A Redesigned Support Experience
My Role
User Experience Designer
Employer
The Creative Alliance
Client
Berkshire Hathaway's biBerk
Responsive design
Desktop & mobile devices
for Small Business Owner's Insurance
User Experience Designer
The Creative Alliance
Berkshire Hathaway's biBerk
Desktop & mobile devices
The previous support tool caused more service calls, higher costs, and customer frustration due to low conversion rates.
An intuitive multi-page support system that efficiently guides users to their specific self-service solutions.
Utilizing tools like FullStory and GA4, we performed user-focused research that influenced our design decisions.
Early support concepts, user testing prototypes, and completed designs.
Unmoderated user testing conducted on two distinct fully interactive prototypes.
Finalized design for mobile and desktop, a fully interactive prototype, and comprehensive functionality notes.
A user-focused analytic platform, we provided user-based research that helped shape our testing prototypes and final designs.
GA4 allows us to track and analyze our client's website traffic and user engagement for the old and updated Support flows.
Where we designed all iterations, our unmoderated user testing prototypes, the final Support designs, and a fully clickable prototype for our client (and dev team).
A platform made for remote usability testing, where we had real users provide feedback on our two prototypes through specific Support tasks and recorded interactions.
A tool used to streamline our handoff of our final Support design files to our client's development team.
Sheets for data collection, user test results & documentation; and Docs for for communicating project progress, updates, and final results to client.
While proto-personas typically rely on assumptions and early observations, our proto-personas are grounded in a deeper understanding of the user's needs, goals, and customer data. This insight comes from extensive time spent on previous research and design projects previously completed for biBerk. As a result, our proto-personas offer a more informed starting point that highlight key opportunities for improvement based on established knowledge.
By studying 6 leading small business insurance companies, we were able to see who leads the industry in support-related tools.
1. State Farm
2. Geico
3. Next Insurance
4. Progressive
5. Travelers Insurance
6. Hiscox
We wanted to know how are these other insurance companies managing and implementing their own support services and do they specifically:
• Easily provide phone number(s), mailing, and email address(es)
• Live Chat/Agent functionality
• Account login/portal
• Frequently Asked Questions (FAQs)
• Useful resources, quick links, etc.
Account Log-In
Most insurance providers include an option to log into an account OR download mobile app. This provides quick and easy access to policy information.
Customer Service Contact
All providers include a specific contact support email address and customer service phone numbers (sometimes several that are department-specific).
Helpful Information
Many of the studied insurance providers include quick access to FAQs, quick-links to specific self-serve tasks, and even live chat capabilities.
.png)
From the 1,527 observed users, just 547 users (37.5%) experienced a successful Support outcome.
Of those 547 users:
While it was encouraging to see a sizable amount of traffic being driven to the Support page, our concern was that the old Support tool wasn't helpful to users. And based on months of documented negative customer feedback (extended call wait times, incorrect information entered via agents on COIs, etc), users aren’t feeling supported post-purchase.
In total, there are 16 preexisting support-related tasks that are available to users. Due to business constraints that are outside of the scope of this UX project, 9 tasks require a confirmation phone call whereas 7 tasks can be accomplished online entirely through the client's website.
For the newly designed Support tool, all 16 tasks need to be available to users at launch.
Our unmoderated user tests involved 10 unmoderated tests, split between 5 desktop/laptop users and 5 mobile users, each lasting 40 minutes. Participants reviewed two prototypes in alternating order to avoid cognitive bias, completing two tasks—changing a business name and obtaining a custom Certificate of Insurance (COI)—with follow-up questions after each. After testing both prototypes, users evaluated three alternative static Support page designs and concluded with post-test questions.
To meet our goals of efficiency, budget, and rapid results, we opted for unmoderated user testing—a remote usability method where participants complete tasks independently, without a facilitator.
Testing for mobile and desktop allowed us to gather insights tailored to the unique experiences and interactions of each platform, ensuring that both desktop and mobile needs were thoroughly addressed.
To accurately capture user insights, each participant will complete 54 "tasks," consisting of a mix of questions and prompts designed to engage users. This includes 7 post-test questions.
My team and I created one high-fidelity, fully clickable prototype for mobile and desktop; while our client's UX team created another. Participants reviewed both prototypes in alternating order to avoid cognitive bias.
We asked all users to complete two essential insurance tasks that most small business owners will need to tackle after purchase: changing a business name and obtaining a custom Certificate of Insurance (COI).
I developed an unbiased script and crafted all 54 questions/tasks to encourage natural responses from the testers. The 10 tests were then set up and launched on our selected testing platform, Userfeel.
All 5 desktop users and all 5 mobile users had no issues completing either main task, regardless of prototype (1. changing business name and 2. getting a custom Certificate of Insurance COI).
All 10 users users intuitively understood that Support cards were clickable / tappable. No users were confused on how to start each task, regardless of which one they started with.
All users intuitively understood that the Support cards were interactive and had no confusion about starting the Support flow. They immediately selected either the My Policy or Certificate of Insurance card to begin their tasks, regardless of the prototype.
At the end of the 40-min test when asked which prototype users preferred, 7 out of 10 users specifically choose our designed prototype compared to our client's in-house UX teams'. The final design pulled successful elements from both prototypes.
After extensive design iterations and in-depth user testing and research, we finalized the updated Support Tool and compiled our deliverables.
Our deliverables included:
1. 2 fully clickable prototypes for both mobile and desktop devices (380px & 1280px widths).
2. Detailed functionality notes for the biBerk development team.
3. All 90 designed Support pages uploaded and organized in Zeplin.
4. Continued data reporting and support for any future design updates.
From overhauling the overall flow to adopting a streamlined design structure, I’ve outlined the key features and design specifications below, showcasing the most significant improvements made to the updated Support Tool.
6 Support Cards that effectively guides users to their Support solution, no matter how specific it is.
In addition to main cards, we included the top-reported Frequently Asked Questions according to insurance agents; along with important policyholder resources.
Through our competitive research, multiple design iterations, and user testing & feedback, we refined the Support design to be as simple as possible. This simplicity applies not only to the visuals but also to the content and wording, all of which have been carefully optimized for user efficiency and clarity. All Support paths lead to either a simple form or a specific CTA (button) taking the user directly to their support solution.
Given the low success and engagement rates observed with the Old Support Tool, we aimed to ensure that every Support path included an "out" — a fallback for users who are unable to find a solution, offering them an alternative route.
Because this new Support Tool lives on biBerk's website, we had to make sure this design works for all device types (380px width - 1440px+ width).
After launching the new Support Tool, we tracked user statistics over the first four weeks. The results were impressive: 92% of visitors engaged with the tool. Out of 10,797 users who accessed the Support page, 9,983 interacted by selecting one of the six Support cards. Additionally, 11% (1,238 users) clicked on either the FAQ or Policyholder Resources links located at the bottom of the page.
From October 1–12, 2023, the old Support page had a completion rate of just 2.7% for request submissions (1,127 users > 31 submissions). In comparison, the updated Support design achieved an 11% completion rate during the same 12-day period (5,103 users > 548 submissions), with traffic to the new page increasing by 4.5x.
