The Complex Effects of Shared Domain Knowledge on IS Development: A Simulation-Based Theory Development

Abstract

Information systems development (ISD) requires the integration of both application domain knowledge and technology knowledge. However, knowledge specialization complicates the IS analysis and design process. The business unit, who is ultimately responsible for making business design choices for the system, holds ample functional domain knowledge but possesses less technology knowledge, while the IS unit, who is responsible for making technical design choices, has deep technology knowledge but less functional domain knowledge. Furthermore, ISD projects are characterized by interdependencies in design choices not only within but also across business and technical domains. Prior research tends to suggest a uniform positive effect of shared domain knowledge between business and IS units on ISD performance. Although a few studies have argued that the effect may not be uniformly positive, little research has explored the complex effect of shared domain knowledge. Since organizations invest substantial resources to develop shared domain knowledge, it is crucial to uncover the conditions under which it affects ISD performance differently. We model the systems analysis and design process as an iterative design problem solving process by which the ISD team, composed of business and IS units with heterogeneous and overlapping knowledge bases, seeks to discover a system design that produces the greatest business value. Using simulations based on the NK fitness landscapes model, we find that shared domain knowledge does not affect ISD performance when ISD complexity is very low or when all interdependencies of design choices exist within the same unit, that it increases robustness of ISD performance across different distributions of design choice interdependencies, and that ISD performance is higher when shared domain knowledge is unevenly distributed between business and IS units or when its distribution matches that of the design choice interdependencies. Based on these findings and our reasoning, we posit eight propositions that advance the theory of shared domain knowledge.