What is Incrementality?

Why It Matters

Incrementality matters because standard attribution metrics like ROAS or conversion count can badly overstate a campaign's real impact, especially for retargeting or branded search campaigns that often capture credit for customers who were already going to buy. A campaign can show an impressive ROAS on paper while contributing almost nothing incremental, because it's mostly reaching people who didn't need the ad to convert. Incrementality testing corrects for this by directly comparing an exposed group's behavior against a comparable unexposed group, isolating the ad's true causal effect. For any team allocating real budget across channels, incrementality is often a more honest signal than attribution-based ROAS, particularly for lower-funnel or retargeting spend where inflated credit is most common.

Example

A retargeting campaign shows a strong 6:1 ROAS based on standard attribution. An incrementality test using a holdout group, where a portion of the retargeting audience sees no ads at all, finds that the holdout group converted at nearly the same rate as the exposed group. The campaign's true incremental contribution is far smaller than the attributed ROAS suggested.

Frequently Asked Questions

• How is Incrementality different from standard attribution?

  Standard attribution assigns credit to touchpoints a converting customer interacted with, without proving the ad caused the conversion. Incrementality directly tests causation by comparing an exposed group against an unexposed control group.

• Why do retargeting campaigns often have low Incrementality despite high ROAS?

  Retargeting reaches people who already showed strong purchase intent, many of whom would have converted anyway, so attribution gives the campaign credit for conversions it didn't actually cause.

• How is Incrementality typically measured?

  Common methods include holdout tests, where a portion of the audience is deliberately excluded from seeing ads, and geo-based experiments, where ad spend is paused in some regions and compared against regions where it continues.

• Does Incrementality replace ROAS as a metric?

  Not entirely. ROAS remains useful for day-to-day optimization, but incrementality testing is typically run periodically to validate whether a channel's attributed results reflect its true causal impact.

• Is Incrementality the same as Marketing Mix Modeling?

  They're related but distinct. Incrementality testing isolates a single campaign or channel's causal effect through controlled experiments. Marketing mix modeling estimates the contribution of multiple channels at once using statistical modeling of historical data.