A/B test sample size calculator
Find out how many visitors a test needs before its result means anything, and how many weeks that takes at your traffic.
Your test
How the page converts today.
10 means 3% becoming 3.3%.
Split evenly across both versions.
What the test needs
That's over three months. Test a bigger change, move the goal up the funnel, or pool several pages into one test.
How the calculator works
It uses the standard formula for comparing two conversion rates at the confidence and power you choose, the same one most sample size calculators run. At 95% confidence and 80% power, a 3% baseline and a 10% relative lift need about 53,200 visitors per version.
The rule of thumb in our article, 16 times p times (1 − p) divided by the absolute lift squared, is the same math with the constants rounded, so it lands a few percent lower: about 51,700. Either way the lesson holds. Halve the lift you want to catch and the sample roughly quadruples.
Fix the sample before you launch, run whole weeks, and resist stopping the moment the result looks good. Why, and what to do when the traffic isn't there, is in how much traffic you need before an A/B test means anything.
Let us find your next growth channel.
Tell us where you want to grow. We will tell you, honestly, whether search is the fastest way to get there.