LIMITACIONES DEL ANÁLISIS DE VALOR GANADO AL INTENTAR PREDECIR RETRASOS EN PROYECTOS

Abstract

A significant amount of construction projects miss their originally planned completion dates, being this a frequent topic in the construction management literature. Many causes for late projects have been identified, but among these, hardly any research effort has been made when it comes to analyzing how the Earned Value Management (EVM) technique handles the project duration variability. Particularly, both activity durations and cost might be highly variable in real project contexts, but the most common implementation of the EVM technique neglects this variability. By doing so, project duration estimates result generally being too optimistic. However, EVM has other shortcomings when it comes to assessing whether the actual order of activity execution will result in higher or lower overall project durations. This paper presents a brief and analytical critique of the EVM technique concerning its most common implementation (assumption of deterministic activity costs and durations). By means of simulating how activities in parallel are executed it will be seen how the ‘merge event bias’ can inadvertently and negatively impact the construction schedule. A fictitious case study will be proposed and some of its components analysed to identify all the shortcomings previously highlighted.