← All work

Shelter Insurance - Customer Portal

Turning a 15-year-old list into a dashboard for 100k+ customers.

I redesigned Shelter Insurance’s logged-in homepage into a priority-based dashboard, so customers see what needs their attention first on desktop and mobile.

RoleJunior UX Designer · first project after promotion
Timeline2.5 months · 2024
ProductLogged-in customer portal
Reach100,000+ monthly customers
ProblemA list of policies, with no sense of what to do.Alerts stacked without priority, bills had no urgency, and key mobile features like Road Assist were hard to find.
SolutionA card-based dashboard ordered by urgency.Overdue bills and open claims come first, details are disclosed progressively, and the layout adapts to the device.
Results30% fewer steps, 18% faster tasks.Launched to 100,000+ monthly customers after 10+ usability tests, then refined when data showed a drop in proxy signatures.
Final designs
Desktop: payments, claims, and policies in one view, with urgent tasks first.
Mobile: Road Assist and Contact Agent up top, four cards instead of six.
01Context

Fifteen years without an update

Shelter’s logged-in experience hadn’t changed in 15 years. Customers could pay bills and pull ID cards, but even simple tasks felt cumbersome.

As a newly promoted designer, I saw a chance to turn a static homepage into a dashboard for 100,000+ monthly customers.

“It feels like I’m just looking at a list. I can see my policies, but I can’t tell what I should do here.”
Usability test participant
24%look for ID cards
35%contact agent + road assist
47%go to payments
FIG 01Google Analytics showed customers heading to features that were several clicks away, which fed support calls.
02Discovery

Everything looked important, so nothing was

Before designing, I went through past research in our repository. Three patterns kept coming up:

PriorityNo hierarchyAlerts stacked at the top, and every policy looked the same regardless of what needed attention.
ContextNo sense of urgencyCustomers couldn’t tell if anything needed action, or an overdue bill from one due soon.
MobileLost on small screensStacked content meant endless scrolling, and Road Assist was hard to find.
FIG 02The original homepage: a static list of policies and stacked alerts.
03Key decision

From a list to a dashboard

The shift was treating the homepage as a dashboard instead of a list. I designed mobile-first, because the smallest screen forced me to decide what mattered most.

FIG 03Early mobile-first explorations of the card system.
04Critique

The hard part was deciding what to hide

My first iteration showed everything in two columns. I brought it to design critique, and the feedback was direct:

“There’s a lot happening, I’m not sure where to look first. Which of these things actually need my attention?”
UX team · design critique

I rebuilt it as a priority-based card system: urgent actions like overdue payments and open claims come first, details are disclosed progressively, and promotional cards (paperless billing, feedback, roadside coverage) only appear after critical tasks.

Before
FIG 04First iteration: everything competing for attention.
After
FIG 05After critique: urgent tasks first, promos last.
05Billing

One bill, one card

Customers with home and auto policies couldn’t tell what they owed in total.

“Wait, so do I owe $150 total, or is that just for my car insurance?”
Usability test participant

My first payments design put an overdue card and a total-due card side by side. The two amounts competed, and two payment buttons left people unsure which one to press.

The final design uses a single card that consolidates every due bill, overdue first. When a payment is already scheduled, a friendly “You’re all set — no action required” removes the doubt. Full details live on the payments page.

Before
FIG 06aBefore: competing totals and two payment actions.
After
FIG 06bAfter: one card for all due bills, plus a clear “all set” state.
06Mobile

Road Assist, one tap away

People open an insurance site on their phone for different reasons, often urgent ones.

AddedEmergency actions up topRoad Assist and Contact Agent became prominent buttons.
Reduced4 cards instead of 6Policy info moved into a scannable vertical layout.
FIG 07The mobile dashboard with emergency actions within reach.
07Collaboration

Designed alongside the team

Another designer was building a dedicated policy management experience. We shared research and brought her policy list into the dashboard so both felt like one product. I pulled policy data and found most customers have 3 or fewer, which set how many the dashboard shows.

Engineering told me each policy needs separate API calls for payments, claims, and details. Instead of a generic spinner, I designed loading skeletons that hold the page’s hierarchy while data arrives.

Policies
FIG 08The policy list from my teammate’s policy management work, brought into the dashboard so both feel like one product.
Loading
FIG 09Skeletons keep the layout stable across separate API calls.
08Results

Launched, then measured

After 10+ usability tests and multiple iterations, the dashboard launched to Shelter’s 100,000+ monthly customers, with ID cards viewable and printable from the policy overview.

Navigation30%Fewer navigation steps to reach key tasks.
Task completion18%Faster task completion after launch.
Monthly users100k+Customers using the new dashboard each month.
09Iteration

The data found a trade-off

Most metrics improved, but proxy signatures dropped year over year. Proxy voting was easier to understand, yet moving it down the page hurt discovery.

I designed a proxy card that appears right after urgent tasks, keeping it visible without breaking the hierarchy. Legal has approved it, and we’ll keep watching signature rates as it launches.

Before
After
FIG 10Before, proxy voting sat below the policy list. After, it’s a card placed right after critical tasks.
10Reflection

Clarity isn’t the same as visibility

The proxy drop taught me that making something easier to understand doesn’t help if people never reach it. Placement is a design decision with its own consequences.

As my first project after promotion, it also taught me to bring work to critique early and to keep measuring after launch, since that’s where this trade-off surfaced.

Next project
Curly Journey: Undergraduate Capstone 2023-2024
Making curly hair care education accessible & empowering
View case study