Fake reviews usually show up as a pattern, not as one clumsy sentence or a missing badge. Look at when the comments landed, whether the wording repeats, what else the reviewer posts, and whether anyone describes a transaction that could actually have happened.
If several of those warning signs pile up on the same listing, treat the star rating as weaker evidence. Compare the product or business with information that isn't sitting in the review thread, and report a review only when you can point to a platform-policy problem. A suspicious comment is not proof that the product is junk. A genuine one can still be short, emotional, or negative.
What counts as a fake review?
A fake review generally claims an experience that never happened, or hides a relationship that would change how you read it. That includes:
- Paying someone, or offering a refund, in exchange for a positive rating
- A seller, employee, competitor, or marketing agency posting as an ordinary customer
- Coordinated praise for one listing, or coordinated attacks on a rival
- Text copied from another product, business, or website
- An account reviewing a place it never visited or a product it never used
- A business pressuring people to take down honest criticism, or asking only happy customers for reviews
AI-written text is not automatically fake. A real buyer may use a tool to tidy a paragraph, and a fabricated review can still have typos. The useful question is whether the account is presenting a real experience fairly.
A review also isn't fake just because it lacks a Verified Purchase label, uses polished grammar, has no photo, or disagrees with you. Those details can be clues. They are not proof.
There is also no trustworthy single percentage for fake reviews across Amazon, Google, Yelp, TripAdvisor, app stores, and every other site. Estimates move around with the platform, the product category, the country, the sample, and what the researcher counted as “fake.” Some studies include only fully invented posts. Others fold in undisclosed incentives, copied text, or reviews that are simply thin. Platform totals may count content that was blocked or suspected, not content independently confirmed as fraud. Use a headline statistic as background, not as a forecast about the page in front of you.
Examples of fake review patterns across platforms
The sketches below are warning patterns, not accusations against a named seller. A platform, regulator, or court needs more than a hunch before a specific review can be called fraudulent.
Amazon and other retail marketplaces
Rebate groups still follow a familiar script. A listing has almost no reviews. Buyers are told to pay full price, leave five stars, and get the money back. Then a cluster of similar comments appears in a short window, often from accounts whose other activity is a string of unrelated five-star ratings.
A missing Verified Purchase label can support that picture. It does not settle the question. Open the reviewer history, check the exact dates, and see whether the comment matches the model, size, or variation you are considering.
A 2021 scheme involving more than 200,000 people linked to third-party vendors was widely reported at the time. That case is a historical example, not a measurement of all Amazon reviews and not a current 2026 snapshot. PCMag's report on spotting suspicious Amazon reviews covers that episode and explains why automated review scores should be treated as signals rather than proof.
The same burst-and-copy pattern shows up on Walmart Marketplace, Best Buy, Etsy, and other retail sites. A sudden jump matters more when the reviews also skip product-specific details or appear across unrelated listings.
Google Maps and local businesses
On Google, manipulation often looks like a ratings spike, recycled phrases, accounts that review many unrelated businesses, or comments that clearly describe the wrong company. A restaurant collecting praise for a “shuttle service” it doesn't run is the kind of mismatch worth a second look. So is a wave of one-star posts from accounts that never mention a visit, an order, or an employee.
Review growth by itself isn't enough. A new location, a seasonal rush, a promotion, or a delayed batch of customer emails can create a real increase. Wait for several matching signals before you report.
A duplicate listing, a false “permanently closed” status, changed contact details, or an unauthorized edit is a different Google Maps problem. Report the profile or listing issue on its own path instead of stuffing it into a fake-review complaint.
Yelp
Suspicious Yelp activity may look like several reviews posted around the same date with nearly identical descriptions, empty praise such as “best service ever,” or criticism that refers to a different business.
A low rating is not evidence of fraud. People have different nights, and a short review can still be genuine. The better test is whether the reviewer describes a plausible interaction and whether the rest of the account supports that story.
Use Yelp's reporting or flagging option when you can point to a policy problem. Don't flag a review just because it hurt a rating or said something you dislike.
TripAdvisor
Travel reviews need dates and details. Watch for comments posted before a hotel, attraction, or tour opened; several accounts using the same odd description; and praise that never mentions the room, route, meal, guide, or other thing being reviewed.
