Metrics that matter: Improving project controls and analytics in construction industry

Due to increasing complexity and inefficiencies in project control system design and implementation, construction projects struggle to achieve initial cost and schedule performance goals. Assessment of project progress and performance is critically important to the successful delivery of capital facility projects. Often, project managers are misled in their perceptions of project performance until the project nears its end. Major challenges are related to the lack of consistent, reliable, and objective metrics and indicators. This research identified several core predictive and diagnostic metrics that can help provide actionable insights into a project's actual progress, performance, and forecast at completion. It also provides information on ways to improve the reliability of these metrics. The research methodology included a broad literature review to identify progress, performance assessment, and forecasting metrics. Next, a survey was distributed to collect data on metrics and reliability concepts that were used on completed projects. In total, 44 surveys were completed, representing mostly large, industrial projects. This was followed up by a Delphi session including 16 subject matter experts with more than 360 years of experience in total, who evaluated and validated the findings from the survey and case studies. The Delphi session further refined a list of metrics and determined 20 core (“must have”), 7 validation (metrics that confirm the validity of the core metrics), 7 innovative (metrics that are not currently in wide use, but are considered potentially beneficial), and 14 other significant (other metrics that fall outside the previous categories, but are perceived to have value). A metric typology and framework defined predictive and diagnostic metrics with the purposes of achieving consistent project control procedures across the industry. Details of various predictive and diagnostic metrics are visualized using metric maps and a network. Network analysis revealed interrelationships among metrics. Statistical analysis of survey responses with Spearman's rank correlation revealed that compared to projects using fewer core metrics, projects that used more core metrics for project controls experienced higher rates of success at meeting their original budgets. A correlation between the use of more core metrics and better project cost outcomes was observed at the 95% confidence level using the Spearman's rank correlation method. At the same confidence level, utilizing more diagnostic metrics was shown to be correlated with better schedule and cost outcomes as well. Further statistical assessment using Multiple Correspondence Analysis demonstrated that usage of certain metrics are more closely associated with better cost and schedule outcomes. Core metrics were initially selected based on the following project characteristics: large, industrial, reimbursable cost, balanced cost and schedule goals, moderate complexity, and contractor perspective. However, when considering core metrics for other project characteristics, it was discovered that the core metrics will be the same - the only differences relate to the frequency of data collection and level of effort involved in collecting and analyzing these data. Additionally, factors for improving metric reliability in several areas such as project scope definition, execution planning, and risk management were also included. Ten projects were selected for case studies, which provided more in-depth analysis on metrics and reliability issues. 15 critical reliability factors and 85 indicators were identified for improving the reliability of project control metrics. An expert panel verified these findings and added phase specific timing details for application of the factors and indicators. Based on the findings of this research, a Project Controls Improvement (PCI) Tool was created to provide a standardized and systematic tool for project controls. Using the PCI Tool, project stakeholders can identify the gaps in their project control systems a d learn more about core metrics and steps they can take to improve metric reliability within a dynamic and interactive software environment.