Problem-oriented CBR: Finding potential problems from lead user communities

Abstract

Online communities, where lead users openly share their experiences and knowledge on product and technology in the form of a post, have become a fruitful source of innovation. While efforts have been made to provide a way of data-driven case-based reasoning (CBR), existing studies have limitations in reflecting lead users’ characteristics into the CBR process. Current research has emphasized the retrieval and adaptation phase only, which retrieves, reuses, and revises cases. However, what is at the core of lead user characteristics is to find out the problem, and solve the problem by themselves before mass customers. This means CBR needs to focus on uncovering and defining problems and finding relevant solutions for designated problems. In response, this research suggests a novel approach for problem-oriented CBR approach to reflect lead user characteristics. First, this study defines problems, by extracting problem–solution sets related to the specific function using sentiment analysis. Second, this study improves case representation and case retrieval using subject-action-object (SAO) analysis and technology tree respectively. This study demonstrates the approach through a case of drone technology using lead user communities (diydrones.com), and our findings suggest that the approach can help firms broaden the knowledge of existing products to make an improvement.