Decoding data analytics capabilities from topic modeling on press releases

In their quest for data-driven insight, firms align their resources to produce information that is actionable. Moreover, the bundling and utilization of these valuable resources is what defines an organizational capability. Thus, in this paper we conceptualize a new type of capability - data analytics capabilities, DAC, as the ability to assemble, coordinate, mobilize, and deploy analytics-based resources with strategic purpose. Using text as data, we explore the use of probabilistic topic modeling on historical press releases, in an attempt to identify types of DAC from successful data analytics investments. Press and news releases frequently articulate a firm´s resource allocation strategy, proving an opportunity to automatically classify these into topics that can suggest categorization of DAC. We explore 8-year historical press releases and apply Latent Dirichlet Allocation topic modeling to 273 press releases.