Analytics & Strategy

What is Incrementality Testing?

Incrementality testing measures the true causal lift of advertising by comparing outcomes between users exposed to ads and a holdout control group.

Incrementality Testing: Full Definition

Incrementality testing determines how much additional value advertising actually creates, beyond what would have happened organically without the ads. It works by randomly assigning users to a 'test' group (exposed to ads) and a 'holdout' group (ads suppressed or replaced with PSAs), then comparing outcomes between the two groups.

The incrementality lift is calculated as the difference in conversion rate between exposed and holdout groups, divided by the holdout group's organic conversion rate. An ad campaign with 40% incrementality means 40% of its attributed conversions would not have happened without the advertising.

Incrementality testing is the antidote to attribution inflation, the tendency of last-click models to credit ads for conversions that would have happened organically anyway. In mobile UA, organic installs and retargeting campaigns are most susceptible to over-attribution without incrementality validation.

Why Incrementality Testing matters

Without incrementality testing, you don't know how much of your 'attributed' performance is real. A retargeting campaign that claims 10x ROAS may actually have near-zero incrementality. It's re-engaging users who would have returned anyway. Incrementality testing reveals the true causal value of each campaign type, enabling rational decisions about where to allocate budget.

Example

A gaming studio runs a Meta retargeting campaign showing 5x ROAS in attribution data. An incrementality test reveals 70% of those 'attributed' conversions happened in the holdout group too, meaning true incremental ROAS is closer to 1.5x, dramatically changing the campaign's budget priority.

Frequently asked questions

Attribution credits conversions to ad interactions using rules (last click, multi-touch) or probabilistic models. Incrementality testing uses controlled experiments (holdout groups) to measure causality, whether the ad actually caused the conversion. Attribution measures correlation; incrementality measures causation.

Related terms

See which creative elements drive your best ROAS

Connect your ad networks and MMP in 5 minutes. AI tags every creative automatically.

Start Free Trial