The management meeting starts. The screen could have shown the dashboard that was built with care last year. Instead, someone has sent round a spreadsheet beforehand with the figures as they really are, and the first quarter of an hour goes to the question why the dashboard says something else. Nobody decided to stop using the dashboard. It just happened.
That is how trust in a dashboard disappears: not by decision, but with the first number nobody can explain. Whoever has experienced that once checks every figure against their own source from then on, and at that point the dashboard has become redundant. Winning the trust back is possible, but not with a new chart.
Trust disappears at the first number nobody can explain
A dashboard falls into disuse for a few recurring reasons, and almost always for more than one at a time:
- The data is wrong or lags behind. One revenue figure that deviates from the accounts is enough to make every other figure suspect.
- A number without context. March revenue is there, but not whether that is good or bad, so the reader has to interpret it themselves.
- Too much at once. Forty charts without priority overwhelm instead of help; nobody knows where to start.
- Built without the user. A dashboard set up by IT or a supplier based on what happened to be in the database rarely matches the decisions someone takes every day.
- Slow or unstable. Anyone who has waited twice for an empty page opens a spreadsheet the third time.
Of these five, the first weighs heaviest. Users forgive design flaws; they do not forgive a number that is wrong. That is why recovery starts with the question whether the data can be trusted, and only then with the design.
How to tell a dashboard is dead
You do not need to ask anyone whether the dashboard is still alive; behaviour tells you. Spreadsheets still circulate with figures the dashboard already contains. Nobody asks for a change or a new insight, which is not satisfaction but indifference. Meetings open with a verbal update on where things stand instead of with the screen. And hardly any feedback has come in since the launch.
Most BI tools record who logs in when and which pages are never opened; switch those statistics on. Add a short survey among the people the dashboard was meant for, with three questions: which decision do you take with this, which number do you not trust, and what do you still get from a spreadsheet. The answers tell you not only how often the dashboard is avoided, but why.
A dashboard that informs versus a dashboard that prompts action
An informing dashboard shows what happened: revenue in March was 120,000 euros. A dashboard that prompts action says what needs to happen now: revenue in March was 120,000 euros, well below target, and the largest deviation is in the North region. The same data, but the second version answers the question the reader actually has: do I need to act, and where.
The difference is in the design. An action-oriented dashboard works with targets and thresholds, colours deviations and shows each role only what is relevant to that role. That requires a different order when building: not starting from the data that happens to be available, but from the decision someone needs to make. One decision per screen is a good rule. Anyone who needs forty charts to answer one question does not have a dashboard but an archive.
Without an owner and a definition, every number stays debatable
A number that is correct is only trusted once everyone calculates it the same way. So agree per KPI, up front, how it is calculated, which system is the source and who is responsible when the figure deviates. Record that logic in a place where users can look it up; a number you can verify is a number you do not need to distrust.
If the data comes from several systems, such as a CRM, an ERP and the accounts, those need to align at the source. A dashboard only makes contradictory figures more visible; why that is and what to do about it is in the risks of deciding on inconsistent data. And if the dashboard lags because the reporting rhythm does not match the decision rhythm, read the board decides faster than the reporting can keep up.
Repair or rebuild
Not every ignored dashboard needs rebuilding. A number without context, a slow page or a missing threshold can be fixed with targeted changes. Rebuilding is called for when the data sources have changed but the structure has not grown with them, when every change breaks something elsewhere, when the original builder has left and nobody understands the logic anymore, or when the organisation’s own KPI structure has changed. And when trust has sunk so far that users have stopped looking, an improved version of the same dashboard no longer helps; something has to come that proves itself anew.
A rebuild is then also an opportunity. Start from the three survey questions, build the first screen around one decision someone takes every week, and only expand once that screen is being used. Trust grows one correct number at a time.
Are you stuck with a dashboard nobody opens anymore? Look first at the first number that was disputed; that is usually where the cause lies, in the data or in the definition. For the work after that, from reliable sources to a screen that serves a decision, our page on dashboards and reports explains how we approach it.