Start With the Question the Report Must Answer

Every reporting project starts with a request that is not a question.

“Can we get a dashboard for inbound?” “I want visibility into follow-through.” “Leadership wants to see the funnel.” Each of those is a request for a surface. None of them says what anybody would do differently depending on what the surface showed.

So a surface gets built. It is not wrong. It gets looked at on Monday. And no decision has ever changed because of it, which is a strange kind of failure because nothing about it is broken.

The test

A business question has an answer that would change something. Not “how is inbound follow-through doing”, which has no answer that changes anything, but something closer to: of the qualified inbound requests we received last quarter, what share reached a booked conversation, and where did the rest stop?

Now ask the follow-up that makes it real: what would you do differently if the answer were much worse than you expect? If the honest reply is “nothing, but it would be good to know”, that is a legitimate thing to want and it is not what this guide is for. If the reply names an action, you have a question worth building for, and you now also know which cut of the data actually matters.

Two metrics, two questions

Two metrics can both be correct and answer completely different questions. A momentum measure answers “is this moving?” A qualification measure answers “is this real?” A record can be maximal on the first and poor on the second, and that is not a contradiction. It is two facts.

The failure is that whichever one is more visible becomes, in practice, the answer to everything. If your specification produces more than one headline number, say next to each what it does not answer.