The skill: Building dashboards that drive decisions, not dashboards that exist because someone asked for one. A dashboard nobody checks is worse than no dashboard, it creates the illusion of being data-driven.
Benchmarks
In a Nutshell
- One dashboard, one audience, one purpose. The CEO dashboard and the engineering dashboard are different things.
- Big number up top. The most important metric gets the most visual weight. Period.
- Every number needs context. Show the comparison: vs last week, vs last month, vs target. A number alone is meaningless.
- 5-8 charts max. If you need more, you're trying to answer too many questions on one page.
- Insight title, not descriptive title. "Revenue dropped 15% this week" beats "Weekly Revenue."
- Push, don't pull. Set up Slack alerts for anomalies. The dashboard should come to people, not the other way around.
- Kill unused dashboards quarterly. Check view counts. Archive anything nobody opened in 30 days.
The dashboard that actually gets used
The difference between a dashboard that drives decisions and one that collects dust comes down to two things: relevance and friction.
Relevance means the dashboard answers a question someone has right now. Not a question they might have someday. Before building, ask the person requesting it: "What decision will you make differently after seeing this?" If they can't answer, don't build it yet.
Friction means load time, navigation, and interpretation effort. Every extra second of load time, every additional click, every chart that requires explanation, these all compound into abandonment. The best dashboards load instantly, fit on one screen, and tell you the answer without making you think.
A practical test: if you can't glance at the dashboard for 5 seconds and know whether things are good or bad, it needs work. The visual hierarchy should make the answer obvious. Green means good, red means investigate, and the biggest number on the page is the one that matters most.
Do's and Don'ts
Written with ❤️ by a human (still)