Employer Dashboard Redesign

2026

Co-led research across the employer admin journey, then redesigned a data-dump homepage into a dashboard that shows admins what needs attention and when.

Employer admin portal, FSA 2025 usage view: enrollment and contribution totals, monthly claim spend chart, and breakdowns by service category and claim type

Context

The admin portal of a U.S. healthcare fintech hadn’t been redesigned in years. Before redesigning anything, we needed to understand how people actually used it.

Daily users
Employer HR admins and partner account managers, managing thousands of benefit accounts.
Regulated accounts
HSAs and FSAs, where mistakes create real tax problems.
Research first
Understand real usage before changing a screen.

Problem

The old homepage was a welcome message followed by payment, invoice and forfeiture tables, all with the same weight. Nothing told admins what to look at first, and a few problems kept repeating.

35
pain points
Found across the full admin year, not only the homepage.
Data hard to find
Admins couldn’t find the data they needed.
Deadlines as surprises
Nothing warned admins before a deadline arrived.
No view of usage
No visibility into how employees used their benefits.
Spreadsheets instead
Admins built Excel files to answer what the platform should have.

My role

I co-led the research with another designer, then designed the new dashboard solo: one year, from discovery through two shipped releases.

Research
Analytics, a full-year journey map and partner interviews.
Prioritize
35 pain points on a Support/Effort Landscape, then a Kano model.
Design
Two releases, designed and shipped solo.
Validate
UMUX-Lite and 5-second tests.

Process

Quantitative

Datadog analytics across 28,000 sessions: scroll maps, click heatmaps, journey flows. Most users skipped the homepage and went straight to specific sections.

Qualitative

A customer journey map of the full admin year, validated in interviews with partner account managers. AI sped up the transcript analysis: recurring themes, clustered pain points, cross-references to the map.

Prioritization

Each pain point went on a Support/Effort Landscape, and the concepts went through a Kano model. The result: two releases, must-haves first.

28,000
sessions analysed
35
pain points mapped
4
priority zones
2
releases planned

Datadog analytics board for the old portal: a scroll map, the top elements by clicks across 28,000 sessions, a click map, a user-journey Sankey diagram and the top entry points — Most sessions skipped the homepage and went straight to specific sections: the first sign it wasn’t doing its job.

Customer journey map of the employer admin year, from setup through maintenance to wrap-up, with rows of sticky notes for actions, goals, pain points and quotes under each step — Mapping the whole admin year, not only the homepage, is what surfaced the 35 pain points.

Support/Effort Landscape: a scatter plot of numbered pain points split into four zones, next to tables of high, moderate, low and very low opportunity pain points and the resulting project scope — Weighing the support load each pain point created against the return of fixing it decided the order of work.

Kano model board: concept thumbnails sorted into must-haves, performance payoffs and attractive features, and the same concepts grouped into Release 1 and Release 2 — The Kano model turned a long list of concepts into two releases, must-haves first.

Solution

A homepage ordered by urgency

Challenge: Admins kept getting surprised by deadlines, and the homepage gave every table the same weight.
Solution: What needs attention comes first: account health, compliance indicators, upcoming funding. Below it, a calendar surfaces deadlines and scheduled events before they become problems.

Old employer admin homepage: a welcome banner and notifications, then stacked payment and invoice summary tables, all with the same visual weight — Drag to compare the old homepage with the redesign.

Redesigned employer homepage: enrollment counts at the top, then identity-verification status and the next invoice accrual, with the deadline calendar starting below — Drag to compare the old homepage with the redesign.

A usage view for each plan type

Challenge: Admins had no way to see how employees used their benefits without pulling reports and cross-referencing them by hand.
Solution: Each plan type gets its own view with the metrics admins asked for in research, answering in one screen what their spreadsheets used to.

FSA 2025 usage page: enrollment, contribution, election and rollover totals, a monthly claim spend chart, service category and claim type breakdowns, and plan dates and deadlines — One screen per plan type, answering the questions admins had been building spreadsheets for.

Outcome

Both releases shipped: the restructured homepage, then the account-level usage views.

82.2

UMUX-Lite score: Excellent (Grade A)

8 out of 10 participants scored above 80.

Source: UMUX-Lite usability test, 10 participants

100%

rated the visual design 4 or 5 out of 5

67% gave it a perfect score, and every participant correctly identified the page's purpose.

Source: 5-second test, 9 participants

Research became the roadmap.

Product management planned the roadmap from the journey map and the Support/Effort Landscape, which outlived the project.

Every feature traced back to a pain point.

When stakeholders asked “why this feature?”, we pointed to the data.

What I learned

Research is the design
By the time I opened Figma, most of the hard decisions were already made.
AI as a design multiplier
AI handled the work around the design: interview analysis, layout exploration, test scenarios. The thinking stayed mine.
Prioritization earns trust
Being able to say why we build this first changed how the team saw my role.