In 2017, the phrase “retail media” showed up in 19 of the stories we collected at Advertising Perspectives. In 2024 it showed up in 195.
Search was the first wave of digital advertising money. Social was the second. Retail media, or commerce media if you prefer the newer and broader label, is the third, and our archive caught the whole arc: the wake-up call, the gold rush, the growing pains, and now the part where everyone starts asking for fewer of them.
We have been building marketing technology long enough to have watched programmatic go through exactly this cycle. Retail media is following the same script, almost scene for scene. That’s useful, because it means we already know how the next act goes.
The count
Stories in our archive that mention retail media, by year:
| Year | Stories |
|---|---|
| 2017 | 19 |
| 2018 | 13 |
| 2019 | 21 |
| 2020 | 26 |
| 2021 | 40 |
| 2022 | 85 |
| 2023 | 138 |
| 2024 | 195 |
| 2025 | 191 |
The mix of outlets we collect changes from year to year, so read this as a shape rather than a precise measure. The shape is clear enough: flat for four years, then a steep climb from 2021 to a plateau. The plateau is where it gets interesting.
2017–2019: Amazon wakes everyone up
Retail media started as an Amazon story. In August 2017, Digiday reported that L’Oréal UK was shifting some of its search budget to Amazon. A month later it published Amazon’s pitch deck to ad buyers, and that fall it described Amazon readying its blitz on the ad industry.
The idea was simple and, in hindsight, obvious. A search on Amazon is a shopping search. The person typing it is closer to buying than almost anyone a brand can reach anywhere else, and Amazon knows whether they bought.
Other retailers noticed. In 2017 Walmart quietly grew as an advertising force, and by December Digiday was warning Amazon and Walmart that Kroger was building an advertising business too. AdExchanger put the ambition in plain terms a few days later: Kroger wanted to be a walled garden, and it would be an uphill battle.
By 2019 the pattern was set. Walmart was bulking up its ad business. Target was courting advertisers with first-party data. And Amazon’s ad sales leader, Seth Dallaire, jumped to Instacart. When the people who built the first network start leaving to build the next one, you know a category has formed.
2020–2022: Everyone builds a network
The next three years were the build-out, and the timing wasn’t a coincidence. This is the same stretch in which Google announced the end of the third-party cookie, and identity got harder for everyone who didn’t have logged-in customers. Retailers had logged-in customers and purchase data. They had exactly the asset the rest of the industry was afraid of losing.
The early versions were rough. In January 2020, Walmart launched a self-serve ad platform, but without its retail and audience data, “yet.” That one headline captures the whole era: the store front opened before the plumbing behind it was finished.
By 2021 the networks had names and leaders with a point of view. Kroger Precision Marketing’s Cara Pratt told Digiday that “content and commerce are converging.” Instacart added display ad units and programmatic supply in early 2022. That summer, Omnicom closed out Cannes with a Kroger Precision Marketing deal. The agencies had arrived, which meant the budgets had too.
2023–2024: The growing pains
Then came the part every new channel goes through, the part where buyers start asking hard questions.
The first was measurement. In April 2023, GfK’s Bhaumik said in a Beet TV interview that retail media networks should not mark their own homework. That phrase stuck because it named the problem exactly. A retailer sells you the ad, runs the ad and then tells you how well the ad worked, using data only it can see.
The second was fragmentation. Every retailer with a loyalty program now had a media network, each with its own ad formats, data rules and reporting. By April 2024 Digiday’s research found that retail media networks were competing for ad dollars, but marketers were unsure where to spend.
The independents moved to fill the gap. Criteo, which had built its business on cookie-based retargeting, set out to unify retail media with a DSP and a set of data tools, and in 2024 it split out retail media revenue for the first time. LiveRamp argued that data collaboration would drive true scale.
The biggest players went the other way and got bigger. In February 2024, Walmart agreed to buy the TV maker Vizio, and AdExchanger read it as Walmart finally placing its stake in the ad industry. The same day, AdExchanger’s headline for the bigger picture was when walled gardens collide.
2025–2026: The consolidation talk
By 2025, story counts flattened at the top and the tone of the coverage changed. The question was no longer whether to buy retail media. It was how to buy less of it, better.
The Home Depot introduced another acronym, “ROMO,” as it renegotiated how retail media is measured. Target’s Roundel showed ad revenue growing even as the broader business struggled, which tells you how much retailers now depend on it.
And then the word everyone was using was consolidation. Instacart’s Castelli said the cure for retail media’s growing pains is consolidation. Instacart’s Silverblatt said retail media must prove its value through incremental reach and outcomes. WPP Media’s Lauren Lavin said networks face consolidation as the “grading your own homework” era ends. Digitas’ Chelsea Monaco called it programmatic-style consolidation.
Into 2026, the themes are self-serve and off-site. Criteo’s Sherry Smith called 2026 a turning point for unified, self-serve retail media. Instacart is targeting fragmentation with an off-platform push, and Walmart Connect is arguing that retail media must live off-site. Just this month, AdExchanger’s take on DoorDash was that it had bought the ad business it couldn’t build for itself.
We’ve seen this movie
“Programmatic-style consolidation” is the right comparison, and it’s worth taking seriously, because programmatic tells us how this ends.
Programmatic started with a flood of new supply, each source with its own rules. Buyers couldn’t work with hundreds of them, so aggregators and standards emerged, and the market sorted itself into a few large platforms plus specialists. Measurement started self-reported and moved, slowly and under pressure, to independent verification. The companies that came out ahead weren’t the ones with the most inventory. They were the ones whose data buyers trusted and whose inventory was easy to buy.
Retail media is at the same point programmatic reached about a decade ago. Here’s what we expect to carry over.
Fewer networks will get most of the money. Not because the smaller ones lack good data, but because buyers can’t manage fifty of them. Expect more partnerships, more shared platforms and more acquisitions like the ones we’ve just seen.
Independent measurement will stop being optional. “Grading your own homework” worked while budgets were experimental. It doesn’t work once retail media is a line item a CFO looks at. Clean rooms and third-party verification are how that trust gets built.
Off-site is where the growth is. A retailer’s own site and app only hold so many ads. The bigger opportunity is taking purchase-based audiences out into CTV, social and the open web, which is why Walmart bought a TV company and why “off-site” is the word of the year.
The real competition is data operations, not ad formats. This is the part we think gets the least attention. Every serious network has a sponsored-product slot and a self-serve console now. What separates them is how well they understand their own customers: which audiences exist, how fresh they are, what they’re worth to which advertiser, and how quickly a sales team can turn that into a package. That’s a customer data problem, and most retailers still answer it with spreadsheets, analysts and a two-week turnaround.
The part nobody puts on a slide
We spend most of our time inside the data platforms behind this market, and the pattern is consistent. Retailers, publishers and consumer brands have spent the last few years moving loyalty, transaction and engagement data into a CDP or a cloud data platform. The data is unified, consented and governed. That’s the hard engineering, and it’s mostly done.
What isn’t done is the commercial step on top of it. A network’s sales lead wants to know what the loyalty audience is worth this quarter and which segments fit the brands in the pipeline. Today that question becomes a ticket, the ticket becomes a query, and the answer shows up after the deal has moved on.
Retail media’s next act will reward the networks that can answer those questions in minutes instead of weeks.
Every headline in this piece links to the original story. Browse the full archive of retail media coverage, or the related coverage of commerce media, Criteo and Walmart Connect.
Kevin Reynolds is the founder of Advertising Perspectives, MadTech Consulting and data4orce.ai. He is a marketing technology architect and C-level advisor, with a focus on Salesforce and data-driven advertising.