The riskiest assumption in product work usually isn't 'will this feature work' — it's 'is this even the right problem to solve.'
1Why Teams Skip This Step
Stakeholders often arrive with solutions already in mind — 'we need a dashboard,' 'add a filter.' A PM's job is to gently pull the conversation back to the underlying problem the request is trying to solve, because the same problem often has a cheaper answer than the one first proposed.
2From Vague Pain to a Testable Problem Statement
A strong problem statement names the user, the specific pain, and the context it happens in — e.g. 'New trial users abandon setup within 2 minutes because the first screen asks for a credit card before showing any value.' That's testable and falsifiable; 'onboarding needs work' is not.
3Step-by-Step Breakdown
Introduction. Before a single mockup or sprint gets planned, a PM's job is to prove a problem is real, painful, and worth solving. Teams that skip this step build polished solutions to problems nobody actually has.
Symptom or Root Cause?. A support ticket saying 'the export button is broken' is a symptom. The root cause might be that users don't understand what the export actually does. The 5 Whys technique — asking 'why' repeatedly — pushes past the surface complaint to what's actually driving it.
Sizing the Opportunity. Not every real problem deserves a roadmap slot. Size it: how many users hit this, how often, and how painful is it when they do? A rare-but-catastrophic issue and a frequent-but-minor annoyance can score the same — the math forces an honest comparison.
Knowledge Check. Why should a PM resist jumping straight to a proposed solution before the problem is clearly framed?
- →Committing to a solution early anchors the team on one implementation and hides cheaper alternatives that solve the same root problem
- →Because engineering always prefers to be handed a finished spec
Summary. Problem and opportunity identification isn't a formality before the 'real' PM work starts — it is the real work. Get the problem wrong and no amount of excellent execution saves the product.
