The Lumo Lift was a wearable device that tracked and reinforced good posture (now discontinued). It was one of the earliest posture correcting devices on the market, and it debuted with a basic app that had some prominent usability flaws. As an advocate for posture correction and a fan of the Lumo Lift, I conducted usability research on the Lumo Lift and redesigned its iOS and Android apps to be more user friendly and to follow platform conventions.
UX & UI Design
The Lumo Lift app had some core usability issues that undermined its effectiveness:
- Data was difficult to compare, making it hard to make progress
- Key app functionality (comparing trends, configuring the posture coach, settings) was either hidden or undiscoverable
- The app uses "good posture hours" as its key metric, but it is an unclear and ineffective metric that is not explained
I redesigned the Lumo Lift iOS and Android apps with both aesthetic and functional improvements:
Before redesigning the Lumo Lift, I wanted to understand the most prominent and pressing issues that users of the current app faced. To that end, I conducted in-person user testing and gathered app feedback from a variety of inputs.
- In-person testing: 3 testers who expressed interest in correcting their posture and potentially buying the Lumo Lift. I observed them as they set up the Lumo Lift app on their own devices, and they borrowed and used the Lumo Lift for at least 2 full days. At the end of that period, they reported any usability issues that they encountered and answered questions regarding their usage of the main features of the app.
- Online customer reviews: I looked at all of the reviews for the Lumo Lift app on iTunes and the Google Play store up to the end of February 2016. I also surveyed reviews from Amazon, but due to sheer volume and the fact that most reviews were for the Lumo Lift hardware as opposed to the app, I only looked at the month of February 2016.
After gathering in-person and online comments on the Lumo Lift, I categorized the issues, ranked them by priority, and then presented solutions to work into the redesign.See full Lumo Lift user research report
The trends page showed a bar graph of the percentage of good posture within the hour, but did not give specific details. 2 testers mentioned that they wanted to see a breakdown of good and bad posture by the minute, and 2 customers directly commented on it in their reviews: (1) “My main complaint is that I can’t get more detail about my activity. It would be great to know if there are times when I slouch more, or what my activity curve looks like.” (2) “I would be more inclined to use the app if it showed more stats, like my current posture or how many minutes of good posture I’ve had so far that day.”
Solution: Give a breakdown of good and bad posture for each minute of the hour.
The “Trends” button was located at the bottom left-hand corner of the screen. 2 testers were unaware of the trends page. Before being guided to it, one expressed the desire to see something exactly like the trends page, with posture by the hour. (Appendix)
Solution: Surface the trends page by placing the link somewhere more prominent, or putting some data directly on the main page with a link to easily access the rest of it.
9 customer reviews commented that the app lacks a way to view or compare data over long stretches of time, in order to see progress. One mentioned, “Some history display aka FitBit would be useful,” and another wrote, “Chart comparisons would be great or more details.” (Appendix)
Solution: The trends page had graphs that compared hourly posture and daily posture, but it appears that customers may not have been discovering the latter, or may still have found it difficult to use for comparisons. Progress could also be tracked via other data like the duration of continuous good posture or the number of buzzes per hour, as suggested by a tester.
An hour was classified as a "good posture hour" when the user had good posture for at least 40% of the hour. That means that somebody could reach 39% good posture for 12 hours, but still have 0 good posture hours. In addition, good posture for 40% and 99% of the hour were both recorded the same way, as 1 good posture hour.
All 3 testers remarked that the good posture hour metric could be improved: (1) “posture hours is a bit ambiguous;” (2) “it’s not a very intuitive metric;” (3) it’s “just the total, I want something more granular." 4 customer reviews expressed frustration at the use of “good posture hours” as a metric. One pointed out the lack of context: “…it will tell you '5 hours of good posture'. Swell, but five hours compared to what? Did I only wear the thing for five hours, or did I use it for sixteen hours?” Another lamented that it records 0 hours if one is below the 40% goal line: “I love the concept of the device but only seeing zero hours of good posture despite having the coach turned on is discouraging. I would like to see more minute-to-minute data to stay motivated.” (Appendix)
Solution: Another more intuitive metric for tracking posture could be introduced, such as a count of the total number of minutes of good posture, the longest string of continuous good posture, number of buzzes per hour, etc. This would give more motivation to those who are starting out and struggle with reaching the targeted 40% good posture in an hour, and it would also offer those who consistently exceed the target a way to quantify their achievements (distinguishing between 40% and 99% of the hour).
