A/B Testing

A/B testing compares two versions of a web page to see which one performs better. You compare two web pages, version A and version B, by showing the designs to different visitors at the same time. The way each version performs against your set criteria will help you make design decisions based on actual user behaviour.


  1. Construct a hypothesis based on user behaviour or website analytics. For example, “increasing the size of the call to action button will make it more prominent and will increase conversions.”

  2. Create two versions of a page design that will be able to test your hypothesis. One version will be the original design, the other version will be the experiment

  3. Use a software tool that can handle A/B testing on your site. If this isn’t available, you can use your new design and compare any changes with previous analytics data.

  4. Analyze, draw conclusions, and implement the strongest option based on the analytics you receive.


  • A/B testing should only present a single design change in the new version. If there are many different changes it becomes difficult to determine which changes influenced the user's behaviour.