Linking Modularity with Problem Solving and Coordination Efforts

We explore the relationship between process architecture decisions, problem solving and coordination efforts during distributed product development, i.e., when development work is dispersed across organizational boundaries. Process architecture refers to the manner in which the tasks and interfaces within a project are organized so that a project can be executed in an efficient manner. The theoretical foundation for our investigation draws upon the information processing (IP) view of product development and transaction cost economics. The IP view calls for manipulation of the architecture by making problem-solving tasks modular, such that reducing task interdependencies minimizes the development time. The transaction cost view calls for transactional efficiency, for example, reduction in the coordination burden for each productive task. We present several measures of modularity based on a design structure matrix (DSM) mapping of the development task structure. Then, we analyze the relation between these measures and transactional efficacy and task completion times. Our empirical evidence comes from 71 tasks carried out in 11 software development projects. Data analyses show that the improvement in the transactional efficacy is associated with increase in task modularity. Our results establish that increasing modularity reduces the development time and the coordination effort. However, these results also suggest trade-offs: modularity measures do not contribute equally to this gain in efficiency and increasing modularity is negatively associated with the technical problem-solving efforts. Copyright (C) 2008 John Wiley & Sons, Ltd.