Set as preferred source
Insights from Retail Media Report 2026.
Retail media has spent the past several years being framed as one of advertising’s clearest growth stories. The 2026 data complicates that narrative.
Across the US retailers tracked by Sensor Tower, retail media ad impressions fell 17% year over year to 223 billion in the first half of 2026. Amazon remained the dominant platform, but its impressions declined 16%. Walmart and Target each fell 7%.
Taken alone, those numbers suggest a market losing momentum. The underlying picture is more consequential: the decline conceals sharply uneven growth across advertiser categories, retail networks and media channels.
That distinction matters because the next phase will be shaped less by who can create the most impressions and more by who controls the most useful combination of purchase intent, first-party data, distribution and monetizable inventory.
Growth is becoming more selective
The clearest evidence comes from inside Amazon itself.
On Amazon, Shopping impressions fell 31% year over year in H1 2026, while the report’s “other CPG” grouping declined 60%. At the same time, Telecom impressions expanded 2,056%, Investing & Financial Management grew 215%, and Travel Booking Services rose 170%.
Those gains came from smaller categories. Across the tracked US networks in Q2 2026, Shopping and CPG still accounted for nearly two-thirds of impressions. CPG impressions increased 4% year over year, while Telecom grew 133%, Travel & Tourism 49%, and Financial Services 33%.
Non-endemic advertising is expanding quickly. Endemic brands still provide most of the volume.
Retail media was built around brands selling products through the retailer. It is increasingly being used by companies selling subscriptions, financial relationships, connectivity and travel outside the retailer.
That changes the economics of the market. Inventory once competed primarily for packaged-goods and retail budgets. It is now being opened to sectors with very different customer-acquisition models.
The same uneven growth appears across retail media networks (RMNs).
In H1 2026, Amazon, Walmart and Target recorded impression declines, while Lowe’s increased impressions 151% year over year. Home Depot grew 43%, Walgreens 41%, PetSmart 35% and Sephora 23%. In April 2026 alone, Home Depot impressions jumped 336% month over month and Lowe’s 441% as seasonal home-improvement demand accelerated.
Scale still matters. But scale is no longer the whole story. Specialist networks can offer something general platforms cannot always replicate: a concentrated moment of purchase intent.
For advertisers, that makes the market less about maximizing reach and more about identifying where intent is unusually dense.
Retail media is becoming acquisition infrastructure
The second structural shift is less visible but potentially more important.
Retail media was originally attractive because it created a relatively closed loop: a retailer could identify an audience, show an ad and connect exposure with a transaction occurring within the same commerce environment.
Sensor Tower’s 2026 data points toward a broader model.
Amazon devoted 35% of its offsite impressions to streaming OTT in Q2 2026, up 10 percentage points year over year. Instacart reached 23%, Target 17%, while DoorDash and Home Depot were both at 15%. Ten of the top 12 retailers studied increased their OTT share.
Retail data is effectively leaving the digital shelf.
A retailer can identify valuable audiences using shopping behavior and then activate those audiences across streaming environments where the consumer may have no immediate interaction with the retailer itself.
That starts to blur the distinction between retail media, performance advertising and traditional brand media. The strategic question shifts from where an ad appears to what the retailer’s audience data adds—and whether that advantage improves campaign economics.
Travel provides an even more interesting example.
In Q2 2026, Expedia’s Amazon retail media impressions increased 179% year over year, accounting for 75% of impressions in Amazon’s Travel Booking Services subcategory. Its share of outbound traffic from Amazon to travel-booking sites also increased year over year.
The report does not establish that advertising caused the traffic increase. The parallel trend is consistent with retail platforms serving as acquisition channels for businesses that complete transactions elsewhere.
A retail platform no longer needs to own the final transaction to monetize intent.
It can identify the customer, sell access to that audience and redirect traffic into someone else’s transaction funnel. That creates a customer-acquisition path that sits alongside search engines, social platforms and affiliate channels rather than remaining confined to retail advertising.
Paid-ad headroom is only one source of growth
Not every network follows the same commercialization strategy.
Sensor Tower reports that paid brand advertising accounts for 97–99% of Amazon’s measured onsite impressions. Walmart reduced its house-ad share from 13% in Q2 2025 to 8% a year later. Target, by contrast, devoted 34–45% of measured onsite impressions to house promotions.
Target therefore has more potential room to replace self-promotion with paid advertising. But those placements already serve a commercial purpose: promoting the retailer’s own brands, services and offers.
House-ad inventory is a trade-off, not idle capacity.
Channel strategy matters too. The report puts Walmart’s increase in onsite impression share at 27 percentage points year over year in H1 2026, while roughly 92% of Target’s measured impressions were offsite.
The report’s onsite measure covers tracked display advertising and excludes search, sponsored results and proprietary retailer ad units. Its offsite measure includes advertising on other websites, social platforms and streaming services. Amazon’s high paid-ad share therefore does not establish that its overall advertising inventory is saturated.
Networks can grow through several routes: expanding their audience, changing the paid-versus-house mix, developing formats or extending offsite reach. The executive question is which route improves advertising economics without weakening the commerce experience.
Fragmentation is becoming an operating cost
Large advertisers already spread their activity across multiple networks.
