Background

I recommend reading my other post about my time at Garner first as it helps set some context. This post is about year two of the strategy I laid out in that post, and focuses in on our mobile app redesign.

Here’s a quick recap:

In April of 2022, I joined Garner Health as their Head of Design and first full time designer with the charge to build a high performing design team to create a usable, understandable, and trustworthy member experience.

I joined Garner at a really exciting time. The company was starting to get real traction in the market and growing rapidly, but member experience was a blind spot for the organization. I set up systems to measure it through an NPS survey and user research program, and the results showed it was a significant problem. Our NPS was 0 (on a scale from -100 to 100), and our client attrition due to product dissatisfaction and low engagement was too high.

Based on the data and feedback, we developed a strategy with 3 major pillars:

• Increase transparency and speed of claims and reimbursements

• Improve trust in our recommendations

• Improve program understandability

Year 1 was mostly spent fixing foundational issues with our claims, payments, and other data systems. These investments were crucial for the user experience and I pushed hard for them, despite most not being ‘design’ problems. At the end of year 1 we saw minimal movement in our metrics, so refreshed our strategy with a specific focus on driving member engagement and client retention.

I was tasked with defining a UX vision for the app focused around engagement and recommending a 6 month product roadmap that would allow us to launch a redesign of the app in Q3 before open enrollment.

Approach

1. Creating shared context

2. Defining the vision

3. Building buy-in

4. Planning the year

5. Outcomes

Creating shared context

An existing foundation of research gave us a jump start

When we kicked this project off, we already had a lot of both qualitative and quantitative research and insights. Most key stakeholders and project partners already had strong perspectives. We needed now to synthesize those perspectives and bring them together into one coherent vision.

Here’s a summary of what we knew going in:

We knew we had an engagement and member satisfaction problem and it was affecting client retention.

We knew the biggest actionable blockers to both engagement and satisfaction for members were:

• Not enough visibility and transparency with claims undermined trust and reliability

• Poor program understandability made it difficult to get started

• No in app mechanisms to surface contextual notifications to members limited our ability to engage members at key moments

We also knew we probably wanted to redesign the home screen. It looked amateur and undermined trust in our brand which created risk for engagement. It also wasted space — The entire front page of the app was taken up by the global navigation. Last, it created a cumbersome hub and spoke navigation system, making context switching within the app hard. We were open to changing our views as we dug in further, but based on what we knew, this seemed like a fairly obvious, high ROI improvement.

With that context in mind, I planned a series of workshops to get our core team aligned around a unified vision

I organized two workshops with key product, design, and engineering partners that stepped through our core user journey that allowed participants to get in sync, get their ideas out, and begin fusing them together into a coherent app experience. Over the course of two sessions, we stepped through:

1. Review of key insights and context

2. Alignment around key member personas and the core user journey we wanted to design around

3. Ideation across key touch points in the journey

4. Loose wireframing of key ideas into a journey

Here’s an example segment of a journey one participant put together during our wireframing. All participants made low fidelity journeys like this that formed the foundation of our concepts.

Then, I paired with a very strong visual designer on my team to go through several tight iteration cycles refining the user journey and wireframes into more polished mockups that told a coherent story and captured our key concepts. We did this in tight coordination of the original group with feedback cycles every few days until we felt aligned on the vision.

Here’s a sample of vision we came up with

Below are three slides from the deck we used to build executive buy in. Our actual vision was a longer, full journey, but this should give you a sense of what it was like. The vision was designed for voice over, with me narrating the journey our member, Allison, was experiencing, whilet the slides highlighted key concepts. This is the same part of the journey as the lower fidelity mockups from above, so you can see how it evolved.

Across our vision, we highlighted several key ideas that we proposed taking steps toward this year

• More personalization and proactive messaging, manifested through a new dynamic home screen

• Improved program understandability through better onboarding, more personalized guidance, and a fix to a structural usability problem with the Care Team (Shown below).

• Better visibility into claims and payments manifested through a new benefit tracker and expanded set of claims statuses to provide more transparency.

• Better contextual support along the journey, manifested through in app scheduling that allowed us to understand when an appointment was scheduled (This part was a bit aspirational).

I worked with my PM partners to break the vision down into smaller pieces we could prioritize

We identified and prioritized following core pieces and built a more stripped down, MVP version of the vision that was stricter with elements we specifically planned to build that year as a part of one of these components.

• Improved in-app onboarding

• A new home screen that allowed us to surface personalized, contextually relevant information to drive engagement (Shown on the right)

• A new bottom navigation system (Shown on the right)

• A new feature called Care Team, that allowed members to save providers and created more clarity on how to use the program.

• Enabling members to see pending and denied claims

• A benefit tracker showing the total and remaining benefit

• An AI chatbot to help answer member questions faster

We circulated the more stripped down vision, as well as a drafted roadmap plan that would allow us to launch in August ahead of Open Enrollment, circulated the proposal with key stakeholders and leadership around the company, solicited input, and ultimately won buy in to move forward with the redesign a month ahead of schedule.

We won buy in from all key stakeholders and leadership

The projects above were handed off to various teams. My team at this point was understaffed, with two yet-to-be-filled headcount, so I took on Onboarding and the Benefit Tracker personally, while overseeing the other projects with the support of other designers.

I won’t walk through each project in details, but I’ll do a quick zoom in on Care Team to give a sense of more specific project work we did.

Deep dive on Care Team

Garner’s program has an engagement requirement, meaning members must search a provider in the app before they can qualify for reimbursement. The way it works was that when a member searched in the Garner app, Garner automatically saved the providers that showed up in the search results to a list of Approved Providers.

