When a problem arrives, it almost never arrives as a problem. It arrives as a feeling that something is wrong, a meeting that keeps getting rescheduled, a dashboard nobody trusts. The instinct is to reach for a solution. The more useful first move is to give the mess a shape.
Structure is the cheapest leverage I know. Before you have built anything, before you have spent a dollar, you can change how solvable a problem is just by naming its parts and the relationships between them. A list of twenty worries becomes three questions. A vague anxiety about a launch becomes a decision with two options and a deadline.
A shape is a commitment
The reason people avoid this step is that a shape is a commitment. The moment you say the problem has these three parts and not those five, you are exposed. You might be wrong. The blur was safer.
But the blur is exactly what keeps a problem unsolved. You cannot build against a feeling. You can build against a structure, even a slightly wrong one, because a wrong structure is something you can correct. A feeling just persists.
So I treat the first version of the structure as disposable. Its job is not to be right. Its job is to be specific enough that being wrong teaches you something. That is what makes it a feature and not just a preamble: it is the part of the work that turns an unsolvable thing into a solvable one.