How to turn a vague goal into an actual strategy

“Grow faster,” “become the trusted platform,” and “improve customer experience” describe desired outcomes. They do not explain what prevents the outcome, which approach should receive concentrated effort, or what attractive work should wait. Calling the goal a strategy postpones the conflict strategy exists to resolve.

Convert the goal into a causal argument. Diagnose the obstacle that matters most now, choose a guiding policy for addressing it, and select actions that reinforce one another because of that policy. Then expose the simplifications in the diagnosis to people with local knowledge and define evidence that will force an update.

Good Strategy Bad Strategy supplies the kernel of diagnosis, guiding policy, and coherent action. Seeing Like a State supplies the warning: a clean strategic model can erase the local facts and informal systems on which execution depends.

A goal names the destination; diagnosis explains the difficulty

Rumelt’s diagnosis reduces a complex situation to an explanation of the central challenge. “Revenue is below plan” is a symptom. A diagnosis identifies the mechanism: perhaps target users do not reach value before setup cost becomes visible, buyers cannot justify switching risk, or acquisition brings a segment the product is not designed to retain.

A useful diagnosis is specific enough to rule actions in and out, broad enough to connect multiple observations, and contestable enough that evidence can challenge it. It should distinguish the obstacle from the absence of the desired result.

Begin with observations, not strategic vocabulary. Which behavior fails? Where does the process stop? What workaround exists? Which constraint recurs across cases? What changed when performance was better? Preserve facts separately from the inference that connects them.

Competing diagnoses should be written before they become competing roadmaps. One group may believe demand is weak; another may believe demand exists but adoption fails. Funding both full plans avoids the disagreement while dispersing effort. Identify the evidence that would distinguish the models.

A guiding policy creates a rule for exclusion

The guiding policy states the overall approach to the diagnosed obstacle. It is more constraining than a value and less detailed than a project plan. It should help people make local decisions without returning every choice to leadership.

Test the policy with plausible work. If every existing initiative can be reframed as supporting it, the policy has not concentrated anything. A strategy becomes real when resources move and a reasonable activity is delayed or stopped.

The policy should also express the chosen mechanism. “Focus on onboarding” remains a topic. A policy might prioritize letting users experience the core outcome before asking them to complete migration or setup. That criterion can shape product, messaging, service, and measurement coherently.

Policies are provisional. Their authority comes from the current diagnosis, so evidence that weakens the diagnosis should weaken the policy. Loyalty to a slogan after its causal basis fails is not strategic consistency.

Coherent actions create more than a ranked portfolio

Actions are coherent when each implements the guiding policy and makes the other chosen actions more effective. A collection of high-priority initiatives can remain strategically weak if each responds to a different theory of progress.

For every action, state which part of the policy it expresses and what output or capability it creates. Then ask whether removing it breaks the strategic chain. Work can be necessary for maintenance, compliance, or another operating lane without belonging to the strategy; label it honestly rather than forcing every obligation into the narrative.

Coherence also requires sequencing. A proximate objective creates a challenge close enough to solve with current knowledge and valuable enough to improve the next position. The organization may not know the complete path to the goal, but it should know which uncertainty or bottleneck must be resolved first.

Capacity is part of the argument. A strategy with more coherent actions than the organization can execute is still a list of aspirations. Concentration means matching the set to real people, time, and dependencies.

Local knowledge should challenge the diagnosis before scale

Scott’s legibility critique applies because strategy simplifies. Metrics, segments, workflow diagrams, and operating models make the situation visible to decision-makers while omitting variation. The people closest to customers and execution may know which exception is normal, which workaround carries the process, or which metric hides displaced cost.

Run a local-knowledge review before commitment. Ask operators what the diagnosis gets wrong, which behavior does not fit, what would break if the policy were applied literally, and which informal process needs protection or deliberate replacement.

This is not a request for unanimous approval. Local experience can also overgeneralize from salient cases. Preserve observations, compare across sources, and distinguish a contradiction of the mechanism from a preference for existing work.

Use reversible pilots where uncertainty is high. The pilot should exercise the core mechanism and include the hard dependency rather than demonstrate an easy subset. Define rollback and the next commitment before results arrive.

A strategy needs update conditions, not only targets

A target says what success looks like. An update condition says what evidence would change the diagnosis, policy, or actions. Without it, a missed target can always produce more effort inside the same model.

Choose leading observations tied to the causal chain. If the policy is correct, what behavior should change first? Which segment should respond? What should become easier or stop occurring? Add a guardrail for the cost the policy might shift elsewhere.

Record the baseline, expected direction and magnitude, review point, and authority to revise. Strategy is a hypothesis that earns concentrated commitment; it should be strong enough to guide action and explicit enough to be proven wrong.

The next move is a one-page strategy kernel

Write the goal at the top, then complete five paragraphs. First, the diagnosis: the central obstacle and causal mechanism. Second, the evidence: observations supporting it and the strongest competing explanation. Third, the guiding policy: the approach and the attractive work it excludes. Fourth, the coherent actions: the few linked commitments, owners, and sequence. Fifth, the update condition: what result changes the model.

Review the page with people closest to execution. Add the local facts the clean model omitted and identify which decision is reversible enough to test first. Remove any action that cannot explain its relationship to the policy or move it to the appropriate non-strategic operating lane.

The page is ready when a colleague can use it to reject a plausible request, explain why the remaining actions belong together, and name the observation that would make the organization change direction. Until then, the document contains goals and activity, not strategy.