Abstract:
Analyzing customer feedback, accessible on the Internet via social media platforms and tourism-related travel websites, empowers tourism service providers to pinpoint areas of success and concern. A notable approach to identifying factors determining customer satisfaction is the Kano method, which is nowadays typically applied to big data contexts to detect quality features through the analysis of online customer feedback. The present study introduces a novel fine-grained approach for classifying quality factors based on topic areas identified through unsupervised learning techniques. Specifically, a keyword clustering-based topic detection and lexicon-based sentiment analysis is followed by a regression analysis, to identify factors influencing customer satisfaction, which is finally validated using ANOVA. Technically, the proposed approach aims to discern the positive and negative impacts that various topic areas, identified through unsupervised learning in online feedback data, may have on overall customer satisfaction. Findings show that automatically identified topic areas within customer feedback can be meaningfully categorized as Must-Be, One-Dimensional or Attractive Qualities. More specifically, while most identified topic areas exhibit One-Dimensional Quality characteristics, influencing overall customer satisfaction positively when fulfilled and negatively when not, nuanced variations emerge across different attributes. Furthermore, the regression models revealed significant influences between attribute performance and overall customer satisfaction, underscoring the statistical reliability of the models applied. To summarize, our study contributes to theory by providing a refined method for detecting Kano factors on a more fine-grained level and by identifying statistically significant and practically reliable Kano factors that asymmetrically influence overall customer satisfaction.
