Online reviews provide additional product information to reduce uncertainty. Hence, consumers often rely on online reviews to form purchase decisions. However, an explosion of online reviews brings the problem of information overload to individuals. Identifying reviews containing valuable information from large numbers of reviews becomes increasingly important to both consumers and companies, especially for experience products, such as attractions. Several online review platforms provide a function for readers to rate a review as “helpful” when it contains valuable information. Different from consumers, companies want to detect potential valuable reviews before they are rated to avoid or promote their negative or positive influence, respectively. Using online attraction review data retrieved from TripAdvisor, we conduct a two-level empirical analysis to explore factors that affect the value of reviews. We introduc a negative binomial regression model at a review level to explore the effects of the actual reviews. Subsequently, we apply a Tobit regression model at the reviewer level to investigate the effects of reviewer characteristics inferred from properties of historical rating distribution. The empirical analysis results indicate that both text readability and reviewer characteristics affect the perceived value of reviews. These findings have direct implications for attraction managers in their improved identification of potential valuable reviews.