Retail media without the two-year build
The strategy deck is not the hard part. Ad serving, a commercial team and measurement brands will accept are the hard part — and none of them need to be built from scratch.
The retail media business case writes itself. Traffic you already have, brands you already buy from, margins that look nothing like grocery margins. Every retailer we speak with has already made this argument internally and won it.
Then the project meets the roadmap.
Three queues, one team
Standing up a retail media network from nothing means adding three things to a business that already has a full plate.
Ad serving. Placements, pacing, frequency, creative approval, reporting. This is a product with its own roadmap, competing directly against checkout improvements, inventory accuracy and whatever the app team already promised for Q3.
A commercial function. Selling media to a brand is not the same motion as negotiating trade terms with them. Different buyer, different budget, different cycle, different vocabulary. The existing commercial team is excellent at the job they have and mostly does not want this one.
Measurement. Brands will accept a retailer’s own numbers for a pilot. They will not accept them for a nine-figure annual commitment. Independent attribution is the price of moving from test budget to planned budget, and it is not something you can credibly build for yourself.
Most retail media projects stall not because anyone lost faith in the strategy, but because these three queues never clear at the same time.
What is actually specific to you
It is worth separating what genuinely has to come from the retailer from what does not.
Genuinely yours: the audience, the moment of purchase intent, the category data, the merchandising judgement about what belongs on a page, and the supplier relationships. Nobody can supply those and nobody should try.
Not specific to you: the ad server, the demand relationships with brands who are already buying retail media across four other retailers, the measurement methodology, the yield operations, the creative specs, the reporting infrastructure.
The second list is where two years and a lot of capital usually go, and it is entirely rentable.
A sequence that works
The pattern we see succeed looks roughly like this.
Quarter one: model and design. Size the opportunity from real traffic and category mix. Design two or three formats that fit the existing templates — not a format library, two or three. Agree what the ceiling on ad density will be, in writing, before there is revenue pressure.
Quarter two: first cohort. Sell to five or six advertisers in your strongest categories. Small commitments, real money, honest reporting including the parts that underperform.
Quarter three: extend the reach. Take the same promotions off-site through a publisher network with store-level targeting. This is usually where the numbers stop being interesting and start being material, because the audience is no longer capped by your own sessions.
Quarter four: renew and expand. The renewal conversation is the real test. Advertisers who renew without being pushed are the proof; advertisers who need convincing are telling you something about the measurement.
The part worth being stubborn about
Set the density ceiling early and treat it as a constraint, not a target.
Every retail media network eventually discovers that adding one more placement produces immediate revenue and a conversion cost that shows up two quarters later, attributed to something else entirely. The retailers who hold the line end up with more sellable inventory over time, because their pages stay worth visiting.
The ones who do not spend a lot of energy, later, explaining why traffic is down.