A dashboard is not an answer. A dashboard is a tool for someone who already knows what answer they are looking for. When a commercial team asks for a new dashboard, what they are usually asking for is an answer they don’t yet have, in a format that doesn’t actually provide it. The instinct to build the dashboard is wrong. The instinct to ask the question underneath the request is right.
The pattern
It looks like this: a stakeholder comes to the analytics team with a request that begins with “can we get a dashboard that shows…” What follows is some specification — a chart, a table, a filter, a comparison. The analytics team responds with a queue: three weeks, four weeks, eight weeks depending on the data wrangling required. The stakeholder waits. The dashboard ships. The stakeholder opens it once, screenshots the relevant cell, and never looks at it again.
The problem is not that the team built a bad dashboard. The problem is that the request was never really for a dashboard.
What the request actually was
The request was for an answer. The stakeholder had a specific question — usually something pointed, often something they needed for an upcoming meeting — and the question got dressed up in dashboard language because dashboards are how the analytics team is set up to deliver things. The dashboard becomes the work product. The answer becomes a side effect.
This is upside down. The answer is what was needed. The dashboard is an artifact that can sometimes contain the answer, but the dashboard’s reusability — its capacity to answer future questions you haven’t had yet — is usually grossly oversold.
The dashboard is the artifact. The answer is the deliverable. Most analytics teams have those backwards. Working principle, Commercial Intelligence Agency decision interfaces
What to ask instead
When a request comes in for a dashboard, three questions reframe the conversation.
What decision will this dashboard support? If the stakeholder cannot name the decision, the dashboard probably should not get built. The most common honest answer is “none yet, I just want to see the data,” which is fine but is a different kind of request — an exploratory data review, not a production dashboard.
What will you do differently if the number is high vs. low? This is the diagnostic question. If the answer is “nothing,” the dashboard is decoration. If the answer is “I would re-evaluate the Q3 plan,” you can probably skip the dashboard entirely and just deliver a single number with a recommendation.
How often will you actually look at this? Honestly. Most dashboards get opened twice: when they ship, and the next time someone forwards a link in a meeting. If the honest answer is “twice,” you are building something with the wrong shape.
When a dashboard is the right answer
Dashboards earn their keep when three conditions hold: the question recurs on a stable cadence (weekly trade review, monthly demand planning), the answer requires multiple inputs that change frequently, and the audience has been trained to interpret the visualization without supervision. Trade execution dashboards meet all three. Most strategic dashboards meet none of them.
For strategic questions — should we drop the price, should we cut the SKU, should we re-allocate the budget — the right deliverable is almost never a dashboard. It is a memo with a recommendation, a number, the credible interval on the number, and the assumptions the recommendation depends on. The memo can be six pages or six paragraphs. It cannot be a dashboard, because the dashboard is structurally incapable of containing the recommendation.
The translation
Next quarter, when someone comes to the team asking for a dashboard, do the translation: ask the three questions, find the answer underneath the request, and deliver that. The dashboard backlog will shrink. The stakeholders will be happier. The analytics team will spend its time on the substantive questions instead of the visualization layer.
This is not anti-dashboard. Dashboards are useful. It is anti-defaulting to dashboards as the answer to every analytics question, when most commercial questions deserve a memo, a model, or a meeting — not a chart.