A-B split testing is a comparison study of designs. Main purpose is to compare designs to conclude which is more usable. It can be used to compare new design with old one or two completely different design options. Sometimes new features can also be tested with this test by making it available to limited number of Users. Design option “A” can be the existing design which is presented to most of the users and option “B” can be existing design with some new experimental features which is served to small number of users..
A-B testing is done on live site, thus it gives you the live impact of the new design or new features. Two options are compared with the help of statistical data (e.g. no. of clicks).
- It measures the actual behavior of your customer under the real world. Based on the test results, you can determine option A works better than option B.
- It can measure very small performance differences between design options.
- Its cheap.
- It works for fully implemented designs.
- Can only test one page element or design at a time
A-B test can be conducted in many ways. Different web scripts can be written to present different designs to different User groups. Some third party tools and scripts are also available.