Data-Driven Business Decisions

CHAPTER 8

Making Complex Decisions with Trees

8.0. Introduction: What Is the Issue?

Earlier chapters introduced some basic tools for analyzing simple business decisions. The key ones were mean value (MV), conditional probability, and some standard distributions like binomial or normal for obtaining these probabilities. In a real situation, a manager will have a business process that evolves over time. At various points of this process there are sources of uncertainty that are outside the manager's control but also opportunities to make decisions that make better outcomes more likely and the worst outcomes less likely. And to be quantitative, the manager will have to carefully account for the various costs and revenues that may accrue along the way. Without a formal structure for setting out the problem, it is very easy to mix up the accounting, or the treatment of uncertainty or the logical relationships between the decisions and uncertain outcomes. Decision trees provide such a formal structure.

Let's introduce the main ideas through the GamerCorp case. This is quite a complex case so we will use a slightly simplified version. But by the end of the topic, you might like to read the full case and attempt an analysis.

Case Briefing: GamerCorp Releases Their New Gaming Console

GamerCorp designs platforms for electronic games. They are at the final stages of developing the new platform and could easily release it by July 2009, summer being the standard release period for gaming consoles. ...