A/B test sample size and duration calculator
How many visitors your test needs, and how many days it will run.
Enter your conversion rate
Questions
How is the sample size calculated?
With the standard formula for comparing two proportions, two-sided, at 95% confidence and 80% power. It is the formula of statistics textbooks and of the reference calculators.
What is the minimum detectable effect?
The smallest relative lift your test can reliably spot. Going from 4% to 5% is a 25% lift. Halving the lift you want to detect roughly quadruples the visitors you need.
What if my traffic is low?
Test bolder changes. A new promise in a headline has a better chance of moving conversions than a new button colour, and a bigger lift needs far fewer visitors. The table shows which lifts your traffic can read, and how fast.
Should a test run for whole weeks?
Yes, round up to whole weeks when you can. Visitors on a Monday morning and on a Saturday evening rarely behave the same way, and a test that covers part of a week measures part of your audience.
How do more versions change the duration?
The duration assumes your traffic is split evenly between the versions in the test. Each extra version adds a full share of visitors to the total, so the test runs longer.
Your website rewrites itself until it converts.
Start your 3-day free trial