Article Preview
TopIntroduction
The rapid advancement of the Internet has increased convenience and intangibly altered consumer behavior patterns. Nowadays, more consumers are now sharing information of product or service experiences by posting on online forums, and for companies, such information can understand customers’ satisfaction level toward products or services. To investigate product or service deficiencies, online forums have become a major channel used by companies for acquiring feedback from customers who are actually using products or services. Nguyen and Coudounaris (2015) indicated that online hotel reviews not only have a great influence on consumer decision-making but also provide hotel management with valuable insights into consumer service ratings and preferences. According to the Nielsen Global Survey of Trust in Advertising Report (2011), 70% of 28,000 online survey respondents from 56 countries stated that online consumer reviews are a trustworthy source of information. Therefore, online posts have a considerable impact on the service industry. Online posts accumulate rapidly because of the exponential growth of the Internet usage population. Nonetheless, many online posts contain unstructured data; that is, text without a consistent format. The currently available software and methods are thus not useful for those wishing to automatically and systematically analyze service ratings and reviews. Hence, how to effectively extract key information from online posts and interpret that information has become the critical competitive advantage. Because of the unstructured nature and large volume of data on online forums, hotel managers must adopt effective and efficient methods to capture the essence of online customer reviews and analyze the data for detecting valuable but hidden information that enables managers to make better business decisions.
To resolve the above-mentioned problems, the present study focuses on the hotel industry and applies a hierarchical ontology-based back-propagation neural network and data envelopment analysis to develop an online customer reviews-based performance evaluation and deficiency discovery methodology for hotel service. First, a criteria framework for performance evaluation of hotel service is established using information from a hotel booking website. According to the evaluation criteria, this study proposes a benchmark-based performance evaluation model for hotel service to understand the service performance of the hotels. Next, this study develops an identification and improvement model for non-benchmark performance criteria to provide non-benchmark hotels with the required quantities of performance improvements for non-benchmark criteria. To gain insight into the causes of non-benchmark service, this study extracts keywords from customers’ online posts by using text mining. Moreover, this research aggregates these keywords and then establishes a synonym database to reduce the number of keywords substantially so as to simplify the data analysis. Through using the keyword data, this study proposes a hierarchical ontology for service deficiencies of hotels, which not only clearly shows the classes and causes of service deficiencies but also lists the keywords that characterize the causes of each service deficiency. Subsequently, the back-propagation neural network and the hierarchical ontology for service deficiencies of hotel are integrated to develop a hierarchical ontology-based automatic classification system for detecting the causes of service deficiencies.
The analysis results enable hotel managers to effectively and accurately identify and understand the core issues causing customers to rate hotel service as poor, and by understanding the root causes, concrete improvement plans can be developed to boost a hotel’s competitiveness by enhancing its service performance.