Retail Media Networks in 2026: The $73 Billion Channel, Explained for Brands

The forecasts disagree at the edges and agree on the shape, so cite them with their scopes:

The growth rate is halving from its 2023 peak even as the base doubles, which is the signature of a channel entering maturity: the easy land-grab is over, and what remains is consolidation (the duopoly absorbing new dollars), format expansion (sponsored products giving way to display, video, and CTV through Prime Video's 315 million ad-supported viewers and Walmart's Vizio acquisition), and offsite extension (retailer data powering ads on the open web).

Amazon and Walmart take 89% of the new money for one structural reason: closed-loop attribution at national scale. An ad on Amazon can be tied to a purchase on Amazon within the same session graph, which is the measurement dream every other channel approximates. Both networks also possess what advertisers increasingly cannot buy elsewhere: deterministic first-party purchase data on hundreds of millions of households, usable for targeting onsite, for offsite extensions, and for clean measurement partnerships. Walmart adds the omnichannel loop (online ads, in-store pickup purchases attributed), and its 46% global ad growth shows the model replicating beyond grocery. The second tier is real but situational: Target Roundel owns a high-loyalty guest base strong in style, home, and beauty; Instacart's network is the door for grocery CPG without direct retailer scale; Kroger's 84.51° and the marketplace networks (eBay, Etsy, and the quick-commerce apps covered in our quick commerce guide) each own a vertical. The planning rule the concentration implies: anchor one or two scaled networks where your shoppers actually buy, and add second-tier networks only where distribution or category fit justifies another measurement scheme and another minimum commitment.

If you sell exclusively through your own store, it is tempting to file retail media under "Amazon's problem." Three reasons it is not:

1. You are already inside the auction. Competitors and resellers bid on your brand terms and your category keywords on retail networks every day. Where your products (or close substitutes) are sold by anyone on Amazon or Walmart, brand-term defense is retail media spend whether you planned it or not, and the alternative to paying is ceding your own name to a competitor's sponsored slot at the moment of highest intent.
2. The data gravity affects your demand. As purchase behavior concentrates on retail platforms, category demand forms there first; shoppers who discover a product type on Amazon increasingly start even DTC-bound journeys with marketplace search. Your organic strategy and your marketplace presence now share a query universe.
3. Wholesale and multichannel expansion inherits the ad stack. The moment a DTC brand adds marketplace or big-box distribution (the paths in the international marketplaces guide and the B2B guide), retail media becomes the launch mechanism: new listings without sponsored support start on page nine, and the sell-through data those campaigns produce is what earns shelf conversations offline.

The honest counterweight: retail media is pay-to-play at the shelf, and every dollar spent defending position is a dollar not spent on brand building that would lower the defense cost. For DTC-first businesses the channel belongs in the budget as defense and expansion insurance, not as the core acquisition engine, which remains your own site, search, email, and the social channels where your brand story still travels.

Retail media's pitch is measurement; its practice requires discipline, because platform-reported ROAS is a self-graded exam. The stack that keeps it honest:

• Closed-loop first, blended second. Judge onsite sponsored products on the platform's attributed ROAS (its strongest claim), but reconcile against your own sell-through and total revenue in the channel; attributed sales include purchases that would have happened organically, especially on brand terms.
• Incrementality testing. Run paused-market or geo-holdout tests quarterly (campaigns off in matched markets for two to four weeks) to measure the true lift. Amazon Marketing Cloud and Walmart's maturing tooling support custom analysis, but a simple holdout answers the only question that matters: what would have sold anyway.
• iROAS by campaign type. Separate brand-term defense (high reported ROAS, low incrementality), category conquest (moderate both), and launch support (low ROAS, strategic purpose) into distinct budgets with distinct success metrics, because averaging them hides every real decision.
• New-to-brand metrics. Amazon's new-to-brand attribution is the cleanest built-in signal of whether retail media is acquiring or just intercepting; weight budgets toward campaigns that score well on it.

Stage one, marketplace-present but unadvertised: start with brand-term defense and top-three SKU category campaigns; 5% to 10% of channel revenue reinvested is the conventional floor, judged on sell-through lift within 60 days. Stage two, advertising with flat returns: run the first incrementality test before adding budget; most programs find brand terms over-credited and category conquest under-invested, and the reallocation, not the increase, produces the lift. Stage three, expanding: add offsite (retailer-audience display and CTV) only where onsite is saturated and the incrementality discipline already exists, because offsite retail media inherits the same self-grading problem with weaker attribution. Stage four, multichannel mature: negotiate with the second tier (Roundel, Instacart) using your closed-loop data from the majors, and treat retail media line items as one pool with your Google Ads and social budgets, since by 2026 the same shopper is reachable in all three with overlapping intent signals.

How big is retail media in 2026?

US spend is forecast near $73 billion (eMarketer's June 2026 figure: $72.97B, up 20.3%), with global retail media around $200 billion (WARC). Commerce media passed television globally in 2025 at about 15.6% of all ad revenue.

Who dominates retail media?

Amazon, with 75%+ of US spend and $68.6 billion in 2025 ad revenue projected past $75 billion by 2028, and Walmart, whose global ad business grew 46% to $6.4 billion. Together they absorb roughly 89% of 2026's incremental US dollars.

Should a DTC-only brand spend on retail media?

For defense and expansion, yes in most categories: competitors bid on your brand terms wherever substitutes are sold, and marketplace launches without sponsored support start invisible. Keep the core acquisition engine on your own channels and hold retail media to incrementality-tested budgets.

How do I know retail media spend actually works?

Platform-reported ROAS overstates, especially on brand terms. Run quarterly holdout tests (campaigns paused in matched markets), track new-to-brand metrics, and judge onsite campaigns on sell-through lift within 60 days rather than attributed revenue alone.