Decision Analysis Explained
Decision makers consider various alternatives to select the best approach. Such decision-making however remains complex. Each such course of action remains fraught with uncertainties, such as changes in external environment, competitor’s actions, and other events, besides the assessment of the pros and cons remaining subjective. The decision analysis process provides a clearer insight into the differences among various alternatives, helps resolve complexities, and improves quality.
Decision analysis evaluates each option by assigning quantitative values, such as a monetary sum to the value of the outcome and probability values to quantify uncertainty. People compare numerical values much better than they compare descriptions and as such, these numerical values simplify comparisons and the selection of the superior approach.
Assume a company wants to launch a new product. The first course of action is to select an appropriate model, and list the alternatives. The two alternatives here are “launch the product” or “shelve the product,” and the appropriate tool in this context is an influence diagram.
The analysis begins by assigning quantitative values to each alternative. Assume market research shows a 70 percent chance of high demand for the product, and a 30 percent change of low demand, with the high demand leading to net profit of $400,000, and the low demand causing a net loss of $200,000. Thus, “launch a product” has a value of $400,000 with 70 percent probability of occurrence, and “shelve the product” has a value of $200,000, the opportunity cost of possible loss, with a 30 percent probability of occurrence.
The final step is to analyze the options to select the best one. “Launch the product” has the numerical value of $400,000 x 70% = $280,000 and “shelve the project” has the numerical value of $200,000 x 30% = $60,000. Thus, the preferred course of action is to proceed with the product launch.
The Steps for Making Better Decisions
Real-life decision analysis is a complex exercise, and usually requires the deployment of various mathematical models and statistical techniques. Follow these basic steps:
- Create a model structure. While there is no hard and fast rule on the best model structure, decision trees, influence diagrams, and payoff matrices find common use.
- List each possible alternative in the model structure. Possible alternatives are a finite number of possible future events, denoted as “States of Nature” identified and grouped in sets of mutually exclusive events. Represent uncertainties through probabilities and probability distributions.
- Assign numerical values to the probability of the action taking place, and the money value expected as the outcome. Assign value weights for measuring the trade-off objectives, and the risk preference to lend clarity.
- Analyze for payoffs, or expected returns. Different combinations generate different payoffs. Represent the outcome for each combination in a table, with positive (+) value for net revenue, income, or profit, and negative (-) value for expense, cost or net loss.
- Consider the payoff of each combination together with the sensitivity of the outcome, the weighted utility for key probabilities, and the risk tolerance for the combination to select the most preferred alternative.
Undertaking these steps is a complex exercise and requires some practice. Although the decision analysis process lends a structure and a methodology to evaluate choices, the basis of determining numerical values still remains subjective. False assumptions, not having an accurate estimation of the probabilities, relying on expectations, and difficulties in measuring the utility function, or forecasting errors can distort the calculation greatly.
This method nevertheless finds widespread use in many fields including but not limited to business planning, marketing, health care, environment and ecology, litigation, negotiation, dispute resolution, and more, for it provides decision-makers with “algorithms” that provide a structured and analytical alternative to “unaided intuition.”
- University of Baltimore. “Tools for Decision Analysis.” https://home.ubalt.edu/ntsbarsh/opre640a/partix.htm#rwida. Retrieved May 27, 2011.
- “Decision Analysis.” https://www.dynamic-ideas.com/Books/097591460/097591460-ch01-ex.pdf. Retrieved May 27, 2011.
- “Decision Analysis: Influence Diagram Example.” https://www.doc.ic.ac.uk/~frk/frank/da/exercises/inf.pdf. Retrieved May 27, 2011.
Image Credit: freedigitalphotos.net/renjith krishnan