Check whether the review matches the property and the travel period. Management can change, rooms can be renovated, and services can be dropped, so an older negative review and a newer positive one can both be real.
App stores
A mobile app may pick up a wave of generic five-star ratings after an update or just before it climbs a chart. “Great app” and “best ever,” repeated across new accounts, tell you almost nothing.
Read the one-, two-, and three-star reviews for concrete problems: subscriptions, crashes, permissions, unexpected charges, account lockouts, or lost data. Screenshots can help, and they can also be edited. Confirm the important claims through the app's support page, its privacy disclosures, and the platform's billing records.
Trustpilot, Fiverr, Upwork, and other service sites
Service reviews get gamed through coordinated accounts, paid testimonials, and posts from people tied to the provider. Look for what was ordered, when the work happened, what was delivered, and how a problem was handled.
A high rating on a freelance platform does not prove every job or client is genuine. Repeated wording across different company profiles is a stronger warning than one enthusiastic comment.
Reviews used to sell financial scams
Treat reviews that promise guaranteed returns, instant withdrawals, no-risk profits, or a secret investment as advertising. Glowing comments may be invented. Even a testimonial that looks real does not make a financial service safe.
Verify the company somewhere other than the review thread. Don't send money by wire or cryptocurrency just because strangers praise the offer, and check withdrawal terms, fees, and identity details before you share funds or personal data.
A practical checklist for spotting fake reviews
| Signal | Why it may be suspicious | What to check next |
|---|---|---|
| A sudden burst of reviews | Coordinated campaigns often create a visible posting spike | Sort by newest and compare the dates with the product launch or business opening |
| Repeated wording | Several accounts may be using supplied text | Search an unusual phrase on the page and compare other reviews |
| Generic praise | It may provide no evidence of actual use | Look for model, size, location, service date, delivery, or performance details |
| New accounts with one type of rating | A campaign may use disposable profiles | Review the account's other activity; a new account is not proof by itself |
| Identical or staged photos | Images may be copied or supplied by a seller | Check whether the same photo appears on other listings or websites |
| Extreme star imbalance | A seller may be pumping praise or attacking a rival | Read recent middle-rated reviews and compare recurring complaints |
| No purchase badge | The platform may not have confirmed a transaction | Treat it as one limited signal; absence does not prove fraud |
| Claims that don't match the listing | The reviewer may have copied text or reviewed another item | Check the exact model, size, variation, location, and date |
| Perfectly polished but vague writing | AI or templated text may be involved | Look for verifiable facts rather than trying to catch AI by style |
| Reviews posted before launch | The account could not have used the product then | Confirm the actual release, opening, or service date |
One row on that table does not decide the issue. A real launch can produce many reviews, and a real customer may write a bland sentence. Timing plus overlap is what matters.
How to check a listing before you buy
When the reviews are doing most of the selling, work through the page in this order:
- Start with recent reviews. Older comments may describe a different seller, product version, menu, or manager.
- Leave the top-rated pile. Read a mix of positive, middle, and negative comments.
- Look for an actual experience. Useful reviews say what happened, not only how the writer felt.
- Open a few reviewer profiles. Repeated bursts, identical language, or activity across unrelated businesses are more informative than one odd username.
- Check the dates. A long flat stretch, then a sudden jump, deserves a closer look.
- Confirm the exact item or service. Marketplace pages often bundle variations, sellers, or versions under one listing.
- Step off the review thread. Compare the return policy, warranty, contact details, independent testing, and complaint history.
- Weigh the downside. For health, safety, financial, or expensive purchases, pause if most of the evidence is anonymous praise.
- Save records before you pay. Keep the listing, price, seller name, order confirmation, and any reviews you relied on.
- Treat review-scoring tools as extras. A third-party checker can flag unusual language or account patterns. It cannot confirm that a transaction happened, and some well-known tools have shut down or stopped loading. Read the privacy permissions before you install an extension.
You don't need a perfect review page. You need enough independent detail that one rating or one testimonial isn't making a high-stakes decision for you.
