We are here to help!

< All Topics
Print

How to structure a decision problem

A typical analysis process starts with identifying the most important decisions to be made and the associated uncertainties. Using a decision tree, this is put into system to compare alternative strategies, typically with regard to economic value and risk.

Most projects are unique in one aspect or another. A common denominator, however, is that a large part of the basis for value creation is laid through decisions made in the early stages of a project. A good basis for decision-making in the early phase increases the probability that the right choices will be made later on.

Structuring of the decisions and uncertainties in the early phase, is therefor the right focus point in most projects.

Issues to be dealt with

The first step is to identify all issues that must be taken into account in the analysis. Decisions, risks, uncertainties and milestones (if relevant) can be noted down on post-it notes for easy sharing and sorting.

Red is typically used for decisions, green for uncertainties and yellow for any other aspect that need to be dealt with.

Focus decisions

After all the issues have been identified, the next step is to agree on which decisions are most critical and which can be dealt with later.

Decisions can be divided into two groups:

  1. Focus (strategy) decisions are those decisions that clearly stand out as important and that the project should focus on in the first stage. These decisions give the overall direction of the project.
  2. Tactical decisions are the decisions that can be made after the main direction has been given, and which help to optimise the way forward.

Strategies

Different strategic alternatives can be generated by tying different decision alternatives together, starting with the focus decisions.

Depending on the context, a handful of possible strategies can usually be defined that are significantly different from the others.

Each strategy represents a set of actions (decisions) that fit together. There may be a multitude of possible combinations, but not all of them go logically together.

Constructing the tree

Decision trees use a flowchart like structure to show relationships between decisions and uncertain outcomes.

Start with the first focus decision (represented by a red square), and build the different strategies where Decision nodes represent a user choice, like choosing whether to go for a project or not. Event nodes (green circles) represent points of uncertainty during a process or project, like whether a product is accepted by a store. And Terminal nodes (yellow triangles) are the last node in a tree. It will show the payoff and probability for the associated branch in the tree.

Revising the tree

The tree may need to be revised later as decision analysis is often an iterative process, where intermediate results can shift the focus of the team or uncover the need for additional information before proceeding with the evaluation. New information may also become available in the middle of the decision process, and this needs to be reflected in updated models.

Information points

Each of the event nodes in the decision tree (green circles) represents a point were information is revealed and uncertainty is (partly or fully) resolved. This is an important distinction from points were uncertainty exists, but is not resolved.

Remaining uncertainty

The remaining unresolved uncertainty is represented by expected cash flows at the end nodes. The project will have to live with this remaining uncertainty. Note that the remaining uncertainty may differ between the different end nodes in the tree.

Next When are decision trees relevant to use?
Table of Contents