Don't treat one glowing or angry review as the whole story. Fake reviews can dress up a weak product or a dishonest seller, but an unusual post can also come from a real customer who was brief, upset, or just wrong. For U.S. shoppers, the useful habit is to look for a pattern, check the claims that would change the purchase, and report specific evidence through the platform.
Platform rules and consumer-protection processes differ outside the United States. A suspicious review still isn't automatically fake.
What counts as a fake review?
A fake review generally misrepresents a real customer experience or is posted to steer shoppers. Common cases include a review from someone who never used the product or service; a paid or incentivized review that hides the relationship; posts from the business, a competitor, a broker, or a coordinated account network; copied, automated, or AI-assisted text presented as an independent personal experience; and reviews meant to bury legitimate criticism or inflate a listing's rank.
Negativity, typos, or a seller's dislike of the post don't make it fake. A "verified purchase" label can be useful, but it doesn't prove the opinion is independent or accurate.
How fake-review campaigns work
Most schemes mix people, extra accounts, and automation rather than one odd sentence. An operator may recruit reviewers or tap a network of existing accounts, point them at a product, business, or sentiment, then reuse templates, copied language, or AI drafts. The same wording can show up across related listings. Volume, timing, and multiple identities help hide the pattern.
Fluent writing doesn't prove a genuine experience. A polished review can be real, and an awkward one can still be honest.
Be cautious with viral claims about exact prices, specific countries, forced-labor operations, or a single fake-review percentage. Those details often get repeated without a reliable primary source.
Why there isn't one reliable fake-review percentage
There's no dependable, all-platform figure for how many online reviews are fake. Sources count different things: reviews an algorithm flagged, reviews removed after posting, reviews blocked before customers saw them, a small sample from one category or site, or posts that broke a platform rule even if fraud wasn't proven.
Those measurements aren't interchangeable. A company's enforcement total also doesn't show how many fake reviews remain visible.
Amazon reported that it stopped more than 200 million suspected fake reviews before customers saw them in 2020. That's a historical, company-reported enforcement figure, not a current share of fake reviews on Amazon or across marketplaces. Amazon describes its process in Creating a trustworthy reviews experience.
How to spot suspicious reviews
One clue rarely settles it. Several independent signals pointing the same way deserve a closer look.
| Possible signal | What to check | Why it isn't conclusive |
|---|---|---|
| A burst of reviews on the same day | Compare posting dates and wording across the group | A launch, sale, or popular event can create a genuine burst |
| Generic praise | Look for details about use, fit, performance, location, or service | Short reviews can be authentic, especially for simple purchases |
| Repeated phrases | Compare reviews for copied sentences, unusual wording, or the same structure | Customers may echo language from the product description |
| Unusual reviewer history | Check whether several accounts review unrelated products or businesses in the same pattern | New accounts and occasional reviewers can be legitimate |
| AI-like writing | See whether the review makes specific, verifiable claims | Detectors can produce false positives and can't prove who wrote a post |
| A mismatch with the product | Compare features, model, location, or service details with the actual listing | The reviewer may have used a different version or seller |
| A "verified" label | Treat it as one platform signal, not a guarantee of honesty | A purchase connection doesn't establish that the opinion is independent |
| Sudden rating changes | Look at recent reviews separately from older ones | A real change in quality, ownership, or inventory can move ratings |
A five-minute check before you buy
Read a mix of recent positive, negative, and middle ratings instead of trusting the average. Recurring specifics matter more than a pile of vague compliments: several independent posts describing the same defect, delay, sizing problem, or service issue are usually more useful.
Then check dates and profiles. Clusters, copied language, and accounts that review unrelated listings in the same style are worth noticing. Match specs, ingredients, photos, delivery promises, and return terms to the seller's actual information. For an expensive or safety-sensitive purchase, look at manufacturer details, independent testing, or another retailer.
Bad grammar, a five-star tone, a one-star rant, or "AI-like" phrasing are prompts to investigate, not proof.
What detection tools can and can't tell you
Platforms may scan review text, account behavior, timing, purchase information, and connections between accounts. Their methods are usually private, and no system catches everything.
Third-party analyzers can flag unusual patterns, but the scores are screening tools. They generally can't establish whether someone used the product or described it accurately. AI-writing detectors have the same limit and can mislabel human writing.
A removed review may have broken a rule about spam, disclosure, personal information, or promotional content rather than being proven fraud. A review that's still online may simply not have been reviewed yet.
Don't send private order records, phone numbers, or personal correspondence to an unfamiliar detection service. If you use one, read its privacy terms first.
What rules control fake-review disputes?
Two different systems may apply.
The platform's content policy decides whether a review is displayed, labeled, or removed. Sites may prohibit undisclosed incentives, impersonation, spam, conflicts of interest, or coordinated manipulation.
