A/B Test Sample Size Calculator
Calculate the sample size needed per variation for a statistically valid A/B test. Set confidence level, power, baseline rate and MDE. Free.
| Baseline CR | 5% MDE | 10% MDE | 20% MDE |
| 1% | 637,008 | 163,092 | 42,691 |
| 3% | 207,936 | 53,208 | 13,911 |
| 5% | 122,121 | 31,231 | 8,155 |
| 10% | 57,760 | 14,749 | 3,839 |
| 20% | 25,580 | 6,507 | 1,680 |
How to Use
- 1 Set your Confidence Level (95% is the standard for most A/B tests).
- 2 Set your Statistical Power (80% is the standard minimum — higher power means lower risk of missing real effects).
- 3 Enter your Baseline Conversion Rate — the current conversion rate of your control page or email.
- 4 Enter the Minimum Detectable Effect (MDE) — the smallest relative improvement that would be worth acting on (e.g., 10% means detecting a 5% → 5.5% change).
- 5 Optionally, enter your Daily Traffic to get an estimated test duration in days.
Use Cases
Before starting a test, calculate the exact number of visitors you need per variation. With a 3% baseline conversion rate and 10% MDE, you need ~53,000 visitors per variant at 95% confidence.
Enter your daily traffic to see how long your test needs to run. This prevents stopping tests too early (peeking) which inflates false positive rates significantly.
When proposing an A/B test to stakeholders, use this calculator to show exactly how much traffic you need and how long it will take — making the case for the right resource allocation.
Use the quick reference table to understand the trade-off between MDE and sample size. Detecting a 5% improvement requires 4× more traffic than detecting a 20% improvement.
FAQ
MDE is the smallest relative improvement you want to be able to detect. An MDE of 10% on a 5% baseline means you want to detect if conversion rate improves from 5% to 5.5% (a 0.5 percentage point, or 10% relative increase). Smaller MDEs require much larger sample sizes.
At 95% confidence (α = 0.05), you accept a 5% chance of a false positive — concluding a winner when there's no real difference. For most product decisions, this is an acceptable risk. Use 99% for high-stakes changes like pricing. Use 90% only for low-risk, rapid iteration contexts.
80% power (β = 0.20) means you have an 80% chance of detecting a real effect if it exists, and a 20% chance of a false negative (missing a real winner). For important tests, use 90% power to reduce that risk — but it requires more sample size.
Use your current conversion rate from analytics. Look at the last 2–4 weeks of data for the specific page, email, or flow you're testing. Accuracy here is critical — even a 1% error in baseline rate significantly changes the required sample size.
No. All calculations run entirely in your browser. No data is sent to any server.
Related Tools
A/B Test Calculator
Check if your A/B test results are statistically significant. Calculates p-value, Z-score, uplift, and statistical power. Free.
Open →ACOS Calculator
Calculate your Amazon ACOS, ROAS, and break-even ACOS instantly. Free tool for Amazon sellers to measure ad efficiency and profitability.
Open →Ad Revenue Calculator
Calculate your estimated ad revenue from pageviews or clicks. Free tool for bloggers and publishers using RPM or CPC models. No signup needed.
Open →