Key Driver Analysis with Very Small Samples: Using the Analytic Hierarchy Process Method

• Activity Number: 137
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and Marketing
• Author(s): Shon Magnan*+ and Stan Lipovetsky
• Companies: GfK Custom Research North America and GfK Custom Research North America
• Keywords: AHP ; Key driver analysis ; customer satisfaction ; marketing research

Abstract

In customer satisfaction, one of the key research methods is a derived importance or key driver analysis. With it, organizations determine how to set priorities for increasing customer loyalty. The most common statistical approaches to key driver analysis are correlation, regression and structural equation modeling. All of these methods, however, can be flawed or even impossible when one needs to derive importance for a very small group. One approach to solving this problem is to use analytic hierarchy process (AHP). This presentation discusses AHP's application to solving this problem and compares several different statistical models for turning local preferences into global priorities including both classic methods and emerging techniques such as multi-response multiplicative mode, three-dimensional eigenproblem, orthogonal regression, nonlinear local evaluation, linear and nonlin