The problem is no longer limited to primitive bots and repetitive comments. Real people, old accounts, verified profiles, hijacked domains, paid contributor programs, and entire networks of pseudo-publications are used for promotion.

As a result, a seemingly convincing digital reputation increasingly reflects not a company’s quality, but the size and professionalism of its marketing budget.

Even the largest platforms are unable to filter everything

Google uses automated systems, artificial intelligence, and manual moderation. Nevertheless, in 2024 alone, the company blocked or removed more than 240 million reviews that violated its policies, over 12 million fake business profiles, and restricted posting for more than 900,000 accounts. According to Google, most of these reviews were removed before users had a chance to see them.

Product Hunt prohibits bots, vote buying, voting groups, mass requests for upvotes, and any form of incentivized voting. The platform uses automated anomaly detection and manual moderation, and any attempt to artificially influence ratings may result in the removal of a post or account restrictions.

Trustpilot reported that in 2024, it removed approximately 4.5 million fake reviews, accounting for 7.4% of all reviews submitted that year. About 90% of the detected fake reviews were removed automatically. Of the total number of fake reviews removed, approximately 3.4 million were five-star reviews.

These figures illustrate the scale of the problem but do not reveal how many well-crafted manipulations went undetected. The platforms’ statistics reflect only what was detected.

Platforms have no universal way to determine whether the author actually used the service if the review is posted from a regular or long-standing account that is technically indistinguishable from a genuine one.

A verified account does not necessarily mean a verified customer experience.

How Much Does a Fake Reputation Cost?

At the time this article was written, there were publicly available services offering Product Hunt votes for approximately $1.80–$2 per upvote.

One seller openly listed a price of $2 per vote, promising upvotes from “real accounts.” Another offered votes for $1.80, claiming that the tasks were performed manually by real Product Hunt users and that the votes were spread out over time to avoid sudden spikes. These are the sellers’ own claims, not independently verified guarantees.

On the Trustpilot review marketplace, the following prices were publicly listed:

  • about $7 per review;
  • approximately $10–$12 per individual review;
  • $6.50–$8 per review for large orders;
  • about $12–$15 per review, which the seller refers to as “verified.”

The term “verified” in such offers is a marketing claim by the seller and does not imply official verification by Trustpilot.

Thus, a few hundred dollars can generate dozens of positive posts or significantly alter the visual perception of a rating. Moreover, the service provider can offer not just empty profiles, but accounts with a history, other reviews, photos, and activity.

What Our Research Revealed

During our research, we discovered dozens of “specialists” in reputation management and SEO who offered hundreds of options for building a digital reputation.

Among the methods offered:

  1. Placing sponsored articles, reviews, and testimonials disguised as independent publications.
  2. Using old or hijacked domains belonging to defunct companies from various countries.
  3. Filling such domains with fabricated content and turning them into pseudo-publications.
  4. Publishing commercial content without any indication Sponsored, Advertisement or information about the partnership.
  5. Posting reviews through existing and verified accounts on LinkedIn, Facebook, Google, and Trustpilot.
  6. Generating comments, reactions, and discussions around a sponsored post.
  7. Gradually spreading activity across time, countries, devices, and accounts.
  8. Adding neutral and negative reviews to create a more natural picture.

It’s not just new bots that are used. Accounts may have a history spanning several years, including photos, posts, connections, and normal social activity. Some service providers recruit real people and assign them tasks through private groups or micro-task platforms.

As a result, based on the appearance of an article or review, it is often impossible to reliably determine whether the post was paid for.

How False Credibility Is Created

One of the most effective methods is not directly fabricating a review, but rather borrowing the reputation of a well-known brand.

The company posts logos of major media outlets on its website and writes:

  • “As seen in…”;
  • “Featured in…”;

The average visitor perceives this as confirmation that journalists from a well-known publication independently researched the company and decided to write about it. In reality, the nature of the publication may be completely different.

Example: Forbes Councils

One of the most common examples is Forbes Councils content.

