Incrementality
Incrementality measures the additional conversions an ad campaign actually caused "” outcomes that would not have happened without it "” typically using controlled experiments with a holdout group.
Key takeaways
- Incrementality measures the causal lift an ad drove, not just correlation.
- It uses controlled tests with a holdout (control) group.
- It answers 'would this have happened anyway?' "” which attribution can't.
- It has grown more important as user-level attribution eroded.
Measuring true lift
Incrementality asks the counterfactual question: would these conversions have happened without the ad? By withholding ads from a randomized control group and comparing outcomes to an exposed group, it isolates the conversions the campaign genuinely caused "” its lift "” rather than crediting exposures that coincided with conversions.
Why it beats attribution for budgets
Attribution credits touchpoints that were present, but many of those users would have converted anyway. Incrementality strips out that baseline, making it far more reliable for deciding whether spend is actually working "” which is why it's rising as cookie-based attribution declines.
| Measures | Causal lift (added conversions) |
|---|---|
| Method | Randomized holdout / control group |
| Question | Would this have happened anyway? |
| Versus attribution | Causation, not correlation |
Frequently asked questions
What is an incrementality test?
It's a controlled experiment that withholds ads from a randomized control group and compares their conversions to an exposed group, isolating the lift the campaign actually caused.
Why is incrementality better than attribution?
Attribution credits ads that were merely present before a conversion; incrementality measures whether the ad caused conversions that wouldn't have happened otherwise, which is the question budgets actually turn on.