AB Test Sampling
The results of this calculator provide the number of observations required for each alternative in an A/B experiment. You can utilize the this tool to calculate the results of your experiment.
When running an A/B experiment, be careful of confounding variables, such as additional stimuli from the organization during the experiment.
This calculator uses the formula for the Estimation of Sample Size and Power Comparing Two Binomial Proportions. 
For more information, email research@mnoet.com
A/B Test Sampling Calculator
Determine the number of observations required in an A/B Test.
Baseline Conversion Rate: The baseline conversion rate is your current conversion rate, which is the number of conversions divided by the total number of observations. %
Minimum Detectable Effect: Minimum Detectable Effect is equal to the percent change compared with the baseline conversion rate and can be entered as either an absolute or relative value compared with the baseline.
First, choose whether you wish to detect is an absolute or relative value compared with the baseline.
For example, with a baseline of 20% and a desire to detect a lift to 22%, the absolute value you would enter is 2% and the relative value you would enter is 10%. The absolute value is the actual percentage difference and the relative value is equal to the percentage difference divided by the baseline.
%
Statistical Significance: Statistical significance sets your tolerance level for detecting a Type I error, which is rejecting the null hypothesis when it is in fact true - a false positive. 95% is a common value used in the social sciences.
Statistical Power: Statistical power sets your tolerance level for detecting a Type II error- a false negative. 80% is a common value used in the social sciences.
Maximize size of experimental group: Maximize the size of the experimental group by shifting the value of the slider to the right. The slider shifted to the left is for equal size groups. Bear in mind that increasing the size of the experimental group will incrementally increase the total size of the experiment.
Equal groups