Product Design · Fintech
Decisions in hours.17 new deals in year one.
Context
ECSI, a Global Payments company, served university administrators, financial officers, and the executives who answer for enrollment and budget outcomes.
Enrollment and budget calls worth millions used to wait on static PDF reports. Someone had to know the right date ranges, request the run, generate it, and print it for the meeting. Support fielded a steady stream of those requests every month.
Their old workflow had three problems:
- It was slow. Every insight meant requesting and waiting on a report run.
- It was fragile. One wrong parameter meant starting the whole request over.
- It didn't hold up in a room. Everyone showed up with their own printout, so meetings fractured into people reading different pages instead of looking at one shared story.
My role: sole designer, discovery through delivery. I ran research, prototyping, and testing, and I presented the work to a room of university CFOs. I did it alongside a PM, two client advisors, a marketing director, and five developers. The client advisors were my line into what clients actually wanted; I leaned on their relationships for insight I couldn't have gotten cold.
The reporting system wasn't failing on data. It was failing on speed, and on the meeting the data was for.
Objective
What they needed was simple to say and hard to deliver: the numbers, live, in a form they could shape and present together.
Three real-world limits shaped every decision:
- A legacy data model I had to design within, not around.
- A corporate rebrand that landed in the middle of the project.
- A hard print requirement. These are board packets. "Just go digital" was never on the table.
Approach
1. The reports weren't hard to read, they were hard to find
Eye-tracking and clickstream analysis showed users losing a significant share of their time to navigation before they even reached data. They knew what they wanted; the system buried it.
2. A handful of reports carried most of the value
Rather than rebuild everything, I looked at what clients actually ran. A handful of reports accounted for the overwhelming majority of usage. I built the dashboards around those, gave each one a single job, and put the long tail behind search and a catch-all view.
The organizing principle: one dashboard, one story. Each screen makes a single argument instead of dumping every available metric on the page.
3. A design system that could flex without breaking
The rebrand forced a design-system decision early. I built one using atomic design, starting with a content inventory of 30 modules and refining down to 10 that could each carry multiple content types. Storybook plus Figma tokens kept brand and code in sync and meaningfully cut down on QA bugs.
4. Two one-week sprints with real executives
I ran usability sessions with executives across two sprints. Three changes came directly out of what I watched them struggle with:
- Navigation moved to a sidebar. Users expected report tabs to live where the product lived, not in a top bar.
- The date-range picker was rebuilt to behave like the calendar selectors people already knew.
- On-screen print was added, so the board packet came straight off the live view with zero extra dev effort.
5. The roadblock was a wrong mental model, not a labeling tweak
My first wireframes nested all five report areas as tabs under the product. In testing, users kept looking for that navigation somewhere else entirely. It wasn't a labeling tweak, it was a wrong mental model, and I had to move the whole structure into the navigation column and retest before it clicked.
Key decisions and trade-offs
| Decision | Why | Result |
|---|---|---|
| Live dashboards over PDF runs | Kill the request-and-wait cycle | Fewer support requests, faster decisions |
| Customizable widgets per school | Every institution's data tells a different story | Executive satisfaction improved noticeably |
| Built-in print view | Board packets were non-negotiable | Requirement met with no added dev cost |
Outcomes & Impact
- Executives stopped waiting on reports and started closing decisions in the room, pulling live numbers themselves instead of making the call days later.
- Enrollment and budget calls worth millions no longer waited on a report run.
- Support stopped fielding a steady stream of report requests every month.
- Executive satisfaction improved noticeably once each institution could shape its own view.
- Board packets printed straight off the live view, meeting the requirement with no added dev cost.
- The dashboards became a selling point strong enough to lift flagship product sales.
Reflection
Design for the decision before it becomes urgent
I designed for the meeting people were already having. The natural next move is to design for the decision before it becomes urgent: predictive enrollment forecasting and KPI email alerts, so the dashboard pushes the signal instead of waiting to be opened. That was backlog when I left, and it's where I'd take it next.
Limiting interaction patterns accelerated adoption
A few well-designed, familiar patterns got executives productive faster than a richer, more flexible toolkit would have.
