Utility Tree / Utility Matrix
A decision framework for choosing among alternative courses of action by explicitly evaluating the probabilities of different outcomes and the utility (value) of each outcome, often across multiple weighted criteria. It provides a structured, quantitative comparison of complex options.
THE IDEA
Utility Tree / Utility Matrix
The utility tree and utility matrix are techniques for choosing among alternative options (courses of action). They require the analyst to separately evaluate the benefits, costs, and risks of each option, and then systematically combine these evaluations to find the option with the highest overall 'expected utility.'
Key Components
- Alternatives: The specific choices being considered (e.g., installing scrubbers, redesigning a process).
- Outcome Classes (Criteria): The major factors that determine the success or failure of a decision (e.g., cost, environmental impact, political feasibility). Each class is assigned a weight reflecting its relative importance.
- Outcomes: The specific results possible within an outcome class (e.g., for cost: low, medium, high).
- Probabilities: The likelihood of each outcome occurring if a given alternative is chosen.
- Utilities: A numerical value assigned to each outcome representing its desirability.
The Process
- Define Alternatives: List all reasonable courses of action.
- Identify Outcome Classes & Assign Weights: Determine the key dimensions of value. The sum of the weights is typically 1.0 (or 100%).
- Estimate Probabilities: For each alternative, estimate the probability of each outcome within each class. Probabilities for a given class and alternative must sum to 1.0.
- Assign Utilities: Give a utility value to every possible outcome across all classes. The scale is arbitrary but must be consistent (e.g., 0 to 1, or 0 to 10).
- Calculate Expected Utility (Matrix): Multiply the weight of each class, the probability of the outcome, and the utility of the outcome. Sum these values for each alternative.
- Make a Decision: The alternative with the highest total expected utility is the most favorable choice, assuming the inputs are sound.
Illustrative Example: Algona Pollution III
The Algona Fertilizer Company (AFC) must choose how to deal with toxic emissions. The board considers three options:
- Transport toxic waste to federal dumps (short-term, moderate cost).
- Install state-of-the-art chemical scrubbers ($3 million, 18 months).
- Redesign the chemical process (expensive, experimental, long-term).
The major factors (outcome classes) are:
- Elimination of VOCs (Weight: 0.5)
- Effect on State Legislation (Weight: 0.3)
- Cost to AFC (Weight: 0.2)
Probabilities for the outcomes under each alternative are carefully assigned. By assigning utility values to outcomes like 'Full Tax Benefits' or 'Total VOC Removal,' and calculating the weighted expected utility, the analysis presented in the book's solutions points to installing the chemical scrubbers as the best decision. This method forces the decision-maker to explicitly confront trade-offs. A cheaper option with poor environmental results can be compared with an expensive option with better environmental results using explicit assumptions about probabilities, values, and constraints.
When to use it
Strategic Business Decisions: Choosing between entering a new market, launching a product, or acquiring a competitor by analyzing market size, competition, risk, and ROI.
Personal Life Choices: Selecting a college based on cost, location, academic programs, and career prospects, weighted by personal priorities.
Public Policy Analysis: Evaluating different infrastructure projects (e.g., highways vs. rail) by systematically comparing economic impact, environmental sustainability, public support, and long-term maintenance costs.
Project Management: Selecting a technical solution from multiple vendors or design approaches by evaluating performance, cost, ease of implementation, and vendor support.
FROM THEORY TO EVERYDAY LIFE
In practice
Algona Pollution III: Selecting an Environmental Strategy
The Algona Fertilizer Company (AFC) faces a crisis over toxic VOCs. The board considers three options: 1) Transporting waste to federal dumps, 2) Installing chemical scrubbers, or 3) Redesigning the chemical process. The decision is evaluated across three weighted outcome classes: Environmental impact (VOC elimination, weight 0.5), Political impact (tax legislation, weight 0.3), and Financial impact (cost, weight 0.2). Probabilities for specific outcomes are assigned under each option (e.g., for scrubbers: .1 legislation killed, .4 reduced benefits, .5 full benefits). By assigning utility values to outcomes and calculating expected utility, the analysis favors installing scrubbers under the assigned probabilities, weights, and utilities.
A MOMENT TO REFLECT
Put it to work.
The author warns, 'We must resist at all costs being captivated by the numbers... they inevitably tend to take on a life of their own and to drive, not reflect, our analysis.' How can someone using a Utility Matrix prevent the numerical inputs from overriding their qualitative judgment and common sense?
In the Algona exercise, the probabilities for eliminating VOCs were identical across all three options. Was this realistic? How would the decision change if the probabilities differed significantly, and how does the Utility Matrix handle such sensitivity?
The analysis combines environmental, political, and financial factors. If you were on the AFC board, how would you decide on the weights for these outcome classes? Are there other outcome classes (e.g., company reputation, employee safety) that should have been included?
CONTINUE THE CONVERSATION
Source texts
- The Thinkers Toolkit by Morgan D Jones