A/B Test Duration Calculator — How Long Should You Run Your Test?
Get your test end date before you start. Enter your traffic and MDE — never stop a test too early again.
Test parameters
Relative improvement to detect reliably
Test duration
Estimated duration
—
Per-variant visitors
—
Total visitors needed
—
Daily visitors in test
—
Detectable CR
—
Embed with current settings
How to use this calculator
Plan your test end date before running a single visitor through it.
Check your daily traffic
Look at the average daily unique visitors to the specific page you're testing, not your whole site.
Enter your baseline CR
Your current conversion rate on that page. Find this in your analytics for the last 30 days.
Set your MDE
The minimum relative improvement worth detecting. Lower = longer test. 15–20% is a practical starting point for most teams.
Commit to the end date
Write down the date before you start. Don't check results or stop early — peeking inflates false positives significantly.
Related free tools
Frequently asked questions
Why should I run a test for at least 7 days?
A/B tests are affected by day-of-week patterns — Monday visitors often behave very differently from weekend visitors. Running for less than a full week produces biased results that may not hold when you roll out to all traffic.
What happens if I stop my test early?
Stopping when you first see significance is called 'peeking' and inflates your false positive rate well above 5%. If you check significance daily and stop at the first 95% result, your true false positive rate can be 20–30%. Commit to your sample size before starting.
My test would take 6 months — what should I do?
If a test would take longer than 8 weeks, your options are: (1) increase MDE — only test changes you expect to have a large effect, (2) increase the % of traffic in the experiment, (3) add more variants to test multiple ideas in parallel, or (4) use qualitative research (user interviews, heatmaps) to make a confident decision without a test.
Should I expose 100% of traffic to my A/B test?
For most tests, yes — 100% of page traffic split equally between variants is optimal for speed. Reduce this (e.g., 50% holdout) only for very high-risk changes like complete redesigns, accepting a longer test in exchange for a safety net.
Stop guessing.
Start testing.
Your first experiment — A/B test, popup, form, or heatmap — can be live in under 2 minutes. No developers, no contracts, no risk.