Why nobody opens the dashboards: they don't answer a decision question
The difference between a dashboard opened every morning and one forgotten in two weeks isn't design — it's whether it was built from a specific decision question owned by a specific person.
Most BI projects start with "we want a dashboard that shows everything." Two months later an impressive dashboard ships… everyone opens it during launch week, then nobody returns. That's not a tooling problem — it's inverted design.
A successful dashboard starts from a decision question: "Which SKUs do we stop buying this quarter?" "Which branch needs intervention this week?" "Which customer deserves a higher credit limit?" Every question has a specific owner and a specific decision cadence — and those two facts dictate the dashboard's content, refresh rate, and layout.
Then comes the KPI dictionary — the most boring, highest-value deliverable of the project: one written definition per metric. Does "sales" include tax? Before or after returns? At which exchange rate? Every meeting that opens with an argument about numbers is the invoice for an unwritten dictionary.
Finally, trust: attach a visible data-quality indicator to every dashboard — last refresh, source completeness. A manager who once discovered the dashboard was "two days stale" without being told will never trust it again. Honesty about limitations is what builds reliance.