A geo-holdout test splits comparable geographic markets into two groups: a treatment group that receives your advertising and a control group that doesn't. By comparing total business outcomes between the two groups during the test period, you measure the true incremental impact of your campaigns rather than relying on attribution that can't distinguish causation from correlation.
Attribution measures which touchpoints were present when a conversion happened. Incrementality measures whether those touchpoints caused the conversion. People who were going to convert anyway show up in attribution data, inflating the reported contribution of bottom-funnel channels and making it impossible to know what's driving genuine growth.
It depends on conversion volume and the minimum detectable effect you need to observe. High-volume programs can reach statistical significance in two to three weeks. Lower-volume programs may need six to eight weeks. Thrive calculates required run time before every test begins. Calling a test early based on directional results is one of the most common testing mistakes.
Retargeting and branded search, where reported ROAS is highest but incremental contribution is most often overstated. Upper-funnel channels like CTV, display, and brand media, where attribution systematically undercredits. Any channel that's been running for a long time without its contribution being independently validated.