In Sensor Tower’s Q2 2026 analysis of the top 100 advertisers by industry across 32 studied US networks, more than 80% of Shopping and CPG advertisers used at least two RMNs. Half of leading CPG advertisers were active across six or more networks, while more than one-third used more than ten.
Procter & Gamble appeared across 23 RMNs. Coca-Cola used 21. PepsiCo and Nestlé each used 20.
This creates resilience against dependence on a single network, but it introduces a different constraint: coordination.
Media allocation now has to move across retailers, delivery platforms, specialist networks, streaming inventory and onsite placements while preserving a consistent view of performance. Measurement, attribution and creative operations become infrastructure problems in their own right.
There is another reason for discipline. The evidence presented here primarily measures impressions, not advertising revenue, pricing or return on investment. A 2,056% increase in Telecom impressions shows rapid growth in measured advertising activity. It does not establish how many advertisers entered the channel or whether acquisition economics improved.
That makes measurement discipline more important. Changes in impression volume are relatively easy to observe. Determining which placements create incremental demand is harder.
Service brands have to behave differently inside commerce environments
Non-endemic expansion also changes what successful advertising looks like.
Sensor Tower’s creative examples show a surprisingly consistent pattern. Telecom advertising emphasizes monthly pricing and device deals. Travel uses destination imagery paired with immediate discounts. Financial-services campaigns emphasize cashback, free benefits and ecosystem perks.
The common denominator is not creative style. It is commercial compression.
A telecom plan, financial product or travel service can be complicated. The examples reduce that complexity to an immediate, transaction-like benefit.
That creates a subtle competitive constraint for service industries entering retail media. Access to retailer data is only part of the equation. The underlying proposition has to be translated into the language of a shopper who arrived expecting to compare value, not research an abstract service relationship.
Retail media therefore changes more than channel allocation. It changes how non-retail categories package value.
Partnerships can temporarily redraw the competitive map
Food delivery shows how fluid that competitive map can become.
In February 2026, the month of the Super Bowl, Uber Eats captured 85% of Pizza impressions across the delivery networks studied. Grubhub accounted for more than 70% of Chicken Shops impressions in both February and March, which included March Madness.
By June, the picture had reversed.
DoorDash, identified in the report as an official FIFA World Cup supporter, captured 83% of Pizza impressions and 90% of Mexican & Taco impressions in the month the tournament began.
The larger signal is not about pizza.
The pattern suggests that sponsorships, consumer occasions and distribution can combine to shift advertising activity sharply between networks. A platform does not always need a permanent audience advantage if it can control the most commercially relevant context at the right moment.
That makes partnerships part of the media infrastructure rather than a separate branding exercise.
The global read requires more discipline than the headline suggests
The detailed evidence in Retail Media Report 2026 is predominantly US-based. Its overview and appendix reference UK and Canadian coverage, but the report does not present comparable regional analysis for continental Europe or Asia-Pacific.
The appendix also makes clear that the figures cover selected retailers rather than the entire market. Category and advertiser analyses exclude untracked advertisers and advertising for retailers’ own products.
That limits how far the absolute numbers should travel. The US provides a strong proof point for the operating model — first-party commerce data moving into streaming, services and external acquisition funnels — but not a ready-made growth forecast for every market.
For Europe and Asia-Pacific, the useful takeaway is therefore the architecture of the shift, not the current US growth rates.
Risk is moving from adoption to economics
The next phase of retail media carries a different risk profile:
- Inventory competition — Non-endemic expansion could increase competition for placements used by retail and CPG brands. Whether it displaces those brands depends on inventory growth, pricing and allocation.
- Commerce trade-offs — Expanding paid placements can create tension with merchandising and the shopping experience. A high paid-ad share alone does not establish saturation.
- Measurement fragmentation — As campaigns spread across more RMNs and offsite channels, impression growth becomes easier to produce than comparable performance measurement.
- Platform dependence — Reliance on individual networks exposes advertisers to changes in access, pricing and measurement.
- Regional transfer risk — The operating model may travel globally, but the US numbers should not be treated as proxies for markets the report does not measure in comparable depth.
What the 2026 data really changes
Retail media is moving beyond its original category.
It is no longer just a way for retailers to sell advertising around products. Increasingly, it is a system for monetizing commercial intent: identifying people through transaction signals, reaching them across external media and, in some cases, sending them into entirely different transaction ecosystems.
That creates several distinct sources of advantage.
Amazon has scale and a high paid-ad share in tracked onsite display inventory. Walmart is shifting its measured activity toward owned surfaces. Target has more potential room to change its paid-versus-house mix. Specialist networks such as Lowe’s and Home Depot offer access to category-specific shopping occasions. Delivery platforms combine commerce activity with events, partnerships and immediate consumption.
Advertiser activity can spread across more networks while remaining heavily dependent on a few large platforms.
For leadership teams, that argues for evaluating allocations by category, channel and incremental sales or acquisition outcomes, rather than treating impression growth as a proxy for economic performance.
And that may be the most important signal in Sensor Tower’s report. The next contest in retail media is not simply for more ad impressions. It is for control over the points where identity, context, distribution and purchase intent meet.
