During the development of software products, there are some factors impacting over the process. Innovation level of the product, identification level of the requirements, competence level of project manager, customer participation, and preliminary preparations are identified as important factors influencing the product development process. This study exposes the impacts of these destructive factors over the project performance utilizing CHAID (Chi-squared Automatic Interaction Detection) technique. The data set contains seventy-five pieces of project information obtained from the finished software intensive projects. The performance of product development process is measured with respect to time deviation rates of the project. The proposed models help researchers and practitioners to understand the probabilities of failure in software product development projects. The first model estimates project time deviation rate and its frequency of occurrence according to the identification level of requirements. The level of customer participation becomes important if product requirements are complete and accurate in the beginning of the project. The second model taking the competence level of the project manager as a reference shows that there are usually no significant delays for the groups possessing a perfect management experience. The preliminary preparations become important if the projects are not managed by professionals.