Why Retail Media ROAS Is Mostly Paying for People Who Already Wanted Your Product
Written by Elias Oender
November 9, 2026 3 min read
The quick answer
Retail media networks boast massive ROAS figures by taking credit for consumers who were already on their way to checkout. Incrementality testing isolates the actual net-new lift driven by your ad spend versus baseline demand. Without split holdout tests, brands simply subsidize their own existing customers.
Retail media networks pitch a tempting dream. A shopper stands in an aisle with a phone, or browses a digital grocery cart, and a sponsored prompt nudges them to buy. Platforms like Ibotta sell this as frictionless closed-loop attribution, claiming immediate, massive return on ad spend. The dashboard looks stunning. The board smiles. But looking at your total bank balance tells a colder story: top-line revenue barely flinched. You did not acquire a new customer. You merely handed a discount to someone who already had the item in their hand.
What is incrementality testing inside closed-loop networks?
Closed-loop networks take credit for sales that would have happened anyway. If a shopper buys Greek yogurt every Tuesday, triggering an ad two seconds before checkout does not mean the media drove the sale. It means the platform intercepted an inevitable purchase. Marketing teams often discover that their ROAS number lies because platform reporting treats correlation as causation.
Incrementality testing splits your audience into a treatment branch and a holdout branch. You withhold impressions or cashback incentives from a randomized group of shoppers to measure baseline velocity. If the group that sees no ad buys just as much product, your true incremental return is zero. Subsidizing organic demand is not growth strategy. It is an unbudgeted margin leak.
Why performance marketing breaks without strict holdouts
When brands scale spend based on vendor pitches about what brands want from retail media, they fall into the same attribution trap. Many playbooks stumble because marketing tools fail without incrementality. When accounts leak cash into vanity campaigns, money goes in and nobody knows where it leaks. That is the exact friction I solve: the Ad Audit is a Google Ads audit, with Meta Ads and GA4 included, run from inside your accounts to give you an MRI of your budget for €950 one-time (list price €4950). Over 15 years in performance marketing, I look at the numbers without platform bias to separate manufactured credit from net revenue lift.
To find your real marginal return, run clean tests across your media channels:
- Set matched-market holdouts: Exclude specific regions or store clusters from campaigns to measure natural baseline sales.
- Audit brand search overlap: Stop bidding on branded keywords when retail media placements already capture shelf presence.
- Measure new-to-brand buyers: Track whether coupons attract repeat buyers or bring fresh households into the mix.
How to spot vanity returns before signing off on media spend
Agencies love showing retail media dashboards because an 8x ROAS screenshot keeps the retainer safe. They add account managers, coordinators, and layers of slides to justify keeping the machine running. Working with a dedicated senior operator is entirely different: Elias runs automated systems and works directly on the account instead of hiding behind agency overhead and inflated dashboard returns. As an independent operator, the focus stays strictly on driving real profit rather than subsidizing agency headcount. Before locking in your budget, run a free scan to evaluate your real conversion foundations, or book a call to inspect where your media dollars are actually landing.
