Minimum detectable effect
It’s the smallest relative difference the test can tell apart from chance. If your target uplift is smaller, the test may conclude there is no difference when there actually is one.
PRE-TEST CALCULATOR
How long should your A/B test run?
Enter your traffic and the effect you hope to achieve. The calculator estimates the required duration and the smallest uplift your test can detect.
How to read this calculation
Two weeks minimum
Even with lots of traffic, run the test over full weeks. This smooths out behavior differences between weekdays and weekends.
More variants, more traffic
Beyond one variant, the significance threshold is corrected (Bonferroni) to limit false positives. Each extra variant makes the test longer.
Method: two-proportion comparison test, normal approximation. The displayed duration is rounded up to full weeks, with a two-week minimum.
Want an expert eye on your tests?
We’ll frame your hypotheses, test duration and how to read the results, in 30 minutes.
