What Is a Decision Tree Analysis

by Anna Khrupa on May 26, 2022

A decision tree analysis is a graphical representation of a decision-making process. It is used to clarify whether the decision is optimal or not. The analysis compares multiple courses of action against each other and displays the risks and benefits related to them visually. It is both a technique and a tool for risk management, used to support every probable alternative in action. When there are several ways to go for a company, the analysis of which one to follow is required. The alternative paths are considered, and the right one is chosen.

The easiest way to explain the decision tree analysis is to imagine a tree with branches representing different events or decisions. Each of them has its costs and risks (both positive and negative) that should be accepted or avoided where possible.

The analysis is the most effective way to understand what the cost is if something goes wrong after the decision. What does it cost for a company to choose a certain path? Every time a project manager needs to make a decision, he or she appeals to a decision tree analysis, searching for the best business value of a project.

Every branch in a decision tree has an endpoint, which is the outcome of selecting a specific path in a decision-making process. The tree is evaluated by expected monetary or quantitative analysis performed for each branch. It shows the benefits of certain decisions or actions and the probability of the downsides of the project happening along with the actual outcome, positive or negative. For example, buying or upgrading the technology with a detailed analysis of its demand.

How to Do Decision Tree Analysis for Effective Decisions

The decisions to be made and the outcomes associated with them are described graphically. It can be done either manually or digitally with the help of a flowchart or diagram tool. The structure is an arrangement of decision and probability nodes.

Decision Tree Structure

Decision trees are pretty intuitive to work with by starting from one point and going to the next one until you cannot go any further, and the outcome is 100% clear to make the final choice. The basic concept consists of nodes (where choices are made), branches (possible alternatives available at the nodes), and values. It is represented with text boxes, circles, squares, and triangles.

The following elements are applied when using decision tree analysis to create a chart:

Circles indicate probability nodes, which are a set of possible outcomes.
Squares indicate decision nodes, which are the choices made by a decision-maker.
Triangles (or bars) indicate the endpoint on a branch, which is an outcome.
Alternative branches are lines, which indicate the available alternatives.

How to go from raw data to a structured decision tree?

Steps in Decision Tree Analysis

There are a few common steps to follow to create an effective diagram. Here is how to do decision tree analysis in five steps.

A decision-maker has to:

Define a problem;
Model a process with every possible outcome;
Apply financial and probability data;
Perform “sensitivity” analysis;
Present outcomes.

First of all, the factors relevant to the solution should be determined. What is the future behavior of those factors? Then, financial data concerning the outcomes are collected. A decision-maker creates a decision tree demonstrating the existing alternatives in a problem-solution process in the form of a scheme. What is the appropriate probability value? How does it differ from the conditional monetary value for every presented outcome?

The branch with the largest expected value is identified. “Sensitivity” analysis is performed to understand how the solution reacts to changes by changing probability and conditional financial values. In which direction is the reaction moving? What risks are taken if the results of the decision tree are used? What is the best outcome from the expected ones? How much risk is the company willing to take? All of those should be calculated and analyzed. But what is the most commonly used criterion for decision tree analysis? Expected monetary value, or EMV. It specifies the amount of value expected from a particular choice. The calculation can be done directly from a chart by multiplying the probability with the impact of the risk.

Positives and Negatives in Decision Analysis

Why is schematic representation better than a written description? The diagram gives an instant, obvious answer to the problem.

Here are a few more advantages to get from decision analysis.

Advantages

Decision analysis is applicable in almost every area where there is more than one solution available (project management, organization strategies, budget planning, etc.).
The analysis assists not only with current solutions and problems. It estimates the decision after it has been made, providing a view of what to expect in the future under similar conditions.
The chart is easy to create and follow. If a decision-maker structures the diagram correctly, each element flows logically into the other.
It is a part of a project, constantly growing alongside and accommodating any changes and updates.
It requires little time and resources.

Disadvantages

Apart from all the advantages of decision tree analysis, there are some drawbacks too.

If there are too many possible actions to be taken, the process becomes overly complicated.
If a person who chooses between several courses does not involve every possible situation, a false choice can be made.

Example of Decision Tree Analysis

An example of decision tree analysis could be a decision to either build a new application or upgrade the existing one. The decision nodes are the cost of building and upgrading an app. What is the investment required? The chance nodes are potentially large and small revenues, whereas the endpoint nodes are the profits. All of these displayed in one scheme give a decision-maker a clear vision of what is going to result as the expected value and the best outcome for a company.