Project monitoring by dynamic statistical control charts

Abstract

Project monitoring activities are fundamental to assure a timely identification of unacceptable project's deviations from the baseline, so that corrective actions may be taken to bring the project back in line with its objectives. Regarding this, the most used approach is the Earned Value Management (EVM) technique. However, traditional EVM metrics do not allow the Project Manager (PM) to recognize if project deviations are due to the natural project variability or to systemic and undiscovered causes that are moving the project to unacceptable out-of-controls. With this recognition, a statistical project control system based on the use of dynamic Shewhart's and CUmulative SUM (CUSUM) control charts is proposed in the present paper to deal with the project monitoring problem. The dynamic design is based on the update of charts' parameters as new data become available over time and makes charts able to be used since the initial phase of the project execution when available measurements on observed variables are still only a few. The efficaciousness and robustness of the designed statistical-based approach are demonstrated by its application to a set of real projects.