Retail Media Is a Tax You Should Negotiate, Not a Channel You Should Love

Retail media is the fastest-growing ad channel in history, and the growth has one honest explanation: the retailer controls the shelf. When Amazon, Walmart, or the grocer who stocks you invites you to advertise, the invitation carries the quiet weight of your distribution. Brands know the difference between an opportunity and a toll booth, and they are paying the toll at a rate that will pass television this decade.

None of that makes the inventory bad. Onsite retail media is genuinely strong inventory: the shopper is in the store, wallet open, searching your category by name. Sponsored product placements on a retailer's search results are the closest thing commerce has to bidding on intent at the moment of purchase. The conversion data is real, first party, and closed loop, which is more than most channels can say.

The problem is what the closed loop hides. Retail media's flagship metric, ROAS against attributed sales, is measured by the party selling you the ads, against purchases that happen on their property, many of which would have happened anyway. When you bid on your own brand terms inside a store where you hold the top organic slot, the attributed return is spectacular and the incremental return is a rounding error. I made the open-web version of this argument in bidding on your own brand name is not a scandal, it is math, and the retail version has sharper teeth, because the retailer sets the auction, owns the measurement, and stocks your competitor.

There is also the margin spiral nobody puts in the deck. Trade spend used to be a negotiation; retail media makes it a self-serve auction where your competitors set your floor. Every point of margin you route into the retailer's ad platform is a point they do not need to concede in trade terms, and the brands that treat retail media budgets as untouchable have effectively agreed to an unbounded listing fee.

So spend, but spend like a negotiator. Separate branded from category terms and judge them differently; branded defense is a cost of shelf, category conquest is where incrementality lives. Push for the retailer's clean-room or third-party measurement rather than platform-reported ROAS, and where they refuse, run purchase-lift holdouts on your own dime; the math pays for itself the first time it kills a "four times return" line item. Fold retail media into the same ledger as trade spend so someone sees the whole tax, not two halves that each look reasonable. And use offsite retail media, the retailer's audience bought on the open web, sparingly; it reprices ordinary programmatic reach at data-broker premiums, and your match rate is the ceiling on everything else applies to their graphs too.

The test that settles most retail media arguments costs nothing but nerve: pause branded sponsored products for two weeks in half your regions. If total sales in those regions hold, which in my experience they usually do when your organic shelf position is strong, the "return" on that line was rented from your own demand. Category and conquest terms deserve the budget that survives; the brand terms deserve whatever is left after the retailer earns it in trade terms.

The shelf was never free. Retail media just itemized the bill. Negotiate the line items and it is a channel; sign whatever lands in the inbox and it is a tithe.