By Andrea Sipos, Product Manager at Skyscanner Engineering
In 2015 we redesigned and rewrote our Flights application completely. Following a huge amount of research, planning, usability testing sessions and coding hours we were incredibly excited about the change and couldn’t wait to see how users and the market would react to it. After releasing it in October 2015, we saw the first reviews coming in and we were satisfied. We expected a small drop in ratings due to the big change and fortunately after a few days all of our metrics increased to a higher level than they were before.
There was only one issue: the ratings in The Netherlands didn’t seem to recover. We didn’t understand why this was happening so we started to dig deeper. We started to analyse users behaviour, compared the patterns in the old and the new apps and arrived to a conclusion: we made the experience in The Netherlands harder, so their sad and frustrated reviews were completely reasonable.
Believe it or not, the change we introduced was a logical and natural decision. Instead of the leg-by-leg selection, when users first select the outbound and then the inbound flight, we showed return itineraries as one product when outbound and inbound combinations are displayed together as one option. Most travel sites are like this, so what’s wrong? In most cases, this was a much more seamless experience, but as it turned out in a few cases it makes the selection more complicated. Unfortunately in The Netherlands this was the case for most searches .
Looking at The Netherlands, Amsterdam is their busiest airport, which has direct connections to hundreds of airports. The most popular routes have more than 50–100 direct flights a day. For example, a one-way search from Amsterdam to London shows more than 200 results for a casual Monday in October. This is the root cause of our problem: in the leg-by-leg selection layout, users had to choose from a 100+ outbound flights, then 100+ inbound flights. The nightmare started when we listed all the flight combinations, so travellers were forced to choose from more than 1,000 itineraries. Not an easy job. Not a happy traveller.
More cognitive load, more pain
We tend to consider selection a single step in the user journey: the necessary transition between searching and booking. But if we take a closer look at what is going on in a traveller’s mind while scrolling through the results list, we can see it’s quite a heavy task of information processing, analysis and eventually coming to a decision.
To be able to decide, the traveller first has to process and analyse the data, so they scroll up and down on the results list. Due to a general pattern of judgement called information bias, people feel that the more information they acquire, the better their decision will be, so they try to take in as much as possible. However, the use and storage of information — both in memory and in the external environment — requires a significant amount of attention. Basically, your head feels like how your laptop reacts when you have 28 browser tabs open: your brain starts to hurt and slow down when there’s too much to process.
Make it simple
During this research we realised that we needed to understand the nature of our result set much better and also what selection problem a user needs to face on specific searches. A great exercise for it is the following:
How do you explain a flight route to your partner when you’re planning your holiday? You look at the results and outline your options: “There are a lot of direct flights on this route for the same price, so we can choose when we’d like to go.”
Or: “There are only direct flights on Tuesdays and Fridays, so it’s worth leaving on one of those days, rather than wasting a lot of time with transfers.” It’s a better summary than showing a list with 1,000 results, isn’t it?
We started to solve the issue we saw on the Amsterdam-London search and experimented with several solutions until we found the one which really improved the quality of the selection experience.
What does the traveller need to think through on an Amsterdam-London search? First of all, there are more than 1,000 results, fine. Looking at the first 20–30 results the prices are really similar, the route is dominated by low cost carriers. As this is a direct route, the duration of the itineraries is the same. So what makes the difference? Departure times are obviously different and also the different carriers go to different airports in London. Basically the traveller needs to select which airline they would like to go with to which airport and after they need to select the departure times. This is what our solution supports. This so-called Timetable Widget appears above the result list and shows the available airlines, their cheapest price and all the departure times. After selecting an airline, it leads to a leg-by-leg selection view.
The solution above gives a quick overview of the results and highlights the main decision points. After introducing it, our conversion numbers and ratings went up, as we reduced the cognitive load and made the selection less painful.
Understanding what we show
A huge learning from this project was that building a selection experience is not only about crafting the user interface and making the experience as easy and seamless as possible: we also needed to learn as much as we can about our content. The characteristics of the content will bring in unique selection problems that users face with using our product. If one builds a one-size-fits-all solution, you will miss the opportunity to make a user’s life easier. The power of customising the experience to content is that one can provide value to all users who use your product, regardless of how much is known about them.
Knowing what is happening in your users’ mind when they use your product is essential in managing that product well. It can give you new ideas, new inspiration and highlight what are the most critical problems to solve. You can build any new feature, but users might not use it as they are struggling with processing information. In our case, understanding what decisions users need to make about their flight itineraries while using our products was essential.
If you are interested in what change this discovery brought to our engineering approach, please read a previous article, From itineraries to widgets from Zsombor Fuszenecker.
SEE the world with us
Many of our employees have had the opportunity to take advantage of our Skyscanner Employee Experience (SEE) — a self-funded, self-organized programme to work up to 30 days during a 24 month period, in some of our 10 global offices. There is also the opportunity to work for 15 days per year from their home country, if an employee is based in an office outside of the country they call home.
Like the sound of this? Look at our current Skyscanner Product Engineering job roles.
About the author
My name is Andrea Sipos and I work as a Principal Product Manager at Skyscanner on the flight selection experience. I love to do research, dig into the details, discover hidden patterns and come up with solutions which make peoples lives easier.
Outside of my work at Skyscanner, I love playing board and card games, reading or discovering places.