EMMA: A Case Study
Anjini Karthik, Sara Kolb, Esme Nava, Anand Shankar, and Misbah Surani
Exercise, who? Don’t know her. 👀
Despite an abundance of research articles and popular science highlighting the benefits of regular exercise to physical and mental health, only 23% of Americans hit weekly exercise targets. During the pandemic, we personally noticed decreased motivation to exercise among our peers and wondered if we could design an intervention to combat this phenomenon.
How might we encourage users to develop a consistent exercise routine?
But first… who are our target users? 👫
Meet Zahra, a woman in her early twenties who currently works as a software engineer at a Bay Area tech company. After work, she enjoys catching up with her friends, watching TV, and relaxing. She wants to increase her overall level of physical fitness through consistent exercise but finds it challenging to motivate herself to take out time to work out every day. We want to help people like Zahra, between 18–30 years old, who are daunted by exercise.
What exactly is the problem? 😔
From our baseline study, conducted with 5 participants over the course of 5 days, we reaffirmed that our participants viewed exercise as a time commitment and prioritized their social activities or leisure time over working out. They also viewed exercise as a taxing activity and were less likely to engage with it when already tired, bloated, or stressed about work. Boredom and high intrinsic motivation, however, encouraged our participants to exercise. A key insight is that participants did have some motivation to exercise at any given point in time (as they recognized its physical and mental health benefits), but they lacked that extra bit of motivation to overcome their inertia.
Accordingly, we designed our dairy study to catalyze a cognitive shift in our participants, such that they would use our nudges to view exercise as easy and quick instead of hard and time-consuming. Every day for five days, we sent our 5 participants an easy and short at-home workout video to complete and record using the Google form (above, right). The videos started at 10 minutes and increased in length by 2 minutes each day for a final duration of 18 minutes. This intervention was twofold: We nudged our participants to exercise by texting them personalized video workouts and gradually increased the duration so that we could “low-ball” to remove initial activation energy and slowly build up to a strong exercise routine.
100% of our participants:
- Exercised at least once over the 5 days
- Only remembered to exercise due to the video prompt
- Cited feeling better after completing the workout
We identified that removing initial activation energy to exercise is key to a successful intervention and that mood improvement is a side effect we might channel. Without class schedule constraints, it would be interesting to see if participants keep up their new routines for an extended period of time.
How should we try to solve it? 💡
As a team, we generated 35 ideas ranging from the everyday (e.g. a text-based exercise reminder platform) to the fantastical (e.g. a hologram that pokes you if you don’t work out before dinner).
We identified our top 5 ideas through heatmap voting, mocked up those ideas, and subsequently narrowed down to 3 ideas that incorporated repeated themes:
- Competition: We wondered if we could harness social motivation as a nudge to encourage users to exercise regularly.
- Gamification: We felt that we could emphasize the “fun” within our goal of making exercise easy and fun so that the activity feels like a game instead of a chore. Game-based reminders also mimic the reminder nudges that were successful in our diary study.
- Customization: Since diary study participants indicated that receiving a predetermined (albeit personalized) workout simplified their exercise routine, we hoped that this prototype could emphasize the “easy” aspect of making exercise easy and fun so users are less daunted by the activity.
We chose the gamification approach for our prototype because games in the market (e.g. Zombie Run) have successfully maintained user engagement, which our platform needs to ensure long-term adoption, and because games let users visualize slow progress, a feature that may serve as another motivational nudge. After all, an exercise habit is built incrementally. However, games don’t inherently provide opportunities for workout customization to suit individual users’ needs, nor do they necessarily engage users and their close social network in a competition. We sought to incorporate some of these ideas within the gamification approach through our storyboard and wireflows:
Key elements (labeled by number on the wireflow diagram):
- Game-like features and transitions, including customizable avatars, points earned per workout, and a weekly leaderboard
- Personalized metrics where users can see their statistics after each workout
- Social competition such that users can invite their friends to facilitate healthy competition among close connections instead of just relying on the app-wide leaderboard
How will we design the best interface? 🖌
We wanted to create a fun and playful visual vibe to reach people who might otherwise be daunted by exercise. We chose orange as our signature color (since it’s already associated with workout brands, like Strava and Orange Theory) and green as a complement (to add an outdoorsy feel):
We intended that our app be modular and easy to use, so we stuck to four key functions (Workouts, Leaderboard, Progress, and Profile), as reflected in our sitemap, and mostly linear taskflows.
