Incrementality
Also: Incremental lift, Incremental impact, Incrementalité, Mesure d'incrémentalité, Inkrementalität, Incremental measurement, Lift measurement
The share of results (sales, conversions, revenue) that only happened because of a marketing action, not what would have occurred anyway.
What It Is
Incrementality measures the results that happened only because of a specific marketing action, and would not have occurred otherwise. It answers a blunt question every CFO eventually asks: if we turned this campaign off tomorrow, how much revenue would we actually lose? A retargeting ad that shows to someone already walking to the checkout takes credit for a sale that was coming anyway. Incrementality strips out that free-rider effect and isolates the true causal contribution of the spend.
Why it matters
Most reporting flatters marketing. Attribution models assign credit to whatever touchpoint sat closest to the conversion, so channels that intercept existing demand look like heroes. Incrementality reallocates budget from spend that claims credit to spend that creates value. A CMO who runs a geo holdout test on paid search brand terms often finds that a large share of those clicks would have converted through organic results at no cost. That finding can move seven figures across the budget. For the CFO, incrementality turns marketing from a line item defended by narrative into one defended by evidence. For the CDO, it sets the standard that a data platform must support: clean control groups, not just dashboards.
How it works
The method is experimental, not analytical. You create a group exposed to the marketing action and a comparable group that is not, then measure the difference in outcomes. The gap between the two groups is the incremental lift; everything else is baseline that would have happened regardless. Common approaches include geographic holdouts (turn off a channel in some regions), audience holdouts (withhold ads from a random slice of users), and ghost bids in programmatic where the platform records who would have seen an ad but did not. A practical example: a subscription business pauses paid social in five matched cities for six weeks, compares signups against untouched cities, and learns the channel drives far fewer net-new customers than its last-click report suggested. The discipline is harder than reading an attribution report, because it costs real exposure and requires patience, but it is the closest marketing gets to a controlled experiment.