The research done
To validate my hypothesis around the users simply wanting to search for a single hotels available dates I set up screen recording on the hotel information pages. This recording was then set to trigger if the user interacted with the hotel availability search.
From the recordings I’d managed to validate my hypothesis and set about demoing some key videos back to the stakeholders to gain buy in for the problem.
Following the validation of my hypothesis I looked at how other people handle the problem of no availability to gain not only inspiration but a benchmark for the work I was doing. I put this all onto a spreadsheet so it was in an easily consumable format and presented my findings back to the stakeholders.
Stakeholder workshop – User flows
I ran a stakeholder workshop to look at the problem and look at potential ideas and associated user flows. Using the analytics data to further prove out the hypothesis and show the extent of the problem, we looked at the existing flow and conversion rates.
I then spent 5 minutes with the experts and stakeholders doing some blue sky thinking, discussing the potential ways we could solve the issue and decided on 3 possible ways forward.
Sketching and critiques
Following the stakeholder workshops, I started sketching out some ideas, gaining feedback at regular intervals, allowing me to iterate through designs quickly.
Wireframes and usability testing
I created an interactive wireframe for each of the ideas based on the sketches and again presented back to stakeholders regularly to allow me to refine the ideas down to something more polished.
I usability tested the wireframes using usertesting.com to allow me to validate that users understood what was being presented to them and that they knew why they were seeing the various aspects of the designs. I also usability tested the existing journey to gain a comparative benchmark.
High fidelity designs
Presenting back the usability testing to the stakeholders, we agreed on a calendar design approach. I created 4 variants with different layouts and information which I worked with the dev to build into our MVT tool and test against each other.
The MVT testing was set up against a control group and we set it live, 20% to control and 20% to each variant. Within a day we could see that all variants were performing well above the control and we were beginning to see a clear winner which eventually ended up on a huge 73% reduction in drop offs and a 23% improvement in hotel bookings.