Compute required sample size per arm given baseline conversion, MDE, power, and significance. Avoid running underpowered tests that ship the wrong winner.
62,464 total across 2 variants
Formula
n = (zα/2 + zβ)² × (p₁(1−p₁) + p₂(1−p₂)) / (p₂ − p₁)²Two-proportion z-test. Add 10–20% buffer for real-world novelty effects, weekday/weekend variance, and segment imbalance.
Heads-up: don't peek. Decide your sample size before the test starts to keep the false-positive rate honest.
A short guide on running A/B tests that actually move metrics — pre-registration, power, sequential testing, and common pitfalls.
No spam. Unsubscribe anytime.
Discover more utility-driven tools designed to enhance your workflow and technical excellence.