Most A/B tests are stopped at the moment they look good, which is exactly when they mean least. This gives the number of visitors to wait for, decided before the test starts.
A/B test sample size calculator
How many visitors each variant needs before a difference means anything, from your current rate and the improvement worth detecting.
Visitors per variant
7600
Total visitors
15 200
Days to run
30,4
Weeks to run
4,34
Obliczane natychmiast w Twojej przeglądarce · żadne dane nie są wysyłane
- 1Enter the rate you convert at today.
- 2Decide the smallest improvement worth acting on.
- 3Read the duration, then commit to it before starting.
- Planning a landing page test
- Deciding whether a test is even feasible
- Settling an argument about a winner
- A 60/40 split over thirty visitors is noise, and acting on it costs months.
- This is the usual approximation at 95 percent confidence and 80 percent power. It sizes a test; it does not replace a statistician.
Często zadawane pytania
Is this calculator free?+
Yes. It is free and needs no account.
Where is the calculation done?+
In your browser, instantly. Nothing needs to be sent to a server to compute the result.
Can I share my result?+
Yes. Use the share link next to the calculator: it keeps your figures in the address, so whoever opens it sees the same result.
Is this financial or tax advice?+
No. The result is an estimate based on the figures you enter. Rates, taxes and legal rules differ from one country to another, so check the official rules that apply to you before relying on a number.
