Skip to content

The Opportunity

A large national construction company was experiencing a loss of tens of millions of dollars of profit annually in some of their projects.  Their objective was to use data and analytics to:

  1. Identify the projects most at risk for profit erosion, not only as the project progressed, but also prior to deciding on terms of the bid;
  2. Quantify risk factors to better understand them and focus on the most important, actionable issues.

The company had previously created a simple risk index and established seven Critical Success Factors thought to be associated with project profitability.

Our Approach

We compiled data on over one thousand projects to model profit margin percent as a function of the seven Critical Success Factors as well as other variables.  These variables included factors such as contract type, project characteristics, open defects, business unit, project schedule status, safety issues, economic factors, and weather.

We used statistical tools such as logistic regression and tree-based methods.  These methods have the advantage not only of making predictions, but at the same time providing explanations of the drivers of lost profit, the second objective of the work.

The Impact

The final model delivered these insights and business advantages:

  • It determined that four of their original seven Critical Success Factors were predictive in a statistically significant sense.
  • It identified eleven other factors as contributors to profitably, both positively and negatively.  Not all were within their control, but many were.
  • On a subset of recent projects, the model estimated that the top 10% of most risky projects were responsible for $13.2MM of profit erosion.  This was double the amount that their prior simple risk index was able to uncover.
Back To Top