This post describes how to use Episerver A/B testing and Episerver Visitor Groups to test an entire customer journey against a Episerver A/B test KPI. 

Why would you want to personalise based on a running A/B test? 

Sometimes A/B testing an individual piece of content isn’t enough to really prove an outcome. A simple example could be does the text “Continue” or “Next” work best on a button on a checkout path. The button exists on multiple content items so would need to be changed in multiple places. By using Episerver Visitor Groups it should be possible to test multiple pieces of content on a site using visitor groups based on a running A/B test.

Implementation 

By default there are no visitor group criterion for running A/B tests. The code for this can be found over on GitHub and there is also a Nuget package that adds the visitor group to your solution.

Once installed it’s simply a case of setting up the A/B test as normal:

Then set up a visitor group using the "Participating in A/B test” criteria from the new “A/B test criteria” section:

Then simply use visitor groups to personalise as normal based on the running test:

This allows editors to personalise an entire journey based on a running A/B test. This means it’s possible to test a journey against a KPI, not just an individual piece of content which is a powerful feature. 

Outcome

The outcome is that editors can now experiment with multiple changes across a site using visitor groups and A/B testing: 

Conclusion

This shows a simple example of how it’s possible to extend Episerver to allow site-wide A/B tests. 

Comments

Recommendations for you