A legacy analytics tool used daily by 500+ enterprise users across 3 business functions had become impossible to navigate. A research-driven redesign transformed it into a focused, role-based experience — cutting task completion time in half.
The analytics platform had been built feature-by-feature over four years. Every team had requested additions; nothing had ever been removed. The result was a dashboard with 47 visible metrics on the default view, a navigation menu with 6 levels of nesting, and a filter system that required training to use.
Support tickets for the analytics tool accounted for 23% of all internal helpdesk requests. Quarterly NPS for the tool sat at -12. The Head of Data called it "the most important tool we have that everyone hates."
"How might we redesign the analytics experience so that each user can find what matters to them in under 30 seconds — without losing access to the depth that power users need?"
I started by interviewing 22 users across the three major user groups: Sales Operations, Marketing Analytics, and Finance. What emerged was striking — these three groups were fundamentally different in how they used the tool, what they needed from it, and how often they used it.
Used daily. Needed 4–5 specific KPIs. Valued speed above all. Spent 60% of their time dismissing charts that weren't relevant to them.
Used 2–3× per week. Deep exploratory sessions. Needed powerful filtering and data export. Felt constrained by the rigid grid layout.
Monthly review cycles. Needed historical trend data and clean exports for board decks. Found the real-time updates distracting and irrelevant.
One dashboard could never serve all three groups well. The solution wasn't to redesign a single view — it was to introduce role-based views that shared a common data layer.
The architectural concept was a role-based dashboard model: users would be assigned a default view on login (configurable by admins), with the ability to switch views if needed. Each view exposed the same underlying data, just through a lens optimized for that role's primary jobs-to-be-done.
Navigation was collapsed from 6 levels to 2. The IA was rebuilt from scratch using card sorting with 30 participants to validate groupings before any wireframes were drawn.
Usability testing ran across 4 iterations. We measured time-on-task for 6 core scenarios per role. By the fourth round, every scenario met or exceeded our 30-second target.
The new dashboard shipped in phases over 8 weeks — Sales view first, then Analytics, then Finance. Each release was accompanied by a short in-app walkthrough and a role-specific help article.
The role-based architecture has since been extended to two additional teams (Customer Success and Product) without requiring a redesign — it proved to be a scalable pattern.
Lessons learned: When a tool tries to serve everyone equally, it ends up serving no one particularly well. Permission to segment the experience by role was the single most impactful decision of the project — and it only came after we listened carefully enough to hear that the problem wasn't the UI, it was the information architecture.