Not Every Retail Media Network Deserves Your Money
Not Every Retail Media Network Deserves Your Money
There are over 200 retail media networks operating in the US. That number has roughly doubled in the past two years. Every major retailer has one, most mid-size retailers are building one, and a growing number of non-endemic brands — airlines, banks, telecom companies — have launched media networks built on their own customer data.
For a brand trying to decide where to allocate a retail media budget, this isn't opportunity. It's noise. And the instinct to spread budget across as many networks as possible — to be present everywhere, to not miss anything — is one of the most expensive mistakes in commerce media right now.
Not every network deserves your money. Here's how to figure out which ones do.
"The average brand is running 7.2 retail media networks. The ones with the best ROI are running fewer, better-selected networks with more rigorous measurement on each." — Skai, 2026
The three things a network actually needs to offer
Strip away the sales deck, the case studies, and the audience reach claims, and a retail media network needs to provide three things to justify a brand's budget.
- Audience quality: does the network's shopper data actually represent your category's buyers? A network with 50 million loyalty card members is irrelevant if your category's buyers aren't in that base at meaningful scale. Ask for category-level audience sizing before you discuss placements.
- Measurement credibility: can the network tell you what your campaign caused — not just what happened nearby? Networks that can't offer at minimum a clean holdout methodology are selling reach with a conversion story attached. That's not the same as retail media ROI.
- Purchase proximity: is the network actually connected to purchase behavior? Endemic retail media — networks built on a retailer's own transaction data — has a structural measurement advantage over non-endemic networks built on inferred purchase signals. Both have a place in a media plan, but they're different products with different evidence standards.
The endemic vs. non-endemic distinction
Endemic retail media networks — Walmart Connect, Kroger Precision Marketing, Roundel (Target), Albertsons Media Collective — are built on a retailer's own transaction data. When one of these networks reports that your campaign drove conversions, those conversions happened in stores or on a platform where the retailer has direct visibility into the transaction. That's a first-party data advantage that no non-endemic network can replicate.
Non-endemic retail media — networks built by airlines, financial institutions, health systems, or other companies with large customer bases — can offer valuable audience access, particularly for upper-funnel awareness objectives. But their connection to purchase behavior is by definition more distant. The measurement methodology is different, the attribution is harder, and the incremental lift question is significantly more complex to answer.
Neither category is wrong. The mistake is treating them as equivalent products measured on the same ROAS basis. They're not.
→ Network Evaluation Questions — agencyfiveeighty.com/retail-media-network-evaluation-questions
→ Commerce Media — agencyfiveeighty.com/commerce-media
The category-relevance test
The single fastest filter for evaluating a retail media network: what percentage of your category's buyers are in their shopper panel, and what's the purchase frequency of those buyers in your category?
A network with 80 million shoppers is irrelevant if only 2% of them buy in your category. A network with 20 million shoppers where 40% are active category buyers with monthly purchase frequency is far more valuable — and far more likely to generate meaningful lift.
Most networks will provide category-level audience sizing estimates if you ask before the buy. The ones that won't, or that can only give you total network size without category breakdowns, are telling you something important about the quality of their data.
The concentration principle
The brands with the best retail media ROI aren't running the most networks. They're running fewer networks at higher spend levels, with better measurement on each, and compounding their learning over time.
The logic is straightforward: a $100,000 activation split across seven networks averages $14,000 per network — too small to drive statistically significant lift at most networks, too fragmented to optimize meaningfully, and impossible to measure properly in seven different measurement frameworks simultaneously. The same $100,000 concentrated in two networks where you have measurement infrastructure and meaningful scale is a completely different investment.
Start narrow. Measure rigorously. Add networks when you have the budget to activate them at scale and the infrastructure to measure them properly. Five Eighty applies this principle to every commerce media program we build — because the right answer is never "all of them."