A customer-based approach for selecting attributes and levels for preference measurement and new product development

This is the authors' abstract. We don't add key points for this paper.

Preference measurement is commonly used to identify customer needs and to create products that satisfy these needs. Over the past few decades, most researchers have focused on developing new survey or estimation methods. However, researchers were paying little attention to the selection of attributes and levels for preference measurement. This study provides an overview of the common methods currently in use to define attribute sets for conjoint analysis, and it extends previous research by proposing an evoked set-based attribute selection method (EVAS). The EVAS is intended to aid the selection of appropriate attributes and levels by focusing on products that are acceptable from the consumer's perspective. An empirical study of this approach reveals that the applicability of attribute sets increases, and implausible estimates can be avoided.