Noom is a technology company whose aim is to encourage people to live healthier lives by fostering behaviour change. This is accomplished using a mobile app, and users are assisted by human coaching. The coach communicates with the user on a daily basis via messaging, and speaks to them on the phone bi-weekly. In the medical industry, it is important to keep a trace of patients’ records and progress, and for the Noom coaches, this is done by following a call script. The call scripts consist of a series of questions with text inputs that are to be asked of a user on each call. Depending on whether a user is following a diabetes prevention program or a weight loss program, the questions will differ to better suit the condition of the user. Once a call is complete, the coach submits the call script so there is a permanent record of the conversation.
The coaches in the Noom office had been asking for improvements to the call script process for quite some time. As Noom expands and begins to provide its analytics platform to others outside of the office walls, it was important to begin investigating how to make the process a little less painful.
The first thing I decided to do when I took on this project was to sit in and observe a coach call conduct a call with a user. Because she was managing so many users at one time, the call I was observing began only 5 after the previous one ended, and in this time, the coach had to examine the user’s food and exercise data on the Noom Coach Dashboard, find the appropriate call script to begin typing in, open the scheduling engine and find the user’s phone number to actually place the call. Throughout the 20 minute call, I watched her jump between 3 tabs on her browser page regularly. After the call was finished, she quickly summarized her thoughts that she may not have had time to fully flesh out while on the call, then submit the form and carry on to her next call.
Major Pain Points
After observing the coach call and reading our UX researcher’s findings on coaches’ use of call scripts, I summarized the problem into a few major pain points for the coaches that, if resolved, would speed up their workflow and allow for better coaching.
- Before a call, coaches have to locate the relevant script for each user, and find the user’s phone number
- Coaches keep multiple browser tabs or windows, and jump between them for different information while on the call
- Saved call scripts are not easily accessible and time-consuming to review, so coaches do not bother reviewing the notes from the previous call for context before beginning the next
In this first iteration, I took the existing call script and placed it in an overlay that appeared within the Noom Coach Dashboard on the click of a button. Having the call script exist within the dashboard itself allowed me to make use of the user information that is stored in our database, particularly the program in which a user is enrolled, and whether or not this is the first or follow-up call with the user, because the script changes accordingly. This would prevent the coach from having to find and select the correct script and helps alleviate some of the difficulty. Additionally, keeping the script within the dashboard eliminates one of the windows that the coach has to juggle during a call.
A usability test was conducted with one of the coaches suing a clickable prototype, and immediately one problem arose: the call script blocks the charts and data about the user that a coach references. While reducing the number of windows was positively received, we repeatedly observed the coach trying to exit out of the script pane so she could review the graphs.
The next iteration attempted to tackle the problem of viewing the script and the dashboard at the same time. In my observation of the call and in the previous usability test, I had not seen the coaches make use of the “Inbox” and “History” sections that exist on the left portion of the screen, so I decided to make use of this space for the call script. In this iteration, the script would slide over elements of the page to be displayed in a pane on the left, leaving the charts visible, but also keeping it within the dashboard.
Two usability tests were completed using this prototype, and the coaches responded very well to having the script live in the side panel of the page; however, both of these coaches informed us that in all follow-up calls, the coach is supposed to create a task in the History section outlining the user’s “Action Plan” until the next call. In the prototype in front of them, there was no way to do this because tasks are created from the History section I chose to hide.
Iteration 3 required me to further concentrate on ensuring that all the information a coach might need on a call is accessible within a click or two. The conversation can then continue uninterrupted, while the coach remains informed as to what the user has been doing since their last call and can create reminders for him or herself to follow-up on action items.
The majority of the design work completed for the call scripts feature was completed within a week, including initial observations, as well as all the user testing on prototypes. While it was exciting to be able to quickly gather data, make changes to improve the experience, and then watch another coach try a new prototype, this was typically completed within a day. I felt that working this quickly did not leave me quite as much time as I would have liked to make polished mocks and turn them into a clickable prototype. Though important to maintain momentum and excitement for a project, I think that a little more time and less rush would have been very valuable in this case.
- UX and visual design
- Usability testing
- Sketch for wireframing and mockups
- Marvel for prototyping