A/B Test Sample Size Calculator
Determine exactly how many visitors you need before your A/B test results are trustworthy — stop guessing and start testing with confidence.
The most common mistake in e-commerce A/B testing is ending tests too early. A test that shows a 15% uplift after 500 visitors might show zero uplift after 5,000 visitors — the early result was just noise. This calculator tells you the exact number of visitors you need per variation to detect a real difference with statistical confidence, so you never make a decision based on incomplete data.
Sample size depends on three factors: your baseline conversion rate, the minimum detectable effect (the smallest improvement you care about), and your desired statistical power. This calculator handles the math and translates it into practical terms — how many days your test needs to run given your traffic volume, and what minimum improvement is realistic to detect at your traffic level.
How to Use
Enter your baseline conversion rate
Input the current conversion rate of the page or element you are testing. Use at least 30 days of historical data to get an accurate baseline.
Set your minimum detectable effect
Choose the smallest relative improvement you want to detect. A 10% relative improvement on a 2% baseline means detecting a move from 2.0% to 2.2%. Smaller effects require larger samples.
Choose confidence level and power
The default of 95% confidence and 80% power is standard for e-commerce testing. Increase power to 90% if the decision is high-stakes.
Review your required sample and test duration
The calculator outputs the required visitors per variation, total visitors needed, and estimated test duration based on your daily traffic.
Formula
Interpreting Your Results
If the required sample size translates to more than 4-6 weeks of testing, you have two options: increase your minimum detectable effect (look for bigger wins) or increase traffic to the tested page. Tests running longer than 6 weeks are vulnerable to seasonal effects and external confounders.
This is one of the core reasons Eevy AI uses genetic algorithms instead of traditional A/B testing. A standard A/B test compares 2-4 variations and needs weeks of data. Eevy AI tests hundreds of layout combinations simultaneously, using evolutionary selection to converge on winners faster. It achieves statistical rigor not through large per-variation samples but through the mathematical efficiency of genetic optimization across many small experiments.
Let AI do the optimization
Eevy AI's genetic algorithm handles the testing and optimization automatically — no manual calculations needed.
Try Eevy AI Free