Estimating Segment - Oriented Discrete - Choice Model of Customer Differentiation (Buyers Preferences of Saipa Products as a Case Study)

Subject Areas : Jounal of Marketing Management

M. A. Abdolvand ¹ , J. Nazemi ² , H. Ali Momeni ³

1 - ، دانشگاه آزاد اسلامی، واحد علوم و تحقیقات، تهران، ایران
2 - دانشگاه آزاد اسلامی، واحد علوم و تحقیقات، تهران، ایران
3 - دانشگاه آزاد اسلامی، واحد کرج، کرج، ایران

Received: 2014-03-08 Accepted : 2014-03-08 Published : 2014-02-20

Keywords: Discrete Choice, segmentation, principal component analysis, Multinomial logit,

Abstract :

In product choice models, incorporating socio-demographic characteristics is conceptually appealing and has numerous managerial benefits. The main question is how socio-demographic characteristics differentiate customers or products in order to define relevant marketing mix strategies or what a disaggregated analysis reveals about the customer's profile of company products. Traditionally econometrician and marketers apply discrete choice models into a total population or a sample of consumers (naming as consolidated approach) in order to understand, describe, and predict consumer choices for a particular product and assess socio-demographic variation on the utility functions and subsequently analyse the difference between segments behaviour in order to identify differences between target consumers. The Multinomial Logit model developed based on data obtained from new car buyers of SAIPA automotive company and final model including all significant variables and their estimated parameters on utility functions of all products presented. In this study, a segment-oriented approach is introduced and compared with the consolidated approach. The proposed methodology using principal component analysis to segment the population (or sample) into different groups so as the results indicate principal component analysis is a robust method in order to reach to differentiated segments among population as expected and finally the result is compared with consolidated approach which indicates the segmented approach outperforms the consolidated approach in terms of choice predictability and R-square index of the model. The results bring a new insight to car manufacturers as well as marketing planners concerning policy setting about customer characterizing and targeting that affects marketing mix strategies.