. A campaign can therefore look highly efficient in a retailer’s attribution view while delivering less incremental growth than its reported return implies.
This is not an argument for discounting retailer data. It is an argument for treating it as a distinct form of evidence, with clear limits. Retail media reports often answer: *which sales occurred after an exposure under this platform’s rules?* A commercial investment decision needs an additional answer: *what changed because we spent the money?*
That distinction matters most when budgets are being shifted from other activity. If a brand moves money from broad-reach advertising, price investment or search to retail media, it needs to understand the counterfactual. What would sales, penetration, category share or margin have looked like without the campaign? Without that discipline, the channel risks being rewarded for harvesting demand generated elsewhere.
The new ISBA study’s finding that retailers define core metrics differently, including attributed sales, should therefore be read as more than a reporting inconvenience. It goes to the heart of investment allocation. Different attribution windows, treatment of returns, product substitutions, promoted items, repeat buyers and organic search traffic can materially alter a claimed outcome.
Comparison is now a commercial capability
Retail media networks do not need to become identical. A retailer with frequent grocery missions, a beauty specialist and an online-led proposition should not be expected to offer the same audience, inventory or purchase signals. Their differences may be precisely what makes each strategically useful.
But difference in proposition is not a licence for ambiguity in reporting.
Advertisers should be able to see where like-for-like comparison is appropriate and where it is not. That requires a common reporting language, not necessarily a common technology stack. ISBA’s proposed roadmap – including a shared glossary, common reporting principles, transparency around buying tiers and a recurring maturity benchmark – is pragmatic for that reason.
A shared glossary would make the basic units of a campaign legible. What counts as an attributed sale? Is revenue gross or net? Is performance reported at product, brand or category level? What was the attribution window? Which audiences were eligible? Was delivery self-service, managed service or dependent on a minimum spend?
These may sound operational, but they determine whether a media team can place retail media alongside paid search, social, connected television, promotions and shopper activity in a serious planning discussion. They also matter to procurement teams negotiating commercial terms and to finance leaders deciding whether reported media returns warrant additional investment.
The industry’s emerging European measurement standards point in the same direction. A common vocabulary can support comparability without erasing the legitimate variety of retail environments. More importantly, it gives advertisers a basis for asking better questions before a campaign launches, rather than trying to reconcile irreconcilable dashboards after the money has been spent.
The practical ambition should not be a single universal return-on-ad-spend number. That would create a false precision of its own. It should be an agreed minimum disclosure standard, accompanied by a clear account of methodology and confidence.
The ownership problem sits inside the advertiser
Retail media’s measurement challenge is also an organisational one. At many brands, the channel touches teams that have historically worked to different objectives: ecommerce, customer teams, trade marketing, media, data and insight, sales and procurement.
Each brings a valid perspective. Sales teams value retailer relationships and distribution outcomes. Ecommerce teams focus on conversion and digital shelf performance. Brand teams look for reach, consideration and long-term demand. Insight functions should test whether the audience and behavioural assumptions are sound. Finance needs a view of profitable growth rather than sales revenue alone.
When those perspectives are not brought together, retail media gets assessed through whichever metric happens to be closest to the budget holder. It can be treated as a retailer-funded commercial activation, a performance channel, a media buy or a data partnership. In reality, it can be all four – which is exactly why governance matters.
The most effective response is not another steering group with an expansive remit. It is a concise decision framework agreed before planning cycles begin. Brands should identify which business question each retail media investment is intended to answer: recruiting new buyers, protecting category share, launching a product, raising basket value, improving distribution productivity or converting existing demand.
The measurement approach should follow that question. A short-term listing support campaign may reasonably prioritise availability, search visibility and sales velocity. A brand-building campaign aimed at new households needs different evidence: incremental reach, buyer recruitment, repeat behaviour and its relationship to wider media exposure. Neither should be judged solely on the other’s terms.
This is where market research has a particular role. It should not be drafted in merely to validate an end-of-campaign claim. Researchers can clarify the purchase journey, define meaningful audience segments, identify the barriers that advertising can plausibly influence and design experiments that distinguish correlation from effect. Retailer data is powerful because it captures behaviour; research helps explain the behaviour and test its significance.
Build an evidence ladder, not a dashboard contest
For advertisers, the immediate priority is to establish a measurement architecture that is proportionate to spend and decision risk.
At its base should sit delivery evidence: eligible audiences, impressions, reach, frequency, placements, cost and campaign set-up. The next layer is commercial evidence: product availability, sales, basket behaviour, margin where possible, and effects by customer type. Above that comes causal evidence, such as holdout tests, geographically matched tests, carefully designed control groups or econometric analysis. Finally, brands need strategic evidence: whether activity recruited valuable buyers, strengthened consideration or changed the category’s competitive dynamics.
Not every campaign needs the top tier. Experimental designs take time, access and analytical care. But no advertiser should repeatedly make major reallocations using only platform-attributed sales. The appropriate standard rises with the scale, novelty and irreversibility of the decision.
Networks have a role here too. Greater transparency will not weaken their proposition; it will make it more investable. Clear methodology, configurable reporting definitions, explicit limitations and access to robust test-and-learn options give sophisticated advertisers confidence to commit for longer. They also reduce the friction that comes from every brand developing its own private interpretation of results.
The UK market’s next contest will not be won only by the networks with the largest audiences or most prominent digital shelf. It will increasingly favour those that make their evidence easiest to interrogate.
Retail media has already made the leap from a retailer-side commercial opportunity to a central concern for media and brand investment. The new study shows that its infrastructure is becoming more capable. The harder, more valuable work is to make its claims comparable, challengeable and connected to the decisions advertisers actually have to make.
That is the audit stage: not a brake on growth, but the condition for more durable growth.
Comments