Use eligible traffic
Count only visitors who can actually enter the experiment. Exclude traffic outside targeting rules, users who cannot see the tested surface, and repeat exposures when the experiment assigns at the user level.
Free calculator
A/B Test Duration Calculator
Estimate how many days an A/B test needs from required sample size and eligible daily traffic.
Planning approximation for a two-sided two-proportion test using 95% confidence and 80% power. Validate important experiments with your analytics/statistics workflow.
Estimated requirement
17,943
visitors per variation
18 days
at 2,000 eligible visitors/day split 50/50
Traffic requirement too high?
Narrow pre-launch choices with a preference test, then validate behavioral lift once traffic supports it.
Create a preference testMethod and interpretation
How to estimate A/B test duration
Planning formula
estimated days = (sample per variation × 2) ÷ eligible daily visitorsCover normal cycles
A sample can be reached quickly and still be unrepresentative. Run across complete weekly or purchasing cycles when weekday, weekend, campaign, or payday behavior could materially change the outcome.
Do not stop on a favorable day
A fixed-horizon calculation assumes the sample and stopping rule were chosen before examining results. Repeatedly peeking and stopping when the p-value first crosses a threshold increases false-positive risk.
Worked example
A requirement of 8,000 visitors per variation needs 16,000 eligible visitors total. At 2,000 eligible visitors per day, the arithmetic estimate is eight days, but a team may deliberately run for two complete weeks to cover weekday effects.