Summary
If you want to test two or more different expressions (or a part) of a user interface, you can use the A/B Testing method. Some of the users will see version A and some will see version B. After a few users have visited you can evaluate how both versions performed in comparison to each other. There is a variety of tools/resources that help you to carry out the test, as FigPii or trbo.
Step-by-step
- Identify your web page's most-traffic-areas. Which functions are particularly popular? Where did User Testing reveal room for improvement?
- Create at least two different versions (A, B, C etc.) of one feature, which differ in one or more aspects from each other.
- Wait until you got a good amount of impressions (i.e. users have seen it) on all feature variations.
- Evaluate the results, e.g. whether one version was used longer or more effectively than the other ones. Then implement the variation that performed best.
- If you’re using an A/B Testing software the results will be automatically calculated. This can save time and provide a good overview of all uses.