Canceling ideas and projects is an important part of the Innovation Portfolio Management (IPM) process as stopping the unsuccessful ones avoids sunk costs and sets free resources for successful ideas and projects. In this article, we analyzed within IPM the cancellation of ideas and projects by gatekeeping boards as well as the possibilities of applying machine learning. Using the data from a large European Telecom organization, four different Machine Learning models were created to support managers making decisions in all phases of a project. All models have an Area Under the Curve (AUC) of at least 0.8, making them into possibly valuable instruments for predicting project cancellation.
Click here for the paper