A well-established loyalty matrix classifies customers into four quadrants based on their frequency of purchase and attitude toward a hospitality business. These quadrants represent four distinct types of loyalty, according to combinations of patronage frequency and attitude favorability: true, latent, spurious, and low. Using a sample of casino reward club members, we applied cluster analysis to classify customers based on behavioral loyalty, defined as player gaming worth and visitation frequency; and on attitudinal loyalty, defined as emotional commitment and trust. We obtained empirical support for the existence of all four quadrants, and validated these quadrants on additional loyalty-related attributes. Truly loyal customers have high attitudinal and behavioral loyalty on all indicators, whereas customers with latent loyalty have favorable attitudes but low levels of visitation and gaming behavior, possibly due to income limitations. Spuriously loyal customers have high behavioral loyalty but very low attitudinal loyalty. These customers place a high value on reward program benefits, suggesting that such benefits are driving their behavior. Customers with low loyalty are low in both attitudes and behaviors.