Abstract:

It is virtually a fundamental law of nature that people have idiosyncratic likes and dislikes. They may cluster into groups or segments of similar-minded individuals, but it is often difficult to determine what causes segments to exist. In a typical consumer product test, respondents are usually screened and profiled according to a full battery of demographic, psychographic and product usage attributes. Then an analysis of variance is used to study responses to products by identified groups to determine if there is a group product interaction. In the case of a demographic group such as gender, the interaction reveals whether the products were rated differently by males and females. The mere identification of an interaction does not reveal why the interaction occurs or how to design products that are optimal for each subgroup. In this report we discuss how to take the next step towards understanding and using interactions by fitting a model that reveals the location of individual ideal points for demographic groups in a map that identifies the attributes important to liking