At face value, this seems like a reasonable system. Its automated so members don’t have to think about it, so should minimize friction. Instead, we saw a few problems that surfaced time and time again in user interviews and member feedback.

Key problems

• There was no visible confirmation when a provider was added. Members were uncertain if they’d filled the engagement requirement or not when they’d searched. Oftentimes, they just didn’t understand how it worked, and if they did, it was cumbersome to go sift through the full Approved Providers list to verify if the provider you wanted was added.

• The approved providers list quickly filled with hundreds of irrelevant providers. We added all providers who showed up in search results to ensure we gave appropriate credit, but this meant that the select few a member actually saw were hard to find.

• Maybe most importantly, there was no defining action we could educate around. We told members they had to search, but this often created confusion. Members wanted a more specific action to help them understand what they needed to do to use the program successfully.

This was one of those problems that had been festering for quite a while, and a lot of thinking and ideation had been done around it for over a year leading up to this work. We quickly reached consensus this needed to be a key element of the redesign, but there were several key and controversial design decisions we needed to make about the feature.

Key design decision points:

• Should we build a new feature that lets members add and save, or just modify the Approved providers list to be more navigable.

• If we build a new add feature, should we require members to use it in order to get reimbursed, and if not, how do we educate members about it?

• If we build a new feature, where in the information architecture should it live?

• How can we balance the visual hierarchy of the provider profile with this new element?

Let’s deep dive on one of them.

Where should Care Team sit in our information architecture, and how should it relate to the Approved Providers list?

While everyone agreed we needed to build something like the Care Team, the specific shape of the feature had some meaningful sticking points. At the center of it was the question - where should it fit into our architecture, and what do we do with the existing Approved Providers list? Here were our decision criteria and options:

Decision criteria

• Must allow members to easily find the providers they care about again

• Must project a clear mental model for how to use Garner, and simplify the “1,2,3” explanation. (I.e. How would you explain how to use Garner in 3 steps?)

• Should minimize disruption to existing members

1. Wrap the Care Team into the Approved Providers list

This was a simple, incremental improvement, but de-emphasized the Care Team meaningfully, and still blurred it with a list of providers members didn’t care about.

2. Add the Care Team to search section

Our CEO was pushing this option to wrap the manage care functionality into an expanded search section of the app called Find and Manage Care. This conceptually made sense, but we didn’t feel it elevated the Care Team to the level of importance it deserved.

3. Add a new feature to add providers and educate members they should add their providers to ensure get reimbursed. Don’t actually change adjudication policies and keep the approved providers list, but deemphasize it and use it as a backstop to ensure members get credit for their searches.

This was the approach we selected. It allowed us to get the best of all worlds. Minimized disruption and change to our adjudication policies, and an improved 'How to use Garner in 3 steps’, and a simple list for members to find their providers.

The biggest risk in this approach was confusion between the two lists, particularly for existing members. We rigorously tested this in multiple rounds of user research, including evaluative user research in an early access launch later on, to derisk and validate it would work.

We also thoroughly circulated this within Garner, building buy-in all the way up to the executive leadership level as this requirement a meaningful shift in how we demo'ed the product. I partnered tightly with our marketing, sales, and account management teams on how we’d message it to new and existing users and how we’d demo it to clients.

4. Add a new feature that lets members add providers and use it for the engagement requirement, and deprecate the approved providers list.

This was also a clean option that would have allowed us to fully migrate to a new system that solved the problems listed above. There would be a clear, defining action and a simple list of the providers members actually cared about. The problem was it created a huge usability risk for existing members, and created a stricter system that would likely result in misunderstandings and an increase in denials. If a member forgot to add their provider to the list, they wouldn’t be reimbursed. It also created some weird edge cases - for example, what if they removed them from the list?

Hopefully that gives you a sense of how I approach design decisions: Define the problem, establish decision criteria, ideate a range of options, then evaluate and stress test until we’re sufficiently confident its the best approach. Then continue measuring and testing through and past launch to understand impact and inform future iteration.

Rollout

Our teams designed and developed these features over the last 9 months. We bundled several of them together into a bigger splash launch, while rolling others out on their own timing depending on what made most sense.

The bigger splash launch, which included the home screen and navigation redesign, Care Team, and the benefit tracker, was rolled out in a dogfood then an early access program to de-risk the launch. I conducted evaluative user research with the early access cohort, including live interviews and surveys, while the product team monitored our engagement metrics.

We did a full general availability launch in late August, with the other unbundled features rolling out over the rest of the year.

Here’s some key screens of what the new product looks like:

50% of existing members surveyed said the redesign was an improvement

As of writing this, we only have early leading indicators of the outcomes for this work.

In early access launch survey, we received positive feedback from existing members, who we expected a more tepid response from due to early aversion to change and only sporadic product usage. About 50% of members surveyed said the changes were Major Improvements or Improvements, with about 45% neutral and 5% saying they were worse than before. Given change aversion, this was a very positive response.

In interviews, members cited generally positive to neutral responses, saying it was “easier to maneuver through the site” and how helpful it was ‘Knowing how much benefit I have used”.

We expect the major movement in our metrics to come in Q1-Q2 of next year as we onboard new members to the experience and benefit from fresh impressions.

Overall this year, NPS has already risen from 0 to 35 from the changes we made last year, and similarly we have already hit our engagement targets for the year. This is in large part due to the work cited in my other post about Garner, where we saw delayed movement in metrics until the new year.