This study explores the patterns of online reviews of vacation homes from a community-based travel advisory website with a goal of understanding the biases inherent in online word of mouth (WOM) related to tourism and hospitality services. An analysis of nearly 3,200 reviews from “Reviewsite.com,” (a pseudonym) which posts reviews of vacation rental properties across the USA, finds an overwhelming preponderance of favorable reviews. More to the point, relatively few “moderate” reviews are posted, and the second-highest category is extremely negative comments. Using semantic processing techniques on the aggregate review text, the study identifies the nuanced opinions and concerns of the travelers who write reviews. Negative reviews tend to be lengthy and argumentative, often detailing disappointment over expectations not met. Positive reviews, on the other hand, tend to be relatively brief and confirm the overall rating. Consumers who wrote “high” reviews placed greater importance on value for money, cleanliness, and comfort than did those who wrote negative reviews. Those who wrote “low” reviews placed their emphasis on the service provided by the property staff and management. Negative reviews were more likely to involve a higher price accommodation. This analysis indicates that the overall numerical ratings typically used in review systems may not be the ideal indicator of perceived service quality. The results suggest that review sites should develop better methods to aggregate, synthesize, and publish the review contents, particularly the numerical ratings. This and other review sites show the average of all the point-scale ratings, but such simple means do not take into account the biases that are inherent in the rating systems. Instead, the sites should provide more information and heuristics to help the consumers navigate through the clutter and get the information they desire.