Predictive segmentation in action - using CHAID to segment loyalty card holders

Laura Galguera
University of Oviedo

David Luna
Baruch College

M. Paz Méndez
University of Oviedo


To build competitive advantage both in the short term and in the longer term, companies need to segment the market. Market segmentation can aid in better achieving both strategic and tactical objectives. For strategic purposes, market segmentation results can be used to prioritise market segment opportunities. Kotler (1989) has called this the 'four Ps' of strategic marketing (which are additional to the four Ps of tactical marketing): probing, partitioning, prioritising and positioning. For tactical purposes, market segmentation results can be used to further optimise specific product/service offerings, and can help in implementing communication and advertising plans. As one might expect, recent research has explored different conceptual and methodological aspects of segmentation (Wedel & Kamakura 1998), including the use of geodemographic variables (González-Benito & González-Benito 2005) and the use of benefit segmentation (Ahmad 2003). Other recent articles describe specific cases of segmentation in action (e.g. Hindmarch et al. 2005). Therefore, there seems to be a strong interest in segmentation both among academics and practitioners. Our paper seeks to illustrate the use of CHAID (Chi-squared Automatic Interaction Detection), a post hoc predictive segmentation tool (Wedel & Kamakura 1998), in the loyalty card market.