An Analysis of Activity-Level Predictability from UK Construction Projects

Project time-predictability and performance remains low in the construction sector. Any time improvements provide opportunities to increase project efficiency, profitability and success. In project planning and programming, performance is often measured at project-level. However, this work analyses predictability accuracy at activity-level to establish trends to focus future planning efforts and make prediction improvements. To that end; an aggregation of more than 35;000 lines of activity-level programme data were collected for analysis from 27 recent UK construction projects. Analysis of the activity-level programme data was completed relative to project type; project duration; level of detail; project-level performance; and time-predictability metrics. The analysis demonstrated large disparity between activity-level predictability; with more than one-third of all predictions failing to achieve start date; finish date; or duration accuracy; and less than one-third of predictions able to accurately forecast activity start dates. Here 63% of activities were able to predict positive project outcomes; but only 24% of activities were able to achieve perfect time-predictions. Correlation between positive activity-level predictability and overall project-level time-performance is evidenced. Therefore, increased analysis of activity-level planning provides the potential to improve project-level time-performance.