Tuesday 1 February 2011 | 17:00
This post was first published in German on the the German Conversionroom blog
The air travel portal TUIfly.com is part of TUI, the world's largest tourism group. On TUIfly.com travellers can book flights to around 170 destinations in 42 countries. These include popular holiday destinations around the Mediterranean sea as well as city destinations. TUIfly.com offers flights from more than 10 airlines, including their own fleet. Conceptually and technically the TUIfly.com website is being taken care of by GCP SinnerSchrader.
In the beginning of 2009, a newly designed site was launched which is now continuously being optimized based on the wishes and behavior of visitors. The basis for this optimization is Web Analytics data and extensive A / B and multivariate testing on the site. The Web Analytics Team of SinnerSchrader leads this process of continuous improvement.
Requiring Personal details at the beginning of a booking process leads to higher exit rates
In the flight booking process, TUIfly.com offers the consumer the option to reserve seats and book extras such a specific menu or additional luggage. The selection of seats and extras and assigning these to the person(s) traveling should be as easy as possible, so the names of the passengers are required at the beginning of the booking. The visitor can then assign the seats and extras to each passenger.
An analysis of the booking process based on Web Analytics data showed that a considerable percentage of consumers abandoned the booking process early in the process when they were asked to provide passenger details. Based on this data driven insight, an A/B test was designed to reduce the percentage of consumers abandoning the checkout at this point and increase the conversion rate of TUIfly.com
A / B test
To find a better way of combining the entry of personal details with the selection of seats and booking extras the following A/B test was set up:
- In variation A (the original) the names of passengers were required before selecting seats and booking extras
- In variation B, representative names (first passenger, second passenger, ...) were used - personal details including the names of the persons traveling were not required until the end of the booking process.
During a four week testing period, visitors to the site saw one or other of these variations when going into the booking process.
Variation B reduces the exit rate by 25%
After four weeks, the test had delivered a conclusive result: Variation B showed a 25% lower exit rate than variation A (the original). The conversion rate (the percentage of consumers booking a flight) increased by a factor 1.01 (i.e 1%) Based on these results, variation B was implemented and is now live on TUIfly.com. Although variation A, where seats and extras are assigned to named travelers, seems to make more sense from a usability perspective the test result shows that consumers are more likely to complete a booking when personal details are not required until the end of the booking process.
What should you do next?
Important in optimizing a website (in addition to qualitative usability improvements) is carrying out a quantitative test. By simply making changes based on usability considerations, you might overlook implications that your changes have for other parts of your website as well as for the main goal(s) of your site. Run A/B or Multivariate tests with a tool like Google Website Optimizer to verify that your site improvement hypotheses really work.
SinnerSchrader is part of the Google Conversion Professional (GCP) network in Germany. If you are based in Germany, you can learn more about this program and find an expert that can help your online presence to maximize profitability. If you are based outside of Germany, there is a network of Google Analytics Certified Partners and Website Optimizer Certified Partners that can support your efforts to maximise the ROI of your website
Posted by Jos Meijerhof, Google Conversion Team and Alexander Czernay, Head of Web Analytics at SinnerSchrader