A/B Test Sample Size Calculator

Find out the sample size you need and how long to run your test before you launch.

Audience

Audience Size
Total audience available for the test.
Per Week
Baseline Conversion Rate
The current conversion rate of the control audience.
Expected Lift
The minimum % change you want to detect between the test and control audiences. Also called the Minimum Detectable Effect (MDE).

Test Setup

Test Audience
A 50/50 split is statistically most efficient, resulting in the shortest test duration.
Control Audience
50%
Test Type
One-sided: Detects if the test audience is better than control. Cannot detect negative impact.
Two-sided: Detects if the test audience is better OR worse than control. Requires a longer test duration.
The probability that the test will detect the expected lift, if it is present. Standard is 80%.
The probability that a detected lift is not due to random chance. Standard is 95%.
ABTestPro.com

Test Plan

The minimum audience size and test duration required to achieve statistical significance.

Estimated Test Duration
N/A
Minimum Required Audience
N/A
Target Power: 80% / Confidence: 95%
Audience Size
(Weekly)
50,000
Test
Split
50% / 50%
Baseline
Conv. Rate
1.2%
Expected
Lift
10%
Expected
Conv. Rate
-
ABTestPro.com

Statistical Power Sensitivity

Expected Lift
123456789101112
# of Weeks
ABTestPro.com

Technical Summary

Audience50,000 / Weekly
Baseline Conversion Rate1.2%
Expected Lift10%
Expected Conversion RateN/A
Expected ConversionsN/A
Test Split50% / 50%
Test TypeOne-sided
Power80%
Confidence95%
Standard Error
Z-Alpha
Z-Beta
Statistical Power per WeekN/A
Test Efficiency
Minimum Required Audience for TestN/A
Minimum Required Audience for ControlN/A
Total Minimum Required AudienceN/A
Test Durationundefined Weeks

How to use the A/B test sample size calculator

This calculator tells you how many visitors each variation needs and how long the test will run, before you launch. Enter four numbers and read the result instantly. Everything is computed in your browser.

  1. 1

    Enter your audience. Set your audience size and how often it refreshes (daily, weekly, or monthly) so the tool can translate sample size into a realistic timeline.

  2. 2

    Add your baseline conversion rate. This is the current conversion rate of the page or flow you are testing. It anchors the whole calculation.

  3. 3

    Choose the lift you want to detect. Set the minimum detectable effect (MDE), your desired statistical power (usually 80%), and confidence level (usually 95%).

  4. 4

    Read the plan. You get the required sample size per variation and an estimated duration. Use the Statistical Power Sensitivity table to weigh smaller lifts against longer tests.

Key terms explained

Baseline conversion rate
The current conversion rate of your control, before any change. A higher baseline generally needs fewer visitors to detect the same relative lift.
Minimum detectable effect (MDE)
The smallest relative change you want the test to reliably catch. Halving the MDE roughly quadruples the required traffic, so it is the most consequential input.
Statistical power
The probability of detecting a real effect when one exists. 80% is the common standard; higher power needs a larger sample.
Confidence level
How sure you want to be that a detected difference is real, not noise. 95% confidence corresponds to a 5% false-positive rate.
Sample size per variation
The number of visitors each variation (control and treatment) needs before you evaluate the result.
Test duration
How long the test must run to reach the required sample size at your audience volume and traffic split.

How the sample size is calculated

The calculator uses the standard normal approximation for comparing two proportions. The required sample size per variation is n = (zα + zβ)² · [p₁(1−p₁) + p₂(1−p₂)] / (p₂ − p₁)², then adjusted by a traffic-split efficiency factor. A 50/50 split is the most efficient allocation.

For the full derivation, the 50/50-split rule, and how to read the sensitivity table, see the complete guide to A/B test sample size.

Frequently asked questions

How many visitors do I need for an A/B test?

It depends on your baseline conversion rate, the minimum detectable effect you want to catch, and your chosen power and confidence. Enter those above and the calculator returns the exact sample size per variation.

What is a good minimum detectable effect (MDE)?

Use the smallest lift that would still be worth shipping. Smaller MDEs need dramatically more traffic (halving the MDE roughly quadruples the sample size), so balance ambition against the time you can run the test.

Why does a smaller lift need so much more traffic?

The expected difference appears squared in the denominator of the sample-size formula, so the required sample size grows with the inverse square of the effect you want to detect.

Should I split traffic 50/50?

Yes. For a two-group test, an even 50/50 split minimizes the standard error and therefore the sample size. Uneven splits inflate the traffic you need to reach the same confidence.

Is my data sent anywhere?

No. The calculation runs entirely in your browser. The numbers you enter are never transmitted to or stored on our servers.

Related guides and tools