Alternative approaches to effort prediction of ERP projects

There exist many effort prediction systems but none speci®cally devised for enterprise resource planning (ERP) projects, and the empirical evidence is neither convincing nor adequate from a human user perspective. Consequently, this non-empirical evaluation contributes knowledge by investigating: (i) their applicability to ERP projects, (ii) their added value to a human user beyond making a prediction, and (iii) if they make sense. The analysis suggests that regression analysis seems to be the best choice as an ERP prediction system, and that ANGEL, ACE, CART and OSR primarily add value to a user in exploratory data analysis by their ability to identify similar projects. q 2001 Elsevier Science B.V. All rights reserved.