A/B testing

The so-called A/B tests are suitable for identifying the better one between two different variants. To do this, the target group is evenly divided into A and B groups — which also gave rise to the name of the test. One group gets to see the initial version, while the other is shown the version to be tested.

It is then analysed which variant works better. Two factors are decisive in order to be able to obtain meaningful and statistically relevant results: the Traffic, which leads to test pages and the difference between the respective conversion rate.

Example:
You sell DVDs on your website and want to find out which headline works better for your exhibition series. In your original version, the button that leads to the films is called “Large selection of DVDs” and in the second version to be tested, “Buy DVDs cheaply.” As part of your A/B test, the “Large selection of DVDs” button is now delivered to 50 percent of visitors to this website. The other 50 percent see the “buy DVDs cheap” button. If you now analyze the order data within the trial period, you can determine which option works better and use this variant for the entire target group after the test is complete. You can then test other elements on your website and optimize your homepage step by step.

Important: Test just one element!
You should only check one element during A/B testing so that your results remain meaningful. If you test multiple elements at the same time, you can't determine in the end which element was decisive for an improvement.