In 2024, Amazon captured 56 cents of every new dollar that entered the US retail media system. By 2025, that figure had dropped to 46 cents. The market grew — from approximately $51.9 billion to over $60 billion in the same period — but Amazon's share of it contracted by nearly a fifth. The giant had not shrunk in absolute terms. It had simply stopped growing faster than the category it dominated.
Justin Jefferson, Vice President of Strategy and Insights at Keen Decision Systems, put it plainly in a Modern Retail analysis: "The honeymoon phase of retail media is over. After years of explosive, low-hanging-fruit growth, the channel has reached a high-stakes maturity." The phrase landed because it described something marketers were already feeling: the certainty that concentrated spend on a single platform had been the correct strategy, until it suddenly wasn't.
The hard truth is that the more crowded the bottom of the funnel becomes, the more expensive it gets to stay there. — Justin Jefferson, Keen Decision Systems
The share drop is not a crisis. Amazon still holds approximately 77 percent of total US retail media ad spend, according to eMarketer and Statista estimates. It still processes 45 percent of US online shopping queries. Its DSP reaches 200 million customers off-Amazon properties. The structural position is intact. What changed is the marginal utility of the next dollar sent there.
The redistribution followed two paths. The first was Walmart Connect, which reported global ad revenue of $6.4 billion in 2025 — a 46 percent year-over-year increase. Its US business grew 41 percent in the fourth quarter alone. The Vizio acquisition, completed in 2024, gave Walmart a CTV inventory layer that doubled its addressable surface area. Mondelez reported a 53 percent year-over-year increase in ad-attributed sales through Walmart Connect, with incremental ROI up 29 percent. Danone began calling Walmart a "strategic partner for launching new products" — language previously reserved for Amazon.
The second path was more diffuse: a constellation of mid-sized retailer media networks that collectively accounted for 25 percent of all retail media networks operating in 2025. Target Roundel reached $2 billion in revenue trajectory, growing at 35 percent-plus through the third quarter. CVS Media Exchange adopted IAB measurement standards across both digital and in-store formats, positioning itself as the dominant health-and-wellness media network with 90 million ExtraCare loyalty members. Kroger Precision Marketing integrated fully with Google's Display & Video 360 — the first retailer to offer programmatic access to its first-party shopper data through a major DSP. Chewy, with $105 million in ad revenue in the first half of 2025 alone, became the third-largest retail media network by revenue despite operating in a single category.
Historical shopper data tells us a lot more about the consumer than a cookie ever could. — Christine Foster, KPM Group VP
Brands noticed. The share of brands working with four to six retail media networks doubled from 10 percent to 24 percent between 2024 and 2025, according to IAB Europe research. The average brand now manages six retail media networks, a figure projected to reach eleven by 2026, according to Skai's State of Retail Media report. Amazon remained the dominant node. It no longer remained the only one.
What drove the diversification was not dissatisfaction with Amazon per se. It was the economic logic of concentration. When every brand with a product to sell concentrates its performance budget at the bottom of the funnel — search, retargeting, purchase-capable audiences — the cost of occupying that space rises to meet the aggregate demand. Average Amazon Sponsored Products cost-per-click increased from approximately $0.89 in 2020 to over $1.00 by 2025, a 15 to 25 percent increase in five years, driven by algorithmic bid inflation and enterprise brand entry. In competitive categories, Sponsored Brands CPCs regularly exceeded $2.50.
The response was a format migration. Skai's Q4 2025 data showed Sponsored Brands Video delivering stronger ROI than traditional search and display formats in multiple categories. The academic research supported the shift. A paper by Qin, Pauwels, and Zhou published in the Journal of Marketing Analytics — analyzing 122,000 brands on Amazon's US marketplace — found that Sponsored Brands Video was the only ad product that reliably drove all three funnel stages: awareness, consideration, and revenue. Sponsored Display and Streaming TV drove awareness primarily. Sponsored Products drove revenue but not brand consideration. Brands were allocating 0.9 percent of budgets to video-formatted brand-building while elasticity data suggested 5.9 percent was optimal for awareness goals.
Sponsored Brands Video is the only ad product that drives all three funnel metrics simultaneously. Brands are systematically underinvesting in the one format that actually builds brands. — Qin, Pauwels & Zhou, Journal of Marketing Analytics, 2024
The correction was not simply strategic. It was structural. As bottom-funnel crowding drove CPC inflation, the ROI calculus for upper-funnel formats improved in comparison. Brands that had treated retail media as a direct-response channel began to discover, reluctantly, that it could also be a brand-building channel — and that the two uses required different formats, different metrics, and different retail partners.
