Stores Didn't Become Ad Networks. They Always Were One — Amazon Just Showed the Bill.

You scan your loyalty card at the checkout, save eighty cents on cereal, and walk out feeling like you won. What actually changed hands is more interesting. The store now knows you buy that cereal, on Thursdays, with milk and never coffee — and a brand will pay the store to put a different cereal in front of you next time, on the strength of exactly that fact. The discount was never charity. It was the store buying the one asset that turns a grocer into an advertising company: knowing, with near-certainty, what you bought.

The official story is that Amazon invented retail media around 2012 with sponsored product listings — when it launched Amazon Marketing Services for vendors⁹ — and that physical stores are now scrambling to copy a digital-native idea. Almost every part of that is backwards. Grocers were charging brands for shelf placement, trade allowances, and shopper-data targeting long before Amazon ran its first sponsored ad — slotting allowances alone became a standard grocery-industry practice as far back as the 1980s¹¹, and data-driven targeting followed when loyalty programmes matured.⁷ Amazon didn't originate the model. It digitized it, scaled it, and then did the one thing the grocers never had to: it published the number.

The business was always there. Amazon just printed the receipt.

For years, Amazon's advertising revenue lived buried inside a vague "Other" line, an open secret that analysts estimated but the company never confirmed. Then, in early 2022, Amazon broke it out as its own line for the first time — and the number was $31.2 billion.² That wasn't a financial surprise so much as a disclosure event: the market had suspected the scale; now it had the figure in black and white. And a figure does something an estimate never can. It tells every other retailer sitting on a pile of purchase data exactly how much that data is worth — and tells every brand exactly how much someone else is already collecting.

The growth confirmed it was no fluke. By fiscal 2024, Amazon's advertising services revenue had climbed to $56.2 billion, up from $46.9 billion the year before — the first year the business topped $50 billion.¹ That is roughly an 18-to-19% jump in a single year, off a base most companies would kill to have as their entire revenue. The genie was out, and it was wearing a price tag.

Why the checkout data is the product, not the ad

Here is the mechanism most coverage skips. Retail media's value isn't the ad slot — it's the certainty behind it. Every other ad network on earth is guessing at who you are and whether you bought. A retailer isn't guessing. It watches the purchase happen. Kroger's loyalty card alone captures 96% of its in-store transactions⁷, which means it isn't inferring intent from a browsing trail; it's selling brands access to the actual outcome they're paying to influence. That closed loop — show the ad, watch the sale — is the thing advertisers have chased for a century, and the store had it lying in the till all along.

Which is why this looks like such a beautiful money machine. The cost of serving one more sponsored listing against data you already collected for running the store is close to nothing. So the revenue lands almost entirely as profit. At Walmart, the global advertising business reached $4.4 billion in its fiscal 2024, growing 27% year over year⁶ — a business that has grown dramatically faster than Walmart's retail sales and now represents a disproportionate share of the company's profit. A company built on razor-thin grocery margins now earns a fast-growing share of its profit selling ads against data the groceries generate. The store became the loss leader for the ad network.

The pre-Amazon network nobody talks about

If you want the proof that this predates Amazon, look at Kroger. In 2003 — years before sponsored listings — Kroger partnered with Tesco's Dunnhumby to turn loyalty data into a commercial insights business. It acquired those assets in 2015 and built them into 84.51°, its in-house analytics arm.⁷ The premise was identical to what the industry now calls retail media: brands will pay handsomely to target real shoppers based on what those shoppers actually buy. Kroger didn't need Amazon to teach it the model. It needed Amazon to make Wall Street value it.

The gold rush has a problem: everyone is striking the same vein

The disclosure that made retail media visible also made it crowded. The total prize is genuinely enormous — US retail media ad spending reached almost $55 billion in 2024 and is forecast to grow at a 17.2% CAGR to roughly $97.9 billion by 2028, by which point it would be nearly a quarter of all US media ad spend.⁴ Globally it hit about $140 billion in 2024, a 21.8% jump that made it more than a fifth of all digital ad spending worldwide.⁵ Those numbers are exactly why the field is filling up with networks. And that is the threat hiding inside the opportunity.

The concentration tells the real story. In 2023 Amazon captured 75.2% of the US retail media market — more than ten times the share of the next-largest network.³ Everyone else is fighting over the remaining quarter, and there are a lot of everyones. As more retailers stand up networks, advertiser budgets splinter across them faster than the industry can agree on how to measure any of them. A brand that wants the proven-buyer magic now has to buy it one walled garden at a time, with no common yardstick to compare them. The asset that made retail media special — proprietary, closed-loop data — is the same thing that makes it impossible to standardize. Scarcity at the top; commoditization at the bottom.

Aren't stores really becoming full ad networks?

The seductive version of this story says the physical store itself is becoming a media channel — digital shelf screens, in-aisle targeting, the whole building monetized. It's a great image, and the numbers say go slow. In-store retail media remained below 1% of total US retail media ad spend, around $370 million in 2024, and is forecast to stay under that threshold through 2028 — held back by the heavy technology investment retrofitting a store actually requires.⁸ So the honest framing isn't "stores became ad networks." It's that store data became an ad network. The revenue lives overwhelmingly in on-site search and display, where serving an ad costs nothing. The physical aisle is the next frontier, not the current engine — and anyone selling it as today's gold rush is selling the map, not the gold.

Retail media was never invented. It was disclosed. The grocers had been quietly selling certainty for decades — proof of what shoppers buy, dressed up as a discount on cereal — and the only thing Amazon truly changed was making the number public enough that more than two hundred networks worldwide now want in.¹⁰ The asset is real and the margins are real. But the same proprietary data that makes one network priceless makes a hundred of them impossible to compare. The gold was always in the till. The question now is whether everyone rushing to dig it up leaves anything but a fragmented market and a measurement problem nobody can solve.