Customer segmentation / personalisation - multivariate test

The hypothesis: Linking in with the airline partners segmentation data will increase conversion because they will be more highly targeted based on their customer type.

For example, a family travelling with kids would be show the correct car size with free child seats included.

What I did: The top level customer segments were Business, family, couple, silver surfer& individual and each one of these was broken in a seasonal sun, winter and generic segment. Then working with the e-merchandiser we were able to also use the price sensitivity segmentation to display different merchandising messages around customer type and price sensitivity making the cross sell more personalised.

Designs were mocked up for these templates, showing the ideas early in the process to stakeholders and the product owner, quickly iterating until I had group consensus.

Then a multivariate A/B test was setup for the 3 templates against the as is. Within these 3 templates the correct segmentation image and message was shown to the customer while they were booking their flight.

Future iterations: Having this scalable is very important because when another target segment is identified I want to be able to move quickly and easily to add this into the segmentation. User testing to be done on the new segments to gain further insights.

Test currently running... results to follow