Ad calculators · 100% free

Incrementality Test Sample Size Calculator

How long does a geo-lift / holdout incrementality test need to run? Calculate sample size for statistical significance.

Google Ads Meta Ads Any paid channel

Baseline revenue per geo / week

Expected lift %

Test geos available

Control geos available

Confidence %

Power %

Revenue CV (variability)0.20-0.35 typical

Total geo-weeks needed

Roughly 5 weeks at 10 geos per side

Expected lift in revenue

$1,800

per geo per week at +15%

Minimum detectable lift at current size

22.1%

With your current geo count

Why incrementality matters now

Platform-reported ROAS is increasingly unreliable. Geo-lift / holdout tests are the only way to measure true incremental revenue from paid spend. Run them quarterly to catch attribution drift before it costs you 6 months of inefficient spend.

About this tool

Incrementality testing is the only honest way to measure paid ad lift post-iOS. This calculator sizes geo-lift / holdout tests based on baseline revenue, expected lift, statistical power and number of geos.

Why incrementality matters more than ever

Platform-reported ROAS has been progressively less honest since iOS 14.5. Geo-lift / holdout incrementality tests are the only causal way to measure whether your paid spend is creating sales that wouldn't have happened anyway. Without them, you're optimising to inflated platform numbers and slowly defunding the channels that actually drive growth.

Geo-lift basics

• Pick matched test and control geo groups (DMAs in US, regions elsewhere)
• Pause or boost spend in test geos, keep control geos at baseline
• Measure the revenue delta vs baseline expectations
• Calculate lift % and confidence interval

Cadence

Run quarterly per major channel. Mature D2C brands run continuous geo-experiments via tools like Haus, Meta Lift Studies or Google's Conversion Lift. For smaller brands this calculator + a spreadsheet is enough to start.