Defining the Success Status of Construction Projects Based on Quantitative Performance Metrics Thresholds

Abstract

Despite the high spending and high employment, the construction industry struggles to agree how to go about project performance assessment and what constitutes a successful project. This does not allow stakeholders to address the issues in and drawbacks of their respective projects. Hence, this research outlines performance metrics upper limits, which can be used in defining a successful project. Data were collected from Construction Industry Institute (CII) members and multiple construction firms through University of Wisconsin–Madison active research projects. Two classification and regression tree (CART) models were developed using 4-fold randomized cross-validation to provide the metrics thresholds of success with a classification accuracy of 81% and 85%, respectively. CART models revealed that a successful project can be defined as having construction schedule growth less than 10.4%, construction cost growth less than 9.8%, requests for information (RFIs) per $million less than or equal to 8.6, and RFI processing time less than or equal to 7 days. Additional thresholds showed that a high-performing successful project would have a change orders per $million of less than 0.39 and rework of less than 1.5%. These thresholds can serve as a step toward a quantitative definition of success by translating qualitative success status into quantitative success status. Additionally, the established thresholds can be used as benchmarks for acceptable construction performance, which can then be adopted as risk thresholds or tolerances during project planning.