A/B Testing Screen Resolution (Per-User) Calculator
Estimate the number of users needed in each A/B test variation to detect a meaningful change in conversion rate.
Overview
Use this A/B testing screen resolution calculator to estimate how many users each variation needs before you run a conversion-rate experiment. Enter your current baseline conversion rate, the smallest absolute change you want to detect, and your preferred confidence and power levels.
How it works
The calculator estimates sample size for a two-sided test of two conversion rates. It converts the baseline rate and target change into proportions, then combines the selected significance and power thresholds with the expected variability in conversions. The result is rounded up and assumes traffic is divided equally between the control and treatment variations. Smaller effects, higher confidence, and higher power all require more users.
How to use this calculator
- 1Enter the current conversion rate for your control experience.
- 2Set the smallest conversion-rate increase or decrease that would matter to your decision.
- 3Choose a confidence level for limiting false positives.
- 4Choose the statistical power you want for detecting that change.
- 5Review the required users per variation and total sample size.
Example Calculation
Baseline conversion rate
10%
Minimum detectable effect
1%
Statistical significance
1.96
Desired statistical power
0.84
Users needed per variation
14,735 users
With a 10% baseline conversion rate and a goal of detecting a 1 percentage point change at 95% confidence and 80% power, the test needs about 1,470 users in each variation, or about 2,940 users in total.
Frequently asked questions
What does users per variation mean in an A/B test?
It is the estimated number of eligible users needed in the control group and separately in the treatment group. With an even split, the total required traffic is twice this figure.
What is a minimum detectable effect?
It is the smallest conversion-rate difference that would be meaningful enough for your test to identify. This calculator uses an absolute change in percentage points.
Why does a smaller detectable effect need more users?
Small changes are harder to distinguish from ordinary random variation. More observations are needed to detect them with the same confidence and power.
Should I use 95% confidence and 80% power?
They are common planning choices for A/B tests. A higher confidence level or higher power reduces some statistical risk but increases the required sample size.
Does this calculation work for revenue or average order value?
This version is designed for binary conversion outcomes. Revenue and other continuous metrics need a calculation that also accounts for the metric's variation.
Can I use this result if my traffic split is uneven?
The estimate assumes an equal split between two variations. Uneven allocation generally increases the total traffic needed and should be calculated separately.
Explore Related Calculators
Assumptions and warnings
Assumptions
- The test compares two independent variations with an equal 50/50 split of eligible users.
- The primary result is a binary conversion, such as a purchase, signup, or completed form.
- The calculation uses a two-sided normal approximation for comparing conversion rates.
- Users are counted once and each user's conversion opportunity is treated as independent.
- Results are planning estimates; actual requirements can vary with data quality, exclusions, and the analysis method.
Warnings
- This calculator provides an experiment-planning estimate, not a guarantee of a statistically valid outcome.
- Do not use results when the baseline rate plus the selected detectable effect would exceed 100%.
- Avoid stopping or extending a test solely because interim results look favorable; this can affect false-positive risk.