Frame the problem before committing to a solution

The principle

Across an unusually wide range of independently-invented methods, the same move keeps appearing: do the work to establish what the problem is, and who owns that definition, before committing resources to a solution. The recurring failure each method is built to prevent is identical - solution-space work performed while the problem is still unframed, which produces a well-executed answer to the wrong question and surfaces the mismatch only at acceptance, when it is most expensive to fix. That so many methods with no shared ancestry each install a dedicated step for this is the evidence that the objective is a real invariant of how problems are solved, not the house style of any one discipline. The principle is not “plan exhaustively before acting”; it is narrower and more defensible - keep the solution from silently deciding the problem.

Convergent elements

Each member serves the objective through one specific element, not its whole apparatus:

• Double diamond - the Define gate between the two diamonds, and the rule that a team has not “earned the right” to enter solution space until the problem is converged.
• Jobs to be done - naming the job the customer is hiring the product to do before any solution lens is applied; the job is the stable problem statement.
• Rumelt good strategy kernel - the Diagnosis element: a plan without a diagnosis of the challenge is not a strategy, however strategic it sounds. Diagnosis is problem-framing made mandatory.
• Customer development - separating search for a repeatable business model from execution of one; testing the market-risk (problem/customer) hypotheses before building or scaling.
• Podmar - declaring the Outcome before Deliverables, Actions and Resources, so intent is fixed before method; the back-briefing check confirms the problem was understood, not just the task.
• Shape up shaping - shaping work so it is “solved enough to prevent wandering” before it is handed to a team; the problem is framed (with a rough solution boundary) before anyone commits to build.
• Cynefin framework - sense-and-categorise before respond: do not act until you know what kind of problem you face, because the appropriate response differs by domain.
• Requirements baseline before commitment - agreeing and baselining the meaning of each requirement, with an owner, before committing to build against it; the requirements-engineering instance of the same discipline.
• Product development pipeline - the diagnostic “which phase’s thinking was skipped or contaminated by a later phase’s”: solution thinking bleeding back into problem thinking is a named defect.
• Design thinking dschool - the Empathise and Define modes preceding Ideate; building a real insight about the need before generating options.

Independence

The members belong to six or more traditions that developed separately: industrial/service design (Design Council’s Double Diamond, Stanford d.school), lean venture practice (Steve Blank’s Customer Development, Eric Ries’s Lean Startup), military mission-command (Auftragstaktik and its PODMAR briefing descendant), corporate strategy (Richard Rumelt), requirements and systems engineering (Tom Gilb’s Planguage lineage, BAE lifecycle management), and complexity science (Cynefin). Because these lineages did not borrow the move from one another, their convergence is genuine corroboration rather than an echo. The one caution: Double Diamond and d.school share the mid-2000s design-thinking milieu, so they count as closer to one source than to two - the principle does not rest on them.

Differentiation

The members agree on the objective and diverge on how to serve it, and the divergences are the real design trade-offs:

• What counts as “the problem.” A user need (JTBD, design thinking), a strategic diagnosis (Rumelt), a market/business-model risk (Customer Development), a contractual requirement (Planguage/requirements baseline), or a type of situation (Cynefin). The framing you choose determines which member is the right tool.
• How much solution is allowed during framing. Double Diamond and d.school insist on solution-free problem definition; Shape Up deliberately carries a rough solution into shaping because problem and solution co-evolve. This is the sharpest axis: pure problem-first versus problem-framed-with-a-solution-sketch.
• Define up front, or probe to reveal. In ordered situations where commitment is expensive and downstream (requirements baselining, EVM-governed work), you define hard before acting. In complex situations (Cynefin’s complex domain, Customer Development’s unknown market) you cannot define up front, so you probe - run cheap experiments to let the problem emerge - which honours the same objective by refusing to lock a solution while the problem is unframed. Same principle, opposite tactic.

When it applies

Reach for the principle whenever commitment is expensive, irreversible, or far downstream of the framing decision - the leverage is greatest exactly where Cost committed early exceeds cost incurred bites, because a solution chosen against the wrong problem commits whole-life cost before anyone notices. How to converge is gated by the situation type: use Cynefin framework and Method fit is gated by evolution to decide between defining up front (ordered, known, costly to reverse) and probing to reveal (complex, unknown, cheap to test). The converse failure is real and worth naming: over-defining a trivial, reversible problem is waste - in Cynefin’s clear domain, acting first and correcting is cheaper than a discovery phase.

Falsification

The principle is bounded, not universal, and states its own edge. It is refuted, or shown context-bound, if a well-designed method reliably wins by committing to a solution before framing the problem - and one genuine case exists: Cynefin’s chaotic domain inverts the move to act - sense - respond, because in a genuine crisis you must establish any order before you can define anything. That inversion is a boundary condition, not a counterexample to the core, since it applies precisely where framing-first is impossible. The principle would be falsified if the convergence turned out to be lineage (it is not - see Independence) or if solution-first methods succeeded in the ordered, high-commitment contexts where the members claim framing-first is decisive. No such case is on record here.