
Managed the team that improved customer experience 6.9pp and damage estimate accuracy 4.2pp by changing a UI
Mobile app for submitting claim photos
The insurance company's Chief Claims Officer asked us to quickly (two weeks!) identify opportunities to improve customers' experience with a new mobile app that was central to strategy.
The App
The mobile application (iOS and Android) guides users through the process of taking and submitting photos of their damaged vehicles so they can get an initial estimate of the cost to repair.
The Problem
While most claims were achieving high marks for customer satisfaction, survey results had recently declined.
Our challenge was to understand any causes of customer frustration during the experience and develop a plan to address them.
The Team
A cross-functional group of leaders and designers from the Claims, Sales, Customer Experience, Continuous Improvement, Quality Assurance, and Marketing departments
Our Approach
-
An immersion phase to...
-
familiarize ourselves with strategy and context
-
gather data/feedback and develop empathy for customers
-
-
A synthesis phase to...
-
construct persona(s)
-
assess the customer journey
-
-
A prioritization phase to...
-
establish decision criteria
-
identify top opportunities
-
-
A definition phase to...
-
build prototypes of the highest-priority concepts
-
ensure the implementation team would be successful
-
immersion
synthesis
prioritization
definition
My Role
I was hand-picked to serve as project manager, and my job was to co-lead the engagement team alongside the overall business lead. Prior to the two-week project, I defined our objectives, scope, and approach, ensured stakeholder alignment, assembled the team, and coordinated travel. During the project, I guided the project team through our approach, facilitated our design workshops, coordinated prototyping, and compiled the final report. After the project, I led implementation of the top concepts.
immersion
Strategy and Context-Setting
Executive advisors briefed the team on the company's mobility and customer experience (CX) strategies.
Ethnographic Research
The team realized that, while some members understood how the app was supposed to work, it would be important to step outside of their native environment, shed any preconceived notions, and get a better sense of reality. Members spent several days with 'frontline' employees, conducting stakeholder interviews and 'fly-on-the-wall' observations and making contextual inquiries.
Team members also practiced 'walk-a-mile' immersion by using the mobile app themselves, seeing and doing what users experience.

Voice of the Customer (VOC) Data
The team reviewed direct feedback from customers, in the form of survey responses, verbatim comments, and complaint data.
Immersion gave us a better understanding of context and deeper empathy for all involved.
synthesis
People & Systems
To ensure human-centered design, the team pulled insights from pre-existing marketing segmentation data and constructed a persona profile. The team then assessed our hero's journey by conducting experience diagramming.


I oversaw a UX designer, who constructed the journey map for the team. I've blurred some areas of the image to maintain confidentiality.
Synthesis helped us visualize the entire system and assess it from a human's perspective.
prioritization
Patterns & Priorities
The team emerged from synthesis with a comprehensive list of the irritants customers are likely to encounter during their journey.
Affinity clustering and the use of an importance/difficulty matrix, initially based on individual members' judgement and refined through facilitated discussion, enabled the team to identify relationships between irritants and to get a stronger sense of which irritants, if addressed, represented the biggest opportunities.
Team members 'visualized the vote' to determine where to go next.

I've blurred some areas of the image to protect privacy and to maintain confidentiality.
Top Irritant
Many customers felt the repair estimate they received was far too low.
Prioritization resulted in team consensus on a shortlist of opportunities for definition.
definition
Design Rationale
The team quickly analyzed root causes and brainstormed several concepts for addressing them. To aid the implementation team, members developed quick reference guides for select concepts.

I've blurred some areas of the image to protect privacy and to maintain confidentiality.
Modeling & Prototyping
Team members conducted appearance modeling to 'mock up' several of the top concepts.


I oversaw a UX designer, who constructed the mock-ups at my direction. I've blurred some areas of the images to maintain confidentiality.
A handful of simple UI changes to the app itself were among the most promising concepts because of their potential to affect the average repair estimate amount by addressing users' in-app behavior.
solution 1
Repair estimates are based on the vehicle damage that's visible in the photos submitted by users, so generally, more photos = a higher estimate. The app allows users to submit a maximum of 12 photos of damage, yet most users were submitting the minimum number (3) of photos. Customers said in their survey comments that they [erroneously] believed the app wouldn't let them submit more than 3 photos.
The team found that in-app instructions were misleading and came up with a quick fix.
THE CULPRIT


THE SOLUTION


I constructed these mock-ups using the Sketch app.
solution 2
And since glare on the surface of vehicles can inhibit adjusters' ability to see damage, photos from more angles = a higher estimate, generally. Anecdotal feedback from employees indicated that many customers were submitting photos that had been taken from too few angles.
The team found an opportunity in the app to emphasize the importance of taking photos from multiple angles.
THE CULPRIT


THE SOLUTION


I constructed these mock-ups using the Sketch app.
Definition produced actionable solutions, and these UI changes were slotted for implementation.
"You were tremendous to work with on this and I so appreciate your ideas, help, and guidance throughout the two+ weeks."
co-leader feedback
implementation
Result 1: Submissions of photos from customers increased
Before the UI changes described above, just 43.9% of users submitted more than the minimum number of photos of damage; afterward, 63.0% did.

Result 2: Accuracy of damage estimates increased
Before the UI changes described above, repair estimate accuracy was 91.5%; afterward, it was 95.7%.

Result 3: Satisfaction of customers increased
Before the UI changes described above, customer satisfaction was 77.6%. Afterward, it was 84.5%, and when customers did provide negative survey feedback, fewer of those complaints mentioned the estimate amount (47% before, 41% after).

lessons learned
If we'd had more time, I coulda involved users by leveraging participatory research (e.g., 'buy a feature') or evaluative research (e.g., 'think-aloud' testing).
COULDA...
SHOULDA...
Didn't occur to me then, but I shoulda insisted on more resources for the follow-on implementation team, since that team's limited capacity turned out to be a barrier.
Next time :)
WOULDA...
If it had been possible, I woulda conducted an A/B test to control for other potential factors when assessing the effect of changes to the UI.