--- title: "The dashboard nobody opens: BI that answers questions" url: "https://hurukan.com/blog/the-dashboard-nobody-opens" description: "Dashboards are built for the questions someone had last year. Managers have new ones every week. Analytics should start from the decision and the question, not the chart." --- # The dashboard nobody opens: BI that answers questions June 9, 2026·3 min read·Editorial team A manager seen from behind standing in a glass meeting room, looking at a large wall screen with abstract blue-and-grey charts, a notebook with handwritten questions on the table and empty chairs around it The dashboard project took a quarter. It has twelve tabs, a colour scheme agreed with the board and a launch email from the managing director. Three months later the usage log tells the story: the first tab is opened on Monday mornings by two people. The other eleven are opened by the analyst who built them, when something breaks. Meanwhile the sales director still emails that analyst with questions. Why did margin fall in the southern region last month? Is it one customer or all of them? What happens if the supplier price increase goes through? The answers come back two days later, in a spreadsheet, and the dashboard is not involved. The cause is not bad design. A dashboard is a frozen set of answers to questions somebody anticipated when it was specified. Management questions are mostly “why” and “what if”, they come with follow-ups, and they are specific to this week. And a red number, on its own, does not tell anybody what to do. ## Why dashboards go unopened **They answer “what”; people ask “why”.** A chart can show that margin fell. Explaining it means slicing by customer, product and price, and following whichever thread turns out to matter. A fixed layout cannot anticipate which thread that will be. **They are built for a role and used by a person.** “Regional sales manager” is a role. The actual manager is worried about two accounts and a new competitor. The dashboard shows them everything except those. **They require going somewhere.** A separate tool, a separate login, a separate habit. Questions arise in email, in meetings and on the phone, and the dashboard is never where the question is. **Numbers without trust.** If two dashboards disagree once, people stop trusting both and go back to asking the analyst. Consistent definitions are a precondition, covered in [one definition of revenue](https://hurukan.com/blog/one-definition-of-revenue). **Metrics turned into targets.** Goodhart’s law observes that once a measure becomes a target, it stops being a good measure. Dashboards tied to bonuses are watched closely and improved creatively, and they stop describing the business. **No path to action.** Even when the dashboard shows the problem, the next step happens somewhere else, by someone else, if at all. ## Analytics that starts from the question **Collect questions, not requirements.** For a month, keep a log of every real question managers ask about the business, in email, in meetings and in passing. That log is a better specification than any requirements workshop, because it records what people actually need to know. **Define the metrics once, in a shared layer.** Revenue, margin, active customer and backlog are defined in one place on top of the data, and every answer uses those definitions. **Answer questions in language, and show the working.** A manager asks in plain words; the answer comes with the number, the definition used, the filters applied and the sources. Language models are good at turning a question into a query against a well-defined model of the business. They are poor at inventing definitions, which is why the definitions have to exist first. More on where they help and where they do not in [where language models earn their keep](https://hurukan.com/blog/where-language-models-earn-their-keep). **Keep a few dashboards for monitoring.** Some numbers do need watching continuously: cash, backlog, overdue receivables. Those deserve a fixed view. Most other questions are better answered on demand. **Connect the answer to a decision.** If margin fell because of one supplier’s price, the next step, reviewing that contract, becomes a task with an owner and a date, not a comment in a meeting. Hurukan is designed this way. Reports per area sit in the [Finance](https://hurukan.com/modules/finance) module, including budget control as plan, committed, actual and forecast. They read one data model from the [Data layer](https://hurukan.com/system/data), questions are answered against the company’s model with the sources shown, and a decision that follows becomes a proposal in the approval inbox. It is in build, and nothing here describes a deployment. ## A test you can run today Most BI tools record usage. Open those statistics and list every dashboard that nobody has opened in the last thirty days. Then search your inbox for the last ten questions you or your managers sent to an analyst. For each, ask: could an existing dashboard have answered it, without help, in under five minutes? Compare the two lists. The unopened dashboards show what was built that nobody needed. The unanswered questions show what was needed that nobody built. The distance between them is the size of the problem, and the second list is where better analytics should start. If the number everyone argues about cannot even be traced to its source, start one step earlier, with [a data platform that is more than a data lake](https://hurukan.com/blog/a-data-platform-is-not-a-data-lake). - business intelligence - dashboards - metrics - decision-making ## Keep reading - [One identity for people and agents: who may do what](https://hurukan.com/blog/one-identity-for-people-and-agents) - [The live state and the record: two memories of a company operating system](https://hurukan.com/blog/the-live-state-and-the-record-two-memories-of-a-company-operating-system)