Investigation into the Review Selling Industry

Review selling is no longer a hidden service offered online by a select few bad actors. It’s an entire industry built on the demand for fake reviews and a mix of people looking to make easy money on the side and those running entire businesses from it.

This industry has grown exponentially in the last few years as the demand from businesses wanting to build or boost their reputation online keeps growing. Businesses know the value reviews bring with 93% of consumers say they affect their purchasing decision. In 2023, Google blocked or removed more than 170 million reviews violating policy - a figure that was up by 45% on the previous year. They also removed or blocked more than 12 million fake business profiles in the same year. Since Google is the most popular source for consumers looking to find business reviews (81%), their reviews really need to be authentic and genuine.

It’s important to make a clear distinction between reputable Online Reputation Management businesses that legitimately help businesses monitor and address feedback and reviews online, and review sellers who get paid by disreputable businesses to write fabricated reviews.

The challenges presented by the review selling industry are now so great that review platforms have started taking decisive measures to combat the volume of fake reviews coming onto their platforms, including the formation of the Coalition for Trusted Reviews. The aim of this group of 6 members (Amazon, Tripadvisor, Glassdoor, Trustpilot, Expedia and Booking.com) is to ultimately maintain trust and authenticity in online reviews.

‍A rapidly expanding industry

The size of this industry is rapidly expanding due to the need for businesses to build or improve their image, reputation and brand. We investigated multiple sites and forums commonly used by review sellers to offer their services and through this we have been able to determine that there is at least one new review seller every week offering their service. 

Some of these newcomers are individuals who initially requested a review service and realised how lucrative and simple it is. Deciding to give it a go themselves without caring about the implications of their actions, they then use the forums to learn more about the different techniques and best methods to create these fake reviews.

Initially, they will offer free review samples in exchange for feedback on the sites and forums to help build up a customer base and reputation. 

As the review selling industry grows, so too does competition between review sellers. They're all competing with each other on price and service for the same business. Naturally, the lower the price, the more attractive it is to a business, flooding platforms with even more fake reviews.

Differing review quality

From our investigation, it’s clear to see there can be a considerable difference in the quality of service provided by review sellers, marked by price differences and the longevity of a fake review - they refer to this as review ‘stickiness’ - how long a review ‘sticks’ on a platform before being detected. 

Many review sellers offer their review service to cover multiple platforms whilst others focus on specific platforms. 

Review sellers offering their service for multiple platforms usually pick the top platforms shown when people search for reviews via Google’s search engine. This then gives their customers the option of buying reviews to cover most popular review platforms and when people search the reviews in Google for the specific company, they will mostly see positive (fake) reviews.

To make it easier to help identify the different types of review seller and their level of expertise, we categorised the review sellers into 3 different tiers as follows:

Tier 1 - Highly sophisticated techniques and detection evasion methods used, positive reputation for reviews sticking - high cost.

Tier 2 - Somewhat sophisticated but lacking in consistency of the techniques/methods used compared with tier 1, moderately positive reputation for review stickiness - cost slightly more than the basic option.

Tier 3 - Low/non-existent review seller reputation, technique/methods used are very poor - low cost review service.

Review sellers can use several different methods to make their reviews stick, and go undetected, which is reflected in the price of the service offered. The more complex the techniques used, the more expensive the cost.

It’s not all positive

We often think primarily of fake positive reviews when talking about fabricated reviews. However, review sellers also offer negative reviews, as well as negative review removal - something that can be extremely challenging for businesses on certain platforms. 

There are multiple reasons a review seller would be asked to post negative reviews. Below are a few examples.

• Negative reviews are usually requested by businesses to damage their competitors. Review sellers will post low star reviews to bring a business rating down. 

• Negative reviews could also occur when a payment hasn’t been made by the customer and as retaliation the review seller will post negative reviews to bring the customer’s rating down.

• Another emerging trend is blackmailed negative reviews. These are negative reviews created by a review seller for the purpose of blackmailing the customer into paying for the review service. Failure to do so means the review seller will continue to post them.

Costs can vary from $1 for a standard review to $10,000 for a single business listing removal. Review platforms have different levels of detection measures in place; some are harder to evade than others. This is reflected in the price of reviews being offered by sellers claiming it's more challenging to make fake reviews stick for such platforms, so they charge a higher fee.

In an attempt for their fake reviews to go undetected by platforms, review sellers not only learn about each of the platforms and how they operate but also test the systems of each platform to identify what works and what doesn’t. Forums are a popular place for individuals to share best practice when creating reviews and what methods/tools to use when doing so. However, the most experienced and savvy review sellers will avoid disclosing their own techniques.

