welcome to wynd.

Elena Wa
6 min readMar 21, 2021

Promote good sleep habits by winding down with wynd

Who: Elena Wagenmans, Christine Chen, Dhruvik Parikh, Kevin Penner, Khalid Ahmad

Role: UX Researcher and UI Designer

When: Jan 2021 — March 2021

Summary: Our team focused on designing behavior change for sleep with the guidance of Christina Wodtke, Sean Chang (TA), and Matt Bernstein(TA) as part of CS 247B.

The Challenge

What are some behaviors that almost everyone wants to change? A few answers come to mind, from exercising more to eating better to drinking more water. What stood out most to us, though, was sleep.


College is a pivotal time for developing good sleep hygiene. Competing academic and social needs, plus the norm of late nights has reinforced the idea that sleep is low-priority. But the consequences are tough; a poor sleep schedule can result in low-quality sleep which in turn can lead to academic underperformance. Unsurprisingly, there are plenty of research papers on sleep, however, proposed solutions don’t really seem to work. As was confirmed later on when conducting initial interviews, existing tools aren’t sustainable and have mixed results when it comes to actually stimulating behavior change. While “the analysis found that the apps were generally well-designed and easy to navigate, and most helped users set goals and track sleep patterns, only a minority of the apps included features that support behavior change.”

With our solution, we wanted to set up our users up for success through habit building and effective prompts.

The Process

Our process in finding a solution started with various studies to better understand not only the problem, but our users’ behavior. We asked 9 students and recent grads to track numerous data points over the course of a week, including caffeine intake, exercise, naps, and typical bedtime routines. To better visualize this data, we created individual journey maps.

Proto Personas of 5 people, describing their background, sleep behaviors, motivation, setting and environment, attempts to solve, and general tools

This helped us better understand typical pain points, motivations, and factors of daily life which can affect sleep. From there, we were able to integrate and prioritize features to create a synthetic persona and journey map of Jeremy Rest.

During our initial study, we found that school and social anxiety would often cause students to pile on activities, tasks, and distractions late into the night, at the cost of poor sleep, creating a negative feedback loop. While this wasn’t a particularly new finding, we did discover that there is also a positive feedback loop, like the relationship between sleep education, or being aware of one’s sleep habits and quality, having a nighttime routine, and experiencing high-quality sleep. For our intervention study, we decided to focus on this positive loop, which we hypothesized can actually displace the negative feedback loop of nighttime chaos.

The Intervention Study

Our study consisted of two main interventions, both with the goal of establishing a nighttime routine.

  1. Setting reminders on your phone to tell you to make/eat dinner, start wrapping up your work, and go to bed. The goal was to use these context prompts to remind participants about their goal of forming a routine, and to disrupt the cycle of procrastination and overwork that usually happens during these times.
  2. Picking a wind-down activity to conduct every night. This can include things like reading, playing a game, listening to music, or anything that relaxes your mind. We also had participants anchor this activity to something that is already a part of their nightly routine, such as brushing their teeth. By using an action prompt and anchoring, we hoped that participants would be more likely to adopt the wind-down activity as a habit.

Some participants found the context reminders useful in “forcing” them not to work on school work until the moment they get to bed, but others who had better sleep routines during the baseline did not find the reminders particularly helpful. Instead, they were more focused on the wind down activity, choosing to play a particular game or listening to music. Of those that were more activity-focused, the participants who chose various wind-down activities each day tended to have more irregular sleep quality than those who chose to do the same activity each night.

After synthesizing the intervention study, journey maps, and system map, we then proceeded to brainstorm prototype solutions.

We ultimately landed on a watch-based app that can sense if you are too active (according to your heart-rate) too close to the time you intended to go to bed (which would be inputted by the user at the start). We also recognized the importance of having anchors when changing behavior, and wanted to incorporate the customizability of having wind-down activities as anchors in the app as well.

We focused on 4 key tasks:

  • Introductory Onboarding Survey
  • Nudges to start the bedtime routine with wind down activities
  • Overactive notifications for when heart rate exceeds a level too close to bedtime
  • Checking in on wind-down habit tracking every month and re-syncing of settings (same as onboarding survey)

Visual Design

For our visual design, we leveraged a dark mode aesthetic that was fitting for a night-focused app, with a near-black background and avoiding blue light hues in favor of reds, oranges, and yellows. Drawing from each team member’s mood boards, we took calming, lavender colors for the accents, and opted for friendly, readable font with Source Serif Pro and Open Sans.

Our Final Prototype

User Test Results and Conclusion

From user testing with classmates, we found that most of the issues were in our initial survey task and fell under 3 categories, language, user control and user freedom. In our reminder screen for example, it was unclear to our users what the different options were, like “keep it minimal” and “send me everything.” Adding more specific choices and language would make this more easy to use. There were some issues with selections being unclearly pre-selected, causing confusion. In some of the screens that required users to rank from 1 to 5, there was a lack of indication of low vs. high.

More generally, the onboarding process was very long and the overactive and wynd down activities were not explained at the start. Having a mainly notification-based app meant that there wasn’t a very clear home page, so if we had more time, the next things we would work on would be to fix some of the “copy” issues, shorten the survey, and add a walkthrough of how to respond to different notifications and nudges.

Issues and Next Steps (not in order, but 10 most important)

  1. Confusing low to high scale (Fix: add labels, for caffeine sensitivity, add low to high)
  2. Confusing preselected boxes (Fix: Add check marks or make them blank initially)
  3. Very long onboarding survey (Fix: Delete unnecessary questions)
  4. Hard to select or scroll on screen (Fix: Reduce size of back and next buttons, or add swipe arrow functionality)
  5. Confusing “wynd down” or “overactive flow” definitions on second screen (Fix: Add a walkthrough)
  6. Confusing frequency of notifications (Fix: Add actual values, not just text)
  7. Limited number of wynd-down options (Fix: Add more customizable options, which might be easier with a companion phone app)
  8. Washing up page confusing, waiting for it to load (Fix: Delete … copy)
  9. Initial button? (Fix: Fill in box to make it obvious that it’s a button)
  10. Wind-down activities are short (Fix: Add functionality to change timer)