A Wharton research paper by Bale, McDonnell Feit, and Bradlow provided the theoretical architecture for what was happening empirically. Analyzing eleven years of grocery sales data, they found that brand equity moderated retail media effectiveness — brands with higher equity benefited more from in-store display marketing and exhibited lower price sensitivity. The implication was counterintuitive: investing in brand equity first, then amplifying through retail media, produced better returns than using retail media alone to drive volume. Retail media, in other words, amplified whatever brand proposition it encountered. Weak brands got weak returns. Strong brands got strong ones.
This reframed the competitive dynamics of the retail media landscape. Amazon excelled at capturing demand that already existed — the shopper who knew what they wanted and searched for it. But building demand in categories where consumers had not yet formed strong preferences required a different approach. Here, Walmart's physical footprint — 90 percent of Americans within ten miles of a Walmart store — offered something Amazon could not replicate: the ability to reach consumers in a context where price and immediacy competed against digital abstraction. Target's affluent demographic offered precision in beauty, home, and discretionary categories where purchase intent was aspirational rather than transactional. Chewy offered pet owners in a relationship context that Amazon's commodity approach could not approximate.
The data from Sensor Tower's first-half 2025 analysis confirmed the strategic shift among leading brands. Nestlé allocated only 16 percent of its retail media spend to Amazon despite being a top-five global advertiser. L'Oréal prioritized Sephora and Ulta for its prestige beauty portfolio. P&G, historically one of Amazon's closest advertising partners, had shifted significant budget toward Walmart for categories where physical shelf visibility complemented digital targeting. Amazon was not being abandoned. It was being contextualized — used as a conversion engine rather than a brand-building platform.
The fragmentation of retail media spend created a parallel problem: the inability to measure it coherently. A CIMM and 4As study of 197 marketing executives in March 2026 found that despite unprecedented access to analytics platforms, advertisers were overwhelmed by competing measurement systems, platform-specific metrics, and black-box methodologies producing multiple incompatible versions of campaign performance. More data was producing less confidence. Sixty-eight percent of advertisers already tested retail media's upper-funnel capabilities, but only 35 percent had implemented incrementality testing — the methodology required to distinguish actual sales lift from baseline purchase behavior.
Forrester Research estimated that fragmentation was costing the industry approximately $28 billion annually in misallocated spend — nearly 20 percent of total retail media investment. The source of the loss was not the existence of multiple networks. It was the absence of cross-network measurement standards that would allow a brand to understand whether a dollar spent on Walmart Connect was complementing or cannibalizing a dollar spent on Amazon. "In retail media, no one sees the whole wallet," as one industry analysis noted. Even the largest commerce companies could observe only their own transactions. This incomplete view was structurally embedded — retailer media networks competed on their data assets, creating no commercial incentive for interoperability.
Sixty-eight percent of advertisers already test retail media's upper-funnel capabilities. Fewer than 40 percent have implemented incrementality testing. — Bain & Company / Affinity, 2025
The measurement gap was not merely an operational inconvenience. It was a strategic obstacle. Ryan Verklin, Paid and Retail Media Lead at Bayer, described the situation to The Drum: "Right now, all the commerce media networks have their own measurement silo. We need a holistic view of measurement instead of the silos that currently exist." The observation reflected a broader industry anxiety: brands were making budget allocation decisions across six to eleven retail networks without a coherent framework for evaluating return on investment across the portfolio. The result was a flight to the simplest available metric — last-click attribution — which systematically over-credited bottom-funnel channels and under-credited the upper-funnel activities that had actually created the demand being attributed.
The trajectory points toward consolidation. Analysts at Solomon Partners project US retail media reaching $97 billion by 2028, with the market consolidating around five dominant ecosystems: Amazon, Walmart Connect+, Target Roundel plus Shopify, Kroger plus the Independent Alliance Network, and a tier of CTV and off-site specialists. The era of 200-plus competing retail media networks is structurally temporary — the operational complexity of managing eleven networks per brand has already exceeded what most marketing teams can effectively handle, and the measurement crisis is accelerating consolidation around networks that can demonstrate causal ROI rather than correlation.
Amazon's position in this consolidated future remains strong but no longer unassailable. The FTC v. Amazon trial, scheduled for October 2026, carries potential remedies that could constrain how Amazon structures its advertising marketplace — particularly around the fee extraction model that has historically funded its below-cost retail operations. If the government prevails, the economics of Amazon advertising change. If Amazon prevails, the structural dominance continues. Either outcome will shape how the remaining $97 billion in retail media spend gets distributed.
What is clear is that the 56-to-46 drop was not an anomaly. It was a leading indicator. The brands that had concentrated spend on Amazon because concentration was the optimal strategy are now discovering that the optimal strategy has changed — because the bottom of the funnel is crowded, because upper-funnel formats deliver better ROI than the industry understood, because first-party data from multiple retail contexts produces more precise targeting than any single network, and because the measurement crisis is forcing a more sophisticated allocation than "send everything to Amazon" ever required. The money is moving. The giant is shrinking. The category is growing around it.