Utilizing Predictive Analytics to Aid Project Continuity Decision Making Open Access
Downloadable ContentDownload PDF
Although corporations collect project information such as performance indicators, they tend to use this information only to estimate the cost of future projects rather than as a way of predicting potential project failures. By not using performance indicators from prior projects to predict future project success or failure, organizations are not taking advantage of information that can be used to avoid spending time, money, and resources on projects that are likely to fail. If companies were able to accurately predict project failures, they could avoid incurring the opportunity cost of carrying out projects that should be cancelled early in their life cycle. This praxis utilizes predictive analytics based on project performance indicators as inputs to identify potential project failure candidates allowing managers to stop project work and redirect project resources to other potentially successful projects. In this study we demonstrate that predicting project failure based on past performance of similar projects could enable organizations to make scientific, data-driven, and evidence-based decisions on whether a project should continue. The proposed model recommended in this praxis yields an average 98.56% prediction accuracy with 0% False Positive Rate (FPR). Data used in this praxis is from a large size organization with employees numbering in the tens of thousands. By predicting projects likely to failure after phase two based on past project performance data, this study found that the organization would save about $300,000 on average, per project.