Incrementality measures how much a marketing activity actually caused — the additional conversions or revenue that happened because of an ad and would not have happened otherwise. It separates ads that drive genuinely new revenue from ads that merely take credit for conversions that would have occurred anyway. Where attribution asks "who gets credit?", incrementality asks "did this ad change the outcome at all?"
How incrementality is measured.
Incrementality is measured by comparing an exposed group against a control (holdout) that didn't see the ad — the counterfactual. Common methods include:
- Holdout / randomized control testsWithhold ads from a random slice of the audience and compare conversion rates.
- Geo experimentsRun ads in some regions and not others, then compare.
- Conversion-lift / ghost-ad studiesServe placebos or track non-visible ad deliveries to control groups to isolate ad lift.
The quantities you are looking for are calculated as:
Conversions (Exposed) − Conversions (Control)
Incremental Conversions ÷ Control Conversions
Because it relies on a controlled comparison rather than a single tracked path, incrementality is measured periodically through experiments — not calculated continuously from one formula the way ROAS is.
Incrementality example.
Run a campaign to a test group, holding out an equivalent control:
- Exposed (saw ads)
- —
- 1,000
- Control (no ads)
- —
- 800
- Incremental conversions
- —
- 200
- Incremental lift
- 200 ÷ 800
- 25%
The ad was credited with 1,000 conversions — but only 200 were truly incremental. The other 800 would have converted anyway. A channel can look great in attribution and still have low incrementality.
Why incrementality matters for attribution.
Attribution assigns credit for conversions you observe; incrementality tests whether those conversions were caused. The gap between them is where budget gets wasted. Branded search and retargeting are classic examples: they show high attributed ROAS because they intercept people already on their way to converting — but their incremental lift can be low, because much of that revenue would have arrived without the ad.
The practical answer isn't to pick one over the other — it's to use both. Reconciled, revenue-based attribution is what you run day to day to allocate budget across channels; periodic incrementality tests validate that the channels attribution rewards are genuinely adding revenue, not just harvesting demand. Attribution tells you where credit landed; incrementality tells you whether it was earned. (The full attribution guide →)
