This case study showcases: Data visualization design • User testing & iteration • B2B UX • Collaborative design • Complex interaction design
Role: UX Designer | Timeline: 7 months | Team: 3 developers, 2 Data scientists |
Why were data analysts constantly switching between tools?
Data analysts and human resources teams faced a fractured workflow: constant context-switching between dashboards, spreadsheets, and presentations. Each transition risked losing valuable insights, while fragmented tools created isolated data silos that prevented discovery of deeper employee experience patterns.
Working with a highly technical team, we embraced collaborative design, making data analysts active participants in shaping the solution making someone's job easier.
I joined from the very beginning and helped shape the product direction — from defining user flows and functionality to prioritizing what mattered most.
What do employees really need from their data?
The analysis crystallized three core jobs-to-be-done, each representing a critical analysis phase in transforming raw employee data into meaningful narratives about our people's journey.
What if visualization could be navigation?
Breaking from conventional dashboard thinking, we envisioned a system where visualization became the primary means of navigation. Key conceptual breakthroughs emerged:
Graph visualization:
Translating analysts' natural whiteboard mapping behavior into digital experience
Drawer system:
Addressing the need to maintain context across multiple data views
Contextual navigation:
Seamless drilling through organizational layers
How do we know it actually works for users?
We implemented a two-phase testing strategy to validate both foundational assumptions and real-world task completion.
What did stakeholders really understand?
Tested terminology, data interpretation, and role-specific needs with dashboard creators.
Before | After |
|---|---|
"View type" - caused confusion about functionality | "Calculation method" - matched user mental models |
"Compare results against" - too long | "Analyse by" - clearer for organizational filters |
Could users actually complete their real tasks?
Scenario-driven testing with data consumers revealed critical interaction issues.
Metric | Result | Insight |
|---|---|---|
Task completion rate | 78% | Consistent across all users |
Graph interpretation failure | 100% | Network graphs didn't match mental models |
Team comparison struggles | 75% | Interaction pattern needed redesign |
When users fail, what does that teach us?
Time visualization failure: Users took time to interpret discrete time-series charts. With real data showing minimal variance, we completely redesigned with bars chart.
Graph vs. hierarchy: 100% struggling rate on understanding network graph tasks led to adding hierarchical tree view - users needed familiar organizational chart structure alongside innovative network visualization.
Comparison workflow: Redesigned multi-selection mechanism with explicit comparison mode and side-by-side panels.
What does success look like in the real world?
Our final solution evolved into an interactive platform featuring:
Ring organization view: Immediate insight into team health
Contextual drilling: Seamless navigation through organizational layers
Dual visualization modes: Network graphs for pattern discovery, hierarchical trees for familiar navigation
Integrated comparisons: Side-by-side analysis eliminating tool-switching
Historical context: Trends integrated directly into team views
Did it actually solve the problem?
The tool launched with immediate adoption success and positive workshop feedback.
Metric | Result |
|---|---|
Active users in month 1 | 200+ |
Pre-launch demand | Users actively asking for release dates |
"Pretty easy to find the results asked" - User testing participant
What would I do differently next time?
The most effective decisions came from letting go of cleverness and focusing on clarity.
It wasn’t always the elegant solution that worked — it was the one people understood.
Because if no one can use it, it doesn’t matter how clever it is.