Understanding geographical markets of online firms using spatial models of customer choice

As the online channel matures, many firms are finding that an understanding of how their online customers' preferences and choices vary across geographical markets can be very useful. In this paper, we propose a spatial multinomial model of customer choice and illustrate how spatial modeling of choices of online customers across geographical markets provides useful insights in the context of a product mix and pricing decision of an online book publisher. The spatial multinomial model specifically accounts for the spatial correlations among customer choices among different product forms – print and PDF. The estimation results obtained using data generated from an online experiment show that the spatial model accounts for the geographical variation in many of the unobserved effects possibly due to locational differences and price sensitivities. The resultant spatial decision maps provide useful predictions as to how purchase rates vary across geographical markets as a function of the price differential between product forms, with implications for targeting customers through local market advertising, direct marketing, and cross-channel promotion.