Exploring the Impact of Negative Words Used in Online Feedback in Hotel Industry: A Sentiment Analysis, N-gram, and Text Network Analysis Approach

This study examines the words and situations that trigger and those that do not trigger a hotel response when customers post negative online feedback. The research explores through sentiment analysis, bigrams, trigrams, and word networking the valence of online reviews of five important hotels in Las Vegas. Only the feedback that has been categorized as negative by the algorithm is selected. In correspondence to this feedback, the existence of answers from the hotels is checked together with the response style. While the negative valence of the feedback can represent a mixture of subjective and objective emotions, there are common features in their expression. On the responses side from the hotel, not all the reviews receive attention. As such, the negative feedback words are extracted and separated into those that belong to reviews that obtain a response and those that do not. The responses are standard by following an established pattern. This paper aims to contribute to a prominent issue in tourism that is little tackled: responses to feedback. The findings may help the hotel’s management to explore different paths in improving their services and their responses alike. Behavioral marketing researchers might want to use these results to confirm the existence of such patterns in different datasets or in different situations.