How to Protect Yourself from Yelp's Robot Protection

Like all things reviewable on the Internet, Yelp's full of phony reviews, even with a filtration system in place since 2005. Still, as one Yelp user has learned after losing $2,000 to a moving company, it's not always perfect. Here's how to not let the same thing happen to you.

This article is from the archive of our partner .

Like all things reviewable on the Internet, Yelp's full of phony reviews, even with a filtration system in place since 2005. Still, as one Yelp user has learned after losing $2,000 to a moving company, it's not always perfect. Here's how to not let the same thing happen to you.

Yelp is hardly the only site to fall victim to faulty online recommendations. Amazon and TripAdvisor have both had recent fake online review problems. But unlike those incidents, Yelp user Justin Vincent didn't fall prey to robot created promotions, he was duped by Yelp's own system to protect its users, which it put in place after learning that people would go to rather great lengths to game its review process. The phony reviewers are still out there.

In the moving torture story that Vincent writes at his personal site, he used the good reviews he saw on Yelp to decide which company to hire. "I did business with a moving company based on 5 star recommendations that you presented," he writes. "I was strong-armed into paying $2000 more than originally quoted. I spent 40 days without any furniture and quite a few of my belongings have been misplaced – forever." The problem, he says, wasn't that he relied on phony reviews, but rather all the bad reviews were hidden by Yelp's filter. "Your reviewers described exactly what I experienced and warned against this company again and again. But you hid all of those reviews."  He asks them why, but we can explain how to avoid a similar fate.

1. Understand How the Yelp Filter Works

As Vincent explains it, Yelp's spam killer, in trying to find the most trustworthy writers, targets those who don't write too often on the site. "Your algorithm typically hides entries by people who only post one review and who don’t otherwise engage in Yelp. Your assumption is that if a user only posts one review, posts no comments, has no friends etc. then most likely they are fake and trying to game the system," he explains. That has obvious flaws, which Vincent points out.

In each case the one star review was left by someone who would never normally leave a review… they were simply so outraged that they were motivated to signup to Yelp and try to warn others how bad this company is. None of them ever used Yelp again. Furthermore, they didn’t have the knowledge or inclination to try to make their Yelp profile look acceptable to Yelp’s automated suppression systems.

Online reviews tend to attract the outraged and the blown away. If someone has one extreme experience they might be moved to write something on the Internet about it. That doesn't mean they have or ever will review again. In fact, it seems like some of the most genuine reviews would come from these one-off musings, not someone who just reviews because they get off on it.

Yelp describes it system in cheerier terms, explaining that the review filtration algorithm depends on the Credibility of the reviewer. Credibility is exactly what consumers want. But the Yelp definition of Credibility doesn't quite match our expectations, as it relies on the level of activity from the Yelper. This little cartoon video does a good job explaining the whole thing. But basically, the more engaged a Yelp member, the more "credible" the review. It's important to note, however, the filter does not effect the overall star rating of a place. Just the detailed reviews that show up. Even so, one can imagine how a scammer could get around this, boosting star-power with quantity.

Beyond the problem we just discussed, the system actually seems set up that way to help Yelp, rather than the consumer. People who use the site more get rewarded, having their reviews mean more.

2. Read the Filtered Reviews

Rather than eliminate "malicious" reviews, the filtration system buries them at the bottom of a review page under an almost invisible grey link in parentheses.

Click that! After entering a very complicated captcha, under there you'll find reviews the robots say aren't "Credible." Some of those read as credible. While looking up an Atlantic favorite watering hole, Rivers, for example, we found some pretty accurate musings filtered out of the front page. Like this two-star review from writer T.V., which matches the experience of our many evenings spent at the establishment.

One of the stars is entirely for the location, which is fantastic, right next to the Kennedy Center.  Lukewarm crab cake, greens and corn fritter. Just one step above cafeteria food. At least it was served quickly by a nice waiter. Had it been sitting in a warming tray? Could this have been the response to prior criticism about slow service? Calling Chef Ramsey!

Though some of the filtered reviews look pretty spammy, giving it a scan will eliminate the problem Vincent encountered. He would have at least seen all the reviews warning of the scam. How will you know if a filtered review feels genuine or not? That brings us the next point.

3. Read smart

Beware of too many adverbs, adjective, pronouns, exclamation points and too many mentions of the establishment. We learned this from a group of Cornell researchers, who sifted through fake reviews. Robots or fake humans tend to write in a way people never would when describing an experience. Even looking at the positive reviews Vincent saw for his moving company, we see some of these no-nos. There's a lot of "I" and "we" going on. The truthful, negative reviews, on the other hand just read like something a human being would write when wronged.

This just "feeling it out" method isn't science. But, for Vincent, it would have worked better than Yelp's flawed science machine.

This article is from the archive of our partner The Wire.