A/B Test Duration Calculator
Estimate how many days an A/B test needs to run to collect its required sample, from your daily traffic, variant count and allocation.
How it works
Per-variant per day = daily visitors × allocation% ÷ variants · Duration = required sample per variant ÷ per-variant per day
A test ends when each variant has collected enough observations to detect the effect you care about, so duration is the required per-variant sample divided by how fast that variant fills up each day. More variants split the same traffic thinner and stretch the test; lower allocation does the same. Always plan to run at least one or two full business cycles (typically a week) so weekday/weekend behaviour averages out, even if the math says fewer days.
Worked example
Needing 5,000 per variant with 2,000 eligible visitors a day, two variants and 100% allocation: each variant gets 2,000 ÷ 2 = 1,000/day, so 5,000 ÷ 1,000 = 5 days — round up and run a full week to cover the weekly cycle.
Frequently asked questions
Where does the required sample per variant come from?
From a power calculation based on your baseline conversion rate, the minimum effect you want to detect, and your chosen significance and power (commonly 95% and 80%). Use an A/B test sample size calculator to get that number, then bring it here.
Should I stop the test as soon as it looks significant?
No — peeking and stopping early inflates false positives. Fix the sample size and duration in advance and wait until both are met, or use a sequential testing method designed for continuous monitoring.
Why run for full weeks?
User behaviour differs by day of week, so a test that runs three days may over-weight weekday traffic. Running in whole-week increments averages out those cycles and gives a more representative result.
Related calculators
This calculator is for educational purposes. Double-check important results before acting on them.