The home screen was the area with the most potential for improvement. As the first screen that people saw, the home screen needed to be easy to comprehend and navigate. I also surfaced commonly used features that were previously hidden and rethought metrics and interactions that caused confusion.
HOME SCREEN BEFORE & AFTER
- Data was displayed in a circular graph, which is more difficult to approximate
- Arrows to see previous/next days were in the center of the page, removed from the date and easily missed
- Trends link was at the bottom of the page. Testers missed it and tried to tap on the circles instead to look at trends (no link)
- Coach status was displayed, but the user could not turn coaching on or off on the homepage
- Data is displayed in a bar graph along with the percentage completion of the user's goal
- Data from other days is easier to access, and indicators are added showing if goals were completed on previous days
- Trends link is in the tab bar at the bottom. Tapping on the bar graphs also leads to the Trends page
- Coach can be turned on and off directly on the home screen
- A live posture report is added. People liked this functionality but complained that it was hidden
RETHINKING "GOOD POSTURE HOURS"
The primary metric that the Lumo Lift app used was a "good posture hour," which was defined as an hour in which someone maintained good posture for at least 40% of the hour. User research revealed that people were confused with the metric of "good posture hours," and did not find the metric to be useful (see user research findings for more details).
In order to enable people to benchmark and track their posture improvements, the app needed a new metric that was more intuitive and granular. The metric that I propose is total time in good posture for each day. This metric is easy to compare across days, works well with goal setting and tracking, and does not depend on Coach Mode being turned on (as metrics related to buzzes would).
The Lumo Lift was already collecting a form of time in good posture, as it displayed the percentage of each hour in good posture. But development may be required to collect posture data at a more granular level. Further testing could show whether the increased clarity of the new metric would be worth the investment.
PLATFORM DIFFERENCES: DATE PICKER
For the most part, there are not many differences between the iOS and Android versions of the redesigned Lumo Lift app. The date picker is the area with the most prominent difference, so it has two separate designs in order to follow platform conventions.
The date picker on iOS follows the date pattern in the Apple Health app, which consists of a row of dates that the user can individually select or scroll through horizontally. This allows people to get an overview of past days at a glance and makes it easy to browse through the most recent days.
The iOS date pattern is not common on Android, so the Android date picker uses the Android convention of a dropdown calendar instead. This pattern is found in the native Google Calendar app. It has an extra step (tapping to open the calendar) compared to the iOS date pattern, but it allows people to easily browse through a wider range of dates.
I redesigned the Trends screen to make data easier to compare at a glance, and introduced a more granular breakdown of posture by the minute.
TRENDS SCREEN WIREFRAMES
During wireframing, I explored alternative ways of displaying step and posture data on the Trends screen. Horizontal bar graphs leave more space to display numbers, but too many important features were hidden (posture and steps could not be displayed at once, and the minute option could not be in the menu). I went back to vertical bar graphs, but could not display the numbers in a legible way. Finally, I landed on vertical bar graphs with progressive disclosure, which the app already used, but I cleaned up the screen and added more structure to the graph.
BREAKING DOWN POSTURE BY THE MINUTE
User research showed that people wanted a more detailed breakdown of their posture than the app provided. Though people could see their percentage of time in good posture for any given hour, they wanted more granular information on their posture throughout an hour (see user research findings for more details).
To address this need, I added a breakdown of posture by the minute to the Trends screen. This enables people to relate their posture with specific activities and to better pinpoint when and where their posture tends to be the best/worst.
To implement this new feature, a couple of things would have to be considered: first, granularity. Tracking and displaying posture for each second may be unfeasible, but instead the app could track in 10-second blocks (maintaining good posture for the full 10 seconds would result in a "good posture" block, slouching at any point during the 10 seconds would result in a "slouching" block). Second, there may also be development work required to achieve the required level of granularity, and further testing would be needed to show whether it would be worth the investment.
A central part of the Lumo Lift app redesign was updating the app UI and making sure that it was consistent throughout. I selected a new color palette that better matched the Lumo Lift branding and created a style guide that holds true across both iOS and Android.
BEFORE AND AFTER
A comparison of the previous Lumo Lift app and its redesign on iOS and Android:
ORIGINAL HOME SCREEN
iOS HOME SCREEN
ANDROID HOME SCREEN