A/A testing is a method used in the field of website optimization, where the same version of a webpage is shown to two different groups of users. This is done to check the accuracy of the testing tools and to ensure that the data collected is reliable and not due to random chance.

How A/A testing Works

In an A/A test, the same version of a webpage (version A) is shown to two groups of users (Group 1 and Group 2). Since both groups are seeing the same version of the webpage, there should ideally be no significant difference in user behavior between the two groups.

Purpose

The main purpose of A/A testing is to check the accuracy of the testing tools being used. If there is a significant difference in user behavior between the two groups, it could indicate a problem with the testing tools or the data collection process.

Online and Offline A/A Tests

A/A tests can be either online or offline. An online A/A test is run on real users, with an engineer instrumenting your app with the Statsig SDK to check for experiment assignment. Since there is no difference in experience to the user, you expect to only see statistical noise.

An offline A/A test works by querying a representative sample of your data, randomly assigning subjects to Test or Control, computing relevant metrics for Test vs Control and running them through the stats engine. You're looking for the % of false positives.

A/A testing example

For instance, if you're using an A/A test to test a new feature on your website, you would show the same version of the webpage (with the new feature) to both Group 1 and Group 2. If the testing tools are accurate, there should be no significant difference in user behavior between the two groups.