LEARNING THROUGH DESIGN
Crafting Our Solution
LEARNING THROUGH DESIGN
SCOPING
To craft an effective solution, we narrowed down on an audience that we found particularly compelling.
SCOPING
As we found in our spring research, triggers like illnesses are the primary drivers for people to start paying attention to their health. However, at that point, it can be too late – their health may have already declined, leaving them with a big fat bill to pay on top of their health concerns.
Jamie’s current journey begins with a negative health event. Not only does she have to deal with possible changes in her lifestyle associated with this, but she also has to navigate finding a doctor and paying for her treatment. Her plan also doesn’t cover the full cost, which she didn’t plan for and is now frustrated that she has to deal with.
In Jamie’s ideal timeline, she interacts with her doctor and insurance before a negative health event. This allows her to become familiar with the insurance system and the finances around her care. She also now is on top of her health, and can catch any potential issues early.
In Jamie’s ideal timeline, she interacts with her doctor and insurance before a negative health event. This allows her to become familiar with the insurance system and the finances around her care. She also now is on top of her health, and can catch any potential issues early.
SCOPING
We realized that the trigger needs to be appropriately designed in order to create an emotional response, as well as motivate the user to get invested in preventive care.
IDEATION
We ideated, user-tested, and iterated features that would motivate people to see their doctor. After synthesizing our feedback, we discussed our results with Amy Bucher (VP of Behavior Change Design at Mad*Pow) and our clients to eliminate, tweak, and finalize our features.
INITIAL INFORMATION ARCHITECTURE
With our features tested and validated, the task turned to organizing our platform in a way that provided a streamlined experience while also maximizing motivational potential.
FEATURE + USABILITY TESTING
Once we decided on an information architecture, we created a mid-fidelity prototype and tested it with users in our target audience.
After onboarding, members were still uncertain of the intent of our app.
Outcome
We revamped onboarding to create an interactive experience to communicate the health and financial benefits of preventive checkups.
Members were unclear where users should go to get the care they needed, and why there were multiple lists of tests.
Outcome
We focused on creating one full list of tests, while increasing the visibility of the appropriate providers in the “Test Details” page and at the booking level.
The quantity of information and the way it was organized made the purpose of this page unclear to members.
Outcome
We reorganized the information under tabs labeled “General,” “Cost,” and “Impact.”
“How am I saving dollars? Is it because I'm living longer? Is it because my insurance is covering it?”
Many members did not explore the “Comparison” tab, and of those who did, many didn’t understand the distinction between “My Vitals” and “Comparison.”
Outcome
We collapsed the two separate tabs into one, placing the view of personalized health data in the context of the full population.
Without any context, users were unclear how this page differed from “My Tests” and how they could use it to learn about their loved ones care.
Outcome
We wrote a short description for this page, renamed it “Family,” and moved it to a new “Dashboard.”
“This page seems the same as my recommendation."
When members saw their expenses, they were motivated to take advantage of insurance’s services
Outcome
After seeing how motivating expenses was, we broke it out into a separate page so as to emphasize its importance. We also wanted to highlight insurance’s potential spending to quantify what members could gain from getting tests done.
“It looks like I haven’t been getting any tests done, so I would've been getting my money's worth if I got the tests."
FINAL INFORMATION ARCHITECTURE
After few rounds of feature and usability testing, we rearchitected our product to improve its cohesiveness.
FEATURE + USABILITY TESTING
With our new structure, we set out to explore how users navigated our product and responded to its features.
The new information architecture clarified the feature purpose and navigation of the app.
Outcome
In our high-fidelity prototype, we tweaked the order of the features to emphasize “Biometrics” and “Expenses” as motivating factors.
With provider information stored at the “Test Details” level, there was no indication that multiple tests could be completed in one appointment.
Outcome
We grouped preventive tests by providers. This has the added benefit of emphasizing the behavior we were trying to drive: increased visits to a Primary Care Physician and physicians that offer preventive care.
Again and again, members said they would use the “Book Test” functionality, not “Remind Me.”
