Incrementality

Definition

Incrementality: Incrementality measures the revenue a marketing channel actually caused — the conversions that would not have happened without the ad. It is the opposite of attribution, which credits channels for conversions they merely touched.

# Incrementality Incrementality answers the question attribution cannot: how much revenue did this channel actually cause? A branded-search ad may be credited with $50,000 in last-click revenue, but if 90% of those buyers would have typed the brand name into Google anyway, the incremental revenue is $5,000 — and the channel is dramatically over-credited. ## Why it matters Attribution rewards channels that show up at the right moment; incrementality rewards channels that actually move the needle. The two numbers often disagree by 2-10x. Running a geo-lift test on brand search or retargeting is one of the highest-leverage experiments a DTC brand can run — it often reveals that 20-40% of claimed ad revenue is cannibalizing organic demand. ## How Admaxxer surfaces it Admaxxer ships a v0.1 observational incrementality view in the Attribution dashboard that compares paid-cohort conversion rate to organic-cohort conversion rate using a two-proportion z-test. It is labeled as observational (not causal) and is intended to flag channels that deserve a proper geo-lift holdout. Rigorous geo-lift tests are on the v1.5 roadmap. ## Example A brand runs brand-search ads spending $5,000/mo at a reported 12x ROAS ($60,000 revenue). A two-week geo-lift holdout in five matched markets shows organic branded search recovers 80% of the "lost" revenue when ads are paused. True incremental revenue is closer to $12,000 — a 2.4x incremental ROAS, not 12x.

Most related

Continue exploring the DTC ad-analytics vocabulary — every term in this glossary cross-links to the next.

Frequently Asked Questions

How is incrementality different from attribution?

Attribution credits channels for conversions they touched. Incrementality measures conversions caused — the delta between running the ad and not running it.

What is a geo-lift test?

A holdout experiment where the channel is paused in a set of matched geographic markets while it continues in control markets. The revenue delta between the two is the lift attributable to the channel.

Is observational incrementality reliable?

It is a directional signal, not causal evidence. Observational views can flag channels worth testing; only a randomized holdout gives you a real incremental ROAS number.

Try Admaxxer Free