Forbes explicitly labels such content as “Council Post,” noting that it is created by Forbes Councils members, that the program operates under license, that the opinions expressed belong to the authors, and that membership is paid. This is not the same as independent editorial content from Forbes.

The Forbes Councils program itself cites the opportunity to publish author articles on Forbes.com as one of the benefits of paid membership. The program description specifically emphasizes that such publications help increase a member’s visibility, credibility, and professional reputation. Members also receive special badges that can be displayed on websites, in signatures, and on social media.

It is not always technically correct to label every Council Post as a regular advertisement or sponsored article: it is a distinct format for paid membership and author-written publications. However, it also does not constitute independent journalistic endorsement of the company.

The marketing ploy lies in the fact that, as the content is further promoted, the context often disappears. Instead of the phrasing:

“The company’s CEO published an author column through the Forbes Councils paid membership program,”

the user is shown:

“The company is featured in Forbes.”

Many readers overlook the small disclaimer Council Post, information about the paid membership, and the disclaimer often go unnoticed by many readers. At the same time, the Forbes logo creates significant credibility.

The following can be used in a similar way:

  • author columns;
  • guest posts;
  • partner content;
  • contributor platforms;
  • press releases;
  • paid rankings and awards;
  • articles automatically distributed across dozens of websites;
  • badges for members of professional councils;
  • interviews arranged by the subject of the article themselves.

The mere fact that material is posted on a well-known media outlet’s domain does not necessarily mean that the editorial staff conducted an independent investigation, vetted the company, or recommends it to its readers.

You can even buy the appearance of encyclopedic significance

The creation of mentions and pages on Wikipedia has become a separate market segment.

According to offers documented during our research, intermediaries quoted the following prices:

  • starting at $2,000 if the company already has publications in independent, authoritative sources;
  • approximately $8,500–$17,000 if the necessary information base must first be created through paid articles, followed by a 7–10-month process to secure publication and pass verification by Wikipedia contributors.

These are not official rates set by Wikipedia or the Wikimedia Foundation, but rather commercial offers from third-party intermediaries. They do not guarantee that the page will be accepted, retained, or not deleted at a later date.

The Wikimedia Foundation does not sell articles. The rules require disclosure of paid editing, including the employer, client, and other related parties. Concealed paid work violates the platform’s requirements, and promotional editing may additionally conflict with the rules of a specific Wikipedia language edition.

However, to the average reader, a professionally produced Wikipedia page may appear to be independent confirmation of a company’s notability.

This creates a closed system:

  1. Articles are purchased from pseudo-publications or contributor programs.
  2. These articles are used as “independent sources.”
  3. A Wikipedia page is created based on them.
  4. A link to Wikipedia is used to boost customer trust.
  5. The page itself then becomes a basis for new publications and ratings.

The review market has turned into a market for selling reputation

Modern reputation management agencies don’t just sell a single review; they sell an entire information ecosystem:

  • positive reviews;
  • articles and reviews;
  • “client” profiles;
  • social media discussions;
  • posts on old domains;
  • links from pseudo-media outlets;
  • ranking positions;
  • mentions in contributor programs;
  • suppressing negative search results;
  • creating the appearance of fame and public recognition.

At the same time, the major platforms themselves monetize visibility management tools.

The sale of such tools does not in itself prove the sale of fake reviews or more lenient moderation. However, it creates an obvious conflict of interest: the platform simultaneously acts as a measure of trust and sells companies ways to use that trust more effectively in marketing.

During our investigation, after registering on some popular platforms, account managers contacted our researchers. In private conversations and phone calls, they discussed opportunities to improve a company’s standing relative to competitors, manage visibility, and ensure “more favorable handling of disputed situations.”

These incidents relate to specific communications and do not, in and of themselves, prove the existence of an official platform-wide policy.

But such proposals highlight a fundamental problem in the market: those who sell trust simultaneously derive commercial benefit from the companies that take advantage of that trust.

Why the usual signs of fraud no longer work

A domain’s age does not prove its reliability. An old domain can be purchased, acquired after its registration expires, or filled with entirely new content.

A visually appealing website also proves nothing. A template, a team of authors, legal pages, and a multi-year archive can all be created or imported in a short period of time.

