Your Retail Media ROAS Is a Vanity Number

@Hippeas grew from organic chickpea puffs into retail distribution nationwide. Retail media spend is growing. The ROAS numbers look strong. The question is whether those numbers are real.

The reflexive invariant in CPG retail media: "If the platform reports 8:1 ROAS, we are winning." That closure prevents anyone from asking the harder question: how much of that revenue would have happened without the ad spend? Incrementality and ROAS are different polynomials. Solving one does not solve the other.

When I took over omnichannel strategy, we found that 47% of retail media attributed sales were customers who would have bought anyway. We were paying premium prices for demand that already existed.

The Retail Media Incrementality Framework:

1. Test against holdout groups. If you are not measuring lift, you are guessing. A clean holdout test tells you what the ad caused versus what it witnessed.
2. Track new-to-file customers separately. That is your true retail media contribution. Repeat customers attributed to retail media were probably already yours.
3. Measure profit, not revenue. High ROAS with zero incremental profit is vanity. Include wholesale margin, not just retail media revenue.
4. Coordinate with your DTC strategy. Retail media and DTC targeting the same consumer segment creates cannibalization that neither channel reports.
5. Build predictive cannibalization models. We forecast overlap before spending a dollar. If predicted incrementality is below threshold, we do not buy.

The S4 construction: an incrementality-first retail media engine that measures what it causes, not what it claims.

I have cut CPA by 66% and lifted ROAS 67%. Retail media taught me the hardest lesson: vanity metrics kill profitability.

What is your real retail media incrementality? Have you ever measured it with a holdout test?