How marketplace sellers fake reputation

Fake reputation is usually assembled, not invented

A marketplace seller does not need to fabricate an entire identity to look trustworthy. They only need to make the visible evidence point in the right direction.

A high star average. Thousands of feedback entries. An account that appears years old. A product page crowded with five-star reviews. A badge. A long list of completed transactions.

Each signal can be useful. The problem is that each signal answers a narrower question than buyers often assume.

A 2025 NBER working paper on fake reviews in Amazon’s marketplace found that review manipulation can do two kinds of damage at once: it can misinform shoppers about the item in front of them, and it can make the ratings system itself less trustworthy. In the authors’ model, fake reviews shift sales from honest sellers toward dishonest ones and reduce consumer welfare.

That is the economic point of fake reputation. It does not have to make a bad seller look perfect. It only has to move the buyer’s confidence enough to change the click.

The manipulation usually happens in a few recurring ways.

Shill reviews make praise look independent

The simplest version is also the oldest: the seller arranges for praise that appears to come from ordinary customers.

That can mean friends, employees, paid reviewers, review-exchange groups or buyers who receive money, discounts or free products in return for feedback. Etsy’s shilling policy explicitly covers biased reviews from friends or family, compensated third parties and extra accounts posing as independent buyers. Amazon likewise prohibits sellers from paying for reviews, reimbursing buyers after a review or using third-party review clubs that connect sellers with reviewers.

The important detail is that a real transaction can still produce a fake reputation signal.

A buyer may genuinely place an order. The marketplace may genuinely record the purchase. The review may therefore look stronger than a random anonymous comment. But if the seller reimbursed the purchase or conditioned a benefit on the review, the transaction does not make the endorsement independent.

That is why “verified purchase” should be read literally: it can strengthen the evidence that a transaction occurred. It does not, by itself, certify the reviewer’s motive.

In the United States, the FTC’s 2024 rule on consumer reviews targets this broader problem. It prohibits specified fake or false reviews, buying reviews whose required sentiment is positive or negative, certain undisclosed insider reviews and some forms of review suppression. The rule treats the reputation signal itself as something capable of deceiving consumers.

Feedback farming can make a seller look more proven than the current sale deserves

A large feedback number feels like history.

Sometimes it is good history. Sometimes the history was built in a part of the marketplace that says little about the purchase you are considering now.

eBay’s seller-performance policy specifically prohibits selling large numbers of low-dollar items when the purpose is to conceal complaints associated with expensive items. It also prohibits transactions whose purpose is to increase feedback scores, detailed seller ratings or site privileges.

That policy points to a useful buyer question:

What produced this reputation?

Ten thousand successful sales of inexpensive accessories are evidence that an account has completed ten thousand transactions. They are not automatically evidence that a $2,000 camera, collectible or laptop listing is equally trustworthy.

The same logic applies when the seller’s inventory changes sharply. Reputation is strongest when the historical transactions resemble the transaction being used to justify trust now.

An old score can be genuine and still be weak evidence for a new kind of sale.

Ratings can be attached to the wrong thing

Marketplace pages often collapse several different reputations into one screen.

There may be a product rating, a seller rating, a listing history and an account history. Buyers can easily blend them together.

Amazon’s review policy gives a particularly clear example of how that can be abused. The company prohibits sellers from creating variation relationships between products that are not actually related when the purpose is to aggregate reviews and raise a product’s star rating.

The result can be a page where the visible rating looks established while individual reviews refer to a different size, model, design or even a materially different item.

This is sometimes described as review hijacking. The useful test is less dramatic:

Do the reviews actually describe the thing being sold on this page?

If a listing for one product contains reviews talking about another, the overall star average has lost much of its meaning even if many of the underlying reviews were written by real customers.

This distinction also matters on large marketplaces where a product page can be shared by multiple merchants. A highly rated product is not necessarily evidence that the particular seller fulfilling your order has an equally strong record.

Product reputation and seller reputation are separate claims.

Negative feedback can be filtered before it reaches the score

Reputation can also be manipulated by changing who gets encouraged to review.

Amazon prohibits sellers from soliciting reviews only from customers who had positive experiences while diverting unhappy buyers to a different feedback channel. It also prohibits offering compensation to get a reviewer to change or remove a review.

The FTC rule similarly addresses certain forms of review suppression, including misrepresenting a review section as representative when negative reviews have been suppressed because of their sentiment.

This matters because a five-star average does not only depend on how many positive customers exist. It depends on which customers are allowed, encouraged or pressured to become visible.