Outcome
In our high-fidelity prototype, we removed the reminder button.
Members easily navigated this feature now that it was simplified to a single page, but the ability to filter the underlying distribution of metrics was still unclear and complicated.
Outcome
We constrained the filtering options and tried to add more clarity to its purpose in the high-fidelity prototype.
By adding context and framing filtering as adding a family member, members understood the purpose of this section.
Outcome
If a family member is a dependent, we gave the added functionality of being able to see their biometric results (if permission is granted).
“I’d gladly add my father, I will use this to give him the care he needs."
Testers were motivated yet again to use their insurance once they saw this page and that they were not taking advantage of their insurance.
Outcome
In our high-fidelity prototype, we A/B tested two graphs to determine which visualization was clearer and more motivating, and implemented the more effective one.
"It makes me understand what I am supposed to do, makes me feel I need to do more tests to balance it out."
USABILITY TESTING
We conducted one round of high-fidelity testing focused primarily on usability and navigation, with a lesser emphasis on feature validation. We wanted to mimic the experience of the app “in real life,” so our task prompts were less structured and more exploratory in nature.
Overall
We tested an interactive onboarding flow wherein participants sorted a word like “Exercise” into the category Health, Money, Neither, or Both. For the word “Doctor’s Visits,” most participants chose “Both.” It was our assumption that they would choose only “Health.” However, most participants already made the connection between preventive checkups, health, and money.
All participants were able to complete onboarding, so there wasn’t a usability problem. However, when asked if they learned anything new from onboarding, the majority of participants said that they did not.
"If I stay healthy then I save money, so I'll select both"
Outcome
Our final onboarding design focused on creating a cognitive dissonance between the users’ behavior and their knowledge rather than trying to impart knowledge, since most people don’t visit the doctor for preventive checkups regularly.
Test Progress Bar
When asked to find how many tests they’ve completed, the majority of participants did not refer to the test progress bar, instead referring to the My Test page. A significant share of participants even clicked the bar but didn’t answer the question until they got the information from another page.
Informative Popup
The majority of participants did not like that there was a health statistic (viewable at the bottom of the screen), with one participant referring to it as “unnecessary fear-mongering.”
Outcome
The test progress bar did not get redesigned because the team felt that the current design was sufficiently clear and users were able to get the information they needed. We eliminated the informative popup.
Overall
Participants successfully completed all the usability tasks for the My Tests tab, including:
1. Identifying what doctor to visit for a specific task.
2. Identifying the recommended tests.
3. Identifying the cost of a particular test.
Outcome
No changes were made to My Test because users successfully completed the tasks using the current design.
Visualization of Healthy Range
Participants successfully identified whether their blood pressure, cholesterol, and blood sugar levels were within a healthy range or not.
Blood Pressure Reading
Participants were able to read an ideal blood pressure range, but it was difficult for some. We tested this reading specifically because it requires a unique visualization.
Filter
Participants were not able to filter the biometric data, nor did they find the information useful.
“I actually don't know why I would be looking for that information.”
Outcomes
We removed the filter feature from the biometrics page and further emphasized the healthy range within each graph.
Account Information
Participants successfully identified their premium, deductible, and out-of-pocket expenses.
Insurance’s Potential Expenses
We A/B tested two graphs, both of which depicted how much insurance spends per user when they do and don’t get preventive checkups. In general, participants thought that Graph B was clearer.
"I'd feel like I'm getting ripped off by my health insurance!”
Outcome
Our team implemented Graph B and clearly communicated the graph’s purpose by writing a description below it.
Overall
Participants generally understood the purpose of the “Family” section and did not have difficulty navigating it. However, we received mixed feedback as to whether participants would utilize this feature. Participants were unclear if a family member had to be on the same insurance plan to use this feature. Additionally, our team noticed that people usually thought to add family members that were younger than themselves, perhaps because of the mental model that dependents are younger.
Outcome
In the final design, users can view a biometric page for non-dependents (which isn’t populated with personal data, but does come with healthy ranges for the family member’s demographics).
