Capturing Relative Importance of Customer Satisfaction Drivers Using Bayesian Dominance Hierarchy

Abstract

Customer satisfaction (CS) research traditionally focuses on large data sets collected over long periods of time across several business units. Business unit managers or property managers have a different focus in that they need to address dissatisfaction issues on a monthly basis and on a property basis. In search for zero defects, they are often confined to small samples lacking power where they cannot draw the relative importance of each variable responsible for the making of the overall perceived quality in their customer base. We propose to use a Bayesian approach to estimate the relative importance of predictors in the presence of small samples. Based on 12 consecutive months of CS survey data collected in a hotel, we show how the hotel manager can easily prioritize his or her quality management action plan on a monthly basis. The results of our study complement the current CS research methods while managing limited resources.