Free A/B Test Sample Size Calculator
Find out how many visitors each version of a page, ad or email needs before an A/B test can give a reliable answer, and how many days the test will take with your traffic.
How to use the sample size calculator
- Enter your current conversion rate for the page, ad or email you want to test.
- Enter the smallest uplift you care about, as a relative change. A 20% uplift on a 3% rate means detecting 3.6%.
- Keep 95% confidence and 80% power unless you have a reason to change them.
- Add daily visitors to see how long the test will take.
Why sample size matters
With too few visitors, random luck can make one version look better when it is not. Many small businesses stop tests after a day or two and pick a “winner” that was just noise. Deciding the sample size before you start protects you from that mistake.
What the settings mean
| Setting | Meaning |
|---|---|
| Baseline rate | How often visitors convert today |
| Minimum uplift | The smallest improvement worth finding. Smaller uplifts need much more traffic |
| Confidence 95% | About a 5% chance of calling a winner when there is no real difference |
| Power 80% | An 80% chance of spotting a real uplift of the size you set |
Low traffic? Test bigger changes
Halving the uplift you want to detect roughly quadruples the visitors you need. If your site gets a few hundred visitors a day, test bold changes such as a new headline, offer or page layout rather than button colours. When the test is done, check the result with the A/B test significance calculator.
More free tools and guides
Practise tools like this on real projects in our digital marketing course in Faridabad, with weekday, weekend and online batches.
See the courseAll free toolsAsk on WhatsApp
Frequently Asked Questions
Is this sample size calculator free?
Yes. It is free with no sign-up, and the maths runs in your browser.
How many visitors do I need for an A/B test?
It depends on your conversion rate and the uplift you want to detect. Low rates and small uplifts need thousands of visitors per version.
What does minimum detectable effect mean?
It is the smallest relative improvement you want the test to reliably spot, such as a 20% lift from 3% to 3.6%.
Can I stop the test once it looks significant?
It is better not to. Stopping early when results look good raises the chance of a false winner. Reach the planned sample size and run at least a full week.