How long should you run an A/B test?

It depends on your traffic and effect size. Most tests need 1-4 weeks. Always run at least one full week.

Can I stop my test early if I see significant results?

Stopping early inflates false positive rates. Use sequential testing for valid early stopping.

Should I run my test for full weeks?

Yes, always round up to complete weeks to capture weekday and weekend traffic patterns.

A/B test duration calculator

Question: How long should I run my A/B test?

Find out how long your A/B test needs to run to reach statistical significance. Enter your daily traffic and test parameters for an accurate estimate.

How to use this calculator

Enter your average daily visitors, the percentage of traffic allocated to the test, your current baseline conversion rate, and the minimum detectable effect. The calculator computes the required sample size and divides by your daily available traffic to give the number of days needed. Adjust the traffic percentage to see how dedicating more or less traffic affects duration.

How duration is calculated

The duration calculation combines sample size analysis with traffic planning. First, the required sample size per variant is computed using power analysis. Then, the total sample is divided by the effective daily traffic (daily visitors multiplied by the traffic percentage, divided by the number of variants). This gives the minimum number of days required. Best practice is to always run for at least one full business week to capture day-of-week effects.

When to use this calculator

Use this calculator during the planning phase to determine if a test is feasible with your traffic level. If the estimated duration exceeds 4-6 weeks, consider increasing the MDE (detecting only larger effects), increasing the traffic allocation, or choosing a higher-traffic page for testing. Knowing the duration upfront prevents the common mistake of ending tests too early.

Common mistakes in test duration

The biggest mistake is ending a test before reaching the required duration because early results look promising (or discouraging). This is called peeking and inflates your false positive rate. Other mistakes include not accounting for weekday/weekend traffic differences, ignoring seasonal effects that may skew results, and not planning for enough full business cycles.