Peak cubes in service operations: Bringing multidimensionality into decision support systems

Abstract

Companies like Ritz Carlton, Disney and Verizon are among many who have invested in analytics to improve their customers' service experiences with the firms. Extensive data are collected on all aspects of how customers interact or experience the products or services. Research has shown the importance of the “peak-end” rule in service design; that is, providing a customer with good “peak” service levels and “ending” the service experience with high quality can enhance customer satisfaction and build loyalty. However, previous studies have examined this phenomenon only in contexts with unidimensional service levels. We introduce peak cubes, which enable service designers and scholars to pinpoint prominent service levels in multidimensional service experience profiles—thereby extending current research on behavioral economics and service design to more general settings. Results indicate the potential of multidimensional peak-end models to better predict customer satisfaction in various service scenarios. Using Shapley values in coalitional game theory, the resulting models can also inform service designers about the quality dimensions that are critical from the perspective of multidimensional peak-end heuristic and customer satisfaction. Our research contributions and proposed methodology will enhance decision support systems with multidimensional capabilities and have applications to fields as diverse as service operations and healthcare.