A/B Test Duration Calculator

Work out how many weeks your A/B test needs to run before it can reach a statistically valid conclusion. Enter your baseline conversion rate, the smallest lift worth detecting, and the weekly traffic entering the test; the calculator divides the required sample by your traffic and rounds up to whole weeks. Use it before committing a page to a test, if the answer is months, you need a bolder change or a busier page, not more patience.

Worked example

Your checkout page converts at 2%, you want to detect a 20% relative lift, and 5,000 visitors per week enter the test. The statistics require roughly 16,000 visitors total, so the test needs 4 weeks (16,000 / 5,000, rounded up). At 2,000 weekly visitors the same test needs 8 weeks; at 1,000 it needs 16 weeks, past the point where it is worth running as designed.

How the maths works

First we compute the required sample size with the standard two-proportion formula at 95% significance (two-sided) and 80% power, the same calculation as our sample size calculator. Duration is then total sample divided by weekly traffic, rounded up to whole weeks, because partial weeks bias the traffic mix (weekends and weekdays convert differently). Tests projected past 12 weeks are flagged as impractical: over a quarter, seasonality, promotions, and site changes contaminate the comparison more than extra data helps it.

Frequently Asked Questions

How is the test duration calculated?

We take the total sample size the statistics require (both variants, at 95% significance and 80% power) and divide it by your weekly traffic, rounding up to whole weeks.

Why should tests run in whole weeks?

Conversion behaviour differs by day of week. Running whole weeks means every variant sees the same mix of weekday and weekend traffic, which keeps the comparison fair.

What if my test would take longer than 12 weeks?

A test that long usually is not worth running as designed. Test a bigger change (larger expected effect), test on a higher-traffic page, or use a more sensitive metric such as revenue per visitor.

Does more traffic always mean a shorter test?

Yes, proportionally: doubling the weekly visitors entering the test roughly halves the duration, because the required sample size is fixed by your baseline rate and minimum detectable effect.

Related tools

Also try our A/B Test Sample Size Calculator and A/B Test Prioritisation Planner. Not sure what to test or fix first? Run our free CRO audit tool for a scored, 6-pillar report on your page in 3-5 minutes. Explore all free CRO tools.