A/B Test Significance Calculator

Know if your results are real or just random noise. Plan how many visitors you need before you start — or evaluate whether a completed test has a real winner.

Enter your test data

Add visitors and conversions for both Control (A) and Variation (B) to find out if your results are statistically significant.

Quick exampleControl A: 5,000 visitors · 150 conversions (3.0%)Variation B: 5,000 visitors · 185 conversions (3.7%)

How to Use This Calculator

Three modes for every stage of your testing workflow — plan before you start, evaluate once you have data, or compare multiple variants at once.

STEP 01

Choose your mode

Use "A/B Test" to check results you already have. Use "Plan Test" to figure out how many visitors you need. Use "Multi-variant" to compare 3 or 4 versions at once.

STEP 02

Enter visitors and conversions

Conversions can be any goal: purchases, sign-ups, clicks. Just make sure you use the same goal definition for both variants — mixing goals will give misleading results.

STEP 03

Set your confidence level

95% is the industry standard — it means you accept a 1-in-20 chance of a false positive. Use 99% for high-stakes changes like pricing; 90% for fast exploratory tests.

STEP 04

Read the colour-coded result

Green means a statistically significant winner. Red means the variation is significantly worse. Yellow means there's not enough data yet — do not declare a winner early.

STEP 05

Never stop tests early

Peeking at results and stopping when you see significance inflates false-positive rates dramatically. Decide your sample size in advance using "Plan Test" mode.

STEP 06

Plan before you test

Use "Plan Test" to calculate how many visitors you need. Enter your daily traffic to get an estimated test duration — and always run for at least one full week to avoid day-of-week bias.

Ready to build a testing programme that actually works?

Statistical significance is the floor, not the ceiling. Search Indicators connects your SEO, CRO, and testing data into one growth picture — so you know what to test and why it worked.

Talk to us