A/B Test Duration Calculator
A/B test duration = required sample size per variant ÷ daily visitors per variant — run for at least 14 full days. Example: a page with 1,000 daily visitors per variant that needs 17,000 per variant takes 17 days. Enter your traffic, baseline CVR, and minimum detectable effect (MDE) below to get the exact number of days to reach statistical significance — before you start.
Enter your traffic data to see how long your test needs to run.
The #1 reason A/B tests give wrong answers
Stopping a test too early is the most common – and costly – mistake in A/B testing. It creates a 25–30% false positive rate even at "95% confidence." This is called the peeking problem.
The fix: decide your test duration before you start, then commit to it. Never stop early because it "looks like" it's winning (see the most common A/B testing mistakes to avoid). This tool tells you exactly how long you need.
Run tests for at least one full business cycle
Even if the math says 10 days, run for at least 2 full weeks. Weekday/weekend traffic behaves differently – running less than 14 days can create seasonal bias that makes results ungeneralizable.
Duration depends on sample size, not just calendar time
Test duration is really a sample-size question: how many visitors per variant you need, divided by how many you get per day. Work out the visitor count first with the sample size calculator, then check whether your current results already clear the bar with the statistical significance calculator. For the full method, read our A/B testing sample size guide.
Need help designing your test roadmap?
Our team prioritizes and sequences tests so you always have valid, actionable results – not a backlog of inconclusive tests.