Dashboards People Actually Use: A Short Field Guide
Every company has a dashboard nobody opens. It was requested, built, demoed once, and then abandoned. The failure is almost never technical — the data pipeline works fine. The failure is that the dashboard answered no question anyone actually had.
Start from the decision, not the data
Before a single chart, write down the sentence the user will say after looking at the screen. "We need to add a shift on Thursdays." "Refunds are spiking in the EU region." "We can pause the ad spend." If you can't name the decision, you're building a report, and reports get archived.
One screen, ranked by urgency
The most useful dashboards fit on one screen and put the thing you'd act on today at the top-left, where the eye lands first. Everything requiring scroll is implicitly "later." Use that hierarchy deliberately:
- What's on fire right now (act today)
- What's trending the wrong way (act this week)
- Context and totals (reference, not action)
Numbers need a comparison to mean anything
"Revenue: $48,200" tells you nothing. "$48,200 — up 12% vs last week" tells you everything. Every headline metric should carry a comparison: versus target, versus prior period, or versus a threshold. A number without a reference point is decoration.
Make the boring path fast
If the dashboard takes eight seconds to load, people stop opening it, and an unopened dashboard has zero value regardless of how correct it is. Pre-aggregate. Cache. Show a stale-but-instant view with a quiet "updated 4 min ago" rather than a spinner. Speed is a feature here, not an optimization.
A dashboard that loads instantly and answers one real question beats a comprehensive one that loads slowly and answers none.
Build for the decision, rank by urgency, always give a comparison, and make it fast. Do that and the dashboard earns its daily open — which is the only metric that matters for a dashboard.