Data Analytics and Firm Productivity: The Importance of Technological and Organizational Architecture

Research has shown that data analytics contributes substantially to business value creation and firm productivity. Data-rich firms increasingly treat data as a valuable resource, adopting technical solutions and acquiring human capital to develop predictive analytics capabilities using large amounts of data. However, the degree to which these strategic and technical investments effectively generate productivity gains varies widely across firms. Such benefits are even harder to recoup among incumbent corporations in traditional industries. We argue that the organizational design of data architecture is a crucial differentiating factor in the productivity impact of data analytics across firms. We combine archival data at the firm level with a detailed technical survey that measures data architecture around a comprehensive set of technical components and organization-wide practices. Our empirical results show that two data architecture design principles--modularity and integration--are particularly important in complementing data workers to produce useful co-invention--predictive analytics embedded across a variety of business processes and software applications. Hiring data workers is much more productive when the firm adopts a modular and integrated data architecture design. However, architectural inertia prevents firms from redesigning their data systems, and it causes data architecture and data workers to be persistently mismatched.