Online Reviews: A Current-Day Analysis of The Market – By Kaila Krayewski

These days, pretty much everyone and their cat is familiar with online reviews – the modern-day answer to word of mouth. In all likelihood, you’ve probably written one or ten yourself.

Perhaps you enjoyed excellent service when out to dinner the other night and you logged on to TripAdvisor to reward the restaurant with a positive review. Perhaps you treated yourself to a new sound system on Amazon and when you turned it on post-unboxing, the audio quality was terrible. You have to let people know, right? You’re likely on Amazon reviews already, warming up your fingers to let everyone know what a waste of money those speakers are.

Conversely, as a potential consumer of a product or service, you’ll likely spend some time reading what you think are authentic customer reviews before making a purchase. These reviews can at times tempt you into buying a product you didn’t know you wanted, or other times, sway you from paying for a service you thought you needed.

But what if we told you that not all online reviews are authentic? What if we told you that there are robots and underground operatives out there, programmed or paid (or both) to craft positive or negative reviews that are completely false, serving the needs of greedy companies willing to employ black-hat tactics to boost their profits?

Let’s back up for a second. What is an online review? An online review is a critical evaluation of a product or service published on an online platform, usually accompanied by a rating system in which the reviewer can rank said product or service according to its relative merit.

Online review platforms were created to bring the power back to the consumer by creating a democratic ranking system for customers to provide authentic feedback about a particular product or service they have bought and/or tried. They can be good, bad, neutral, or anything in between. And just like many ideas, though online reviews sound great in theory, they have such a powerful impact on businesses that a good portion of the market has become become abused and corrupted through covert tactics.

Fear not though, clever consumer. It is often possible to distinguish a fake online review from an authentic one. It’s simply a matter of educating yourself – and that is exactly what we’re here to help you do.

Fake or inauthentic online reviews

First off, not all “inauthentic” reviews are necessarily “fake”. Many people ask their friends and family to write reviews for their product or service – and these friends and relatives often have tested what they’re reviewing, and genuinely feel their review is authentic. Still, if you knew that one of the positive reviews you were reading about a book you’re considering purchasing on your Kindle was written by a close friend of the author’s, you might think twice about its authenticity, wouldn’t you? Similarly, that author may have enemies with an agenda to lower his or her book sales. And though these “enemy” reviews may be authentic – perhaps the composer genuinely thinks the book is awful – the bias remains firmly in place. Friend or foe, we would categorise these reviews in the gray area: not quite completely authentic, but not fake by definition.

So what would we categorise as a straight-up fake online review? These come in many forms, namely those written by people who are paid to write positive (or negative reviews), those written by robots, or overzealous owners (or competitors) who create fake profiles to write made-up reviews about a product or service. Another form of fake review is written by someone who received or enjoyed the product or service for free, in exchange for a positive review, something that goes against the policies of pretty much every online review platform out there.

How to distinguish a fake online review from an authentic one

Fake reviews are usually pretty easy to spot. They’re often short, with vague details in the write-up, or lack any accompanying text whatsoever. This is because fake reviews are often created simply to drive up the product or service’s overall star rating, which a simple five-star rating can accomplish quickly and easily. Look for generalities in the writing; ambiguous adjectives like “great” “awesome” and “fantastic” can usually be spotted in fake reviews, as they are easy to incorporate into text when the reviewer hasn’t actually used the product or service to be able to write anything more specific.

The more advanced online review platforms also now require the reviewer to set up a profile, through which they themselves are ranked according to how helpful their reviews have been to date. Checking the reviewer’s ranking will go a long way towards revealing if their review is genuine or not. The higher their ranking, the likelier chance the review is genuine.

If the review is paid, this can often be detected by checking out their previous reviews. If the reviewer in question wrote a bunch of reviews all in the same time period, about similar products or services with similar content, the chances of these reviews being fake is greatly heightened.

Another way to identify an authentic review, though not all sites offer this, is by checking to see if the review is accompanied by a “verified purchase”. Amazon now offers this for all of its products, though of course, this would only apply to products purchased through its e-commerce platform.

Robotic reviews

Just a few months ago, it was revealed by researchers at the University of Chicago that scientists have developed software intelligent enough to write extremely believable online reviews. The robots are trained with thousands of real online reviews, through a deep learning technique. The resultant robotic reviews are extremely difficult to detect, since they are written character-by-character, rather than using outdated methods of swapping out words from a text template here and there, which can be easily detected by plagiarism software. The paper goes on to describe how this form of AI could completely disrupt the covert paid reviews industry, though the researchers do admit they have yet to see examples of this happening already in the real world. Still, the threat looms. Industrial-scale fake reviews pose a huge threat to online review sites. Luckily, there are some tactics that can help.

How online review platforms flag fake reviews

Given how powerful online reviews can be at either making or breaking a product or service, it’s little wonder that many of the top online review platforms have a stealthy moderation board that reviews all of its, well, reviews.

TripAdvisor’s substantive moderation team employs AI – in addition to their dedicated audience – to screen their reviews, which are reviewed by a person when flagged as potentially inauthentic. According to its Review Process FAQ on its website, TripAdvisor wisely won’t reveal its tactics but, “We can tell you that we dedicate significant time and resources ensuring that the content on TripAdvisor reflects the real experiences of real travelers. We have quality assurance specialists who have brought a wide range of professional experience to enhance our prevention methods and our team spends thousands of hours every year ensuring the integrity of content on TripAdvisor.”

In 2015, Amazon sued 1,114 people for writing fake reviews on its site. The company revealed that many reviews were traced back to ads offering to create fake customer reviews on budget freelancer hiring sites like Fiverr.com. The defendants’ profiles openly revealed that they would write a five-star review for a product or service.

Yelp’s spokesperson, Rachel Youngblade, gave the following telling statement about the company’s own internal review tactics: “Yelp’s recommendation software…uses many signals beyond text-content alone to determine whether to recommend a review, and can not-recommend reviews that may be from a real person, but lack value or contain bias.”

When it comes to AI-generated fake reviews, however, online review platforms will be forced to fight fire with fire. Subtle differences in AI-generated reviews are not visible to the human eye, but can be flagged with the right software. For example, the distribution of characters is different when a review is written by a robot, than it is when written by a human.

How to respond if your company receives a negative online review

It’s never enjoyable to read a negative review about your business, particularly as the owner or founder of that business. It can be tempting to get defensive about an especially angry review for your product that you feel is unjust, but take a step back and evaluate how your response on behalf of the business will come across to others reading the review and reply.

Let’s rewind for a second to the moment you read a negative online review about your product or service. The first thing to do is check it for the ways listed above for differentiating a fake review from a real one. Scepticism is your best friend here: was the review perhaps written by your competition? Or by someone who has a vendetta against you? Read between the lines, and if you determine that it could be a fake review (genuinely, not just because you want it to be), flag it. Too many false flags, however, could decrease your ranking on the online review platform, so be careful with how often you do this and for what reasons.

If you determine the review to be genuine, it’s best to respond in a professional manner. Leave out any abusive language, even if the reviewer did not do you the same courtesy, and assure them that their feedback is important to your team and you are reviewing the matter to ensure it doesn’t happen again. Then, request a chance at redemption – this could be inviting them back to your driving range, or offering to exchange their faulty product. This will make the reviewer feel important, and it will also show your potential customers reading the negative review that your company places high value on customer service, which could sway them back in your favour.

Join the discussion