A method of comparing two versions of a website, page or app to figure out which performs better on the same audience. In this experiment, users are shown two or more variants of the same page/app, and with statistical analysis you can identify which version works better in terms of conversions.
It’s important to take into account the timeline for running the test. It should cover full weeks and last for a minimum of 2 weeks. If traffic volumes are small, the tests should be longer.
Each test should be supported by a specified thesis.
More about A/B testing on Piwik PRO Blog: