What Is Incremental Attribution in Meta Ads?

Standard attribution answers “did a conversion happen within a set time after someone clicked or saw the ad?” Incremental attribution asks a narrower question: “would this conversion have happened without the ad?” Meta answers it with models that predict how likely each conversion was caused by the ad, and both delivery and reporting follow that prediction. You pick it when you create an ad set, as an alternative to the standard model.

How is it different from standard attribution?

Meta says it currently offers two attribution models, standard and incremental, chosen at the ad set level:

Compared on	Standard attribution	Incremental attribution
What it credits	Conversions inside the time window and engagement type you choose	Conversions its models predict were caused by the ad
What you choose	Click-through (1 or 7 days), view-through (1 day) and engage-through (1 day) settings	The model itself; attribution settings apply to standard only
What delivery optimizes for	Conversions inside your selected windows and behaviors	Incremental conversions
What reporting shows	Conversions in your window	Incremental conversions

Meta also describes a third option, custom attribution, which lets you share attribution data from your own analytics tool so delivery optimizes toward your attribution logic.

What does “incremental conversion” mean?

Meta defines incremental conversions as additional conversions that would not have happened without the ads being tested. That is the same idea a lift test measures directly: lift tests compare people who had the chance to see your ads with a holdout group that did not. Incremental attribution predicts it with models.

Can you compare incremental and standard ad sets in one table?

Not directly. Meta warns that results cannot be compared in the campaign overview table across ad sets with different attribution models, because each model counts conversions differently and comparing them leads to inaccurate conclusions. Meta’s guidance:

1. Use the compare attribution settings feature in Ads Manager to compare accurately across models.
2. Evaluate performance in your external analytics tool, if you use one.

Where do you see which model an ad set uses?

Ads Manager has a column for the attribution setting that lists the settings and model chosen when the ad set was created. For incremental, Meta describes that column as reporting incremental conversions by using models to predict their incrementality.

How is it different from a conversion lift study?

A lift study is an experiment with a test group and a holdout, and Meta reports metrics such as cost per conversion lift: amount spent divided by the estimated incremental conversions. Incremental attribution is a model you choose when creating an ad set, and it uses models to predict whether a conversion was caused by an ad.

FAQ

Will incremental attribution show fewer conversions?

It counts a different thing. Standard attribution reports conversions within your chosen window; incremental reports only conversions predicted to be caused by the ad. Because the counting differs, Meta says comparing the two in the overview table leads to inaccurate conclusions, so don’t read a lower number as worse performance.

Can I choose a click or view window with incremental attribution?

No. Meta says attribution settings, the click-through, view-through and engage-through windows, are for the standard model only. With incremental attribution, the model decides which conversions to credit based on its prediction of whether the ad caused them.

Is incremental attribution the same as a holdout test?

No. A holdout or lift test compares people who could see your ads with a group that couldn’t, and measures the difference. Incremental attribution is an attribution model you pick when creating an ad set, and it uses models to predict whether each conversion was caused by an ad.