The Challenge
The existing analytics tool was built by engineers for engineers. Product managers and executives — the people who needed the data most — were bouncing to Excel and building their own reports. That's a signal the tool had failed them.
Research
I spent two weeks shadowing five different user types: a growth PM, a finance director, a customer success manager, a data analyst, and a C-suite exec. Each had fundamentally different mental models of what "a dashboard" should do.
"I just want to know: are we growing? Is anything broken? That's it."
The analyst wanted granularity. The exec wanted a single number. The solution had to serve both without compromise.
What I Designed
The key insight was context layering — a summary layer that answered the exec's question in 5 seconds, with progressive depth available for anyone who needed to dig. I introduced:
- KPI cards with inline trend sparklines and anomaly flags
- Drill-down panels that opened contextually without navigation
- Saved views so each team could set their own default layout
Results
Adoption among non-technical users increased by 60% in the quarter following launch. Monthly active users for the dashboard tool doubled. The product team reported a significant drop in ad-hoc data requests to the engineering team.