What Is Meta Incremental Attribution? (2026)

Incremental attribution changes what an ad set counts as a result. Under the standard model, any conversion inside your click or view window is credited to the ad. Under the incremental model, Meta counts only the conversions its models predict the ad caused, and it steers delivery toward people likely to produce those. The result columns get smaller and closer to the question owners care about. They remain Meta’s modelled estimate, not a controlled experiment.

What does incremental attribution do?

Meta defines it as “an attribution model that optimises ad delivery for incremental conversions”, using “machine learning models that predict whether a conversion is caused by an ad” (About incremental attribution). It does two jobs at once:

• Optimisation. The ad set’s delivery aims at conversions that are more likely to be driven by the ad, not at people who would have bought anyway.
• Reporting. The Results column, and metrics derived from it such as cost per result, include only the conversions Meta considers incremental.

It sits beside two other attribution models on the ad set (About attribution models and attribution settings). Standard is the familiar one. Custom attribution, which Meta says may not be available to you yet, optimises toward credit you send from your own analytics tool (About custom attribution).

How is it different from 7-day click attribution?

Meta’s own example (How to view results for incremental attribution): a campaign with 100 total conversions, of which Meta considers 70 incremental, shows 70 results under incremental attribution and 100 under standard. Cost per result rises to match, because the same spend now buys fewer counted results.

For the standard windows and what each one credits, see which Meta attribution window should I use?

How do I turn on incremental attribution?

Meta’s steps (How to use incremental attribution in Meta Ads Manager):

1. Go to Ads Manager and click Create.
2. Select Sales, Engagement or Leads as the campaign objective, then click Continue.
3. Choose your campaign setup, review the campaign settings, and click Next.
4. In the ad set, choose Website or Website and app as the conversion location.
5. Pick Maximise number of conversions or Maximise value of conversions under Conversion goals, then click Show more options.
6. Under Attribution model, open the Options drop-down and select Incremental.
7. Finish the ad set without bid controls, which this model does not allow, then create the ad and click Publish.

Decide before step 7. Meta says you cannot change the attribution model after you publish; to switch, you build a new campaign.

How do I see incremental results for my existing campaigns?

You do not need to rebuild anything to look. Meta says you can view modelled incremental results even when incremental attribution is not your ad set’s model (How to view results):

1. Go to Ads Manager.
2. Click the Columns: Performance drop-down menu.
3. Scroll down and select Compare attribution models.
4. Select Incremental and click Apply.

Two limits apply. The column has no data for date ranges before 1 April 2025. And Meta recommends comparing campaigns within the same attribution model, not incremental results against standard ones, because each model counts differently. Its attribution settings page adds that the campaign overview table cannot compare ad sets that use different models (About attribution models).

Does incremental attribution replace a lift test?

No. The two answer the question in different ways. Incremental attribution predicts, conversion by conversion, whether an ad caused it. A lift test measures it: Meta randomly splits the audience into a test group and a holdout group that is kept from seeing the ads, then compares what each group did (About lift and holdouts). Meta describes the randomised holdout as what lets a lift test show causal impact.

That makes the two complementary. Use incremental attribution for day-to-day delivery and reporting, and a lift test when you need to check a large budget decision. For what a lift study can and cannot prove, read is Meta Conversion Lift a reliable incrementality test? For the wider choice of method, read what is the difference between attribution and incrementality?

Does it actually get more incremental sales?

Meta says so, in aggregate. In its 28 January 2026 post on 2026 performance, Meta wrote that its latest Q4 model rollout “drove a 24% increase in incremental conversions compared to our standard attribution model”, and that the product reached a multi-billion-dollar annual run-rate seven months after launch (Meta Newsroom). That is Meta’s figure across advertisers, measured by Meta, not a promise for your account.

On its help page Meta is more careful: incremental attribution can “potentially drive more incremental conversions compared to your existing campaigns”. To find out for your own account, run a new incremental campaign alongside a standard one for the same product, judge both on your store’s orders and margin, and add a lift test if the budget justifies it. The ROAS and attribution guide covers how to read those store-side numbers.

When is incremental attribution the wrong choice?

• You rely on bid controls, such as a bid cap. Meta does not allow bid controls with this model.
• Your conversions happen in an app only, or offline. Meta’s steps require Website or Website and app as the conversion location.
• You want reporting continuity. Results drop against your standard history, and Meta advises against comparing the two models side by side.
• You need a causal answer for a big decision. A modelled estimate is not a randomised holdout.

FAQ

Is incremental attribution the same as 1-day click?

No. A 1-day click window counts every conversion within a day of a click, whether or not the ad caused it. Incremental attribution has no window to choose: Meta’s models predict which conversions the ad caused and count only those. Meta says comparing the two models directly leads to inaccurate conclusions.

Why are my results lower with incremental attribution?

Because the model excludes conversions Meta predicts would have happened anyway. Meta’s example is 100 standard results becoming 70 incremental ones. Cost per result rises by the same logic. That drop is the model working as described, not a delivery problem, so judge it against your store orders rather than against old standard results.

Can I switch an existing campaign to incremental attribution?

No. Meta says you cannot change the attribution model after you publish a campaign. Create a new campaign with Incremental selected under Attribution model in the ad set. You can still preview modelled incremental results for existing campaigns with the Compare attribution models column.

Which objectives support incremental attribution?

Meta’s setup steps list Sales, Engagement and Leads, with Website or Website and app as the conversion location and a conversion goal of maximising the number or value of conversions. Meta notes it is rolling out a new setup experience for some objectives, so the exact screens in your Ads Manager can differ.

Sources

Sources: Meta Business Help Center, About incremental attribution, How to use incremental attribution in Meta Ads Manager, How to view results for incremental attribution, About attribution models and attribution settings, About custom attribution, About lift and holdouts; Meta Newsroom, 2026: AI drives performance, 28 January 2026 (checked 2 Oct 2026).