Incrementality testing template
Most incrementality tests fail before they start: no pre-agreed decision rules, a sample size that was never feasible, or results reinterpreted after the fact. A good template forces those decisions up front. Here is what it needs to contain, what you can get for free, and where the complete pack lives.
What a complete test template covers
- Test design doc: hypothesis, primary metric, unit of randomization, duration, holdout share
- Sample size and feasibility check (before you pick regions)
- Pre-registration doc: decision rules agreed before launch, so nobody moves the goalposts
- Holdout calendar: which regions are dark when, and what else runs during the test
- Interpretation guide: what to conclude per outcome, including the awkward ones
Start free
- Geo test sample size calculator — check feasibility in two minutes, no signup.
- How to run a geo experiment — the full method, free.
- Designing a holdout test for paid social
The complete template pack
The test design doc, pre-registration doc, holdout calendar and interpretation guide ship as working files inside The Marketing Triangulation Playbook, alongside the step-by-step weeks that use them. One-time purchase, lifetime updates.
See what's in the Playbook →Frequently asked questions
Is there a free incrementality testing template?
Partly. The free geo test sample size calculator on this site covers the feasibility step, and the how-to guide walks through the full design. The complete template pack (test design doc, pre-registration doc, holdout calendar, interpretation guide) is part of the paid Playbook.
What should an incrementality test template include?
At minimum: the hypothesis and primary metric, the unit of randomization (usually geos), sample size and expected detectable lift, test duration, holdout share, decision rules agreed before launch, and a pre-registration section so results can't be reinterpreted after the fact.
Does this work for Meta, Google and TikTok?
Yes. Geo holdout testing is platform-independent by design: you measure outcomes in your own sales data per region, not in platform attribution, so the same template works for any channel including offline.