The gaming industry is the largest entertainment industry in the United States, with more than $80 billion in revenue annually. Because of the stochasticity of gambling outcomes and the complexity of the casino context, forecasting individual-level revenues in a casino setting is extremely challenging, and yet crucial for customer relationship management. Current approaches for customer base analysis are usually too general to handle the unique context of the casino setting. To fill this gap between research and practice, this paper develops a stochastic model that incorporates visitation, wagering, and gambling outcomes to forecast gamers' revenues for a major casino operator. The proposed model is parsimonious and can be scaled to handle massive casino customer databases. Despite its parsimony, a holdout prediction test shows that the proposed model provides more accurate individual-level revenue predictions than other forecasting methods that are based only on the observed data. Copyright © 2012 John Wiley & Sons, Ltd.