Guide
The Retail Media & Commerce Media Playbook (2026)
Retail media is the fastest-growing major ad channel in 2026, but Amazon and Walmart are absorbing nearly all the new money — this playbook covers network selection, onsite versus offsite budgets, iROAS measurement, the agentic-commerce connection, and how mid-size brands survive fragmentation.
The Retail Media & Commerce Media Playbook (2026)
Short answer: US retail media ad spend will hit roughly $71 billion in 2026, up about 18% year over year — faster growth than search or social. But the money is concentrating hard: eMarketer projects Amazon and Walmart will capture 89% of incremental retail media spend this year. The playbook for brands is straightforward. Anchor on one or two scaled networks where your shoppers actually buy. Fund onsite search to defend share, then extend offsite only where incrementality data supports it. Demand iROAS, not the network's self-graded ROAS. And treat retail media as the training ground for agentic commerce, because AI-referred shoppers already convert 40% better than traditional channels.
Retail media is no longer an experiment line. It is a market with two winners, a scaled second tier, and a long tail you should mostly ignore.
How big is retail media in 2026, and why is it still growing?
The numbers are unambiguous. eMarketer's H1 2026 forecast puts US retail media at $71.09 billion in 2026, up from $60.32 billion in 2025 — roughly 18% growth while search and social grow slower. The IAB/PwC Internet Advertising Revenue Report (April 2026) counted commerce media at $63.4 billion for full-year 2025, up 18% inside a record $294.6 billion US digital ad market. Globally, WARC projects retail media climbs 12.4% in 2026 to $196.7 billion — about 16% of all ad spend worldwide. Dentsu's May 2026 forecast lands in the same range: 12.3% global retail media growth against 5.0% growth for advertising overall.
Three things drive this. First, retail media sits at the point of purchase, so budgets shifting from linear TV and mid-funnel display find measurable outcomes here. Second, retailer first-party data survived signal loss better than anything else in the ecosystem. Third, retailers need the margin: at Walmart, advertising and memberships now contribute roughly a third of operating income.
Who actually wins — and why concentration changes your plan
Growth is not evenly distributed. Amazon remains dominant: eMarketer projects its retail media revenue will exceed $75 billion by 2028, more than $65 billion ahead of the next-largest network. Walmart is the clear number two and growing faster — global ad revenue rose 37% in Q1 FY27 (the quarter ended April 30, 2026), with Walmart Connect US up 44% excluding Vizio. That is roughly double Amazon Ads' growth rate in the same period.
Meanwhile, the long tail is falling behind. Dozens of mid-tier and specialty RMNs are fighting over the ~11% of incremental spend the top two don't take. Many will consolidate, plug into Amazon or Criteo infrastructure, or quietly become inventory resellers. For advertisers, the implication is blunt: every additional network adds ops overhead, minimum spend commitments, and another self-interested measurement methodology. Concentration is the market telling you where the efficient frontier is.
Which networks deserve your budget?
| Network | Scale (2026) | Strengths | Measurement | Best for |
|---|---|---|---|---|
| Amazon Ads | ~77% of US retail media (eMarketer); $75B+ projected by 2028 | Largest closed loop; DSP reach across Prime Video, Netflix, Roku, Disney, Spotify; authenticated graph claims 90% of US households | Amazon Marketing Cloud (AMC) enables custom incrementality work; strongest self-serve tooling | Any brand sold on Amazon; increasingly full-funnel via CTV |
| Walmart Connect | #2 US network; US ads +44% ex-Vizio in Q1 FY27 | Grocery + mass reach, in-store network, Vizio CTV inventory, strong omnichannel (pickup/delivery) data | Improving; Walmart Luminate/Scintilla data access; incrementality tooling maturing | CPG, grocery, household; brands with heavy Walmart shelf presence |
| Target Roundel | Mid-single-digit share, scaled second tier | High-loyalty guest base, strong for style/home/beauty; good offsite via partners | Solid closed-loop reporting; smaller data footprint than top two | Brands with strong Target distribution and design-led categories |
| Instacart Ads | Leading pure-play grocery marketplace network | High-intent basket data across hundreds of retail banners; strong for launches and share-of-search | Category-leading grocery attribution; limited offsite reach | CPG food/beverage without direct retailer scale |
| Kroger Precision Marketing / regional grocers | Long tail | Loyalty-card data depth in grocery regions | Variable; often third-party-verified | Regional CPG plays; test-and-learn only |
The rule for most brands: one anchor network (usually Amazon or Walmart, dictated by where your revenue lives), one secondary network with genuine distribution overlap, and nothing else until both are saturated with proven incremental return.
Onsite or offsite: where should the money go first?
Onsite first. Sponsored products and onsite search defend digital shelf share where purchase intent is highest, and they are the table stakes retailers increasingly expect from suppliers. Fund onsite to the point of diminishing incremental returns — which you can only find by testing, not by chasing a ROAS ceiling.
Offsite is where the growth (and the risk) is. Amazon's 2026 Upfront (May 11) made the direction obvious: Dynamic TV Creative personalizes interactive Prime Video ads against shopping signals at the moment of impression, AI agents now handle campaign planning and optimization inside the DSP, and a LinkedIn partnership brings B2B audience targeting into Amazon's CTV inventory. Retail media and CTV are converging into one buy. That is genuinely useful — closed-loop measurement on upper-funnel video is something linear TV never offered — but offsite is also where wasted spend hides, because "retail media" pricing gets applied to what is functionally programmatic display. Extend offsite only with an incrementality test attached.
