Project: Usability study and re-design of SJs ticket machine interface
Course: Usability – metoder och verktyg
Project team: Sara Lindgren, William Falkengren, Billy Astorsson, Nils Arvidsson och Karin Lagercrantz.
SJ (Scandinavian Railways) has ticket machines at 80 train stations throughout Scandinavia. The studied and current version was introduced to the users during 2016 and lets the users buy and collect tickets by using a touch screen interface.
The ticket machine is placed at train stations of different sizes and traveler load. This leads to a user interaction that happens more or less frequent depending on the placement. In general train stations are stressful environments with a lot of people, which can effect the use of the ticket machine.
The ticket machine is used to search, book and pay trips with SJ and other companies that collaborate with SJ, as well as for collecting pre-payed tickets bought through some of SJ’s other booking services.
These choices can be made on the ticket machine:
- Destination – to/from
- Type of ticket – single/return
- Date of departure/arrival
- Number and age of travelers (maximum 4 people)
- Specified seat (by categories, not specific seat number etc)
- Reward Account/Travel card number
- First and second name
The user can also see their travel plan.
The ticket machine was assumed to be used to buy tickets quick and spontaneous, not when users have planned ahead as their is other more convenient ways to buy train tickets (smart phone, internet etc). The users were assumed to be of mixed age spans and have different levels of technical knowledge. The users were assumed to be of different nationalities and therefore speaking different languages.
The study did focused on the usability of the ticket machine’s digital interface when the user search for and purchase a ticket.
To evaluate the current interface and its possible problems and improvable areas a theoretical evaluation was carried out. The methods used were hierarchical task analysis, cognitive walkthrough and Predictive Human Error Analysis as well as usability theory.
For the empirical evaluation of the interface twelve test users were chosen, two of these tested both the original and the re-designed interface.
The test users performed four specific tasks during direct observation of the test leaders and the users were told to think out load during the tests. The users also filled out a survey where they rated their experience with the interface. The tasks, environment and survey were the same in both of the tests (original/re-design).
During the test subjective measurements as effectivity and clarity were evaluated, as well as objective measurements as observed interactions to complete tasks, task time and deviation from optimal interaction.
The tests were carried out on a product representation and not on the real product, this so the experiences could be compared without a difference in intractability. The product representation consisted of a digital prototype made in Adobe XD presented on an iPad placed in a real sized paper model. To mimic the stress at a train station sounds from Gothenburg train station were played during the tests.
Usability – original interface
In conclusion the original interface performed well both in the theoretical and empirical evaluation. The users mental model was coherent with the model presented in the interface. The users received good feedback and could easily correct errors that happened.
The problems were insufficient overview, inconsequent interaction with pop-up windows, easily missable options and the large amount of process steps.
With the goal to increase the usability of the interface it was redesigned. In the following section you can see and read about some of the changes.
Insufficient overview – in the original interface (left) the progress bar at the top of the page didn’t show all steps of the process which confused the users. To increase the overview all steps of the process were included in the progress bar of the re-designed interface (right). Another problem contributing to the lack of overview was the fact that you didn’t get a summary of your travel information until the end of the process. To tackle this problem a side bar was added that present the information as the process proceeds.
Inconsequent interaction with pop-up windows – the original interface (left) confused the users as the pop-up windows only had an exit button and no confirm button. Because of this the users didn’t understand if their choices were implemented or not. In the re-designed interface (right) a ‘done’ button was added to make this clear.
Easily missable options – in the original interface (left) some options were easy to miss, and in this picture it is the seat specification option. This option was hidden in the travel summary and almost no user found it without help. To minimize this frustration this option got to be a standard step in the purchase process of the re-designed interface (right).
Usability – re-designed interface
The re-designed interface and purchase process were experienced as more time effective than the original, even though it gained a few steps. The interface gave a better overview and the users didn’t miss available options.
Comparison time to task completion (left) and numbers of interactions to task completion (right). Both measurements improved with the redesigned interface.
Comparison of the deviation of optimal number of task interactions between the two interfaces. With the redesigned interface the users almost completed all tasks with the minimum numbers of interactions needed. Task number one had a larger deviation in test number two, this might be due to a combination of more process steps and a simple task.
In conclusion the re-designed interface offered a improved efficiency and clarity, as the time to task completion was decreased and the number of faulty or unnecessary interactions were a lot fewer.