•  
  •  
 

Keywords

Database marketing; Data mining; Decision tree analysis; Classification; Casino; Predictive analytics; Cross-selling

Document Type

Original Research Article

Abstract

Cross-selling different types of games could provide an opportunity for casino operators to generate additional time and money spent on gaming from existing patrons. One way to identify the patrons who are likely to cross-play is mining individual players’ gaming data using predictive analytics. Hence, this study aims to predict casino patrons’ propensity to play both slots and table games, also known as cross-gaming, by applying a data-mining algorithm to patrons’ gaming data. The Exhaustive Chi-squared Automatic Interaction Detector (E-CHAID) method was employed to predict cross-gaming propensity. The E-CHAID models based on the gaming-related behavioral data produced actionable model accuracy rates for classifying cross-gamers and non-cross gamers along with the cross-gaming propensity scores for each patron. Using these scores, casino managers can accurately identify likely cross-gamers and develop a more targeted approach to market to them. Furthermore, the results of this study would enable casino managers to estimate incremental gaming revenues through cross-gaming. This, in turn, will assist them in spending marketing dollars more efficiently while maximizing gaming revenues.