User Review Analytics – Part 1: Are Positive or Negative User Reviews More Influential?

5 star 550x101 1

How important are user reviews? Do positive or negative reviews matter more? And which reviews are more influential – rants or raves?

Web-savvy consumers are increasingly looking to ratings from fellow buyers and fellow travelers before making purchasing decisions, which gives businesses an incentive to try to influence the ratings. So we decided to put our communication analytics tools to work and analyze the persuasive language of reviews. What we found, surprisingly, is a significant difference in influence between positive and negative reviews.

We used our quantified communications tools to compare the language used in positive reviews against the language in negative reviews and uncovered some fascinating results. Knowing that most professional reviewers tend to write positive reviews, we hypothesized that there would be a difference in the language characteristics between positive and negative analyses. Using our natural language processing tools, we analyzed hotel, book, and car reviews from the Simon Fraser University Review Corpus – a collection of online user reviews, separated into positive and negative responses.

So how do you tell if a review is positive or negative? See if it is persuasive. We found the language in positive reviews to be 1.5 times as persuasive as the language in negative reviews. This is due primarily to reviewers speaking more directly to their audience of potential new buyers/customers instead of just about themselves. This language characteristic establishes a common ground between the author and the reader, which communication science tells us is essential for persuasion.

Our outcome coincides with findings from other studies regarding the influence of positive opinions. An article by Kenneth Chang from the New York Times describes a study in which artificial positive feedback on an unnamed Web site solicited more positive feedback, creating a “snowball effect” of positive responses. However, artificial negative opinions did not encourage additional negative responses. A different study from Cornell University looked at the characteristics of reviews from the top thousand customer reviewers at One of their key findings was that 88% of respondents reported that they gave either “all” or “most” positive reviews. They shied away from posting negative reviews.

Why does this matter? Businesses (particularly small businesses) often see the effects of user reviews on their bottom line. According to a study from Harvard Business School, a one-star increase on Yelp, a popular rating website, led to a 5 to 9 percent increase in revenue for businesses. Because reviews can be so influential, some companies will hire people to write reviews for their products or services. They generally ask for honest reviews, but will often encourage positive feedback through incentives. According to an article by Sandra Parker, a former professional review writer, “While they didn’t require me to write lies or tell me exactly what to write, if the review wasn’t five-star, they didn’t pay the typical $10-20 fee.”

Understanding that positive reviews tend to be persuasive and create snowball effects of more positive reviews and increased sales, the logical follow-up question is “how do you spot a fake review?” We will reveal in part 2 next week.