Why retail media stopped being a channel and became a problem of questions
There is an uncomfortable moment that repeats itself in the conference rooms of major consumer goods companies: someone presents a dashboard filled with hundreds of retail media metrics, everyone nods, and nobody knows exactly what decision to make with any of it. The panel that CVS Media Exchange and Adweek staged at Cannes Lions this year was not a product presentation or an investment announcement. It was, rather, the public acknowledgment of that uncomfortable moment, elevated to the status of an industry diagnosis.
What the executives present had to say deserves to be read carefully, because beneath the usual diplomacy of festival panels lies an admission that is rarely heard so directly: the problem with retail media is no longer access to data. It is that nobody really knows what to do with it.
Too much data, far too few correct questions
Parbinder Dhariwal, vice president and general manager of CVS Media Exchange, framed it without any equivocation: brands already have more data than they can process. What is missing is not volume of information, but rather clarity about what one wants to answer before a campaign is even switched on.
That distinction matters more than it might appear. When a brand launches a retail media campaign without defining upfront whether it is seeking incrementality, acquisition of new buyers, or retention, what it receives at the end is a report that can justify any conclusion. The data becomes a mirror that reflects back whatever one wants to see. And in an environment where retail media spending is projected to capture close to 30% of the digital budget over the next two or three years, that ambiguity carries concrete budgetary consequences.
Cara Pratt, global president of retail and media at Circana, added another layer to the problem. There is an entrenched tendency to treat retail media exclusively as a short-term performance channel, as if its only job were to move units this week. That underestimates what the purchase signal can do when it is used to build a brand with precision, not merely to close transactions. The very same buyer behavior database can inform both a tactical activation in the store aisle and a positioning strategy that takes two years to bear fruit. The problem is that very few organizations possess the internal structure needed to use the same source toward both of those objectives simultaneously.
What the panel articulated in Cannes is, at its core, a shift in emphasis: mature retail media is not measured by how much data a network accumulates, but by how well-formulated the questions are that those data are meant to answer. It is a distinction that sounds simple and that, in practice, completely reorganizes how an investment in this channel is planned, executed, and evaluated.
The problem that no retailer can solve alone
There is a structural limitation embedded in the current retail media model that the panel named with unusual honesty. Dhariwal described it with a metaphor worth retaining: CVS is very good at measuring its own house, but it has no visibility into how the neighborhood behaves. That is to say, the CVS network can track in fine detail the behavior of a shopper within its own stores — their replenishment cycles, their categories of interest. But the moment that same person walks out the door and makes a purchase at another retailer, the signal is cut.
That is not a minor technical defect. It is the reason why attribution in retail media remains, in large part, a convenient fiction. If a brand launches a campaign through CVS Media Exchange and then registers an increase in sales, part of that increase may be occurring at Walgreens, at Target, or at Amazon, with no way for the CVS network to see it. The result is a systematic underestimation of the real effect of the campaign — or, worse, an overestimation if the shopper would have purchased at CVS anyway, regardless of whether the ad had ever existed.
The work that Circana is doing with CMX points directly at that blind spot. According to what was shared at the panel, a significant portion of the incremental return on investment generated by campaigns on CVS Media Exchange materialized in purchases made at other retailers. That is important for two reasons. First, it confirms that the effect of a retail media campaign spills beyond the boundaries of the retailer hosting it, which has direct implications for how that inventory is valued and purchased. Second, it suggests that siloed measurement by network — which remains the industry standard — is rendering invisible a substantial portion of the value that these campaigns actually generate.
The implication for brands is direct: a retailer that can only show its own slice of the story has a structural incentive to present that slice in the most favorable light possible. Cross-retailer measurement is not merely a technical improvement. It is a discipline mechanism that changes the incentives of the entire chain.
The physical store as an unconquered frontier
There is a third thread in the Cannes conversation that receives less attention than data and measurement, but that will likely move more money over the long term. The majority of CVS shoppers leave the store with a purchase in hand, and their visits last only a matter of minutes. That context — a shopper physically present, actively moving through the store, with purchase intent already activated — is one of the most valuable advertising environments in existence. And it remains, in large measure, uninstrumented.
Retail media was born in the digital world. The first networks were built on inventory in e-commerce websites, sponsored search, banners on product pages. The physical store was left behind because instrumenting it is more complex and more expensive: it requires digital signage infrastructure, real-time data connectivity, and measurement systems capable of connecting ad exposure to a purchase made at the register. It is not an insurmountable technological barrier, but it is a capital and operational design barrier that most retailers are still negotiating internally.
Dhariwal and Pratt were explicit in stating that this potential still lies ahead, not in the present. Retailers are working through the design requirements and investment implications involved in converting a physical store into a measurable medium with the same rigor as an e-commerce site. That tension between the potential value of the physical point of sale and the cost of activating it defines one of the most consequential investment debates for any retail chain over the next three years.
What makes this point particularly interesting from a consumer behavior perspective is that the physical store is not a residual channel that e-commerce is replacing. It is an environment where purchase intent is already present and exposure time is short, which is precisely the opposite of the problem that digital advertising faces — where the shopper may be in any state of mind whatsoever and attention is fundamentally dispersed. An ad in the aisle of a pharmacy reaches someone who has already decided to buy something. That carries a value that current retail media models are only just beginning to capture with rigorous data.
The next cycle will not be won by the networks with the most inventory
The retail media industry has spent years growing on the basis of a relatively simple argument: we have purchase data, we have inventory close to the point of decision, and advertising works better here than in mass media without data. That argument was sufficient to build networks worth several tens of billions of dollars over the last decade.
What the Cannes panel signals — and what analyses from firms such as Bain and Dunnhumby support from different angles — is that this argument no longer differentiates. Today, nearly every retailer with scale has a network. Nearly every network offers audiences segmented by purchase behavior. The volume of available data has already surpassed the capacity of organizations to process it with purpose.
The next consolidation cycle in retail media will not be won by the networks with the most impressions available. It will be won by those capable of answering three questions with data verified by a third party: how much of what they attribute is truly incremental, what happens to the shopper after they leave the store, and how exposure to an ad connects with a purchasing decision that takes place days or weeks later in another channel entirely.
Those are not technical questions. They are business questions that brands should have been asking from the very beginning and that, under pressure to scale budgets rapidly, were systematically deferred. What the work between Circana and CVS Media Exchange places on the table is that there is a methodological path for answering them. The question that remains — and that the market will resolve over the next two or three years — is how many retailers have a genuine interest in subjecting themselves to that level of scrutiny when the more comfortable option continues to be showing only the numbers that their own network is capable of measuring.
The brands that formulate the right questions before signing the next retail media contract are going to receive very different answers. And those different answers are going to redirect budget. That, more than any technological advance, is what defines the next phase of the sector.