U.S. consumer-protection law can address deceptive review practices by businesses and intermediaries. The FTC warns companies about fake and misleading reviews and asks people to report suspicious activity.
A moderation decision is not a court finding that every statement was false. There's also no general guarantee that a consumer can force a platform to delete an unfavorable opinion just because a business disputes it.
The FTC's warning about fake reviews is aimed at deceptive practices. It doesn't replace the platform's reporting process or guarantee that one complaint will lead to removal, an investigation, or a refund.
How to report a fake review
1. Save the relevant evidence
Before you report, record the review page and the product, business, or listing it concerns; the reviewer's displayed name or profile link; the date, rating, and exact wording; screenshots of the review and any related pattern; evidence of a disclosed or hidden incentive, if you have it; and your order confirmation or support messages if you were a customer.
Keep it factual. A screenshot of five similar posts is more useful than calling the reviewer "obviously a bot."
2. Use the platform's own report function
Open the review's menu, flag icon, or reporting link and choose the reason that best matches what you saw. Menu names change, so follow the instructions on the site.
The FTC's How to report suspicious online reviews gives reporting directions for popular services, including Amazon, Google, Yelp, and Trustpilot.
Submit one clear report rather than organizing repeated or false complaints. Don't contact, threaten, or publicly identify the reviewer. Harassment can hit a legitimate customer and make it harder for moderators to assess the evidence.
3. Report a broader scheme
If several reviews appear connected, include listing links, account names, dates, repeated language, and any evidence of coordination in the platform's abuse or support channel. Explain what you observed without stating unverified conclusions as fact.
For suspected business deception affecting U.S. consumers, submit a complaint at ReportFraud.ftc.gov. The FTC uses complaints for enforcement and consumer warnings. That form is not a personal refund request or a guaranteed takedown route.
If a fake review influenced your purchase
A fake review can affect your decision. Its existence doesn't automatically create a refund or chargeback right. The remedy may depend on the seller's return policy, what was advertised, what you received, and how you paid.
Preserve the listing, the review, the receipt, the product description, and any promise that influenced the purchase. Contact the seller or marketplace about the product or service problem and ask for the available return, replacement, or refund process. Follow the stated requirements and keep shipping records, photos, and confirmation numbers.
If the item wasn't delivered, was materially different from the description, or a promised refund never arrived, ask your payment provider about a dispute, the deadlines, and the documents it needs. A fake review alone is not an unauthorized charge. Don't describe a transaction as unauthorized if you made the purchase.
Report suspected deception separately. A platform complaint and an FTC complaint serve different purposes from a refund request.
How fake reviews affect consumers and honest businesses
Consumers may pay for an unsuitable product, miss a better option, or trust a business that wouldn't otherwise meet their needs. The damage can be greater for purchases involving health, safety, travel, housing, or substantial money.
Honest businesses can lose visibility when coordinated reviews distort rankings or drown out genuine feedback. A flood of suspicious praise can undermine trust in legitimate reviews, while an organized negative campaign can make a fair business look unreliable. A poor rating by itself still isn't evidence of sabotage.
The useful response isn't chasing a perfect star average. Look for consistent, detailed experiences and verify the claims that matter most to your decision.
If you're reporting a review about your business
Don't publish a customer's personal information or make an accusation you can't support. A calm response can invite the reviewer to contact the business privately while you submit evidence through the platform.
Order numbers, appointment logs, support tickets, delivery records, and messages can help. A missing customer record can justify a platform review, but it doesn't always prove the person had no experience. They may have used a different name, received a gift, or dealt with a franchise or third-party seller.
Report the specific policy concern, such as impersonation, undisclosed compensation, or coordinated spam. Don't ask customers to post false praise or organize mass reports against criticism.
Common questions
Can an AI detector prove that a review is fake?
No. It may identify unusual wording, but it can't establish whether the reviewer had a real experience or who wrote the text. Use it, if at all, as one clue among several.
Is a verified-purchase review always genuine?
No. The label generally means the platform connected the review to a purchase. It doesn't prove the review is accurate, independent, or free from manipulation.
How many online reviews are fake?
No single number reliably applies to every platform. Treat claims such as "one-third of all reviews are fake" as estimates unless the source clearly explains its sample, definition, date, and method.
Will the FTC remove a fake review or get my money back?
Treat an FTC complaint as a report about possible deceptive conduct, not a guaranteed removal or refund process. Contact the platform about the post and the seller or payment provider about your purchase.
If you're about to buy, read recent mixed ratings and verify one or two facts that would change your mind. If you already see a coordinated pattern, save screenshots, use the site's report link, and add a complaint at ReportFraud.ftc.gov when it looks like business deception.