Gatwick Airport approached us wanting to improve conversions. In particular, they asked: “how do we increase bookings on mobile?”
With mobile being an ever-growing area of online traffic, Gatwick Airport appreciated that work was required to make more of this audience. Having such a broad question, however, meant we needed to understand how users were behaving to be able to identify where the difficulties lay. By having this research in place, we would be able to iron out those issues and encourage more users through the booking funnel.
To assess where the greatest impact within the mobile experience could be, we started by looking at the existing user journey.
With six steps between landing on the site and booking confirmation, we paired each stage with the percentage of users who dropped off. We found the biggest drop off was on the ‘Parking options’ page – with 67% of users leaving the booking funnel. With such a high drop off rate, it was likely we could have the biggest impact over the shortest period of time on this stage in particular.
Exploring user behaviour
Using click maps, scroll maps, user journey recordings and algorithmic eye tracking, we were able to see exactly how users were acting on this page and what they were seeing.
From click maps we saw a lot of people interacting with the date picker. We recognised this as unusual behaviour as users had just entered their date on the previous step; it was unlikely their travel dates had changed in the time between pages, so why were they changing their dates?
On looking at the page, we noticed the only call to action visible on page load was the ‘Apply’ button – users were being asked to change their dates before they saw a product name or price; this needing adjusting to help retain user focus.
Generating a hypothesis
We hypothesised that by reducing the date picker we’d give focus back to the products, bringing the ‘Select’ button into view. This would therefore raise the click-through rate on the ‘Select’ buttons and ultimately increase bookings.
Creating variations and testing
To test our hypothesis, we produced a variation so we could compare the original against what we believed would perform better.
Our variation reduced the size of the date picker, allowing the content of the first product to come into the window. We then ran this as an A/B split test, sending half the traffic to the original design and half to our optimised variation.
This variation had a substantial 6.9% increase in bookings – a fantastic outcome! But it left us wondering if more could be done. If we were even more economical with space, could we raise bookings further?
Iterating for extra uplift
Building on our first test, we hypothesised that changing the product panel layout so they took up less space would allow users to compare products. If we could get the key information for more than one product on the screen at a time, we believed even more users would continue through the booking funnel.
We built a new variation, keeping the date picker collapsed but adjusting the product panels underneath.
We managed to rearrange the content so the parking options’ USPs, price and ‘Select’ button were now visible on page load – with the second parking option name, price and button visible on most devices as well, indicating more content on the page.
To test this variation, we were confident our initial findings were statistically significant, and so decided to only compare the newest variations. Rather than comparing our new design against variation A from the previous test, we compared it to variation B instead.
This experiment iteration saw bookings increase by 11.1%! This meant that if we compared the conversion rate of our latest variation to the original design, we had a strong combined bookings uplift of 21% on mobile.