Nobody reads the dashboard, reporting should tell you to do something, not just show you numbers

A dashboard gets built. There is a meeting where it is shown, and it looks good, because dashboards always look good on the day they are demonstrated. For two weeks people open it. By the second month, nobody does.

This happens so reliably that it is worth treating as a design problem rather than a discipline problem. People are not failing to check the dashboard because they lack rigour. They are not checking it because it never asked them to do anything.

A number with no decision attached is decoration

Most reporting answers the question how are we doing. That is a comfortable question and it has no end. You can look at the answer every day for a year and never be required to act, which means the cost of not looking is zero, which means eventually you stop.

Useful reporting answers a narrower question. Do I need to do something today, and if so, what. That question has a threshold in it, and a threshold is the thing that turns a number into an instruction. Enquiries this week is a fact. Enquiries this week are down by a third on the same week last month is the beginning of a decision.

Push, do not pull

The second failure is delivery. A dashboard waits for you to visit it. Every visit is a small act of will, competing with everything else in the day, and the visits that matter most are the ones during the busy weeks when nobody has the time.

So the reports we build come to people. A short weekly summary in an inbox, on the same morning every week. An alert when something crosses a line that was agreed in advance, sent to the person who can act on it, with enough context that they do not have to go looking. The full detail still lives somewhere for the moments you want to dig, but nothing important depends on anyone remembering to go and look.

Three questions before anything gets built

Before a report is built, we want three answers. What decision does this inform. What number would change that decision, and in which direction. Who is the one person who acts when it does.

If any of those three cannot be answered, the report does not get built, and that is usually the most valuable outcome of the conversation. A surprising amount of requested reporting evaporates at question one, which saves everybody the ritual of building something that will be quietly abandoned by October.

What good looks like

Good reporting is boring. It is a small number of things, delivered without being asked for, each attached to an action and an owner. It tends to be less impressive in a demo than a wall of charts, and considerably more useful in the eleventh month.

If you already have a dashboard nobody opens, do not add to it. Ask what it was supposed to make somebody do, and build that instead.