Retail media networks just crossed a threshold that's forcing every major brand marketer to rethink their budgets. What started as sponsored product listings on Amazon and Walmart.com has quietly - and then very suddenly - become a $52 billion advertising category, with the most aggressive growth happening entirely off the retailer's own digital properties.

Editorial Analysis: Our analysis suggests that off-site retail media now accounts for roughly 34% of total retail media revenue - a share that has more than doubled since 2021 - and will likely surpass on-site inventory spend by 2026. The brands moving fastest aren't waiting for standardized measurement frameworks; they're building proprietary clean room infrastructure to justify the shift internally. AdWire estimates that the top ten retail media networks collectively control first-party purchase data covering over 180 million unique U.S. households, a data asset no traditional digital channel can credibly replicate.

The shift is structural. Retailers are no longer content to monetize attention only when shoppers are already on their platforms. They're extending their first-party data into connected TV, programmatic display, social media, and physical in-store screens - turning the closed-loop attribution model that made retail media so attractive into something that reaches consumers across their entire media diet. It's a fundamentally different pitch to brand advertisers, and most of them are listening.

"We're seeing retail media evolve from a performance channel into a full-funnel media solution," said Maya Chen, VP of Commerce Media Strategy at GroupM. "Brands that were allocating retail media budgets in a silo are now asking us to unify that spend with their broader brand campaigns because the data quality is simply better than what they're getting from traditional digital channels."

The economics here aren't complicated. On-site inventory - those sponsored slots at the top of a search results page - is finite. There are only so many positions on a digital shelf, and the biggest consumer packaged goods players have already bid that inventory to uncomfortably high CPMs. The margin pressure is real, and smaller brands can't compete. Off-site solves the supply problem by giving retailers effectively unlimited inventory to sell against their own audience data.

Kroger Precision Marketing, Albertsons Media Collective, and CVS Media Exchange have all announced significant expansions of their off-site programmatic capabilities in the past eighteen months. The mechanism is straightforward - a retailer's purchase-based audience segments get activated through demand-side platforms and matched against broader web inventory, CTV supply, or even streaming audio. The brand gets the attribution precision of a loyalty card database. The retailer collects a data licensing and media fee. Nobody's complaining about the arrangement. At least not publicly.

Here's the uncomfortable truth, though: many brands are paying twice for the same consumer. A CPG company might already be buying CTV inventory through a traditional DSP, reaching an audience modeled from third-party data, and then separately activating a retailer's audience extension program that's reaching overlapping households. The de-duplication problem is real, it's largely unsolved, and the retailers selling these programs have very little incentive to fix it.

"The industry needs to get honest about frequency and overlap," said James Okafor, Head of Measurement Innovation at Circana. "Clients are asking us to validate whether retail media off-site buys are truly incremental, and the answer is often more complicated than the pitch decks suggest."

"Off-site retail media is the most oversold and underleveraged product in digital advertising right now - oversold because networks are papering over serious overlap and frequency problems, and underleveraged because the brands actually doing the measurement work are finding legitimate incrementality that justifies a 20 to 30 percent budget shift away from traditional programmatic. The gap between those two groups of advertisers is only going to widen." - Rachel Voss, SVP of Commerce Strategy at Publicis Media Exchange

Nowhere is the expansion more aggressive - or more interesting - than in quick commerce. Instacart, DoorDash, and Gopuff have built retail media networks on top of delivery infrastructure, and the attribution loop they can offer is genuinely tighter than almost anything else in digital advertising. Someone sees a sponsored product in the app, adds it to their cart, and takes delivery within thirty minutes. The conversion signal is clean. The path to purchase is essentially one click long.

Instacart's Ads platform reported that brands using its off-platform ad products - display and video inventory that runs outside the app - saw measurable basket lift among households who then shopped on Instacart within seven days. That kind of closed-loop measurement, running across off-site inventory, is what separates quick commerce networks from legacy retail players still stitching together panel data and modeled attribution.

DoorDash has been particularly bold, pushing into CPG sponsorships that look more like brand partnerships than traditional performance media - co-branded campaigns, first-to-market product launches, exclusive window placements. It's not cheap. But brands that sell through convenience channels are finding that quick commerce audiences skew younger, urban, and higher-income in ways that justify the premium.

The next frontier - and nobody quite knows how big this gets - is in-store. Physical retail media, meaning digital screens in aisle, at checkout, and on shopping carts, is projected to grow faster than any other retail media format through 2027. Brands can target based on real-time shelf availability and local promotional pricing, which makes the creative execution genuinely dynamic in ways that online formats have claimed to be for years but rarely delivered. The idea that a brand could suppress an ad for a product that's actually out of stock on that specific store's shelf, in real time, is the kind of thing that sounds like a feature but might quietly become the standard that every other media channel gets measured against.

  • For Brands and Advertisers: The overlap and de-duplication problem won't solve itself. Brands investing in off-site retail media without a clean room strategy are almost certainly funding audience redundancy - and the retailers on the other side of that transaction have no structural incentive to flag it.
  • For DSPs: The acceleration of retail media off-site spend is a direct competitive threat to standard programmatic buying. Platforms that don't build native integrations with major retail media networks risk losing budget to walled-garden audience extension products that offer comparable reach with demonstrably better attribution.
  • For Publishers: Premium publishers are increasingly becoming the supply layer for off-site retail media activation. That's a short-term revenue opportunity and a long-term using problem - once a publisher's inventory is bundled into a retailer's audience extension product, the retailer controls the pricing narrative.
  • For SSPs: The in-store digital screen opportunity is largely unaddressed by existing supply-side infrastructure. Whoever builds credible measurement and transaction standards for physical retail media inventory first will have a meaningful first-mover advantage in a format that could represent $8 billion or more by 2027.

The $52 billion figure will look conservative inside of three years. The retailers that move aggressively on in-store digitization and CTV integration - Walmart Connect and Kroger Precision Marketing are the names to watch - will pull further ahead of the field while mid-tier grocery chains scramble to license the infrastructure they should have been building since 2022. The more immediate inflection point is measurement: expect at least two major industry coalitions to release competing retail media attribution standards before the end of 2025, and expect neither of them to fully satisfy the buy side. Brands that wait for consensus before committing will have already ceded ground to competitors who made peace with imperfect data and moved.