Retail Media Measurement: Incrementality, Attribution, Metrics | Particular Audience
How to measure retail media: closed-loop attribution, holdout incrementality, halo, new-to-brand and the effect of ads on the retailer's own conversion.
Retail media measurement: incrementality, attribution and the metrics that matter
Retail media measurement connects ad exposure to purchases using the retailer's own sales data. Attribution reports sales linked to an ad; incrementality tests compare exposed shoppers with a holdout group to show the extra sales the ad caused. Strong programmes report both, plus brand halo, new-to-brand buyers and the ads' effect on the retailer's overall conversion.
Why measurement is the selling point
Brands invest in retail media because a retailer can show what happened after an ad was seen: whether the shopper bought, what else they bought and whether they came back. The quality of that evidence decides how much budget a network wins.
Attribution
Closed-loop attribution matches ad exposure (impressions and clicks) to orders, within an agreed window. Decide the window, whether views count or only clicks, and how sales are credited when several ads were seen. Attribution is fast and continuous but counts sales that would have happened anyway. See closed-loop attribution.
Incrementality
Incrementality asks: how many extra sales did the ad cause? The standard method is a holdout test — a random group of eligible shoppers is not shown the ad, and their purchases are compared with the exposed group. See incrementality, holdout tests and how to measure incrementality.
Good practice: randomise at the shopper level, size the test so a real difference can be detected (the free Bayesian A/B test calculator helps), run through normal weekly patterns, and agree the metric before the test starts.
The metrics that matter
- ROAS — attributed sales ÷ spend. See ROAS.
- Incremental ROAS — incremental sales ÷ spend.
- Brand halo — sales of the brand's other products after exposure. See brand halo.
- New-to-brand — buyers who hadn't bought the brand in a set period.
- Share of voice / share of category — visibility relative to the category.
- Retailer-side impact — conversion and basket value on pages carrying ads, versus a control. This protects the shopping experience. See ad load and conversion.
Measuring the retailer's side
A programme can raise media income while lowering product sales if ads crowd out useful results. Measure total page value — product margin plus media income — on a holdout, not media income alone. Relevance-ranked ads are designed to keep both moving together. See Unified Ranking.
Delivery and reconciliation
Measurement also covers what was delivered: impressions, clicks, pacing against plan and the numbers used for invoices. AdOps teams should be notified early when delivery falls behind; how to respond, including any make-good, is the retailer's decision. See billing reconciliation.
Related guides
Sources
- Bayesian A/B test calculator, checked 2026-11-13
- Unified Ranking, checked 2026-11-13
How Particular Audience does it
Frequently asked questions
What is the difference between attribution and incrementality?
Attribution counts sales linked to an ad. Incrementality compares with a holdout group to count only the extra sales the ad caused.
How long should a retail media holdout test run?
Long enough to cover normal weekly shopping patterns and reach enough shoppers to detect a real difference; a test calculator helps size it.
Should retailers measure their own conversion when running ads?
Yes. Measure total page value — product margin plus media income — against a control, so ads don't quietly reduce sales.
Related guides
- How to measure retail media incrementality: Plan a retail media holdout with stable assignment, comparable outcomes and visible uncertainty. Separate incremental sales from attributed ROAS and brand halo.
- How to implement closed-loop retail media attribution: Connect retail media events to purchases with explicit identity, scope, windows, deduplication and returns. Reconcile attributed ROAS without calling it iROAS.
- Incrementality in retail media: Incrementality estimates what advertising changes beyond the counterfactual. Learn why exposure comparisons, attributed ROAS and sequential periods need care.
- Holdout tests in advertising: A holdout assigns an eligible group to a comparison experience. Learn how stable assignment, valid outcomes and uncertainty support advertising measurement.
- Retail media: the complete guide for retailers and brands: Retail media explained: what it is, onsite vs offsite, ad formats, how the money flows, how to measure it and how retailers start. A practical guide.