Fuzzy consideration sets: a new approach based on direct use of consumer preferences

This paper offers a new approach to building fuzzy consideration sets, a well-established practical tool for marketers. Previous studies have been based on the standard (McFadden) model of consumer choice. In practical terms their approach is a variant of logistic regression, with fuzzy consideration utilities being imputed rather than treated directly. Here we demonstrate the feasibility of a direct method, in which consideration sets are constructed from the ground upwards using consumer preferences as expressed in questionnaires. The method and its computations are fully based on fuzzy mathematics and make use for example of linguistic variables. As well as demonstrating the feasibility of our method we also justify it as a potential tool for practitioners. In addition we consider some general issues relating to 'fuzzy' techniques.