Willemsen et al’s article discusses the judgments people make about products being sold online and how such judgments are influenced according to the nature of the reviews for that product. Essentially, negative content for products of experience has a stronger impact on choice, and positive content for search products is more influential. Looking at argumentation, the density and diversity of the kinds of arguments surrounding a product are important for a consumer. If there is a star rating and also descriptions, it’s more likely that a person will make better judgments. Lastly, the expertise claims didn’t show a very strong correlation to the perceived usefulness of reviews.
I enjoyed reading the article because it made me think about what I look for in reviews when I shop online, and I think the authors’ findings are consistent with my own experience. For example, when I moved here, I needed a bike to get to campus, so I looked at different department store websites like Target and Walmart for something cheap but with good reviews. I found myself taking more seriously the negative reviews because I didn’t want a piece of junk, even though I wanted it cheap (*irony*). The more reviews I read, the more leery I noticed I was becoming of positive reviews. Is it possible for online reviews to be fake, a deliberate attempt by the company to entice the consumer by providing a positive review? How could I discern between the authentic reviews and the fake ones?
A study presented at conference for computational linguistics provides the answers to these questions. In this study, Hancock et al develop a system that detects “deception opinion spam,” testing the veracity of reviews from hotels and other companies. An example of a deception opinion spam is based on yelp.com: “My husband and I stayed at the James Chicago Hotel for our anniversary. This place is fantastic! We knew as soon as we arrived we made the right choice! The rooms are BEAUTIFUL and the staff very attentive and wonderful!! The area of the hotel is great, since I love to shop I couldn’t ask for more!! We will definatly be back to Chicago and we will for sure be back to the James Chicago.”
Some of their findings show that truthful reviews commonly speak about “spatial configurations (e.g. bathroom, on location)” as opposed to the liars. The liars tend to speak more about “aspects external to the hotel being reviewed (e.g., husband, business, vacation). Deception also includes less usage of first person singular, and reviews tend to be more positive than negative.
Here’s a link to the article: http://aclweb.org/anthology/P/P11/P11-1032.pdf
and how it is discussed in the New York Times: http://www.nytimes.com/2011/08/20/technology/finding-fake-reviews-online.html
I’m thinking that fake reviewers could learn a lot from both articles. I feel weird about making this conclusion because it’s almost as if I’m encouraging deception in reviews. But I’m glad to be better informed about what to look out for when I shop online.
Its a double edge sword that by figuring out key indicators in fake reviews we also tip off the sales people to what they should avoid. I wonder if websites could require the user to enter a receipt number in order to review a product on sites like Target and Amazon. That might cut down on the fake reviews.
The receipt idea sounds good, but I’m not sure that companies would go along with it. As we have read, it is the companies themselves that have been writting fake reviews. Also, what would stop them from generating fake receipt numbers and posting reviews of their own products?
The idea is that large retailers like Amazon and Target that sell a variety of brands would use this system to create more customer loyalty, by protecting customers from cloaked sales pitches. I don’t know how you could stop companies from coming up with fake receipt numbers, but it can’t be that hard to verify them since you need that for all basic returns now.
If companies (and the people they hire to write fake reviews) start reading academic research in order to write better reviews, I think I might take that as a win! But more to the point, given how vague these kinds of fake reviews are they are unlikely to be of much use to people, so I’m not sure what companies think they are gaining.