Structure Thinking

Overview

Use this skill to turn a messy situation into a clear decision path. You will model the system to find real levers, then build a compact argument that enables action. The focus is practical: define the decision, diagnose the system behavior, choose interventions, and communicate a decisive recommendation.

Preferred Inputs

• Decision owner and deadline.
• Success definition (metric, threshold, or observable change).
• Constraints (budget, time, policy, technical limits).
• Behavior over time (trend, seasonality, oscillation).

If any are missing and the user wants an answer now, proceed with explicit assumptions and mark them as Assumed.

Workflow

1) Define the Decision and Question

Goal: one clear governing question and a provisional answer.

Do:

• Write a one-sentence decision statement: “Decide whether to X by date Y to achieve Z.”
• Capture Situation, Complication, Question, Answer.
• List assumptions and unknowns explicitly.

Output:

• Governing question.
• Provisional answer in one sentence.

2) Describe Behavior Over Time

Goal: pin the problem to a trend, not a feeling.

Do:

• Summarize how the key metric changes over time.
• Note seasonality, spikes, or oscillations.
• State the time horizon that matters.

Output:

• Behavior-over-time summary (2-4 bullets).

3) Model the System

Goal: explain why the behavior persists.

Do:

• Define system boundary and stakeholders.
• Identify 1-3 critical stocks and their flows.
• Draw reinforcing and balancing loops.
• Mark delays and missing information.

Output:

• System map notes: stocks, flows, loops, delays.

4) Generate Hypotheses (MECE)

Goal: create testable explanations or options.

Do:

• Build an issue tree with 3-5 MECE branches.
• Label each branch as an assertion (not a topic).
• Rank branches by impact and evidence availability.

Output:

• Issue tree with ranked branches.

5) Select Leverage Points and Interventions

Goal: choose a small set of actions that change structure, not just parameters.

Do:

• Map top branches to leverage points.
• Propose 1-3 interventions and how they change the system.
• Identify risks, side effects, and where resistance will appear.

Output:

• Intervention shortlist with mechanism + risks.

6) Build the Argument Hierarchy

Goal: make the decision obvious and actionable.

Do:

• Lead with the answer.
• Add 2-5 support points, each an assertion.
• Place evidence under each support.
• Keep each layer MECE and parallel.

Output:

• Decision-ready outline (top-line + supports + evidence).

7) Validate and Iterate

Goal: avoid false confidence.

Do:

• Run counterfactuals and ask what would disprove the answer.
• Check for feedback delays and unintended consequences.
• Update the system model and argument as evidence changes.

Output:

• Final recommendation with confidence level and known gaps.

When Inputs Are Missing

Deliver a best-effort output with explicit assumptions.