Quick answer: Kevel, Dollar General and The Trade Desk have launched a partnership that connects on-site and off-site retail media into a single, measurable campaign. Matthew Fantazier, VP of Retail Data Partnerships at The Trade Desk, told Kevel CEO James Avery that brands buying across retail media today are managing far more complexity than the industry usually admits, closer to 1,500 separate platforms once every on-site, off-site and in-store login inside each network gets counted. He also argues that share of category, not raw sales, is the metric that actually shows whether a brand is winning, and that “retail media” as a category, may not survive the next five years.
For years, brands buying retail media have lived with a familiar trade-off: reach a shopper on the open internet through one system, then hope to reach them again once they land on a retailer's own site through an entirely different one, with no way to tie the two together. Kevel CEO James Avery has joked for a while that the industry is building a thousand walled gardens. On a recent episode of Unlocking Retail Media, he sat down with Matthew Fantazier, VP of Retail Data Partnerships at The Trade Desk, to talk about a new partnership between Kevel, Dollar General and The Trade Desk built to close exactly that gap.
Fantazier spent nineteen years on the advertiser side at J&J and Kenvue before moving to The Trade Desk, where his job is helping brands unlock and activate first-party commerce data at scale. Below are the questions his conversation with Avery actually answers.
What Is the New Partnership Between Kevel, Dollar General and The Trade Desk?
Quick answer: It lets a brand run one coordinated campaign that reaches a shopper across the open internet and again inside Dollar General's own environment, with shared measurement and identity instead of two disconnected buys.
The idea started with something Fantazier calls an obvious gap. Retailers like Amazon already let advertisers buy upper-funnel media and then follow a shopper onto product pages and search results inside one connected system. Most of the industry hasn’t attempted to build that type of ecosystem for anyone else. Fantazier raised the idea in his first week at The Trade Desk and found a ready partner in Austin Leonard at Dollar General, who had been chasing the same idea from the retailer's side. Leonard has called it his white whale.
The two connected in September of last year, and Fantazier met with Kevel soon after over coffee in London. All three sides agreed on the same non-negotiables: the campaign had to tie identity together, measure performance across both environments, and actually change how a brand invests, not just add a thirty-ninth place to buy the same inventory.
The result, now live, lets a brand target an audience once, reach them through The Trade Desk's open internet inventory, and then serve that same campaign when the shopper lands on dollargeneral.com, with one shared view of what worked.
“It's all one campaign.” - Matthew Fantazier
That is a meaningful shift from how most retail media budgets get spent today, split across a shopper team, a brand team, and whatever system each one happens to use.
Why Is Retail Media's Fragmentation Problem Bigger Than People Think?
Quick answer: It's not 300 retail media networks. Once you count the separate on-site, off-site, audio, and in-store systems inside each one, the real number is closer to 1,500 platforms.
Fantazier and Avery agree that most conversations about retail media fragmentation understate the problem. The industry counts networks. It rarely counts the systems inside each network. Avery pointed out that a single retailer often requires a separate login for on-site display, another for off-site, another for audio, and another for in-store, meaning the real number of platforms a brand has to manage is several times the number of retailers.
For most brands, that complexity does necessarily get solved. It's outsourced, to a managed service arrangement where the brand hands over a budget and hopes for the best, which is its own kind of inefficiency.
Why Should Brands Measure Share Instead of Sales?
Quick answer: Sales rise and fall with seasonality and weather. Share tells a brand whether it's actually winning or losing against competitors in the same aisle or on the same digital shelf.
Fantazier pushes brands to think in terms of category share rather than raw sales volume. Sales numbers move for reasons that have nothing to do with a brand's own marketing: a cold winter, a competitor's stockout, a pricing change nobody controls. Share strips that noise out. It answers a simpler question: when a shopper walks down the aisle or scrolls a digital shelf, are they choosing you more often than before, relative to everyone else on that shelf.
