Path Dependence and Organizational Choices in Product Architecture Evolution

Extant research highlights the significance of earlier design in shaping product architectural evolution and the tendency of product designs to become overly complex over time. However, the underlying mechanisms of such path dependence, as well as the impact of different organizational choices remain largely under-investigated. This study of an industrial software architectural evolution analyzes how past design characteristics affect the complexity of future product design, examining also how software development work organization choices influence such evolution. Our findings demonstrate that product architectures characterized by design modularity and hierarchy are less likely to evolve towards unnecessary complexity. Moreover, we show that, in order to contain future design complexity, team work organization is more effective than individual task assignment, especially for cyclical (non-hierarchical) design elements. We test our hypotheses on a unique dataset, which includes micro-level information about the architectural properties of 13 versions of an industrial software. Our study contributes to management of innovation literature, highlighting how path dependence and organizational choices jointly shape product architecture evolution, as well as how designers might prevent that products and services evolve overly complex.