Customer Complaint Analysis via Review-Based Control Charts and Dynamic Importance-Performance Analysis

Abstract

Review-based control charts that integrate sentiment analysis with traditional control charts are emerging as effective tools for analyzing customer complaints. However, existing approaches face significant challenges due to their lack of alignment with the natural characteristics of online reviews, such as limited review volumes or imbalanced mentions of attributes. To address these challenges, which are commonly encountered in real-world applications, this study proposes a novel framework for review-based complaint analysis that combines a bi-level control chart with dynamic importance-performance analysis. The proposed method converts unstructured reviews into a set of statistical control charts, analyzes how the average of and variation in attribute importance and performance change over time, and identifies whether the changes are a natural event, a temporal aberration, or a gradual trend. This allows for conducting apples-to-apples quality comparison among multiple attributes and establishing appropriate improvement strategies in real-time and proactive manners. To demonstrate the applicability and effectiveness of the method, an empirical case study using online hotel reviews is presented.