Incrementality

Incrementality asks a harder question than attribution. Attribution asks "which ad did this buyer touch?". Incrementality asks "would this buyer have bought without the ad?". A retargeting ad shown to people already at checkout can be credited with many sales and cause very few.

It cannot be read from tracking data alone, because every buyer either saw the ad or did not; nobody saw both versions of their own life. It is measured by comparison: a lift test that holds a random group back from the ads, a geo test that switches ads off in some regions, or a marketing mix model calibrated with such tests. Google describes incrementality experiments as the way to measure the causal impact of ads (Google).

A practical reading: attribution tells you where to look, and incrementality tells you what the money did. When an attributed channel looks too good to be true, test it. See attribution models explained.