A/B test sample size calculator
How many views each version needs before a test can reliably detect the change you care about. Nothing you enter is saved.
Needed per version
13,809
- Total across both versions
- 27,618
- Change you're testing for
- 25% relative
How it works
- n per version = (z₁₋α/₂ × √(2p̄(1 − p̄)) + z₁₋β × √(p₁(1 − p₁) + p₂(1 − p₂)))² ÷ (p₂ − p₁)²
- p₁ = current rate, p₂ = rate to detect, p̄ = their average; rounded up.
How to read it
Small changes need far more traffic than big ones: detecting 2% → 2.5% takes roughly 13,800 views per version, while 2% → 3% takes roughly 3,800. If the number is out of reach, test a bolder change.
The full explanation is in our guide: A/B Testing for Higher CTR: How to Run a Fair Test.