Free A/B test calculator
An A/B test compares two versions of a page to see which converts better. Use the first calculator to check whether your result is likely real or just chance, and the second to know how many visitors you need before you start. Both run in your browser and store nothing.
Calculate your test
1. Is your test result real?
2. How many visitors do I need?
Uses 95% confidence and 80% power. Everything runs in your browser.
How to read the result
- p-value below 0.05: the difference is unlikely to be chance at 95% confidence. It does not tell you how big or how valuable the difference is.
- p-value above 0.05: you do not have enough evidence yet. It does not prove the versions are equal.
- Relative difference: how much better (or worse) B is compared with A, in percent.
Rules for a test you can trust
- Decide the sample size first (the second calculator) and wait until you reach it.
- Do not stop at the first good-looking result. Checking over and over and stopping early inflates false wins.
- Change one thing per test, so you know what caused the difference.
- Split traffic randomly and run both versions at the same time.
- Run full weeks, so weekday and weekend behavior are both included.
- Look by device. A win on desktop can be a loss on mobile.
What to test first
Use a heatmap and session replays to choose what is worth testing, instead of testing at random.
- The headline and the first screen of your top landing page.
- The label and position of your main button.
- The number of fields in a form.
- Shipping and price messages near the buy button.
- The order of sections on mobile.
See Visiflow in action
Short demo videos on sample data (not a real customer). Each one shows a task from start to finish.
Find where visitors drop off
Open the funnel, spot the worst step and see which pages people were on.
- Open Funnel
- Hover the biggest drop
- Click it to see where they went
Videos use fictional sample data.
When you do not have enough traffic
Small sites often cannot reach the sample size for a small lift. Then test bigger changes (a new offer, a shorter form), or skip the formal test: make the change that the data supports, and compare the same funnel step before and after.
Visiflow does not run A/B tests yet
Visiflow shows you what visitors do, so you can find what to change and compare before and after. It does not split traffic between versions. Use a testing tool for that, and use this calculator to read the result.
Questions
How do I know if an A/B test is significant?
Compare the p-value with 0.05. A lower value means the difference is unlikely to be chance at 95% confidence.
How many visitors do I need?
It depends on your current rate and the smallest lift you want to detect. Use the second calculator.
Can I stop the test early?
Stopping when it first looks good raises the chance of a false win. Decide the sample size first.
Is the calculator free?
Yes. It runs in your browser and stores nothing.