Case study · UL Solutions · User Experience Manager
How I got a stretched design team out of reactive mode on UL's customer portal
- Problem
- A portal with 100,000+ registered users, a sign-up flow customers couldn't finish, and a design team taking work from whoever asked loudest.
- What I did
- Set up written intake, capacity sized in points, and a UX review before every UAT. Then used the sizing to push back on risky requests.
- Result
- The team shipped an account-access redesign and a new navigation model. With the automated approvals it supports, customer wait time for access fell from about 70 hours to about 30 seconds.
- My role
- UX design lead for the portal from October 2022. UX lead for the next-generation portal from April 2024, directing three designers.
The wait-time figures are how long customers waited for access. Engineering built the automation. My team designed the flow it runs on.
The problem
When I became UX design lead for myUL in October 2022, the portal had more than 100,000 registered users and plenty of complaints. Our usability study that fall, 14 customer sessions, showed how hard it was just to get in. Not one account administrator we tested could finish inviting a colleague. More than half the participants didn't know the site number the portal asked for, and most hit slow pages along the way.
The design team had its own problem. We supported more products than we had designers, and there was no shared way to size requests. Whoever asked loudest got the work.
My role
I led UX for the portal and managed the designers on it. From April 2024 I also led UX for the next-generation portal and directed three designers there. Product owned scope and priority. Engineering owned performance and the approval automation.
What I tried first
Two months in, I got the product team to agree to written intake rules: four weeks' notice for planned work, Director approval for anything unplanned, and acceptance criteria on every request.
It helped, but it didn't hold when things got busy. In 2024 I still had to cut planned customer care research to fit in a custom report link for one customer. The rules told people how to ask for design work. They didn't show anyone what it cost.
What changed
In August 2024 I started sizing UX work in points, the way engineering already did, at 32 points per designer per quarter. Nothing was accepted until it had a size. Once the numbers were in front of the product team, they could see our real capacity had shrunk as designers got spread across more products. In January 2025 the Director of Product Experience and I agreed that larger requests would go through product leadership. I didn't have to cut research again.
The same habit helped with bigger calls. In June 2025 leadership wanted to speed up the next-generation rollout, and one idea was letting customers switch back and forth between the old and new portals. I wrote up what that would do to access, entitlements, and support, and walked the product and engineering leads through it. In July the program decided against it. Once a customer moves, they stay moved.
Around the same time, new homepage rules pushed our estimate from 44 to 72 points. I broke the extra work down with product, we cut it to an MVP, and the epic split across two milestones instead of missing its date.
What we delivered
- A UX review on every backlog item before acceptance testing, plus a Figma component library with handoff standards. Our December 2024 UX maturity assessment scored Operations 16 out of 20.
- A redesigned account-access approval flow. With the automated approvals it supports, 1,929 requests between November 2024 and September 2026 cleared in about 30 seconds each, down from about 70 hours.
- The navigation model for the next-generation portal. It passed all four of our stress tests, beat the other three models, and became the shell the first portal on the new platform launched on in March 2025.
- UL's first fully responsive app standard, down to a 360 pixel screen, and notification guidelines that product turned into backlog items.
Not everything went my way. We were over capacity when a vendor's access design came in, and it brought back problems we had already fixed. We accepted it and planned to fix it after rollout. Then the program changed direction in October 2025, before the full portal reached UL's certification customers.
What I'd do differently
I'd start sizing in points on day one, right next to the intake rules. The rules got people asking the right way. The sizes are what got the product team making trade-offs with us.