Predicting the expected service level and the realistic lead time of rfi process using binary logistic regression

This research describes an investigation into the development of predictive Request-For-Information (RFI) performance models, particularly to aid contractors and owners in predicting the expected service level of RFIs and establishing realistic review times. Two quantitative models have been developed and are discussed. Four input variables were developed and employed to construct binary regression logistic models. Results show large discrepancies on average between the contractor-want-time and actual lead time and between the expected CWT and realistic lead time to achieve a particular desired service level. Contractors can use the model to predict the service level for more reliable planning. Owners in turn can use the model to verify the expected RFI process time and to gain a valuable internal benchmark for process improvement. With higher awareness of the possible outcomes of a project and how likely each is to occur, project teams can better determine which options amongst those available are likely to yield the most favourable results.