Project Details

Background

The idea of booking a holiday is exciting for most people, but when it comes to working out all the fine details of the trip, a lot of people find it stressful and prefer package holidays.

But even finding the best deal on a package holiday can be tricky, and that’s where TravelSuperMarket fits into the picture.

TravelSuperMarket is the one stop shop for package holiday comparison. Comparing across leading travel agents, the company shows users the best deals to fit their needs.

TravelSuperMarket is a totally free service and is available on mobile web and iOS.

I joined the business as UX/UI specialist to tackle some problems with the iOS app.

Project Highlights

  • 13.5% overall increase in conversion rate –
  • Pre-filtering: +2%
  • Days either side: +2%
  • Results card MVT: +5%
  • Map view: +2.5%
  • Filter prompts: +2%

Project goal

The goal of this project was to identify issues with the current design and propose solutions to optimise the performance of the checkout journey.

I had a number of different means for getting user data to base my proposals from.

Available data sources:

  • User Testing.com: for remote user testing
  • Mixpanel: for iOS app analytics
  • Google Analytics: for website analytics (I used these as a secondary source of quantitative data)

User profiling

To get started I wanted to run baseline user testing with a small group of users. This would help me understand which features users enjoyed most, and which features were missing or needed improvement. But before I could do this I needed reliable data about the users of the products.

Working with the Product Owner and Head of Data Science, I was able to get real user data which I could use to create define the demographics of the test participants.

Main user segments:

  • Couples living together, approximately 30 years old, income approximately £49 000 per year, take about 6 holidays per year
  • Families with 2 school-aged children, income approximately £49 000 per year, approximately 40 years old, take about 6 holidays per year

Baseline user testing

I set up a test that asked users to find a package holiday to the Canary Islands, which my research showed was the top destination for UK holiday makers. I also asked users to tell us which features they found to be the most useful.

I tested 5 participants, aged between 30 and 40 to get users representing both segments (couples and families).

Main issues:

  • 4 out of 5 users commented that they would usually just choose the first deal offered on a room as they knew that this was the best value (users didn’t want to manually compare deals)
  • 3 out of 5 users commented that they would usually shop by price, and would like to see prices of days either side of their chosen date to find the cheapest dates
  • 2 out of 5 users wanted to see hotels on a map so they could make an informed choice about the location of their accommodation
  • 2 out of 5 users commented that the results page was really long and they struggled to make a choice with so many different hotels shown to them

Features that were liked:

  • 3 out of 5 users liked seeing specials and offers on rooms
  • 2 out of 5 user said they liked the clean and modern layout
  • 2 out of 5 users liked how easy it was to perform a search

Solutions

I condensed these finding into job stories. These are an effective way to communicate the issue and the context in which they become a barrier to user task completion.

Results Card Multivariate test

Job story: When I want to go on holiday and I am busy, tired, and time poor I want to find the best value deal, that has good ratings and meets my needs, quickly and easily so that I can book a package holiday with confidence.

Solution: I proposed a small change to the design that allowed users who just wanted to see the top deal, to view that one on the provider’s website. For users who still wanted to manually compare all deals, they had the option to ‘View All’ deals.

This tweak to the interface reduced the journey by 1 click, but increased conversion by a whopping 5%!

Days either side

Job Story: When I want to go on holiday and I am busy, tired, and time poor I want to see what prices are available across different days so that I can find the best value deal and book with confidence.

Solution: Looking around other popular travel websites and mobile apps I noted what the best competitors were doing, and designed a TravelSuperMarket version for the app.

The Days Either Side graph was released on an A/B test and produced a 2% increase in conversion.

Map view

Job Story: When I want to book a package holiday I want to see where my accommodation is situated on a map so that I can book a hotel that is near local tourist attractions, shops and transport.

Solution: Using a Google Maps service, I visualised how hotels should be shown in a card-style format on a map.

Adding the map view increased conversion by 2.5%.

Apart from these 3 issues, there was potentially a fourth opportunity for improvement. 2 users had commented that the page was really long making it difficult to choose a hotel to stay at.

But despite the comments, these users had still managed to complete the test task. So while there were signs of difficulty in the test, it wasn’t enough to justify redesigning the page.

Looking into Mix Panel (App analytics) I could see that there was a 43% dropoff rate on that page. Using this and users comments from testing, I formed a hypothesis that users were being given too many irrelevant results and needed to be nudged to use their filters to narrow down their choices.

This problem of showing too many results became known as the ‘Infinite scrolling problem’.

Infinite scrolling problem

I proposed a solution that was inspired from other products I had seen. After scrolling past 10 results, a filter prompt is shown suggesting that users try filtering to get better results. If they choose to filter they are shown all available filters, ordered by popularity.

In this case I was able to determine filter popularity from Google Analytics data from the TravelSuperMarket website. Whoever had set up the tracking had done a great job, because we could identify the most commonly used filters and use that data to tweak the app’s interface to meet UK traveller’s preferences.

Adding the Filter prompt, and optimising the display order of the filters increased conversion by 2%.

The final optimisation I had in mind was the ability to pre-filter from the app’s home screen.

Home screen pre-filtering

Google Analytics data had already shown us which filters were most popular. Also, judging from the success of adding filter prompts, it seemed like a worthwhile test to allow users to pre-filter their searches from the home screen.

I designed a version that allowed users to pre-filter by the 2 most poplar criteria; Star Rating and Board Basis.

We set this version live via an A/B testing tool and were pleased to watch the conversion rate rise by a further 2%.