In accordance with good design principles, we generated sketchy screens (not pictured) before translating our ideas into a medium-fi Figma prototype. Our medium-fi prototype is designed to present a fun and energizing interface that makes it easy for the user to carry out key tasks.
What should we improve? ⚙️
We conducted usability tests with three classmates to identify aspects of our medium-fi prototype that we could improve. With additional time, we would test with at least 10–15 users as well as conduct a heuristic evaluation to categorize suggested improvements.
- Confusing points system: Multiple participants thought points would be redeemed to “buy something” in-app and didn’t realize that points were just our metric to track progress on the leaderboard. We plan to change the word “points” to something less transactional (e.g. “stars”) to indicate that we’re focusing on exercise progress, not purchases, with this concept.
- Unclear challenges: Participants found the language of “challenges” to be misleading, since starting a challenge on our app simply means doing a workout. We plan to change the language of “challenges” to reflect the fact that users’ choosing to start a challenge simply means letting the platform pick a workout instead of manually going through videos to select one.
- Purpose of friends: If users didn’t start on the Leaderboard tab, they didn’t always realize they could work out with friends since the logic for “Add Friend” is embedded into the “Start a Challenge” flow. We plan to integrate our social component more thoroughly in all aspects of the app to make the social nudge more salient.
- Ambiguity upon opening the app: In this iteration of the prototype, we didn’t include onboarding instructions. We very quickly learned that we need to do so in order to make users feel supported and guided from their very first interaction with the app, even if the home page seems self-explanatory.
- Relevance of statistics: One of our participants expressed a desire to see more relevant statistics, such as calories burned and time spent exercising, on the My Progress page. We plan to conduct more research to understand what statistics are standard for workout apps (and how we might do better than products currently on the market!) before adding statistics to this page.
- Additional nudge: One of our participants expressed a desire to see a “weekly target” field and progress bar that helps users recognize how far they are from achieving that weekly goal. We plan to potentially conduct a mini-intervention study for this type of nudge and, if it proves valuable, incorporate this field into our interface design.
Where do we go from here? 🚀
It’s been a great ride working on EMMA this quarter! If we had more time and/or access to participants, we would have loved to conduct another intervention study where our app (as represented by our medium-fi prototype) served as the intervention. It would be interesting to see how EMMA compared to basic text messaging reminders or other workout platforms. We also look forward to prototyping the social aspect of our app (i.e. features allowing users to add friends and complete workouts with them), which we will do once we gather data from initial user studies as to what an ideal social-exercise intervention looks like. Finally, we would be excited to consult with experts in coming up with the algorithm to generate personalized workout recommendations and ensure that the datasets we use to train our model are unbiased and representative of our target population.
Key takeaways from this project include the following. Firstly, and perhaps obviously, behavior design is challenging! While we were successful in modifying participants’ behavior through our diary study, encouraging users to maintain their behavior changes over time and without constant external motivation is a very different ballgame. Secondly, attention to detail is paramount. A well-designed nudge can be distinguished from an ineffective one by just a few minor differences, and as behavior designers, it’s our job to identify and harness these details. Thirdly, behavior design involves much more than UI/UX work or frontend development. It’s an interdisciplinary field that sits at the intersection of psychology, ethics, and computer science and is informed by perspectives in each of these fields.
Thank you to the CS 247B teaching team for giving us the opportunity to deeply explore this domain and for providing guidance along the way! ✨
