A decision tree is a useful construct for visualizing a firm’s business decision and setting out all logical possibilities associated with it. A decision tree presents in an ordered form the decisions faced by a company in a given situation by tracking the options available to the decision maker and the expected payoffs and probabilities associated with the potential outcome of each decision.
- Enables a thorough evaluation of available alternatives based on a review of sequences of decisions with possibilities attached to them
- Decision nodes indicate open possibilities, i.e., events where the decision maker is in control
- Outcome nodes mark situations which are out of the control of the decision maker, but of great importance
- Every outcome node bears the “expected” pay-off as the result of the sum of the different pay-offs times the respective probabilities
- Critical chance value is the indifference point between two outcomes: it is used for sensitivity analysis
- Gather and group data. All critical decisions in the planning period for a specific issue must be included—the data should be exhaustive. Apply probabilities and values to alternatives. Group related issues.
- Sequence decisions. The main issue is the first node, with each branch stemming from the node. The process is applied sequentially from left to right, until all of the possible final outcomes and related pay-offs are reached. Probabilities assigned to branches stemming from one node must add up to 1 (and alternatives must be exhaustive).
- Interpret. For every outcome node compute the expected pay-offs as the product of the expected value and probability. For every decision node choose the outcome node with the highest expected pay-off. Determine the route with the highest pay-off.
- Sensitivity Analysis. Study the variability of the judgments made along the tree. Re-examine the most subjective components. Alter the various pay-offs and probabilities to determine the point at which a different alternative would be chosen.
- Is based on the “maximum expected value” criterion. It focuses on maximizing expected pay-offs
- Standard approach to decision trees is based on pursuing alternatives that provide “maximum expected value” in terms of expected pay-offs
- Expected value criteria does not account for the decision-maker’s attitude toward risk
- Expected value criteria is appropriate when carrying out financial calculations to arrive at investment decisions. Its application to business unit strategy must be employed carefully, as some alternatives might not have a direct quantifiable payoff
- Simplifies the decision-making process by providing a systematic approach
- Forces you to set out all of the options and ask the critical what-if questions
- Shows temporal causal relationships
- Helps to judge the nature of information which needs to be gathered
- Is a “structured thinking” tool to clarify sequential steps of a decision process and to share the analysis with other decision makers
- Aids in considering a longitudinal sequence of decisions
- Can be used non-quantitatively to illustrate the decisions at hand
- Uncertainty is associated with many elements of the decision model and can easily be neglected
- Value and probability figures are difficult to estimate accurately
- Highly variable as different decision-makers might apply different probabilities and values, which lead to different results
- Size of a decision tree increases exponentially according to the complexity of the problem explored