The making of data commodities: data analytics as an embedded process

This paper studies the process by which data are generated, managed, and assembled into tradable objects we call data commodities. We link the making of such objects to the open and editable nature of digital data and to the emerging big data industry in which they are diffused items of exchange, repurposing, and aggregation. We empirically investigate the making of data commodities in the context of an innovative telecommunications operator, analyzing its efforts to produce advertising audiences by repurposing data from the network infrastructure. The analysis unpacks the processes by which data are repurposed and aggregated into novel data-based objects that acquire organizational and industry relevance through carefully maintained metrics and practices of data management and interpretation. Building from our findings, we develop a process theory that explains the transformations data undergo on their way to becoming commodities and shows how these transformations are related to organizational practices and to the editable, portable, and recontextualizable attributes of data. The theory complements the standard picture of data encountered in data science and analytics, and renews and extends the promise of a constructivist Information Systems (IS) research into the age of datafication. The results provide practitioners, regulators included, vital insights concerning data management practices that produce commodities from data.