Sample Size Calculator
Set your baseline conversion rate and the smallest lift worth detecting to see how many visitors each variant needs before you start the test.
Visitors per variant
31,234
Total visitors
62,468
2 variants
Estimated duration
32 days
at 2,000 visitors/day
This detects a change from 5.00% to 5.50% with a 80% chance of finding it if it's real, and a 5% false-positive rate (two-sided test). Run for at least one full week to cover weekday and weekend behaviour, and don't stop early when the result looks good.
How to use the Sample Size Calculator
- 1Enter your current (baseline) conversion rate.
- 2Enter the smallest relative lift worth detecting.
- 3Choose significance and power, and optionally add daily traffic to estimate how long the test runs.
Frequently asked questions
What is the minimum detectable effect?
The smallest change you want the test to reliably detect, as a relative lift. Smaller effects need far more visitors — halving the effect roughly quadruples the sample size.
What power should I use?
80% is the common default: a real effect of your chosen size will be detected 8 times out of 10. Use 90% when missing an effect would be costly.
Why run a test for at least a week?
Behaviour differs between weekdays and weekends. Running whole weeks avoids results skewed by the day you started.
More free tools
All tools →SQL Formatter
Format messy SQL into clean, readable queries instantly.
A/B Test Calculator
Check statistical significance, lift and experiment results.
Descriptive Statistics Calculator
Mean, median, mode, standard deviation and quartiles from a paste of numbers.
Correlation Calculator
Pearson r, r², a scatter plot and a significance test from two columns.
Percentage Change & CAGR
Percentage change between two values, or compound growth over many periods.
JSON ⇄ CSV Converter
Convert CSV to JSON or JSON to CSV, both directions, in your browser.
Put your skills to work
Thousands of open roles from companies hiring now — apply directly with the employer.