SaaS Analytics Dashboard
How do you help 200 people understand 10 million data points?
- Senior UX/UI Designer, Design Systems Lead
- 2024
The Problem
Ops teams spent 40% of their day just *finding* data. Dashboards existed, but they were fragile—1 person owned each, and everything broke when data changed. Decisions were slow, firefighting was constant.
Context
Mid-market SaaS. 18-month engagement. Cross-functional team: 3 engineers, data lead, product lead, and me. Budget for enterprise-grade design. The ask was 'make it faster'—but the real problem was 'make it clearer.'
Key Learning
Spent 2 weeks on-site with operations teams. Found: they didn't need more data; they needed clarity on *what changed and why*. One quote stuck: 'I spend 20 minutes looking at a dashboard just to answer a yes/no question.' That changed everything.
What We Did
Built a 'decision first' model. Every dashboard started with the one thing you needed to decide. Supporting data was layered below, not alongside. Added a 'compare' tool (this week vs. baseline) and alert rules (tell me only what's broken). Introduced a design system so teams could self-serve new dashboards without breaking things.
Outcome
Decision latency dropped from 40 minutes to 8 minutes. Self-serve dashboards launched; central team unblocked. Customer retention went from 82% to 94% YoY. NPS jumped 15 points.
What's Next
We should have built more collaborative features earlier—teams were emailing screenshots and losing context. Also, the 'why' explanations (powered by data science) came late; they should have been a core feature from day one, not a bolt-on.