Decision Tree Construction Protocol

A step-by-step protocol for building a visual decision tree to structure complex decisions with multiple stages and uncertain outcomes. Step 1 — Define the initial decision: Draw a square node and label each branch with a distinct option. Step 2 — For each option, identify the key uncertain outcomes: Draw circle nodes and label each branch with a possible outcome and its estimated probability (probabilities from each circle must sum to 100%). Step 3 — For each outcome, determine if another decision follows: If yes, draw another square node. If no, estimate the final value or payoff at that endpoint. Step 4 — Calculate Expected Value by working backward: Multiply each endpoint value by the probability of reaching it. At each circle node, sum the probability-weighted values. At each square node, select the option with the highest expected value. Step 5 — Sensitivity check: Vary your probability estimates by +/- 20% and see if the best decision changes. If it does, your decision is highly sensitive to that assumption — invest in better information before deciding.

When to use it

Multi-stage decisions where one choice opens or closes future options. Investment decisions with uncertain outcomes. Strategic planning with multiple possible market scenarios. Medical decisions with branching treatment paths. Any decision where you find yourself saying 'Well, it depends on...' — each 'it depends' becomes a branch in your tree.

How it can help

Makes the invisible structure of complex decisions visible. Many decisions that feel overwhelming become surprisingly clear when drawn out — you often discover that the decision hinges on one or two key assumptions, not the dozens of factors swirling in your head. The sensitivity check is particularly powerful because it tells you where additional research has the highest payoff. Works on paper, whiteboard, or any simple drawing tool — no special software required.