Business Intelligence
Why Nobody Opens Your Dashboard
It loads, the numbers are right, and usage is twelve views a month. The usual cause is that it answers a question nobody has to act on.
Most abandoned dashboards are technically fine. The data is accurate, the refresh works, the layout is tidy. They are abandoned because they were built from a request for information rather than a decision that needed making.
Start from the decision
Before building anything, get an answer to three questions: who acts on this, what are the actions available to them, and what would the number have to say to change which action they take. If the third question has no answer, the dashboard is decoration and its usage will reflect that.
If no plausible value of the metric changes anyone's behaviour, you are building a report, not a tool.
The failure modes
• Too many tiles — thirty charts mean the reader has to do the prioritising you were supposed to do.
• No comparison — a number with no target, prior period, or benchmark cannot be judged.
• Wrong grain — a regional director cannot act on a company-wide total.
• Latency mismatch — a daily refresh for a decision made hourly, or vice versa.
• No owner — nobody is embarrassed when it breaks, so it stays broken.
Answer the first question in the first screen
A good dashboard says whether things are normal before it says anything else. Lead with two or three numbers against their targets, then supporting detail, then the breakdowns for people who are diagnosing rather than monitoring. If the reader has to scroll or filter to find out whether they should be worried, the design has inverted.
Push it to where the work happens
Requiring someone to remember to visit a URL is a weak mechanism. The dashboards that survive are the ones embedded in a routine — the first slide of a standing meeting, a scheduled summary in the channel the team already reads, an alert that fires only when a threshold breaks. Delivery beats availability.
Instrument and prune
Log views by user and review the list quarterly. Dashboards with no viewers in ninety days should be archived, not maintained — and the archiving is a favour to everyone, because a long list of stale dashboards is why people stop trusting the catalogue and go back to asking an analyst directly.
The uncomfortable version of this test: retire something quietly and see who notices. If nobody does within a month, you have your answer, and you have freed the refresh capacity for something people use.
We ran the retirement test and archived nineteen of thirty-one dashboards. Two people noticed. Both were rebuilt in an afternoon and are better than the originals because we finally asked what decision they served.
The latency mismatch point is underrated. Ours refreshed nightly for a team making intraday calls, so they built their own spreadsheet. Not a trust problem — a timing problem, and we spent a year misdiagnosing it.
From the consumer side: the grain issue is the main reason I stopped opening ours. Company totals told me nothing I could act on and there was no way to filter to my own region without three clicks I never remembered.