Quick Look at the PMBOK® Guide: Decision Tree Diagrams

5PMBokQL-Decision-TreeIn this blog series, we’ll get you up to speed on using the key tools listed in A Guide to the Project Management Body of Knowledge, (PMBOK® Guide) – Fifth Edition, including Decision Tree Diagrams.

PMI wants the project manager to be able to compute decision tree diagrams manually for the PMI exam, however there are tools that automate the decision tree process rather easily. Decision trees are essentially a predictive analysis tool: they can be used for either build or buy scenarios, or for return on investment scenarios. They are generally used when a lot of factors have to be considered simultaneously in a risk analysis.

An example of a build-or-buy scenario is shown below:


For a simple decision tree with just a few factors, the average business user get up to speed in about an hour learning how to do a decision tree manually. For a decision tree with multiple decision factors and even more uncertainty factors, it is generally best to let an automated tool handle that. Learning how to use an automated decision tree will take about an hour of user practice and time.

How useful is the decision tree? If you’re doing a project with a lot of predictive analysis required, the decision tree might be a great tool for you on which to get up to speed.

This is an excerpt from the Global Knowledge white paper, Are All Those Project Management Tools Really Needed?

Related Courses
PMP® Exam Prep Boot Camp
IT Project Management

In this article

Join the Conversation


  1. Peter Chan Reply

    I don’t understand your diagram, why would you sum up the decision node and chance node value before multiplying with the chance %? Should you not first multiple the chance node value with the chance % then add up the decision node base value?

    1. Kelsey Garner Reply

      Hi Peter – here is the response from the author:

      If I understand what is being asked, you’re asking why we add the chance and decision nodes prior to performing the multiplications.

      It’s actually the opposite…

      The answer is, on the purchase software side, the costs of the software are sunk costs (The decision node). We have to consider these costs. The chance that the need is fulfilled by buying the software is 40% (The chance node). 40% of $50,000 is 20K ( The net path value). BUT you also have to consider the chance that it all goes wrong at an additional cost of $200,000. We have to consider the sunk costs of 50K PLUS the 200K of additional cost for a total of 250K. That’s 60% of 250K which is 150K. 150K + 20K = 170K. We have to add both chance nodes together to get a final “chance” (the 40% number and the 60% number)
      You do the same thing on the ‘build’ side (A decision node). In this case the sunk costs are 120K. Chances the software works is 80% (a different chance node).. 80% of 120K is 96K (A net path value). Chances the software doesn’t work is 20% at an additional cost of 30 K. The sunk costs of 120K + the 30K yields 150K * 20 % = 30K. 96K + 30K is 126K.