In context
Your Meta says 4x ROAS. You run a Haus incrementality test: 1.8x incremental. The 2.2x gap is organic / branded / cross-channel that would have converted regardless. This doesn't mean Meta is 'bad' — it means 1.8x is the real marginal dollar return, and you budget against that.
Why it matters
- Platform-reported ROAS always overstates true incrementality.
- Branded search and direct converters inflate Meta attribution.
- Incrementality tests are the CFO's only honest view.
Related terms
Frequently asked questions
How often should I test incrementality?
Quarterly at minimum; monthly at scale. Incrementality drifts as audience saturates.
Incrementality multiplier — is 0.6 typical?
For mature Meta accounts: 0.5-0.7 of reported ROAS. Early-stage: sometimes 0.8-0.9 (less branded interference).
What's the cheapest incrementality test?
Geo holdout (turn off Meta in 2-3 states for 4 weeks). Imperfect but near-free.