K. Ravikumar
GM R&D- India Science Laboratory
BANGALORE
Abstract
Demand sensing refers to sensing customers’ purchase behavior or more generally, their choice behavior. Its scope can range from estimating the price a potential customer would be willing to pay to understanding customer’s latent consideration set, the set of new products or the set of new features in products that the customer will be interested in. Choice-based trade-off analysis is the most prominent marketing research technique used to determine which features a new product should have by conjointly measuring consumers trade-offs between discrete attributes. Such models are extremely useful, for example, for describing the purchase behavior for large durable items of which, at most, one alternative is selected at a given purchase occasion and the features of the items are well defined. Consumers are exposed to different competitive choice sets in which item features are systematically varied. After viewing each set of options, the respondent selects one item or none, depending on his or her most likely purchase behavior if confronted in the marketplace with the options presented.
Most methods for analyzing choice-based trade-off data do so by aggregating data for all individuals, which could obscure important individual aspects of the data. In this talk, we review some of the recent advances in choice modeling that address this issue and highlight applications of machine learning in these models.