Atlanta is projected to reach 10 million residents over the next 40 years. Today, however, the city’s transit system suffers from uniquely low ridership, even among American cities. The Metropolitan Atlanta Rapid Transit Authority (MARTA) faces challenges that particularly discourage infrequent or first-time users.
Through User Research, we identified that MARTA had a discovery problem, hiding the system’s scale and amenities. The system is difficult to navigate and understand despite its relatively simple structure.
We developed an interactive station pylon system to encapsulate navigation, discovery, and schedules in a single interaction that would require no other devices. A pylon such as this could be installed in any station, and provides an affordable solution to quickly reduce confusion on MARTA.
Such a system would provide scalability to support the city’s new population, primarily implants from Northeastern cities eager to use public transit.
We spent a cumulative six hours in MARTA stations, carefully observing user behavior across three stations. We paid particular attention to how riders navigated the platforms and station environment more broadly, particularly at major transfer points. This helped us formulate our interview strategy.
Due to the nature of train stations, extended interviews with MARTA riders in context was not possible. As a result, we conducted fourteen 3-5 minute intercept interviews with consenting participants while they waited for their train to arrive.
These interviews focused on the users's familiarity with MARTA, as well as how they navigated the system, and provided an opportunity to voice their complaints and provide general feedback with regard to the transport system.
MARTA users were split somewhat evenly into three categories:
Frequent user (Commuter, or usage at least once a week)
Infrequent user (Monthly or even less frequent usage)
First-time user (Out of town visitor or one-off usage)
Additionally, two key complaints emerged:
MARTA schedules and stations are confusing
MARTA feels unsafe
Based on the user types, we created three primary personas:
All persona details have been directly adapted from individuals interviewed or observed.
User Flow and Scenarios
Our early scenarios centered around improving the payment system, but the emphasis on physical improvements as opposed to mobile designs arose early on. After more iteration, we gradually moved away from payment and towards navigation, which we found was a more significant problem.
By later milestones, we had fleshed out the concept of supplementing in-station navigation options.
An early storyboard that illustrates a simplified, lower friction path to reload a breeze card.
Early illustrations of improved signage
Our first ideas of dynamic signage and the pylon, which would soon take a life of its own.
Initial Prototypes for Signboards
We chose to elaborate upon our Interactive Station Pylon prototype because it had the most potential as an interactive tool for navigation assistance which exists outside the confines of a mobile application.
We evaluated our prototype on a touchscreen display with multiple MARTA users, asking them to complete the following tasks:
Find out how long you will have to wait for a train to Lindbergh Center
Identify which bus routes you can take from Lindbergh Center Station
Find a place to eat near Sandy Springs, and how long it will take to walk there.
Describe when and where the #30-eastbound bus terminates
Identify how long it will take to get to East Lake Station
Along with observing and taking notes, we asked for their thoughts after their evaluation. This allowed us to evolve our design in a human-centric way, making sure that any changes were in service of usability.
We received valuable feedback and noticed key usability problems with our prototype design.
Our key insights:
Users were confused by why a station would appear north of them on one screen and "south" of them on another.
Enlarged station icons could represent too many things — current location, transfer points, and final destination.
The Home Page focused on only stations you could access, but this meant it became difficult to identify one's spatial location in the system as a whole.
There were multiple usability issues that prevented users from realizing they could tap or scroll.
Pylon in Context
Send To Phone
Taking our evaluation feedback, we made a couple of key updates. This began with an updated visual design that matched MARTA visual guidelines. We replaced several indicators with unique iconography to make current location clear, and replaced collapsible headers and buttons with scroll interactions wherever possible. We also solved the spatial issues by offering a full map view, and replacing navigation views with a horizontal and simplified station map to make it clear that displayed Journeys are not necessarily spatially accurate.
There were three key points we came to realize:
The crux of UI design is direct user research and feedback. Our entire team agreed that the two most important steps of our design process were the intercept interviews we performed, and the task analysis component of user research. These methods best enabled us not only to identify the problems to solve, but also which designs would actually address them.
The implicit assumptions which designers make are numerous and difficult to avoid. We were shocked to see the results of our task analysis because we found ourselves watching potential users struggle to use components of our design which we thought were obvious or otherwise easy to understand. That experience is a testament to the fact that as designers, we’re privy to make assumptions about how our designs will be perceived to a far greater degree than we imagined.
It doesn’t always have to be an app. While there’s plenty we would do differently, we believe that the choice of pursuing a design which was not a mobile app paid off - we didn’t have to justify the download of new software, and we got to experiment with a more unique set of design parameters, which ultimately yielded what we would argue a more thoughtful design than we might’ve produced had we created an app.
We definitely made some mistakes in our process, as well:
We underestimated the importance of user feedback, and by extension iteration, in design. We were impressed by how useful the feedback we received in our task analysis ended up being, but it came too late in the course of the semester to act more substantially upon. Given another chance, we would make a point of going through three or four iterations of our initial prototypes and making changes based on the results of quasi-formal user studies. In this same vein, it would’ve been a good idea to develop more prototypes than we did.
We should have considered using a wider array of user research techniques. The intercept interviews we performed within the MARTA system were interesting and yielded some unique insights, but they only allowed us a surface level glance at any given individual’s MARTA usage and habits. Finding two or three MARTA users to perform a long-form (30-60 minute) semi structured interview with would’ve been a great way to get more context about the system. Moreover, we believe it would’ve supplemented the quick interviews we did nicely.