Importance-performance analysis to develop product/service improvement strategies through online reviews with reliability

Online reviews are important data for developing product/service improvement strategies. Relevant studies treated different online reviews as equally important, and the validity of the results was vulnerable to unreliable online reviews. To solve this challenge, this study proposes an importance-performance analysis model that considers the reliability of online reviews. First, the reliability degree of online reviews is defined based on the quality and timeliness of online reviews and the credibility of reviewers. To estimate the importance of product/service attributes from online reviews, a preference learning model is designed based on the reliability degrees of online reviews, where the online reviews with higher reliability have a greater impact on the learning results. In addition, the attribute performance is determined by aggregating the satisfaction of online reviews for the attribute. Finally, we verify the practicability of the proposed importance-performance analysis model by a case study on four five-star hotels.