The final project in course MMT031 - Usability at Chalmers University of Technology, each group of 5 to 6 students had to redesign a user interface of a public system or interactive display. Our group chose Styr & Ställ - a renting bike system, which consists of two screens: a large screen with information about how to rent and choosing a bike, and a samller screen where you type in your password when checking out or when you register for the first time.

The system today is not very usability-friendly, according to a user test we conducted. We built a new physical system with a new interactive display to test if the usability has changed.

Danny Lam, Frida Hagelberg, Lily Ekman, Patricia Paulsson, Mikael Wiberg & Sara Nadi 

WHAT'S STYR & STÄLL?

Styr & Ställ is a rental cycle system located in central Gothenburg with bicycle stations that promote rapid accessibility and a flexible way of transport. Styr & Ställ is for everyone and has at present two forms of subscription: one seasonal card (SEK 75) and one three-day card (SEK 25).

 

Styr & Ställ is a large system consisting of 60 stations outdoors in central Gothenburg and 1000 bicycles where a bicycle can be borrowed in one place and left on another. The station consists partly of bicycles and a station terminal. Many functions can be performed directly at the station, such as buying 3-day cards and picking up a bicycle. Styr & Ställ also has a website for buying seasonal cards and an app for further information.

Styr & Ställ station at Stenpiren in Gotheburg.

PROJECT SCOPE

The empirical and theoretical evaluation showed clear deficiencies regarding the system's usability. The existing interface was perceived as cluttered and untrusted. Users experienced difficulties in absorbing the presented information, which often led to confusion. Furthermore, the empirical tests showed that the system had shortcomings in its interaction with the user, which was primarily considered to be due to the system being operated with a keypad separated from the screen where the information was displayed. The interaction was also hampered when the user was forced to divide his attention between two different screens.

The results of the evaluation of the existing interface were compiled and then lay as a basis for a redesign of the existing interface. Subsequently, the redesigned interface was tested and evaluated in the same manner and compiled to be compared with the existing interface.

I am trying out our recreation of Styr & Ställ.

PHASE 1: Theoretical evaluation

The theoretical evaluations were carried out in order to identify possible errors and problem.

To structure which parts the system contains and which steps the user needs to complete in order to rent a bike a Hierarchical Task Analysis (HTA) was performed (Osvalder et al., 2010). The method first formulates a main goal that is broken down into smaller sub-goals until a desired abstraction level is obtained. At the lowest level of abstraction, the document that needs to be performed for the sub-target is also described (Bligård, 2014). This method gave a good overview of the system's functions and their structure.

Source: pixabay.com / @goumbik

Based on the results in HTA, the methods Cognitive Walkthrough (CW) and Predictive Human Error Analysis (PHEA) were used to investigate which problems and errors that usage of Styr & Ställ can lead to.

 

A CW investigates whether the user is trying to perform the task correctly and why. Each document contains four questions that describe how well the system works based on the concepts of Gulf of Execution and Gulf of Evaluation:

 

    Will the user try to achieve the right effect?

    Will the user discover that the right action is available?

    Will the user associate the right action with the desired goal?

    If the right action is performed, does the user receive feedback on this?

 

PHEA, instead, analyzes the faults that can be performed during use. The method answers the cause of the error, if and how this is detected, its consistency and recovery. Using four questions, possible use errors can be identified:

    Will the user try to achieve the right effect?

    Will the user discover that the right action is available?

    Will the user associate the right action with the desired goal?

    If the right action is performed, does the user receive feedback on this?

PHASE 2: Empirical evaluation

A usability test were conducted, with 10 first-time users. The emphasis of the test was on the criteria Effectiveness, Efficiency, Satisfaction (ISO DIS 9241-11) and Guessability and Learnability (Jordan, 1998). The criterion Guessability was considered important since the system has many first-time users and, among other things, there are people who have never come in contact with the system who need to understand it directly. We built a physical Styr & Ställ terminal and let participants try out two different subscriptions in a controlled environment, the Usability Lab.

The two different subscriptions consisted of one seasonal card and one 3-day card. The latter option can be purchased on-site through interaction with Styr & Ställ's station terminal and is the most likely alternative for first-time users and tourists. Therefore, the usability tests have been based on data included in the purchase and use of a 3-day card. The scenarios A and B:

    Change language and buy 3-day cards.

    Change language and rent bike with 3-day card.

¨The Usability Lab is where we tried out the usability tests.

The usability test consists of an introduction about Styr & Ställ, Scenario A and B, a form that the participant will fill out of how they felt during the Scenario A and B, another Scenario C (it's the same as Scenario B, this is because we wanted to measure their Learnability), ending with an interview about their overall thoughts about the three Scenarios.

Here's how the flow looks like in the current (old) Styr & Ställ graphical user interface.

PHASE 3: Redesign

A redesigned UI with new way of interaction.

Based on the theoretical evaluation and the empirical tests, the greatest problems were identified with respect to the chosen usability criteria. Before designing the new interface, design principles were specified to take into account when designing the new interface.

To overcome overall usability problems and user errors, the Idea Switching Method was used (Johannesson et al., 2004). Each group member gets five minutes to sketch a solution to a problem. After that, the group members changed ideas with each other and continued to generate ideas on a started solution. When an idea passed a whole round in the group, it was put aside and the same procedure was repeated on the next problem.

The results from Idea Switching Method were used as a basis for brainstorming. By not allowing criticism and letting quantity go before quality, a large number of solutions can be created (Johannesson et al., 2004).

Next step was to design the new user interface (see flow below). A new way of interaction were also designed, where you now can tap on the sides of the screen instead of on integer buttons below screen. A usability test was conducted, with same scenarios:

    Change language and buy 3-day cards.

    Change language and rent bike with 3-day card.

The participants were invited individually into the lab room. Two of project members were presented - one who acted as a pilot and one who acted as Wizard-of-Oz. In order to increase reliability, only two acted as Wizard-of-Oz during all tests. Remaining group members rotated between acting pilot and sitting in the control room to observe, record, and film the tests. During the tests, the test leader followed a script and all tests were performed in English for increased comparability. User test 1 and 2 followed the same procedure.

PHASE 4: Evaluation

A lot of comparisons were made after usability test 2. For example, the median values of the time have decreased between the different tests mainly for scenario A and B. In scenario A, the time decreased by 55% while the time decreased by 64% in scenario B. This points to greatly increased efficiency but also improved guessability. The time reduction in Scenario C was measured at 22%, which is not an equally significant difference.

Another example of comparison is the number of clicks. The number of clicks has been reduced in all three scenarios between the original (old UI) and the redesigned interface (new UI). Unnecessary steps have been eliminated and through, among other things, a more consistent and intuitive design, the users succeeded in achieving their goals both more efficiently and effectively. To sum up, it can be mentioned that the time improvement, the reduction of clicks and the improvement of System Potential show that the changes made in the redesign actually have improved the usability of Styr & Ställ.

Y minutes it takes in Scenario X, comparison between Test 1 and 2.

+46765871821

©2019 by Danny Lam.