Lore

Incrementality and the counterfactual problem

Incrementality asks: would this person have converted anyway without seeing the ad? This differs from attribution, which asks who converted after exposure (see Deterministic (in-platform) attribution).

The counterfactual problem

You cannot observe both the world where the ad ran and the world where it did not for the same person. Causal measurement therefore needs a stand-in for the missing world, a comparable group that was not exposed. See test-vs-control-incrementality-experiment.

Motivating case

eBay turned off its brand search ads. They looked wildly profitable by conventional measurement, but traffic and sales stayed flat and the change saved millions per year. Related: Incrementality testing of brand content.