You’ve Data Mined. Now What?

Abstract

Data-mining technologies are within the grasp of many organizations. Commercially available data-mining packages make it relatively easy for firms to transform their data resources into predictive models. Yet, despite technological advances, the precise manner in which data-mining output should be incorporated into an organization’s decision-making processes remains unclear. This paper attempts to clarify the role of data mining by situating it within the context of Simon’s model of decision making. We use a complex decision problem from the video game development industry to illustrate several practical challenges managers face when using data-mining output as a decision making input. We then show how some of these challenges can be overcome by incorporating data-mined predictive models into a conventional decision-analytic formulation of the problem.