During Spring 2017, I participated in a UX Research project for an HCI course (INFO 214) at the University of California, Berkeley, under the advisement of Steve Fadden, Head of UX Research at Google. This was a Needs and Usability Assessment class and the final project was an investigation of user needs for a proposed product, in this case, the BART ticketing machine.
My responsibilities included:
Our objective is to build a detailed understanding of the BART's ticketing machine problem from users to create a foundation for future design work.
My team and I were to evaluate the usability of the BART's outdated ticketing machine interface and address how to decrease the time users spend using it. Also, we sought out strategies on how to make the machines more usable though the feedback from our interviews.
The questions we had were:
We delivered a usability report with a 2x2 framework, 8 insightful findings, and 10 recommendations about the target users' behaviors and needs about the BART ticketing machine backed by data through our research.
The impact was to provide insightful findings about user pain points and behaviors by synthesizing data into user groups and design recommendations.
Our team was fortunate to use the HCI department classrooms at UC Berkeley to conduct our interviews. Our field studies were carried out at the Downtown Berkeley BART station. Both qualitative and quantitative research methods helped us gather an in-depth understanding of user behaviors.
Over the course of a week, our team observed the Downtown Berkeley BART station using the AEIOU framework during rush hour times (5:00pm-6:00pm). We were interested in seeing how users behave with the machine in their natural setting.
We found that ticket buyers spend an average of 1 minute more than Clipper Card users at the machine. Users spent the most time looking at the price sheet, which was a sticker on the top right of the ticket machine. Majority of ticket buyers purchased tickets using cash rather than credit cards and all change was dispensed in coins. We also noticed that the users who spent the longest time at the machines were people who were purchasing multiple tickets because the machine only allows one purchase at a time.
To discover what parts of the ticket buying flow were memorable, we asked 8 users at the station (4 paper ticket users/4 Clipper Card holders) a set of questions about why they use either a ticket/Clipper Card, what they like/dislike about the machine, and demographic information. We found that frequent riders (3-4x a week) purchased Clipper Cards and described it as their "debit card for transportation".
We discovered that 100% of the users brought up the map with prices on the upper-right side of the machine, but few remembered any of the functions on the screen other than purchasing a ticket. They also mentioned that decreasing the value was second nature to them now.
"It sucks that we can only change the ticket price increment by a dollar or 5 cents; it doesn't make sense. But I'm already used to it, so I press the button a lot until I get to around 4 dollars."
- UC Berkeley Junior (paper ticket buyer)
I led the data analysis for a week. We focused on addressing our 2 research questions of how to decrease the time users spend at the ticketing machines, and how to make the machines more convenient to use. Another pattern between payment method and single vs. multiple use of tickets emerged in our data that informed many of the findings towards our research goals, so we decided to frame our results with in that pattern.
BART ticket machine users can be categorized in 2 ways: how they pay for their tickets (card or cash) and whether or not they use the ticket for more than one ride. We saw that single-use tickets were usually loaded with cash, and to the specific amount needed for that ticket. People who usually keep and reuse their tickets would do so because of the extra money left over from using their card to pay.
Currently, there is no function on the ticketing machine that allows the user to get their leftover money back from the ticket. While card is faster and easier to use for many riders, it leaves them with extra balance on their paper ticket, that they eventually toss out. Users stated that having a $20 default on the machine is an inconvenience for many people because it slows them down.
Because there is no easy access to a map directly from the machines and the chart lists the stations in alphabetical order, rather than geographical or by line, riders who are not familiar with the area must do their own research and find out which stop they need to get to.
The cognitive walkthrough was excellent for observing behaviors through the user's explanation of each step. This helped me to see where users spent the most time during the ticket buying process and how they go about it. I learned that this is an effective way to gather data within the context of use instead of relying on interviews for qualitative insight.
I also recognized that usability problems can change depending on how the user interacts with the machine day by day. This helped me realize that as UX designers/researchers, diving into the users' natural environment to talk and understand how they think is important for improving usability.