Retail media networks: when they destroy margin
September 20, 2026 · 5 min · Free, with full transcript
Retailer ad businesses now run at 70 to 80 percent margin against 2 to 3 percent net margin on groceries, which makes retail media a profit line rather than a side project. The position here is blunt: most mid-tier retailers copying Kroger, Walmart and Tesco are copying badly, stuffing sponsored products into search until their own private label falls below the fold and basket size shrinks while the media dashboard glows green. You leave knowing the scale test of roughly a few million active weekly shoppers, why identity resolution across loyalty card, app login and email sets the ceiling, how HubSpot and SEMrush data readiness and attribution claims need cross-checking against MIT Sloan, and how to prove incrementality with a holdout group.
Listen to the podcast
4 min
Chapters
Key takeaways
- Give one person accountability for both ad revenue and shopper loyalty so the tension between them stays visible.
- Check whether sponsored placements are pushing high margin private label below the fold before scaling ad inventory.
- Run identity resolution first: match loyalty card, app login and email to one person, because that match rate caps everything else.
- Do not build a network below roughly a few million active customers, since thin inventory cannot command a premium.
- Measure incrementality with a deliberate holdout group rather than trusting vendor attribution models.
Read the full transcript
Host:Leaders' Insights, today turn your first-party data into a margin line.
Expert:A grocery buyer I know spent last Tuesday morning not haggling over pallet prices, but staring at a dashboard showing which brands were bidding for the digital shelf above the pasta aisle. That's the room now.
Host:And that room is where the money is.
Expert:The thing that happened this year, retail media networks, meaning the ad businesses retailers run on their own sites and apps, crossed from novelty into structural profit. We're talking margins that rival a grocer's entire net margin on selling food.
Host:Rival it. That's a big claim. Put a number on it.
Expert:A supermarket nets somewhere around 2-3% on the groceries themselves. The advertising line inside the same business runs at 70-80% margin. So a retailer doing a billion in media revenue is throwing off more profit from ads than from half its stores.
Host:Amazon proved that years ago. Why is a regional grocer suddenly relevant?
Expert:Because the asset finally matured. First-party data. What your own shoppers actually bought from you, not what a third-party guest, is the only clean signal left after the privacy rules gutted the tracking cookie. Kroger, Walmart, Tesco built the muscle. Now the mid-tier is copying, and most of them are copying badly.
Host:Define badly. I want the field report, not the brochure.
Expert:Quietly means a CMO who launches a media network to hit a revenue target and quietly ruins the customer experience. I watched a chain. I won't name them. They're a client. Stuff so many sponsored products into search results that their own private label range, the high margin stuff they own, vanished below the fold. They sold ad slots and cannibalized their best profit line. Net negative.
Host:So the ad money masked the loss.
Expert:Exactly. The dashboard glowed green on media revenue, while basket size shrank. Nobody connected the two for two quarters.
Host:What separates the ones getting it right?
Expert:They treat the media network as a margin line the CMO owns, not a side hustle for the ad sales team. That means one person accountable for both ad revenue and shopper loyalty, so the tension is visible. When those live in separate silos, the ad side always wins and the customer always loses.
Host:You keep saying data's the asset. Everyone says data's the asset. How much is actually usable?
Expert:Less than you'd hope. HubSpot puts marketers' data readiness at something like 60-odd percent. Though worth flagging, they sell the software that fixes it. So cross-check that against something independent. MIT Sloan's work is more sobering. Most retailers can't reconcile a loyalty card with an online account for the same human. If you can't join those two records, you don't have first-party data, you have two piles of noise.
Host:So the first job isn't selling ads. It's plumbing. The plumbing?
Expert:Identity resolution. Matching the card, the app login, the email to one actual person. Unsexy, and it's where the value either exists or doesn't.
Host:Let's talk about who shouldn't build one. Because the article's blunt. This can destroy more value than it creates.
Expert:If you don't have scale, don't start. You need enough weekly shoppers that an advertiser can reach a real audience. Below roughly a few million active customers, the ad inventory's too thin to charge a premium. And you burn engineering budget building a shop nobody wants a stall in.
Host:Give me the threshold moment. When does a CMO know it's working versus vanity?
Expert:Watch incrementality, whether the ad actually caused a sale that wouldn't have happened anyway. SEMrush and the analytics vendors will sell you attribution models that claim everything worked. Naturally, they're selling the measurement tool. Scale incrementality means running a holdout, a group you deliberately don't advertise to, and comparing. If the advertised group doesn't outperform the holdout, you're charging brands for sales they'd have made regardless. That fraud gets found out fast, and the ad budget leaves.
Host:One thing the listener does Monday morning.
Expert:Pull one week of loyalty data and one week of online orders, and ask your team to count how many customers appear in both, matched to a single ID. Whatever that number is, it's the ceiling on everything you're about to build. Know it before you sell a single ad.
Host:Sources for today's episode. SEMrush vendor SEO analytics tools, DigiDay HubSpot vendor CRM marketing automation, MIT Sloan Management Review. That's your five. The full CMO track, structured and free of fluff, is at MBA-training.com.
Read and go deeper
- The article behind this episode: Turn your first-party data into a margin line
- The free Marketing track
- Watch on YouTube · the full playlist