Objective:Extensive evidence in the customer satisfaction literature indicates that the relationship between the performance of quality attributes and customer satisfaction is neither linear nor symmetric. This asymmetry arises from inherent differences in the nature of quality attributes as well as heterogeneity in customers’ perceptions, and levels of theirs expectation. Despite the development of various analytical approaches, many studies rely primarily on aggregate-level quantitative analyses or qualitative interpretations, paying limited attention to individual-level differences and their behavioral implications. Moreover, although the Multicriteria Satisfaction Analysis (MUSA) model is widely recognized as a robust quantitative framework based on ordinal data and offers important analytical and managerial advantages, it implicitly assumes symmetry between attribute performance and satisfaction and exhibits limited sensitivity to customer heterogeneity. Accordingly, this study aims to extend the analytical logic of the MUSA model and enhance its capability to analyze asymmetric customer satisfaction by focusing on individual customers and differences in their levels of expectation.

Method:First, the analytical logic of the MUSA model was revised to enable the analysis of customers’ ordinal judgments at the individual level, leading to the extraction of an index representing each customer’s level of expectation. Based on this index, customers were classified into three homogeneous groups: low-, moderate-, and high-expectation customers. Subsequently, using the overall and partial satisfaction value functions derived from the model, a penalty–reward(PRCA) analysis was applied to assess the asymmetric effects of increases and decreases in the performance of quality attributes on customer satisfaction and dissatisfaction. This approach allows for the distinction between positive and negative performance effects and facilitates the examination of the behavioral roles of attributes across different expectation levels. Finally, the satisfaction analysis was conducted separately for each customer group to identify differences in behavioral patterns among heterogeneous segments.

Findings:The results show that the PRCA analysis based on the outputs of the MUSA model enables an empirical classification of quality attributes within the Kano framework and reveals meaningful differences in the roles of attributes in creating satisfaction, preventing dissatisfaction, and prioritizing improvement actions. In addition, the magnitude and nature of the effects of quality attributes on customer satisfaction are significantly dependent on customers’ levels of expectation, such that the same attribute may assume different behavioral roles across customer groups.

Conclusion: the findings suggest that the proposed approach, while preserving the core advantages of the MUSA model, provides a coherent and data-driven framework for analyzing asymmetric customer satisfaction by extending its interpretive capacity to the individual level and integrating the logic of the Kano model with PRCA analysis. This framework supports a more accurate understanding of customer satisfaction behavior, enables the identification of improvement priorities aligned with customers’ levels of expectation, and facilitates the extraction of meaningful decision-oriented insights for managers.