Marketing & Product

A/B Test Sample Size Calculator

Work out how many visitors per variant you need before running an A/B test, given the effect size you want to detect.

Confidence levelStatistical powerThe chance of detecting the effect if it's really there.Baseline conversion rate (%)Minimum detectable effect (pts)Absolute percentage-point lift worth detecting, e.g. 5% → 6% is 1 point.

Required sample size per variant

8,156

16,312 total across both variants

This is a pre-test planning estimate — how many visitors you need before you start, not a test of results you already have. For checking significance on a completed test, use the A/B Test Significance Calculator.