Prediction of Project Performance Based on Degree of Interaction

Abstract

This paper demonstrates a direct relationship between designer/builder interaction and project performance, based on a large database. It also presents a predictive model to forecast project performance based on the planned level of interaction. Degree of interaction (DOI) is defined as the extent of interaction among designers, builders, and project team members during all project phases. As DOI scores rise, project performance quickly improves and becomes more consistent. Regression analysis of 38 recently completed projects shows a modest but significant correlation between DOI and project performance, including user satisfaction. Beyond a certain DOI score, performance tends to level off. Achieving this score takes only a modest increase in interaction. Threshold analysis separates the projects into those with DOI scores above and below this score. Performance is significantly better for projects with higher DOI in all areas except design deficiencies. Finally, these results are combined with statistical analysis to predict the performance of future projects based on their DOI scores. The probability of improved average performance on future projects with higher DOI scores is very strong. DOI scores can be used to predict ranges for expected cost growth, schedule growth, number of modifications, design deficiencies, and user satisfaction.