Hayes and Jarvis – Booking flow design sprint

Hayes and Jarvis Booking Flow – Design Sprint

Continuation rates through the booking flow were identified as a cause for concern by the Product Owner. We saw a particularly large drop off on the last stage of booking.

We agreed as a Sprint team that we would embark on a design sprint to explore alternative ways of stepping customers though the booking flow which includes selecting alternative flights and transfers.

Ideation session

Following some expert interviews we asked the whole delivery team to sketch alternative ideas for the booking flow. We wanted to get as many ideas out there in the shortest time possible.

Ideas were voted on at face value and then as a team we ran through each idea to understand what people had liked about them and why.


After choosing to go for ‘a rumble’ (setting two ideas against each other we developed the two best ideas into storyboards which were used to to test against one another (and the live site) during usability testing sessions.

Prototype 1

This prototype turned the process into a very linear process. Giving customers smaller and more discreet steps to go through in order to book. We also tried using a horizontal form layout (placing the fields side by side) and utilised interstitials between pages.

Prototype 2

This prototype kept the existing ‘hub’ page for flights and transfers but aimed to provide a clearer progression through these steps. We also tried a much more visible and sticky progress bar in the hub to see whether people understood it better than the live site.

User testing

We ran two rounds of user testing with an iteration day in-between so any pinch points in either prototype could be quickly addressed. We also tested the live site alongside to see how the process compared.

We recruited and tested 5 participants on the first day and 5 participants on the second day. Participants were given a scenario and then asked to complete a set of tasks for each prototype. At the end of the tasks they were asked for ease of use scores for each prototype.


Overall people did not find the prototypes significantly easier to use than Hayes and Jarvis.

Where 1 = Very easy. 5 = Very difficult

Control – 2.2
Prototype 1 – 2.9
Prototype 2 – 2.6

Although the fact the control was more finished than the two prototypes we did not feel that there was string enough evidence to move toward either model which would have been a big and potentially risky change to the booking flow.

Several features and issues were highlighted with the control ,which – if corrected – could address some of the key concerns from the business, fulfil missing user needs and potentially improve the ease of use scores further.

Key takeaways

This process clearly demonstrates the importance of testing and validating our ideas. In a few days we had evidence that perhaps the problems identified initially could be resolved through iteration and identifying key improvements. Had we followed a full redesign route this could have potentially been very wasteful and even may have had a negative impact on the continuation rates.