Retail Media Networks: The Operator's Strategy Guide to Buying Across 200+ RMNs

Retail media is the fastest-growing line item on most brand media plans—and the most operationally chaotic. If you're still treating it as a sponsored-product add-on to your Amazon budget, you're behind; if you've already expanded to six or more networks without a measurement framework, you're spending blind.

Retail media networks are now the third-largest advertising channel in digital marketing, and the channel is projected to cross $165 billion globally in 2026. The money is moving. The structural problem is that what started as a handful of networks built around Amazon and a few big grocers has turned into a fragmented map of more than 200 retail media networks, each asking a brand to log in, build a campaign, and report results on its own terms. This guide is for operators who need to buy across that landscape efficiently, not just survive it.

Why Retail Media Is Structurally Different from Programmatic Display

If you came to retail media from programmatic display, your instincts will mislead you on several key points.

First, the data asset is fundamentally different. Marketplace operators hold a structural data advantage over traditional ad networks because they see cross-vertical shopper behavior across thousands of third-party sellers—a deterministic shopper graph with SKU-level basket data, repeat-purchase cadence, return behavior, and category co-occurrence. Where ad networks model intent probabilistically from clickstreams, marketplace operators see it in transaction records, which enables search-grade targeting fidelity at programmatic-scale inventory.

Second, the auction mechanics differ by retailer and placement type. Most onsite sponsored search runs on second-price auction logic; offsite programmatic inventory flowing through open DSPs runs on first-price. Treating them the same in your bidding model is a common tax operators pay without realizing it.

Third, attribution is vendor-controlled in ways that disadvantage you. Retail media networks are positioning themselves as the source of truth. They pitch themselves as measurement solutions, not just media sellers. The conflict of interest is identical to the one programmatic buyers learned to distrust a decade ago with DSP-reported viewability.

The Fragmentation Problem Is Getting Worse, Not Better

The concentration is stark: industry estimates suggest Amazon and Walmart will absorb more than 89% of all incremental US retail media dollars in 2026. At the same time, brands are managing more networks, not fewer—a Skai/Stratably survey of 166 retail media advertisers found the average brand works with six retail media networks today and expects that to reach 11 by end of 2026.

Eleven networks means eleven login portals, eleven reporting schemas, eleven naming conventions, eleven attribution methodologies, and eleven salespeople telling you their ROAS numbers are real. Despite strong performance in transactional scenarios, retail media networks present major challenges for marketing operations and analytics teams, particularly in large enterprises managing five or more networks simultaneously. The biggest pain points are measurement and attribution gaps, fragmented non-standardized reporting across walled-garden retail media networks, and workflow bottlenecks in data access, integration, and optimization.

The measurement gap is the most acute. Skai and Stratably's 2026 State of Retail Media Measurement and Incrementality study found that 75% of advertisers cite incrementality as their biggest measurement challenge, yet only 15% feel very or extremely effective at measuring it. Analysts struggle to distinguish incremental sales from baseline sales that would have occurred anyway, making ROI and optimization models unreliable.

How to Structure Your Multi-RMN Strategy

Start with a tiered network model

Not every network deserves equal investment or equal operational overhead. A working tiered model:

Tier 1 — Scale networks: Amazon Ads and Walmart Connect. These are table-stakes. Amazon commands approximately 79.7% of all US retail media ad spend, which shows the scale of the opportunity it pioneered. You cannot ignore either network if you sell through mass retail. Treat them as performance channels with rigorous holdout-based measurement.

Tier 2 — Category-relevant networks: Kroger Precision Marketing, Target Roundel, Instacart Ads, CVS Media Exchange, and others that index well for your vertical. Winning retail media advertisers in 2026 spread investment across multiple networks strategically. Amazon provides catalog scale and transaction volume. Walmart delivers value-oriented shopper access. Category-specific networks provide vertical-relevant audience precision.

Tier 3 — Test networks: Emerging or regional networks where you're buying primarily for audience quality and incremental reach, not conversion volume. Allocate 5–10% of budget, run proper holdout tests, and exit quickly if you can't get independent measurement access.

Onsite vs. offsite — knowing where inventory actually sits

The scope of retail media in 2026 extends across three primary channels: onsite (sponsored product listings, display ads, video ads served on the retailer's own website or app), offsite (retailer first-party data used to target brand campaigns across social, programmatic display, connected TV, and email), and in-store (digital screens, QR-enabled activations, audio, and point-of-sale integrations within physical retail locations).

Onsite retail media inventory is finite. Offsite activation extends retailer audience segments across programmatic exchanges, connected TV, and the open web at significantly larger scale and lower CPMs than premium onsite placements. A well-structured retail media plan combines onsite for high-intent conversion capture with offsite for scale and consideration building across the broader purchase journey.

The offsite opportunity is accelerating fast. The next $20+ billion of retail media spend is moving offsite—into CTV, programmatic open web, and paid social—because onsite inventory is saturating and brand advertisers want the retailer audience exposed wherever shoppers are watching, reading, and scrolling. Walmart's June 2026 acquisition of Vibe.co (a self-serve CTV platform) is a direct signal that offsite is no longer optional infrastructure for major networks. The vendors that win are likely those that can plug retailer first-party data into off-site inventory with credible attribution. For agencies, the near-term consequence is a talent and capability scramble—reporting through 2026 has described agencies rushing to acquire Amazon and Walmart retail-media specialists as budgets shift, and an off-site expansion widens the skill set required to include CTV planning, programmatic buying, and cross-network measurement.

