Data design: A constitutive rules perspective

Management research and practice commonly assume that data are assets that represent facts about an organization or its environment. Yet, organizational data do not come into existence by themselves. Information systems research has addressed the creation of data mainly as a problem of conceptual modeling, while how novel data are created in practice remains undertheorized. We call this data design and articulate it as a sociotechnical process through which organizations configure data-producing arrangements that allow them to produce data points about events, entities, and people. We then propose that the impact of data design on organizational data can be analyzed in terms of constitutive rules . Constitutive rules are conceptual statements that describe how cultural conventions, institutional orders, and technical constraints embedded in a data-producing arrangement make it possible to turn an aspect of an organization or its environment into data. Using a case vignette of a gender identity data project at a large U.S. university, we illustrate how analyzing the reconfiguration of a data-producing arrangement in terms of constitutive rules helps specify the ways in which data mediate organizations and their actions. We further argue that data design can amount to innovation if it enables the production of data that are uniquely valuable to an organization.