LEARN. PRACTICE. GROW.Concept and mechanism
A decision tree separates controllable choices from uncertain outcomes. For each option, calculate each branch’s cost or benefit and weight it by the relevant probability. In the teaching example, A costs EUR30,000 with a 25% chance of another EUR20,000: expected cost is EUR35,000. B costs EUR34,000 without that addition, so it minimizes expected cost within this model. The conclusion does not prove universal superiority of B; other objectives and consequences may change the decision. Sensitivity analysis helps identify which data and assumptions change results most, guiding evidence gathering and response investigation.
Guided application
Avoid precision unsupported by data. Three observations do not by themselves justify a probability reported to two decimal places. Explain origin, ranges or scenarios, and test whether the decision changes. For explicitly independent events with 20% probability each, the probability of at least one is 1 − 0.8 × 0.8 = 36%; adding 20% and 20% counts the intersection twice. For reserve, distinguish total from difference: if base cost excluding reserve is EUR400,000 and total-cost P80 is EUR460,000, the difference is EUR60,000. Check that both use the same scope, reserve is not already included, and funding decisions follow defined authority.
A reserve derived from P80 still allows outcomes above the modeled value; it is not maximum-cost insurance.
Common pitfalls
Forgotten fixed cost; small sample with false precision; improper probability addition; double-counted reserve.
Related topics: Choose and execute responses · Monitor, close, and hand over to operations
Check units, branches, assumptions, and criteria before recommending an option.
Reference: Decision tree analysis for the risk averse organization · PMI-RMP five-domain ECO, updated-2024 public document