LAUSR

Abstract The purposes of the paper are (1) to examine the dynamic properties of online reviews, focusing on whether previous review trends contribute to herding or reactant behavior in subsequent review rating generation (dynamic flow), and (2) to explore the business value of management responses in the dynamic flow of online reviews based on Social Impact Theory and Rational Action Theory as the foundation. To this end, we analyze a series of regression and logistic models with quasi-experimental cases from a large online review dataset, collected from a leading online travel website in China. We find that both types of previous trends of reviews, positive and negative, contribute to reactant behavior in subsequent review rating generation. When the review trends are considered with management responses, we find that management responses have a positive impact on subsequent review ratings in the negative review trend, but not in the positive review trend.