In brief: Incrementality is the additional result caused by a specific intervention compared with what would probably have happened without it.
How I use Incrementality in practice
I do not use the term as another presentation number. I define the decision it should clarify, the data behind it, and who will change something based on the result. Before a campaign I define control and test groups, the primary metric, test period, possible overlap, and the minimum practically meaningful effect. I record the baseline, measurement date, and interpretation boundaries so later changes in tool, sample, or question wording are not mistaken for progress.
What to watch out for
The greatest risk is precision that only looks real. An attributed conversion is not automatically incremental; the customer may have bought without the ad and the platform merely claimed credit. I compare results over time, on a stable sample, and with business context. If the term leads to no concrete next step, there was no analysis, only a new label.
Questions for decisions
- What is the credible counterfactual without intervention?
- Can overlap between groups be prevented?
- Does the test have sufficient power and duration?
- Is the measured difference economically meaningful?