How do you measure this without grading the network's homework?
Every RMN reports its own ROAS using its own attribution rules — typically 14-day, last-touch, view-through included. That is the network grading its own homework, and it systematically credits ads for sales that would have happened anyway. A branded sponsored-product ad shown to a shopper already searching your brand can report a 12x ROAS while delivering near-zero incremental revenue.
The standard sophisticated buyers demand in 2026 is iROAS: incremental return measured against a holdout. In practice:
- Run geo or audience holdouts on your biggest lines at least twice a year. Platform-reported ROAS becomes a directional optimization signal, not a truth source.
- Use clean rooms (AMC, Walmart's data tooling) to separate branded from non-branded, and new-to-brand from repeat.
- Normalize across networks. IAB standardization work is ongoing but incomplete, so build your own cross-network scorecard with consistent windows and definitions. This is where an independent measurement and attribution layer earns its keep — no network will do this for you.
- Set different iROAS bars for onsite defense (lower bar, share-protection value), offsite prospecting (higher bar), and CTV (brand-plus-incrementality composite).
A useful heuristic: if reported ROAS is far above your blended average and the tactic targets people already close to your product, assume low incrementality until a test proves otherwise.
What does agentic commerce have to do with retail media?
More than most planning decks admit. During Prime Day 2026 (June 23–26, $26.4 billion in US online spend per Adobe), shoppers arriving via AI channels converted 40% better than non-AI channels — a reversal from the prior year, when AI traffic converted worse. Shopify reported AI-driven traffic to its stores up 8x and orders from AI-powered search up nearly 13x year over year in Q1 2026 (vendor-reported, but consistent with Adobe's independent panel).
The connection: retail media networks own the structured product data, reviews, availability, and pricing feeds that shopping agents read. The same digital-shelf hygiene that wins onsite search — complete attributes, competitive pricing, strong review velocity, accurate inventory — is what gets your product surfaced when an agent assembles a recommendation. Treat retail media content ops and agent readiness as one workstream: build the asset once, win in both channels.
How do mid-size brands avoid getting eaten by fragmentation?
Fragmentation taxes small teams hardest: every network is another UI, rate card, and reporting export. The counter-moves:
- Concentrate ruthlessly. Two networks run well beat five run poorly. The market is concentrating; your budget should too.
- Buy through interoperable demand paths (Amazon DSP for offsite, Criteo/Epsilon-powered networks) rather than direct IOs with every mid-tier RMN.
- Negotiate joint business plans. Retailers want committed ad spend; trade it for data access, category insights, and merchandising support — not just impressions.
- Centralize measurement outside the networks so budget shifts are driven by your iROAS scorecard, not by whichever network's dashboard flatters itself most.
- Unify the team. Retail media touches trade spend, paid media, content, and analytics. Silos between sales and marketing are the single most common reason RMN budgets underperform.
The 2026 retail media checklist
- Map revenue by retailer and match network investment to actual distribution
- Designate one anchor network and cap total networks at two until both are saturated
- Fund onsite search to defend digital shelf share on top SKUs
- Attach an incrementality test to every offsite and CTV extension
- Run geo or audience holdouts on your two largest budget lines this half
- Build a cross-network iROAS scorecard with consistent windows and definitions
- Separate branded vs. non-branded and new-to-brand vs. repeat in clean rooms
- Audit product content, attributes, and reviews for both onsite search and AI shopping agents
- Fold retail media commitments into retailer joint business plans
- Kill any network line item that can't show incremental return within two quarters
Retail media rewards operators who treat it as a system — media, measurement, and commerce data working together — not as another line on the channel plan. The Matchbox runs retail media inside a full-funnel paid media program, builds the independent incrementality layer through analytics and attribution, and connects retailer, CTV, and owned-channel data through omnichannel integration so the budget follows proof, not dashboards.
Sources
- eMarketer — Retail Media Ad Spending Forecast H1 2026
- IAB/PwC — Internet Advertising Revenue Report, Full Year 2025 (April 2026)
- WARC — Future of Commerce Media (global retail media forecast)
- Dentsu — Global Ad Spend Forecasts, May 2026
- Walmart — Q1 FY27 earnings release (May 21, 2026)
- Amazon Ads — Upfront 2026 recap announcements
- Amazon Ads — Dynamic TV Creative for Prime Video
- Adobe — 2026 Prime Day insights (June 29, 2026)
- PYMNTS — Shopify Q1 2026: AI-driven orders up 13x (vendor-reported)
FAQ
Quick
answers.
Benchmark against category norms, then let distribution decide. US retail media grows ~18% in 2026 to roughly $71 billion, so most CPG and marketplace-heavy brands are moving 15–30% of digital budgets there. Anchor spend on the one or two networks where your revenue actually sits, fund onsite search first, and only expand offsite once incrementality tests clear your bar.
Your growth starts here
Let's build the
growth engine.
Tell us where growth is stuck. We'll show you what one integrated team can move — and how fast.