Free tool

A/B Test Sample-Size Planner

Before you launch a test, find out how many visitors each variant needs to detect the lift you care about — and roughly how long that takes at your traffic. Stops you from calling a test too early.

Confidence
Statistical power

Visitors needed per variant

8,158

16,316 total · detecting 5% → 6.00%

3.3 weeks to reach significance

at 5,000 visitors/week across both variants

Smaller lifts need dramatically more traffic. Run the test the full planned duration (ideally whole weeks, to cover weekday/weekend patterns) — don't stop the moment it looks significant.