What is incrementality?
Why does incrementality matter?
Ad platforms report the conversions they can connect to an ad. That is attribution: assigning credit for a purchase to the ads and clicks on the path before it. Credit is not cause. Someone who was already going to buy can still click an ad on the way to the checkout.
Incrementality tests that doubt. It compares what happened with the ads to what you would have seen without them. That second number is your baseline: the sales you expect with no activity at all. Only the gap above the baseline counts.
Finance asks "so what?" This is the number that answers. Everything else is an argument about credit.
How is incrementality measured?
You need a comparison group that does not see the ads. Vendor guides describe control groups and regional splits as the usual set-up2. Two common designs:
- A holdout test: part of the audience, or some regions, deliberately do not see the ads.
- A geo test: ads are switched off or changed in selected regions, while control regions show what would have happened.
The difference between the groups is the lift. Divide the incremental revenue by ad spend and you get incremental ROAS (iROAS).
A test is one way to measure incrementality. It is not the only way, and it is not always worth the cost. We cover that in incrementality for mid-size ecommerce brands.
Worked example
Easily confused with
Incrementality vs Attribution
Attribution shares out credit. Incrementality asks whether there was any cause to share. The full split is on incrementality vs attribution.
Incrementality vs Lift
Lift is the measured difference between groups. Incrementality is the idea. Lift is the reading.
Incrementality vs ROAS
Reported ROAS counts attributed revenue. iROAS counts only the revenue the ads caused.
Buzzword alert
Incrementality gets used as a sales word with no measurement behind it. One trade-media piece still lists it among buzzwords worth keeping, on one condition: say how the lift was estimated, for example a holdout, a geo test or a lift study3. We agree. No method named, no incrementality claimed.
| Phrase | What it should mean | What to ask |
|---|---|---|
| "Our campaigns are incremental" | Compared against a control, ads added sales | Which control, and over what period? |
| "Incremental lift" | The measured gap between ad and control groups | Gap in what, against whom? |
FAQ
Is incrementality the same as incremental lift?
Close, not identical. Incrementality is the concept: the effect that would not exist without the activity. Lift is the measured difference between an exposed group and a control group.
Do I need a test to know my incrementality?
A test with a control group measures it most directly. It is not the only way. Whether a test is worth running depends on the decision it would change.
Let’s find what’s actually working.
Next step. Put your own numbers through the Is your ROAS real? calculator. No forms, no call.