Explaining the concept of A/B testing

We may sometimes update something in our application and want to check its impact on end users and their behavior. Generally, these kinds of updates relate to the usability or popularity of an application and are associated with UI changes. To allow us to check the impact of an update on an application, we need to create groups of end users, who will be thrown the new updates and asked to evaluate them. The deployment patterns we've already discussed don't solve this problem for us as they're unable to throw a new version of the application to a separate group of end users. Although they allow us to test the functionality of the application in a separated group, that group doesn't persist for very long.

A/B testing, however, is a deployment pattern that allows us to throw a new version of an application to a selected group of end users only. This makes it possible to evaluate the impact of a new version of an application on end users and therefore decide...