The AI layer: what's actually useful vs. vendor theater

Walmart launched Marty (advertiser assistant) and Sparky (shopper assistant) in January 2026—conversational interfaces that let advertisers ask questions like "which campaigns are underperforming" and get auto-generated bid recommendations. Amazon is building the same layer into its DSP.

Treat these AI features as productivity tools, not strategy replacements. For automated bidding, AI agents raise the floor—fewer bad campaigns ship—but do not raise the ceiling. The best operators still beat AI defaults with domain judgment.

AI is reshaping retail media operations. Walmart Connect's Automated Creative Generation reportedly reduced production time by 80%, and industry estimates suggest AI-driven targeting delivers substantially higher ROI than traditional targeting methods. On the measurement side, in June 2026, CVS Media Exchange launched CorIQ, an AI-driven closed-loop measurement platform, and Albertsons Media Collective introduced onsite incrementality measurement—both signals that AI is moving from creative into measurement.

The Measurement Trap: How to Not Get Gamed by Walled-Garden Attribution

This is where most retail media programs quietly bleed money.

Every retail media network's native attribution report has an inherent conflict of interest: the network grades its own homework. When every network runs its own attribution game, brands lose the ability to measure incrementality across platforms using conversion lift studies. It becomes nearly impossible to know what's working outside of what each network says is working.

The pattern is well-documented from the Facebook and Google era, and it's repeating. Organizations are pouring budget into retail media while starving the analytics infrastructure needed to optimize it. Spend grows, but performance insight does not. The pattern is consistent: fragmentation creates distrust, distrust limits investment in capabilities, and limited capabilities perpetuate the cycle.

The operational fix has three components:

1. Independent holdout testing. Run your own geo-based or audience-split holdout experiments, outside the network's measurement tools, for every Tier 1 and Tier 2 network. Incrementality testing, using holdout groups and matched market experiments to isolate the true lift from advertising, is emerging as the standard for sophisticated retail media measurement. For a full framework, see our guide to Incrementality Testing in 2026.

2. Cross-network normalization. Measurement in retail media today operates within the "walled gardens" of each retail media network. Retailers can measure the impact on their own sales, but are completely blind into how advertising drives sales through outside channels. Tools like Amazon Marketing Cloud, Skai's Measurement Hub, and independent MTA platforms are the only way to build a cross-network view. Adopt one and own the data layer yourself.

3. Clean room discipline. Clean rooms are gaining traction as privacy expectations rise. Brands and retailers are collaborating in secure environments to analyze performance without exposing raw customer data.

Amazon's AMC has become a standard feature of mature retail media networks. Expect the number of cleanroom solutions to keep growing, which raises a new problem: brands will soon need to curate and secure data across multiple cleanrooms, not just one.

Network Comparison: Where to Spend What

Note: ROAS benchmarks vary significantly by category, match type, and attribution window. Industry benchmarks from Q1 2026 showed Amazon Sponsored Products spend up 21% year-over-year with ROAS relatively stable, Amazon DSP spend up 41%, and Walmart Sponsored Products spend up 62% with stronger ROAS. Treat any single-network ROAS figure as directional, not as a cross-network comparator.

Building the Operating Layer Above the Networks

The fragmentation problem is the defining story: brands work with an average of six RMNs today and will approach eleven by end of year. The winning brands aren't adopting any single RMN's automated bidding in isolation—they are building cross-RMN operating layers on top.

What that operating layer needs:

• Unified taxonomy. One campaign naming convention and KPI dictionary that translates across every network's native reporting. Without it, every analyst spends 40% of their time reformatting exports.

• A brand-owned data environment. Build a single, brand-owned environment that harmonizes data across every retailer relationship, rather than starting from scratch inside each retailer's cleanroom. Skai's Data Hub is designed to give brands—particularly large CPGs with data scattered across dozens of systems—one unified environment in their own language and structure.

• Incrementality-first budget allocation. Brands will start measuring the combined impact of retail media and upper-funnel channels that feed it—search, social, CTV—not treating them as separate scoreboards. Your retail media budget decisions should flow from incrementality findings, not from native ROAS dashboards.

• Clear ownership. When no team owns incrementality end-to-end, it becomes everyone's priority and no one's responsibility. Assign a named owner for cross-network measurement.

This multi-threaded pressure is pushing retail media to replicate the sophistication of programmatic: omnichannel orchestration, consistent measurement, unified planning, and agnostic attribution that moves seamlessly from brand to demand to conversion. If your retail media operation doesn't look like a mini programmatic desk—with dedicated tooling, measurement rigor, and a unified data layer—it will underperform relative to peers who've built that infrastructure.

You can find vetted tools for managing multi-network retail media at allaspect.com/tools/. And if you're connecting retail media measurement to broader paid creative performance, the principles in our DCO in 2026 guide apply directly to offsite creative testing against retailer audiences.

Bottom Line for Operators

Retail media deserves the same structural rigor as any programmatic channel—tiered network selection, bid strategies matched to auction mechanics, and above all, independent incrementality measurement that doesn't let the network grade its own homework. Without clear incrementality data showing what retail media contributes beyond what other channels would have captured anyway, budget allocation becomes a negotiation rather than a data-driven decision. Brands that invest in proper measurement infrastructure capture this clarity. Brands that skip it stay in perpetual budget uncertainty cycle after cycle. Don't be the latter. Build the operating layer first, then scale spend.

Frequently asked questions