Not every missing negative review implies manipulation. Platforms legitimately remove spam, abuse and policy-violating content. Sellers are also allowed to resolve complaints.

The red flag is not that a negative review disappeared. It is a pattern where praise receives a clear path to publication while criticism is redirected, discouraged or made costly.

Account history can be borrowed as credibility

An account that has existed for years feels safer than one created yesterday. That is reasonable. It is still not proof that every current listing deserves the account’s accumulated trust.

eBay explicitly prohibits buying or selling user accounts as part of its feedback-manipulation policy. That rule exists because reputation has value independent of the inventory currently attached to the account.

Other marketplaces create different weaknesses. Facebook Marketplace, for example, says seller ratings become public after a seller receives five or more ratings — but also states that it does not verify whether the person leaving a rating actually purchased the product or used the service.

That does not make Facebook Marketplace ratings useless. It means the signal is different from a review system tied tightly to a completed order.

Marketplace reputation is not one universal measurement. A five-star score can mean different things depending on the platform that produced it.

The industrial version is a service business

Manipulation is not limited to one seller asking a friend for a five-star review.

There is a market for reputation itself.

Amazon says fake-review brokers solicit reviewers through websites, social platforms and messaging services, offering reimbursement, products or money in exchange for reviews. In its 2025 Trustworthy Shopping Experience Report, Amazon said it proactively blocked hundreds of millions of suspected fake reviews during 2025 and took legal action that helped shut down more than 100 websites involved in fake-review or scam activity.

Those are Amazon’s own enforcement figures, not an independent count of the total fake-review market. But the scale is enough to show why review manipulation should not be treated as an occasional oddity.

When reputation influences search ranking, conversion and access to marketplace privileges, reputation becomes an asset worth manufacturing.

A strong-looking signal can still be weak evidence

The easiest mistake is treating every trust marker as if it proves the same thing.

Signal	What it can support	What it does not establish by itself
High seller rating	Other users left favorable seller feedback	That the current listing is genuine or fairly priced
Large feedback count	The account has accumulated substantial feedback	That the feedback came from similar products or recent activity
Verified purchase	A qualifying transaction was recorded	That the reviewer was not reimbursed or otherwise incentivized
Old account	The account has existed for a long time	That the current operator, inventory or behavior is unchanged
Highly rated product	Buyers rated the product or product page well	That the specific merchant is trustworthy
Marketplace badge	The seller met the platform’s badge criteria	That every future transaction is low-risk
Facebook Marketplace stars	People who interacted with the seller submitted ratings	That each rater completed a purchase

The practical move is not to discard reputation systems. It is to keep the claim proportional to the signal.

How to read a seller’s reputation without pretending you can detect fraud from a glance

There is no magic visual tell for fake reviews.

Amazon itself warns against simplistic detection. A burst of reviews can come from advertising or a genuinely popular product. Poor grammar is not proof of fraud. The company says its own systems use deeper information unavailable to ordinary shoppers, including account relationships, review history, sign-in activity and unusual behavioral patterns.

So the buyer’s job is not to “spot the fake” from one clue. It is to look for consistency.

Start with the newest feedback, not just the lifetime average. Read several middle ratings — two, three and four stars often contain more specific detail than the extremes. Check whether reviews describe the exact product on the page. Look at what the seller historically sold, when the platform makes that visible. Separate product stars from merchant feedback. Notice sudden category changes or a reputation built on transactions unlike the one in front of you.

Most importantly, ask what the marketplace itself verified.

Did it verify a purchase? An interaction? Identity? Delivery? A performance threshold? Or did it simply display user-submitted feedback?

The badge or number is only as meaningful as the system underneath it.

This is the same discipline that applies to screenshots used as evidence: start with what the artifact actually establishes, then resist the temptation to let it prove the larger story by association.

Reputation is evidence, not a warranty

Online marketplaces need reputation systems because buyers and sellers usually begin as strangers. Stars, feedback counts and badges reduce that uncertainty fast.

That speed is exactly what makes them worth manipulating.

The most convincing fake reputation is often not completely fake. It may contain real transactions, real account age, real reviews and real customers. The distortion happens in selection, attribution and context: which transactions built the score, which reviews became visible, which product inherited the stars, which customers were incentivized and which part of the history the buyer assumes applies to the current sale.

A seller’s reputation can therefore be both real data and misleading evidence.

The useful question is not simply, “Does this seller have good reviews?”

It is:

What, exactly, earned this reputation — and how much of that evidence belongs to the transaction I am about to make?