How they operate

A lot of vendors advertise their services on those forums/sites which are used by many of the users (review sellers included) for different reasons, most of which are not good. Some of the services that have been advertised on the forums/sites are things such as proxies (various types), antidetect browsers, accounts and much more. 

A proxy server is commonly used for anonymity. It protects an individual when browsing the internet by hiding their location.

There are various types of proxies that can be obtained online. The two most common ones are residential and mobile proxies:

• Residential proxies are created from residential home IP addresses.
• Mobile proxies are created from mobile IP addresses such as EE or Vodafone. Unlike residential proxies, mobile proxies have a wider pool of IP addresses to choose from.

The cost of obtaining proxies depends on the type of proxies purchased. Mobile proxies are more expensive than other types of proxies but are very effective and very popular for social media use. They enable users to get around IP related restrictions so whilst other types of proxies may be blacklisted, there is a lower chance of mobile proxies also being blacklisted. However, currently the go-to choice for most review sellers are residential proxies as they are cheaper and of good enough quality, so will still allow the user to hide their tracks even though the success rate will not be as high as 3/4/5G mobile proxy.

Added layers of deception

As the demand for fake reviews is increasing, review sellers need to also have the ability to set up multiple accounts to be able to post their fake reviews. To meet this need, review sellers source account sellers.

Accounts of all kinds are available for purchase online - email accounts such as Google, Yahoo or Microsoft, or social media accounts for TikTok, Facebook, Instagram or Discord. All can be found online at a surprisingly small price, starting from as little as $0.02 per account. 

Providers of these accounts offer a wide range of options for each account, such as new accounts (warmed up - the practice of building up the activity on the account to increase the reputation of the email account and to make it look active), and aged accounts. The age of the account and its associated history are regarded as meaning there’s a lower risk of being blocked/suspended. To a platform, these accounts will look more authentic with genuine previous activity.

Anti-detection browsers are also used by review sellers to spoof their digital fingerprints whilst giving them the ability to create and run multiple accounts. This has been making it really challenging for platforms to identify since using fingerprint detection won’t be as effective in that scenario and so platforms will need to use a wide range of techniques to identify bad actors. 

Review seller offers

Review sellers offer sticky reviews and/or a non drop guarantee. As mentioned previously, sticky reviews are reviews made to stick on the platform and therefore go undetected. Review sellers offering non drop guarantees mean they’re uncertain whether their reviews will stick and so instead offer a 30 to 90 day replacement guarantee if the reviews get dropped by the platform. Review platforms need to be wary that when a fake review is taken down, to expect another to come in for the same business. Some may also offer a mixture of sticky reviews and non drop guarantee reviews due to the uncertainty of the security updates of a platform.

‍Reduced manual effort with Generative AI 

As review sellers always look to automate the manual process of setting up accounts and creating reviews, generative AI tools have become part and parcel of their work. They use these tools for various parts of the fake review/account creation process, including sourcing profile pictures where traditionally they would have used random pictures found online, and also for writing the actual review content to reduce the manual effort required.

Uptick in requests to make reviews stick

A recently published blog post by Google mentioned 170 million fake reviews were blocked by Google using AI.

Since then, forums and sites where review sellers operate have seen an increase in users (including digital marketing companies) asking for help to make their Google reviews stick. Google’s detection methods have made it even harder for their reviews to gain traction. Interestingly, this has not really put review sellers off listing fake reviews on Google since it’s  still the number 1 platform and the demand is now even higher. However, with the increased difficulty in getting around Google’s AI algorithm, review sellers are starting to increase their prices specifically for Google reviews. 

With even more trust and safety platforms now utilising AI to combat fraud, review sellers will find it progressively more challenging to bypass the AI algorithms. Nevertheless, as long as the demand for fake reviews is there, review sellers will continue to thrive. 

Evolving techniques require advanced detection methods

We can see this from the evolution of their processes too. Techniques used are becoming more sophisticated and review sellers are getting smarter in their approach and, therefore, even harder to detect. It is down to platforms working tirelessly to combat review fraud in applying the best technology solution possible to detect these fraudsters and safeguard their platforms. 

As review sellers’ practices look to hide their deception, going way beyond fake content creation, advanced detection capabilities need to be at the forefront of identifying and removing  these fraudulent reviews. A proven and effective way of targeting this challenge is to combine AI and behavioral analytics to uncover suspicious patterns of review activity - looking beyond the content to understand the intent and detect the emerging threat at scale. 

This is Pasabi’s approach to detecting fake reviews. We look at behavioural signals, reputational data and patterns of activity to identify review fraud. The graph above shows clusters of suspicious and connected activity to reveal review selling behavior. This graphical representation enables platforms to see connections in their data and highlights the worst offenders.

If you would like to learn more about our fake review detection and how we could help you combat review seller activity, get in touch with a member of our team or book a demo.