How to report a suspected fake review
Report conduct that breaks the platform's rules: invented experiences, undisclosed incentives, spam, conflicts of interest, impersonation, or coordinated abuse. A review that is merely negative, short, or different from your view may still be allowed.
Save evidence first
Before a comment is edited or removed, save:
- A screenshot of the full review, rating, account name, and date
- The direct page or listing link
- The product, business, or service involved
- Your receipt, order record, booking, or other proof of what happened
- Copies of rebate messages, payment requests, threats, or review instructions
- A short timeline showing repeated wording or a burst of similar posts
Don't post a reviewer's personal information in public. Send the evidence through the platform's reporting channel.
Use the platform's report function
On Google Maps or Search, open the review's three-dot menu, choose the reporting option, and pick the reason that best matches the suspected violation. Google says it evaluates reported reviews against its content policies; a report does not guarantee removal. Businesses can follow the official Google instructions for reporting inappropriate reviews, including the appeal process.
For Amazon, Yelp, TripAdvisor, Trustpilot, app stores, and other marketplaces, use the report or flag menu on the review itself. If several comments look coordinated, report the pattern and include examples instead of filing the same complaint over and over. Use the platform's seller, business, or trust-and-safety route when the in-product button doesn't cover coordinated abuse.
If the real problem is a hijacked listing, an extortion attempt, or changed contact information, say that clearly. Don't force every complaint into the “fake review” bucket.
Report broader fraud to the FTC
The platform moderates its own content. The Federal Trade Commission takes reports about suspected fraud and deceptive business practices. Consumers and businesses can read the FTC's guidance on reporting suspicious online reviews and submit details through ReportFraud.gov.
An FTC report will not automatically take a review down, cancel an order, or issue a refund. It creates a record for consumer-protection work. The platform report is what addresses the comment itself.
If a fake review influenced your purchase
Handle the listing and the money as two different problems:
- Ask the seller or platform for the remedy in its terms. Stick to the factual issue and attach your order evidence.
- Contact the payment provider if money or account information is at risk. For a credit-card purchase, ask the issuer about its billing-dispute process. Debit cards, prepaid cards, payment apps, wires, and cryptocurrency follow different procedures and may have different deadlines.
Reporting a review is not the same as disputing a payment. Platform removal does not guarantee a refund. If the transaction looks fraudulent, don't wait on the review decision before you seek payment help.
If you're a business owner
You can report a review that violates platform rules. You generally can't make a platform delete a comment just because a customer is unhappy. A calm public reply can correct a factual error without exposing private customer information.
Don't buy positive reviews, pay only for favorable ratings, ask employees to pose as customers, or threaten people who post criticism. Don't trade a refund or other benefit for the removal of honest negative feedback. U.S. consumer-protection rules can apply to deceptive review practices; the FTC reporting guidance is a practical starting point.
Ask for feedback in a way that reaches customers after both good and bad experiences. Keep records of real transactions and messages in case you later need to report a coordinated campaign.
Frequently asked questions
Does a missing Verified Purchase label mean an Amazon review is fake?
No. It means Amazon hasn't shown that particular purchase-confirmation label. Read it next to timing, reviewer history, wording, and product-specific detail.
Can you identify an AI-generated review with certainty?
No. Detectors and writing style are not reliable proof. A review can be AI-assisted and based on a real purchase, or handwritten and entirely made up. Check the facts and the transaction instead.
Can a platform remove a review just because it is negative?
Usually not. Platforms generally look for a policy violation such as spam, fraud, a conflict of interest, harassment, or an unrelated experience. Use the platform's reporting process and name the specific rule you think was broken.
Should I trust a review-checking tool?
Use it as a risk signal, not a verdict. Tools may cover only part of a review set, use different methods, or disappear. Your own checks should still cover dates, profiles, details, seller terms, and independent evidence.
Where should I report a fake Google review?
Use the review's reporting menu in Google Maps or Search. If the issue is a broader scam, an extortion attempt, or a deceptive business practice, also consider a report to the FTC through ReportFraud.gov.
If you're on a listing right now, sort the reviews by newest, read a few middle-star comments, and open two or three reviewer profiles before you trust the average. Screenshot anything you plan to report, then use the platform's flag tool; add an FTC report only if the pattern looks like fraud rather than one bad night.