Online reviews analysis in product defects and customer requirements via two-stage model

Online reviews offer a valuable source of information for identifying product defects and understanding customer requirements. However, previous research often focused solely on explicit quality issues within a single platform, while overlooking the distinct characteristics of various platforms. In practice, each platform features unique communication styles, and content types, leading to diverse but valuable insights for manufacturers. This study proposes a novel online review-driven modelling framework that utilizes multi-platform with data integration and fully leverages rich user-generated content (UGC) to capture the customer insights. The methodology consists of two stages: first, the analysis of after-sales product complaints to identify specific product defects and construct a Failure Modes and Effects Analysis (FMEA) database through heterogeneous information fusion; and second, the development of a strength-frequency Kano model to classify customer requirements extracted from online forums, with the goal of optimizing customer satisfaction. Case studies on new energy vehicles validate the effectiveness of the proposed method and offer valuable business insights for stakeholders.