Professional responses to negative reviews are not proof of good faith. They can be prepared by an experienced SMM team.

Ranking trends can also be replicated through reputation management campaigns. Service providers can:

  • publish reviews gradually;
  • alternate between five, four, and three stars;
  • include minor criticism;
  • use different writing styles;
  • leave pauses;
  • spread posts across different countries;
  • create a history for the accounts.

Identical phrasing, new profiles, and a sudden surge in five-star ratings remain red flags. However, the absence of these signs does not confirm the authenticity of the reviews.

A good fake is specifically designed not to look like a fake.

Is it possible to reliably determine that a post was purchased?

Based on appearance alone—most often, no.

Without access to payment records, contracts, correspondence, the platform’s internal systems, or acknowledgments from participants, it is impossible to determine with certainty whether a specific article, review, opinion piece, or Wikipedia page was paid for.

It is possible to detect suspicious connections, identical texts, networks of websites, shared analytical identifiers, associated owners, or duplicate accounts. However, this only allows one to assess the likelihood of manipulation, not always to prove that payment was made.

Therefore, the problem cannot be solved by more sophisticated text analysis alone. It is necessary to verify the actual relationship between the reviewer and the company.

How the system can be changed

There is no such thing as a completely reliable verification method. However, platforms can shift from verifying text and accounts to verifying legal, financial, and operational facts.

Verification for B2B

For corporate reviews, it is necessary to verify not only the email address or domain but also the business partner itself:

  • the legal entity’s registration;
  • current legal status;
  • owners and directors;
  • domain ownership;
  • corporate email ownership;
  • bank account ownership;
  • licenses held;
  • operational activities;
  • contract, invoice, payment, or delivery;
  • no obvious connection between the parties.

A domain name should not automatically be considered proof of a business’s existence. It may have been purchased, resold, or registered in the name of a front person.

Legal registration alone is also insufficient. A company may exist formally but have no actual business operations.

Therefore, the status of a corporate review must consist of several independent criteria:

  • the legal entity is verified;
  • operational activity is confirmed;
  • the company representative is verified;
  • business relationships are confirmed;
  • payment or delivery confirmed;
  • the collaboration is ongoing;
  • Cooperation has ended.

The verification must be updated regularly. A company that existed six months or a year ago may now be dissolved or have effectively ceased operations. This usually happens before the first tax return is filed.

Verification for B2C

For B2C, the task is more complex. Even a real person and a genuine account do not prove that the service was received and used.

A photo of a receipt or a screenshot of a banking app is also insufficient: images, documents, and user interfaces can be faked.

A more reliable system should include multiple levels.

1. Independent payment confirmation

The fact of payment must be confirmed not by the company itself, but by the bank or payment processor. Any company acting as an intermediary creates the same risks that currently exist.

The review platform can receive a special token confirming:

  • the transaction itself;
  • the date;
  • the product category;
  • the recipient;
  • that there has been no full or partial return as of the time of verification;
  • no full or partial refund within N days of the verification

You do not need to disclose your card number, the exact amount, or your personal banking information.

2. Confirmation of Actual Use

A one-time payment does not necessarily mean that the customer has actually used the product.

For digital services, you can verify:

  • account activation;
  • login;
  • use of core features;
  • subscription duration;
  • reaching a certain usage level.

For physical goods:

  • handing over the order to the carrier;
  • confirmed delivery;
  • no 14/30-day returns;

For services:

  • confirmation of booking;
  • start of service provision;
  • completion of a stage;
  • signing of a certificate or other confirmation of completion.

3. Evidence Before and After Publication

For some categories, it is necessary to request evidence not only at the time of the review’s publication but also after a certain period of time.

For example:

  • confirmation of status prior to the start of the service;
  • confirmation of activation or first use;
  • a follow-up verification after 30, 60, or 90 days;
  • return information;
  • chargeback information;
  • confirmation that the account has continued to be used.

This allows users to be categorized as:

  • current customer;
  • former customer;
  • a customer who requested a refund;
  • a customer who initiated a chargeback;
  • a user who paid for but did not use the product;
  • a user with confirmed long-term usage.

