A retailer running ads on Meta, Google, and a marketplace's own retail media network usually gets three different answers to one question: did the campaign work?
Each platform's dashboard counts the same sale as its own conversion, and none of them agree on what counts as a result.
Getting a real answer on retail marketing ROI means stepping outside any single platform's numbers and looking at what actually moved revenue across the business. That gets more expensive as retail media grows, since the newest channel is the hardest to reconcile with the rest of the budget.
What retail marketing ROI actually needs to include
Retail media networks, the ad slots inside a retailer's own app or a marketplace, are having a strong run. Kantar's 2026 Marketing Trends report found they deliver 1.8 times better results than digital ads generally, with a bigger lift in purchase intent too.
A net 38% of marketers told Kantar they plan to spend more on retail media in 2026.
That performance is real, and it's exactly why a single-channel view stops working once retail media joins the mix. When one channel performs this well, every other channel a shopper touched along the way claims some credit for the same sale.
A retailer checking in-platform ROAS on each account separately ends up adding numbers together that were never meant to be added.
How retailers are sorting the noise out
The fix isn't a new dashboard. It's a different question: what would have sold anyway, without the ad?
That's what marketing mix modeling and incrementality testing are built to answer, and retail has adopted the first of the two at real scale.
In a Feedvisor survey of more than 1,000 US retail decision-makers, run with Zogby Analytics in November and December 2025, 61% said they already use media mix modeling to measure incrementality, according to eMarketer's reporting on the study.
A model like that needs real spend across several channels to learn from, which is why smaller retailers still lean on simpler comparisons.
A lot of that reporting now runs through the same automation that already handles restocking or an abandoned-cart email, so the ROI figure updates on its own instead of getting rebuilt in a spreadsheet every Monday.
Other retailers fold it into a broader marketing and growth engagement built to report across channels from the start, rather than bolt blended measurement onto reporting that was never designed to combine them.
Why the measurement window changes the number
Photo by Markus Winkler on Pexels
Timing matters as much as method. A Google and WARC analysis of media investment found that every £1 of marketing spend returns £1.87 within the first four months, and the same £1 returns £4.11 once the following twenty months are counted.
Checking ROI too early throws away more than half the return before it has even shown up.
The same analysis points to a wider pattern: businesses putting roughly half to 60% of their marketing budget toward brand building, rather than pure performance spend, saw stronger returns across both windows.
A campaign judged four weeks after launch will always look weaker than the same campaign judged after two seasons, and cutting a channel on the early number can mean cutting the one that was about to pay off.
For an online retailer, this connects straight back to the storefront itself.
If an ecommerce platform change is part of the plan, the same blended view should carry into that case, not just the ad budget. It's one of the questions we hear most from ecommerce clients building that case internally.
Start with one number: total revenue over total marketing spend, checked on a timeline long enough to include the channel that takes longest to show up. Everything more precise than that is worth building later, once that one number stops moving in a direction nobody can explain.
Cover photo by Max Fischer on Pexels
