A/B Test Duration Calculator (Days and Full Weeks)

Enter the sample size per variant, variant count and traffic to get the days and weeks needed.

Use the tool

Why this tool exists

Tests that stop early produce false winners. Knowing the duration up front, and rounding to full weeks so each weekday is covered equally, keeps the test honest.

Get the sample size from the A/B test calculator on this site, using your baseline rate and the smallest lift you care about. Low-traffic pages may need weeks or months.

How to use it

  1. Get the sample size per variant.
  2. Enter the number of variants and daily visitors.
  3. Set the share of traffic in the test.
  4. Read days and full weeks.

Frequently asked questions

Why full weeks?

Behaviour differs by weekday, so full weeks avoid a weekday bias.

Do variants include the control?

Yes, 2 means A and B.

Can I stop early?

Not safely on a significance peek; fix the duration first.

Is data stored?

No.