A/B Test Significance Calculator

Two-proportion z-test with p-value and verdict.

The formula behind it

z = (p̂B − p̂A) ÷ √[ p̂(1−p̂) × (1/nA + 1/nB) ] • p-value = 2 × (1 − Φ(|z|))
p̂A, p̂B
observed conversion rates of the two variants
p̂
pooled rate across both groups

The standard two-proportion z-test asks whether the difference you saw exceeds what sampling noise typically produces. The pooled standard error assumes the null hypothesis (no difference) — the correct test for 'is there any difference?'. The verdict is at your 0.05 threshold: 'not significant' means not proven, not proven equal.

How to use this calculator

  1. 1

    Enter visitors and conversions for variant A.

  2. 2

    Enter visitors and conversions for variant B.

  3. 3

    Read the rates, the z-score, the p-value and the verdict.

Worked examples

A real winner

4% vs 6% with 1,000 each → z ≈ 2.05, p ≈ 0.040 — significant at 0.05.

Same rates, smaller sample

8% vs 14% with 500 each → still significant; the same experiment at 100 visitors would not be.

Frequently asked questions

Explore more categories

Command Palette

Search for a command to run...