The more we researched the quality of services in Myrtle Beach the more we were painfully aware our findings did not always match online review averages. Averages between review sites were so different we often had to confirm business addresses were the same. We found some glaring inconsistencies in reviews posted on Google, Facebook, TripAdvisor and Yelp. While one looked better than another for restaurants, hotels or other services were completely skewed from other review sites. None of this should be a surprise it has been a very well reported issue. Review tampering is a well known and major problem, as even the smallest skew of fakes numbers can foul the entire review system. The result is poor performances go unpunished while stunning performances to go unrewarded. More reviews do not make better quality, only accountability can make reviews better.
A Stanford article moves
Towards a General Rule for Identifying Deceptive Opinion Spam states that deceptive reviews, "tend to exaggerate sentiment" more than what is needed.
While the Stanford article and many others like it are useful guides to help us design automate a filter system for some of the fake reviews. These articles all admit certain flaws in their algorithms are possible.
Big threats to automated review filters
Professional reviewers who offer, "buy good reviews", stay current with changing routines as they continually retool for more natural review designs.
Exaggerations based on emotion or inexperience
Exaggerations based on emotion are most often negative reviews. One of the filtering challenges can occur when a customer feels the incident is a direct personal attack or they feel deeply wronged. If they are then greeted by a stack of recent positive 5 star reviews that describe exactly the opposite of their own experience, frustration may become helplessness or vengeful. Their attitude changes the reviews tone enough to make the review appear fake. We found this practice is used often to intentionally discredit any real negative reviews that draw attention to existing issues well known by owners, employees and supporters. This practice backfires on supporters who go unpunished and freely post fake positive reviews. The source is obvious and tracking it back to the business can also be quite simple. Getting it prosecuted requires resources that rarely exist.
Who writes the fake and deceptive reviews?
Positive fake reviews
- written by owners, employees, contractors, people soliciting favor or friends on the companies behalf with or without the businesses knowledge.
Negative fake reviews
- written by competitors, disgruntled employees, contractors and people soliciting favor of the competition.
- fake reviews disguised as middle of the road positive, they are actually slow burn reviews that are usually small one liners, used in mass by competitors drag down at a completion' good scores with out drawing attention in hopes of leveling the playing field.
How do you spot a deceptive reviews or reviewer?
- reviews may have two or more of the following identifiers.
5 star reviews that directly oppose low scores of the same item, service or time frame
The five or one star reviews using marketing speak
Multiples - reviews that use basically the same words, subject and often the same scores
Multiple perfect review experiences that are glowing but use generalizations
Hyper reviews lots of all caps or excessive punctuation, "THEY ARE SO BAD!!" or "THEY ARE GOOD!!"
Hyper long explanations of products, much better than other near identical products for less price
Exaggerated family or nostalgic references, "Just like my grandmother made! Sure it is pricey, but worth every penny, due to..."
Only a few reviews exist but all are overwhelmingly positive
The reviewer doesn't actually say anything about the product being reviewed
The reviewer includes website links to their own site
search of reviews phrasing returns multiple reviews under different products or reviewer names
A reviewer directly responds to problems from other reviews by saying something like, "Company X wouldn't do that" or "Company X cares". Normally this would be addressed as we did not share your problem or offer a solution of how they got around the problem, even more likely they would point out exactly what other people are doing wrong.
Some reviewers use standard protocols like: create 10 reviews, 6-8 are the same review category, most often 5 stars, the two or more remaining reviews are inert average reviews designed to fool filters, i.e. a local park, large corporations like Walmart or Taco Bell. A insider most often one of the last 5 star reviews is the target. A professional can use up all 5 star reviews to cover paid clients. They often let the profile cool a month or more before they start the process again. A stack of mature profiles provide an almost endless stock of paid reviews.
The FTC has begun cracking down on these practices, it may take some time before a crack down reaches Myrtle Beach but it will.