Decision Tree Analysis: Definition and How to Perform One

What is a decision tree analysis?

A decision tree analysis is a visual representation or flowchart of the potential options and outcomes
when making a decision, together with the cost of each choice. It gives an individual or organization the
opportunity to see each of the possible outcomes alongside the others. It's especially useful when you
have quantitative data to illustrate the impact of each possible choice.

The flowchart starts with the decision and branches out to each of the possible outcomes. Each specific
outcome results from one decision set and its consequences. By comparing each of the outcomes, you can
make the decision that leads to the most desirable one. The purpose of the decision tree method is to
evaluate complex decisions in a simple way.

Benefits of using a decision tree

There are several benefits of using a decision tree to inform important or complex decisions, including
that it:

Is easy to use

A decision tree is an easy-to-use and low-cost way to evaluate a decision. It provides a simple format that
doesn't require extensive illustration to visually represent the options available to an individual or
organization. In most cases, it's easy to represent the data on decision trees without the requirement for
complex computation.

Provides validation

Constructing a decision tree provides a simple way for an individual or organization to validate a
decision. A decision tree can also act as a predictive model. This is because you can use it to do a
quantitative analysis or validate the results from statistical tests. With a little modification, a decision tree
can also assist in problem classification.

Enables comprehensive decisions

Using a decision tree ensures that you make a comprehensive decision by considering all the options. It
avoids rash decisions by encouraging you to take all the options into account and provides a mechanism
to display all these options so that you can compare them easily. This ability to partition the data at a
fundamental level means that the decision tree is more effective than other similar tools.

Provides specificity

Using a decision tree removes ambiguity. The ability to assign a specific value to a specific outcome
enables a quantifiable evaluation of each of the options. Evaluation of the options also becomes easier
because you display the options side by side for the decision-maker to see.

Creates transparency
One of the major benefits of the decision tree approach is that it makes the decision-making process
transparent. The tool demonstrates visually that the decision-maker has carefully considered all the
options as part of the process. Having a physical record of how those making decisions in a business
environment arrived at their outcome can be useful if anyone questions the decision and the other options
that were available for consideration.

Increases efficiency

Using a decision tree can increase the efficiency of decision-making. You clearly identify the assorted
options and the associated costs so that any discussion has a focus. When it's a group decision, the
decision tree ensures that each member of the group has access to the same information.

How to create an effective decision tree

Whether you choose to construct your decision tree on a computer or with a pen and paper, here are some
common steps for creating one:

1. Start with a key decision

Highlighting the key decision is the first step when creating a decision tree. Place the key decision in a
box in the centre of the page, allowing plenty of room on either side to add further information. Create
branches by adding lines from the key decision box. The branches represent each of the possible choices
that you can make. Be sure to label each of the branches. It's up to you whether you incorporate the
relevant data for each outcome in the decision tree.

2. Add chance nodes

As you add information to the decision tree, try to follow each branch through to the outcome that's likely
if you select that option. The outcome a particular branch represents may require additional choices. If
this is the case, add these additional outcomes to the tree. In some instances, there may be uncertainty, or
you might not know the most likely outcome of a particular choice. If this is the case, consider using a
chance node. Use a circle with additional outcomes coming out from it to represent a chance node on your
decision tree.

3. Establish decision nodes

Sometimes an additional decision may be necessary following the key decision. If this is the case, use
squares similar to the one containing the key decision to represent the additional decisions. Add the
details to each of the additional decision nodes.

4. Expand each node to an endpoint

It's important to make sure that you expand each node to an endpoint. This means adding additional
branches or decision nodes to each decision until you reach an endpoint. You typically use a triangle at
the end of the branch to represent the endpoint. Including a prioritization value in each triangle can help
users make a well-informed decision.

5. Calculate tree outcomes

If there's quantitative data available that relates to the different choices, you can add this to the decision
tree. The most common type of quantitative data is monetary values. Another option is to estimate the
expected value you're likely to create for each outcome. The probability of an outcome occurring and the
cost of the outcome are necessary to make this calculation. Once you have these values, you can use them
in the following formula:

Expected value = (First possible outcome × Likelihood of outcome) + (Second possible outcome ×
Likelihood of outcome) − Cost

6. Evaluate the outcomes

Once you've added the relevant information, it's time to evaluate the outcomes and make a decision. You
also consider the amount of risk you're willing to take when making your decision. Therefore, the
outcome with the highest expected value isn't necessarily the best option if it also requires you to take on
the highest level of risk. While the expected value is a useful parameter for making a decision, there's no
definite right or wrong outcome. It's important that the outcome you choose is the one that's best for you.

Advantages of using decision tree analysis

There are multiple tools available to assist organizations when making decisions. Using a decision tree
has several key advantages over other tools, including that it's:

 Flexible: Using a decision tree creates a degree of flexibility around the decision-making process.
It's easy to include additional branches while completing your analysis or when you've completed
the analysis and new information comes to light.
 Independent of quantitative data: A decision tree is effective with or without quantitative data.
 Quick and simple: It's a quick and simple process to create a decision tree once you understand
the different steps. You can complete this process anywhere and all you need is a pen and paper.
 Fair: Once you've created your decision tree, you have each of the outcomes laid out next to each
other. This makes it easy to compare one outcome fairly with another.
 Understandable: The simplicity of a decision tree means that it's easily understood, removing
the requirement for complicated explanations. This is especially useful in a professional situation.

Challenges when using decision tree analysis

If you want to use the decision tree to its maximum potential, it's important to understand where its
weaknesses lie so that you can compensate for them if necessary. These weaknesses include:

 The potential to become overcomplicated

 Variations in data affecting the decision branches
 Challenges when applying to multiple decisions or large organizational decisions
 The potential for bias depending on the opinions of the decision-maker
 The potential for errors when labelling and classifying branches
 The potential unsuitability for large datasets

Although there are challenges associated with the decision tree approach, there are several ways to
minimize these, including:
 Set a maximum length for branches from the key decision to avoid the decision tree from
becoming overly complicated.
 When decision-making as a group, require a majority vote before adding a new branch.
 Engage in pruning, where you remove branches that are of low importance or value.