Modeling of fuzzy-based voice of customer for business decision analytics

Identification, interpretation and response to customer requirements are the key success factors for companies, regardless of their industry. Failing to satisfy customer requirements can damage a company's reputation and cause heavy losses. In this study, we have developed a new approach for properly interpreting and analyzing the fuzzy voice of the customer using association rule learning and text mining. This unique methodology converts textual and qualitative data into a common quantitative format which is then used to develop a mapped Integrated Customer Satisfaction Index (ICSI). ICSI is a framework for measuring customer satisfaction. Previous measures of customer satisfaction ratio failed to incorporate the cost implications of resolving customer complaints/issues and the fuzzy impact of those complaints/issues on the system. In addition to including these important and unique factors in the present study, we have also introduced a dynamic Critical to Quality (CTQ) concept, a novel method that provides a real-time system to monitor the CTQ list through an updated CTQ library. Finally, a procedure for customer feedback mining and sentiment analysis is proposed that handles typographical errors, which are unavoidable in every real database. The results of this study suggest that incorporating the fuzzy level of negativity and positivity of comments into the model instead of treating negative and positive comments as binary variables, leads to more reasonable outcomes. In addition, this study provides a more structured framework for understanding customer requirements.