Estimating brand choice probabilities for air conditioners using logit and probit models

In recent years, improvements in information technology have resulted in the availability of customer transaction data. This trend is closely linked to an ever growing desire on the part of the marketing manager to use the firms as much as possible about his customer database. Now many experts accept an evolution from transaction oriented to customer centric marketing strategies. The data concerned, reveals preference data such as sales and brand choice. The large amount of accurately measured marketing research data implies that simple graphical tool and elementary modeling techniques in most cases simply do not suffice for dealing with present day problems in marketing. In many cases more advanced techniques involve quantitative models, which enable the marketing researcher to examine various correlations between marketing response variables and explanatory variables.

The present study aims at applying probability models for solving problem of consumer brand preferences for an Electronic product, viz.; Air Conditioners. In order to predict the consumer choice probabilities of a particular brand of an Air Conditioner, we have considered three leading brands: Samsung, LG and Onida operating in the Indian markets of Air Conditioners. Initially, Multiple Regression Analysis is carried out for identifying the major factors affecting the Sales of Air Conditioners of each brand and each variant. Next the probabilities of the choices of the brands of Air-Conditioners are estimated using the Logit and Probit Models. The brand choice behavior for Air Conditioners is revealed and the Leading Brand of Air Conditioners is identified through this probabilistic analysis.