No — and conflating them is the single most expensive beginner mistake.
The question: you have a model assigning 30% of credit to Facebook. Does that mean Facebook caused 30% of your sales? Not necessarily.
The distinction: attribution divides credit among touchpoints a converter actually saw. It is fundamentally correlational — it describes the observed path, not what would have happened otherwise. Incrementality asks the causal question: how many of those sales would not have occurred without this channel? It requires a counterfactual — a comparison world.
Why this matters with a number: imagine retargeting gets 25% of attributed credit. You run a geo holdout (some regions see the ads, matched regions don't) and sales differ by only 4%. The other 21 points were people who'd have bought anyway. Attribution counted them; incrementality revealed them as non-incremental.
The nuance: attribution is cheap, continuous, and good for trend-watching across many channels. Incrementality tests are expensive, periodic, and good for truth on a few big bets. They're complements, not substitutes. The mature stack uses attribution for the daily dashboard and incrementality to calibrate it — adjusting attributed numbers downward for channels that fail their holdouts.
Bottom line for practitioners: attribution tells you who was present; incrementality tells you who was responsible. Never report an attributed percentage as a causal claim. When the budget is large enough to matter, spend on a holdout before you spend on the channel.
Credit Where Due
@CreditWhereDue
Is attribution the same thing as incrementality?
This post was published in the Telegram channel Credit Where Due (@CreditWhereDue). All its posts are on the channel page.
Open on Telegram