Incrementality: What Ads Actually Caused

📚 Module 10: Measurement⏱ 6 min read🎯 Advanced

What you’ll learn

The difference between attributed and incremental sales
How Search Incrementality testing works
Why ROAS can overstate your ad’s true impact
How to act on incrementality findings

Here’s the uncomfortable question that separates advanced advertisers from the rest: of the sales your ads got credit for, how many would have happened anyway? That’s incrementality — and once you understand it, you’ll never read ROAS quite the same way.

Attributed vs incremental

ROAS counts all sales attributed to an ad — including the shopper who had already decided, the customer simply reordering, and the person who searched your exact brand name and would have found you regardless. Incremental sales are only the ones your advertising genuinely caused. The gap between the two changes where you should invest.

The core question

As Walmart’s own measurement team frames it: am I changing behaviour, or just capturing what was already happening? A campaign can show strong ROAS, CTR and CVR while mostly reaching people who were already going to buy — money spent confirming a decision, not creating one.

How Search Incrementality works

Walmart’s Search Incrementality product (now generally available) measures true impact using randomized control testing: it compares shoppers who were exposed to your search ad against a control group who weren’t, then isolates the sales that only happened because of the ad. That test-and-control design is the gold standard for proving causation, not just correlation.

Why this reframes your numbers

Incrementality can cut both ways:

Downward — brand-name keywords often show high ROAS but low incrementality, because many of those shoppers would have bought anyway (you’re partly cannibalising organic sales).
Upward — when measured across the full omnichannel journey (in-store plus off-Walmart halo), the true incremental value of a campaign can be substantially higher than a platform-only ROAS suggests.

Little wonder that a large majority of advertisers now rank incrementality as their most important retail-media KPI.

How to act on it

Use ROAS/ACoS for day-to-day, in-flight optimisation (Module 9).
Use incrementality for bigger decisions — annual planning and where to invest across formats.
If brand terms show low incrementality, consider trimming spend there and reallocating to terms that win genuinely new demand.
Judge awareness formats (Display, offsite, Brands) partly on incremental and new-to-brand impact, not last-click ROAS.

Don’t over-rotate

Incrementality testing is powerful but not a daily tool — it needs proper test design and time. Treat it as a periodic lens that corrects your ROAS view, not a replacement for the weekly routine.

Quick recap

Attributed sales include demand you’d have won anyway; incremental sales are what ads truly caused.
Search Incrementality uses randomized test-vs-control to isolate causal impact.
Brand terms often show high ROAS but low incrementality; omnichannel value is often higher.
Use ROAS in-flight and incrementality for planning — and reallocate toward incremental demand.