4. Verification Level Categories

Instead of a single checkmark, Verified it is necessary to show exactly what was verified:

  • identity verified;
  • payment verified;
  • delivery confirmed;
  • activation confirmed;
  • use has been confirmed;
  • Use for more than 30 days confirmed;
  • no return has been recorded;
  • The result has not been independently verified.

This is critically important because a confirmed payment does not prove the accuracy of the text and does not guarantee a result.

5. Protection Against Fraudulent Transactions

The following must be identified:

  • payments between related parties;
  • circular transfers;
  • bulk purchases for the minimum amount;
  • the use of a single payment method by different accounts;
  • payment for a cheap product followed by a review of an expensive service;
  • instant returns after publication;
  • affiliated employees and contractors.

6. Random Audit

A portion of the reviews should undergo additional manual verification, including requests for documents or technical evidence.

The likelihood of an audit can be increased for:

  • new companies;
  • a sharp increase in rating;
  • expensive and high-risk services;
  • financial products;
  • medical services;
  • investment projects;
  • companies with a large number of complaints.

7. Prohibition on Selective Invitations

A company must not be able to send invitations to leave a review only to satisfied customers.

Invitations must be sent automatically to all customers or to an objectively selected random sample.

8. Disclosure of Any Incentives

It must be indicated next to the review whether the author received:

  • a discount;
  • a free product;
  • a gift;
  • a partial refund;
  • an affiliate commission;
  • entry into a drawing;
  • bonus points;
  • a free subscription renewal.

Such a review isn’t necessarily false, but the user should understand that the author received some benefit.

Limitations will still remain

Even this model won’t completely solve the problem.

Payments can be made between affiliated parties. A low-cost service can be used as the basis for a review of an expensive product. After verification, a refund or chargeback can be processed.

Cash payments are harder to verify. The payer could be a relative, an employer, or another company. Some services don’t have a clear end date. Screenshots of usage can also be faked.

In addition, a real customer may intentionally leave a false, paid, or emotionally biased review.

Therefore, the platform’s task is not to label a post as “absolutely true,” but to show which facts have been verified by an independent party and which remain the author’s assertion.

How to Evaluate a Company Today

Star ratings shouldn’t be the main criterion.

For B2B, you need to verify:

  • legal entity;
  • licenses;
  • bank details;
  • actual business operations;
  • contractual obligations;
  • referrals through independently sourced contacts.

A client contact provided by the company itself may be an affiliated party or a participant in a prearranged referral scheme.

For B2C, it is more important to verify:

  • exactly who the money is being transferred to;
  • whether the payee matches the company;
  • the terms of the contract;
  • the return policy;
  • the possibility of a chargeback;
  • licensing;
  • the existence of a regulatory authority;
  • court and regulatory decisions;
  • the actual ability to contact the company;
  • a real dispute resolution process.

The domain’s age, a well-known media outlet’s logo, professional responses to reviews, the Verified and a Wikipedia page can be used as additional indicators, but should not be considered as standalone evidence.

The most reliable indicator is not the number of positive reviews, but the ability to independently verify:

  • the company’s existence;
  • its operational activities;
  • the validity of its payment details;
  • the fact that it fulfills its obligations;
  • genuine relationships with customers;
  • long-term use of the product;
  • no refunds or payment disputes.

Conclusion

Online reviews haven’t become completely useless, but they are no longer sufficient on their own as proof of trustworthiness.

Today, it’s not just a matter of buying a few five-star ratings. You can create customers, accounts, posts, sources, search results, expert status, and an entire company history.

You can publish original content through a paid contributor program and then present it as an independent mention in a well-known media outlet. You can create pseudo-publications on old domains, use them as sources for Wikipedia, and then use Wikipedia as new proof of credibility.

As a result, the review market is increasingly turning into a market for selling reputation.

A future trust system should not be based on star ratings, account age, a well-known logo, or a general score Verified, but on independent verification of the legal entity, payment, actual operations, real-world use of the product, and the resulting outcome.