August 2018




This interface design project added analytics from expert users preparing for their exams.

One of our company goals in fall 2018 was to improve personalization. The focus was on expert users, or users who spend over at least one to two hours per week on the app. After looking at data and gathering user feedback through Intercom, we decided to build a tool to help users better understand their current progress and break down the areas they need to work in.


A few preliminary ideas about the new analytics charts


Early wireframes showing all the changes going into the process

First: The final deliverable to the engineers including all functionality

Second: Sidebar giving a progress update and showing whether the question was answered correctly or not


First: Updated exam reviews that show completion progress

Second: Onboarding to ensure students get the most value from the new feature


Usability Testing

Since this was a larger project, I spent time conducting some usability tests. You can find the details here. The result from that process is a list of findings and recommendations.



  1. The sidebar with the correct and incorrect indicator is really helpful, as it allows students to revise problems before taking exams.
  2. Users go through chapter videos to learn before completing ExamReviews.
  3. The chapter analysis is a little confusing since there is no scale to measure it.
  4. The instructions were too much text to read.
  5. Sometimes, the worksheet and the videos don't align.
  6. Allowing students to retake topic questions they missed would help, since we are showing them how they performed in a section.
  7. It would help to place to ask a question about the ExamReviews, general or specific.


  1. Add helper text (tooltip) to recommendations
  2. Add a scale to the charts
  3. Reduce text on instructions

Based on that feedback, we updated the chapter analysis to include a scale to indicate how many answers were correct and incorrect. We also added a section reset option. This allowed students to retake questions they missed.



The project, like most others, had to be scoped down. Some of our goals were usually too ambitious from a time or a technical standpoint. In this case, it was the "Recommendations," because we had not figured out a sustainable way to make sure that our predictions are meaningful to the users.


A finding from this process was that developers might not always understand how to make graphs and charts or use certain libraries. This may impact the design and how we solve a user problem. Asking developers to spend time upfront and understand if a project is feasible would be productive.

Another realization was, as I made more proposals, attaching relevant data and user-research helps garner confidence from stakeholders. This was important, as I was starting to push for everyone to create briefs for the features they wanted to create.