Analyzing the Effects of Product Architecture on Technical Communication in Product Development Organizations
Author: Sosa R., Manuel E.
Supervisor: Eppinger, Steven D.
Institution: Department of Mechanical Engineering, Massachusetts Institute of Technology
Effective communication in product development organizations has been identified as a key factor of product development performance. Furthermore, understanding how the development organization manages the knowledge associated with the product architecture is broadly recognized as a critical challenge for established firms facing architectural innovation. This thesis presents a research method and statistical analyses intended to enhance understanding of the coupling between the product architecture and the development organization. The research method is summarized by three steps: 1) capture the product architecture by documenting design interfaces, 2) capture the development organization by documenting team interactions, and 3) couple the product architecture with the development organization by comparing design interfaces with team interactions. Our approach is illustrated by analyzing the development of a large commercial aircraft engine. Several hypotheses are formulated to explain the mismatch between design interfaces and team interactions, that is, the cases when: 1) known design interfaces are not matched by teaminteractions, and 2) reported team interactions are not predicted by design interfaces. Effects due to organizational and system boundaries, design interface strength, design interface type, design interface redesign, indirect team interactions, and secondary design interfaces are studied. In addition, through the analysis of the distribution of cross-boundary design interfaces, modular and integrative systems are formally identified, and differences between designing modular versus integrative systems are studied. Two types of statistical analyses were performed. First, categorical data analysis techniques were used to test the mentioned hypothesized effects. Second, a log-linear model, built upon social network analysis methods, was developed to study the association between design interfaces and team interactions controlling for effects of reciprocity, differential attraction, and differential expansiveness of both components and teams. Findings in this research are complemented with the results of another empirical study focused on the effects of distance and communication media use in geographically distributed development organizations. By considering the results presented in this thesis, development organizations can improve the integration process for complex designs. The approach developed is particularly applicable to projects where the architecture of the product is well understood and the development team is organized around the product architecture.