That view gets sharper at the regional or even store level. A brand might be winning share of a category in the Midwest at Dollar General while losing it in the Deep South, which points to a messaging or assortment problem rather than a product problem.
“Ultimately, the measure of a brand's success is your share. It's not your sales.” - Matthew Fantazier
Once a brand knows where it's losing share, retail data becomes the tool for closing that gap rather than just reporting on it.
Is ROAS Still Worth Tracking in Retail Media?
Quick answer: As one signal among several, yes. As the only measure of success, no. Fantazier points brands toward incrementality, new-to-brand data, and brand equity instead.
Fantazier isn't dismissive of ROAS outright, but he's clear about its limits. Used on its own, it risks pushing brands to under-invest in upper-funnel awareness or ignore new-to-brand buyers entirely, since neither shows up cleanly in a return-on-spend calculation. What he wants brands asking instead is what changed because of the investment, which is the definition of incrementality.
New-to-brand data gets a more careful treatment. Fantazier has been skeptical of it in the past, since a buyer who's new to a brand at one retailer may already be loyal to that brand elsewhere. But he sees real value in a narrower version of the idea: a shopper who buys a competing product at a given retailer and has never bought yours is a real opportunity, even if converting them isn't purely incremental in the strictest sense.
None of this replaces ROAS. It just refuses to let ROAS be the whole conversation.
Will “Retail Media” Still Be a Term the Industry Uses in Five Years?
Quick answer: Fantazier doesn't think so. He expects the category to fold into a single, broader media conversation that spans CTV, the open internet, and retail inventory, without a special vocabulary carved out for retail alone.
Fantazier's prediction is a direct extension of what the Kevel, Dollar General and Trade Desk partnership is built to prove: that retail inventory works better as one piece of a brand's full media plan than as a separate budget line with its own rules. He still expects trade-dollar-funded activation aimed squarely at driving sales at a specific retailer to keep existing. What he expects to disappear is the idea that retail media needs to be planned, bought, and measured in total isolation from everything else a brand does in CTV, audio, or the open internet.
Connected TV plays a growing part in that shift. Fantazier has watched CTV go from a channel nobody took seriously to roughly half of all television viewing time, now buyable programmatically in a way that was not possible even a few years ago.
“I don't think we're gonna be talking about retail media in five years. I think that's gonna become an archaic term.” - Matthew Fantazier
Whether or not the label survives, the underlying shift, treating retail inventory as one connected part of a media plan instead of a silo, is already underway.
Key Takeaways at a Glance
What is the new Kevel, Dollar General and Trade Desk partnership?
A single campaign that reaches shoppers on the open internet and again on Dollar General's own site, with shared identity and measurement.
How fragmented is retail media really?
Roughly 1,500 platforms once you count the separate on-site, off-site, audio and in-store systems inside 300-plus networks.
What metric matters more than sales?
Share of category, since it strips out seasonality and shows whether a brand is winning or losing against competitors.
Is ROAS still useful?
As one signal among several, not as the only measure. Incrementality and new-to-brand data matter more.
Will “retail media” still be a category in five years?
Fantazier doesn't think so. He expects it to merge into one broader media conversation spanning CTV, open internet and retail inventory.
Conclusion: Stop Treating Retail Inventory as a Silo
The Kevel, Dollar General and Trade Desk partnership is a bet that retail media works better as a connected part of a brand's media plan than as its own walled-off budget line. Fantazier's broader argument goes further: the vocabulary the industry built around retail media, separate teams, separate systems, separate reporting, may not survive contact with how fragmented that world has actually become.
The retailers and platforms that get ahead of that shift now, connecting on-site and off-site buying, identity, and measurement, will be easier for brands to invest in as the industry consolidates around fewer, better-connected systems.
Listen to the Full Conversation on Unlocking Retail Media
For more insights like these, tune in to the full episode of Unlocking Retail Media, the podcast where Kevel CEO James Avery sits down with industry leaders and innovators shaping the future of retail and commerce media.
Listen to the full conversation with Matthew Fantazier on Unlocking Retail Media.
