Can You Tell a Fake Web Review from a Real One? The web is full of filters and tools for us to streamline our daily correspondence and information gathering.
If you need to know the registered source of a phone number, you use anywho.com to look it up. If you want to make sure the content you’re citing is original, you use Copyscape to scan the web for copies. But how do you find out if a review you’re reading is real or not? Unfortunately there isn’t a way to tell if the review is legitimate or if there’s astroturfing involved.
Fake reviews are a real problem on the World Wide Web. They diminish the web’s power to provide the public with self-determined truths. Luckily, we’re close to creating computer models to better determine real reviews from the fakes.
But until then…can you tell a real review from a fake one?
The Ruse Review: An Art
To see the fake review demonstrated to the extreme degree, we must first go to China. There, the government employs individuals for pennies-per-post to outweigh any complaint against government service made online. These people scour civil sites where someone has posted a complaint on public space.
But when it comes to private businesses, writing a fake review remains an obvious ploy. If you own a restaurant, why wouldn’t you spruce up your online reputation with some flowery high praise coming from a made-up customer?
The problem of course is that such actions are detrimental to the entire point of the online environment. The idea that we can relay real information among each other regarding third party goods and services means the hold that advertising had on society should be slipping away. Yet, through unreal reviews many establishments are luring clientele by tricking them into thinking they’re reading about an authentic experience of someone else, which is in essence false advertisement.
It’s rampant even on sites like Yelp and Urbanspoon, where countless online visitors go every day to get seemingly honest takes on neighborhood businesses. As of today, these sites have not figured out a way to root out the false reviews themselves.
Catching a Fake Reviewer: A Science
Fortunately for these sites and the people who trust them, Ivy League thinking has come to the rescue. According to NPR, Cornell University researchers have recently devised an algorithm they can use to determine whether or not a customer review is real or not.
Researcher Jeff Hancock and his coworkers created fake reviews and accumulated real ones. They then had their computer analyze each. Next, they pitted the computer against random samples of either, and it was 90 percent successful at determining a review’s authenticity. When humans were asked to deduce which were real and which were fake, those who participated in the study performed much worse than the robot.
It’s obvious that fake reviews work for humans, but how does the computer tell the difference?
Apparently, it’s all in the pronouns and their popularity within the text, especially relative to the details. Fake reviews tend to feature many instances wherein general words are used and proper nouns are avoided. It sounds obvious until you’re actually reading different reviews trying to tell the fake ones apart from real ones. The human brain tends to skip over these words, which is why fake reviews are more likely to appear real to us.
Fake Review Finding DIY
Pay attention to whether or not the author includes descriptions relative to space and time. For instance, a restaurant reviewer who raves about “the terrific view from the patio” should be flagged versus someone who remarks about how “neat the southern Seattle skyline looks from the patio.”
Industriously-minded review-manufacturers might be outsourcing fake review writing. This works in your advantage in that oftentimes these reviews are poorly written grammatically and have spelling errors.
But the bottom line is to simply always be aware of the language itself. Sometimes, an online review just doesn’t feel right, and a determined re-read will oust the suspicious sections.
Conclusion – It’s a shame that we should have to worry about online reviews not being real. But then again, the incentive behind fake-review writing is as obvious as the reasons behind not wanting to base your purchase decisions on it. Hopefully, the new Cornell algorithms can be implemented across the web’s many review forums. Until then, keep an eye out for unspecific language and unrealistic mentions. Reviews are worthless without the truth behind them.