TY - CHAP
T1 - A Text Mining Approach to Assessing Company Ratings via User-Generated and Company-Generated Content
T2 - An Abstract
AU - Krey, Nina
AU - Wu, Shuang
AU - Hsiao, Shih Hui
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Retailing and third-party websites serve as online platforms for customers to gather information and interact with other users as well as companies. User-generated content is available through different online platforms, with online reviews being one of the most common type of user-generated content. While previous research has primarily focused on influential characteristics of reviews or on perceived usefulness or helpfulness of reviews from the user perceptive, the current research expands the literature on company-customer online relationship building by investigating the influence of user reviews and manager responses on the overall rating of the corresponding company. User reviews is an important format of user-generated content, while manager responses is a strategically used form of company-generated content. This research implements a text mining approach and sentiment analysis to assess how emotional versus rational user reviews and manager responses impact overall company ratings. Specifically, Linguistic Inquiry and Word Count (LIWC) dictionary is used for the language processing tasks, including sentence segmentation, word tokenization, and lemmatization. Data collection implements a Python-based web crawler to gather a consistent panel of user-level activity from TripAdvisor.com. The final data includes a three-year period from 2016 to 2018 of hotel reviews featuring user information, text reviews, and ratings given by the reviewers. Manager responses and corresponding hotel ratings complete the data set. Current findings provide further insights into how companies can utilize public manager responses as a business strategy to increase online ratings of their firm. Managerial implications include that companies can improve online ratings if managers incorporate emotional responses with long sentences. Future studies should incorporate additional cities beyond the US to enhance generalizability of findings. Furthermore, this research examines online rating data for one service category, namely hotels. Additional industries, service categories, and reviewer characteristics can enhance contributions of future research endeavors. Incorporating additional service categories would allow the robustness assessment of the negative influence of rational online content on company ratings. Lastly, the current time lag variable remains limited due to the three-year period. Assessing how repeated interaction between reviewer and manager influences company ratings over time would provide additional guidance for firms in addressing online content and maintaining positive online reputations.
AB - Retailing and third-party websites serve as online platforms for customers to gather information and interact with other users as well as companies. User-generated content is available through different online platforms, with online reviews being one of the most common type of user-generated content. While previous research has primarily focused on influential characteristics of reviews or on perceived usefulness or helpfulness of reviews from the user perceptive, the current research expands the literature on company-customer online relationship building by investigating the influence of user reviews and manager responses on the overall rating of the corresponding company. User reviews is an important format of user-generated content, while manager responses is a strategically used form of company-generated content. This research implements a text mining approach and sentiment analysis to assess how emotional versus rational user reviews and manager responses impact overall company ratings. Specifically, Linguistic Inquiry and Word Count (LIWC) dictionary is used for the language processing tasks, including sentence segmentation, word tokenization, and lemmatization. Data collection implements a Python-based web crawler to gather a consistent panel of user-level activity from TripAdvisor.com. The final data includes a three-year period from 2016 to 2018 of hotel reviews featuring user information, text reviews, and ratings given by the reviewers. Manager responses and corresponding hotel ratings complete the data set. Current findings provide further insights into how companies can utilize public manager responses as a business strategy to increase online ratings of their firm. Managerial implications include that companies can improve online ratings if managers incorporate emotional responses with long sentences. Future studies should incorporate additional cities beyond the US to enhance generalizability of findings. Furthermore, this research examines online rating data for one service category, namely hotels. Additional industries, service categories, and reviewer characteristics can enhance contributions of future research endeavors. Incorporating additional service categories would allow the robustness assessment of the negative influence of rational online content on company ratings. Lastly, the current time lag variable remains limited due to the three-year period. Assessing how repeated interaction between reviewer and manager influences company ratings over time would provide additional guidance for firms in addressing online content and maintaining positive online reputations.
UR - http://www.scopus.com/inward/record.url?scp=85127944283&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85127944283&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-89883-0_152
DO - 10.1007/978-3-030-89883-0_152
M3 - Chapter
AN - SCOPUS:85127944283
T3 - Developments in Marketing Science: Proceedings of the Academy of Marketing Science
SP - 567
EP - 568
BT - Developments in Marketing Science
PB - Springer Nature
ER -