Understanding the Customer Experience with Seemingly Unrelated Regressions

• Activity Number: 41
• Type: Contributed
• Date/Time: Sunday, July 31, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and Marketing
• Author(s): Jorge Alejandro*+
• Companies: Market Probe
• Keywords: Seemingly Unrelated Regression ; Attitudinal Drivers ; Overall Customer Satisfaction ; generalized least squares ; Bayesian estimation ; multivariate regression

Abstract

Customers interact with a business in different ways via different channels or touch-points. Each one of these interactions contributes to an all-encompassing customer experience that can affect key attitudinal outcomes such as satisfaction and future referrals. A good understanding of what things "drive" this experience as well as satisfaction with each service channel (see the Multi-Attribute Attitude model) are important to managers in guiding resource allocation on product and service offerings. In order to ascertain what these drivers are, practitioners often times model channel satisfaction separately without taking into account the additional information that the correlated errors between these models can provide. In this research, the use of Seemingly Unrelated Regression (SUR) models is used to jointly estimate model parameters when applied to a case study. The results are compared to other frequently used approaches both in terms of explanatory modeling as well as efficiency of the parameter estimates.