Spend a few days working through the paper trail behind storefronts that regulators later shut down, and one detail stops being surprising: almost every one of them was adored. Four point eight here. Four point nine there. Page after page of glowing cards, many posted within days of each other, many praising products that buyers would soon be fighting to refund. The rating was never the warning. The rating was the bait.
That is the uncomfortable thing about the humble gold star. It is the most trusted signal in online shopping and the cheapest one to buy. Read hundreds of these review walls back to back and the fakes rarely scream. They are short, warm, a little generic, and they arrive in clusters that no real customer base ever produces. The polish is the point. A wall of praise reads as consensus, and consensus is exactly what switches off a shopper’s suspicion right before a card number goes in.
The numbers behind that pattern are not subtle, and the enforcement cases naming specific companies are now public record. Here is why praise is the lie of least resistance, what the fakery actually looks like up close, and the sites that kept their stars while cheating the people reading them.
| 46% | ~30% | $770B | $53,088 |
| Share of flagged fake reviews rated five out of five stars | Online reviews estimated to be fake or manipulated in 2025 | Estimated global cost to shoppers from fake reviews in 2025 | Maximum US civil penalty per fake-review violation |
The math that makes praise the path of least resistance
Fraud follows incentives, and the incentives here are lopsided. The FTC has estimated that buying fake reviews can return as much as 1,900 percent on the money spent, which puts a manufactured rating in the same budget line as any other cheap way to pull in customers. Positive reviews are also low friction to produce. A happy five-star blurb needs no target, invites no legal threat, and rarely triggers a complaint, because the person it fools does not realize anything is wrong until the package arrives late, turns out counterfeit, or never ships at all.
Compare that with the alternative. Tearing down a rival with fake one-star reviews is slower, far easier to trace, and carries genuine exposure under defamation and unfair-competition law. Praising yourself is simply the safer crime, so that is the one most scam operators commit.
| Dimension | Fake five-star praise | Fake one-star attack |
| Target needed | None. Aim it at your own listing | A specific, named competitor |
| Legal exposure | Low until a buyer is harmed | High: defamation, interference |
| Who complains | The victim, later and quietly | The rival, loudly and fast |
| Detection speed | Slow, buried in a happy pile | Quick, concentrated and hostile |
| Risk to the operator | Minimal at small scale | Significant and immediate |
The playbook behind a flawless rating
The mechanics are not mysterious, and court filings spell them out step by step. Five tactics show up again and again once you start reading the complaints.
| Tactic | How it works |
| Bought, then dripped | Operators pay a broker for hundreds or thousands of reviews, then release them slowly and from scattered locations so the flood looks organic. In the Roomster case, the FTC said reviews were spread across different countries and timed to appear constant and random. |
| Account farms | The reviews need bodies behind them. The broker in the Roomster matter used more than 2,500 fake iTunes accounts, plus throwaway email addresses, to post ratings that each read like a separate, verified customer. |
| Written by a machine | Generative tools can spin out specific, confident, human-sounding reviews in bulk. In 2024 the FTC charged the AI service Rytr over a product that generated reviews on demand. Subscribers produced tens of thousands, including more than 83,000 for moving services alone. |
| Gating the bad ones | Some sellers do not invent praise so much as bury criticism. Fashion Nova ran review software set to auto-publish four- and five-star ratings while holding anything lower for an approval that never came, hiding hundreds of thousands of negative reviews. |
| Insiders in disguise | The fakes can come straight off the payroll. The FTC found that the CEO of Sunday Riley told staff to post five-star reviews of their own products under invented identities, then use a VPN to hide their location after the retailer caught on. |
Sites that had the stars and still burned people
A high rating and a scam are not opposites. Several operations earned glowing numbers on the way to enforcement actions, refunds, and outright bans. None of this is rumor. Each one is documented in a public FTC action.
| Operation | The stars said | The reality | Outcome |
| Roomster (rental listings) | Tens of thousands of four- and five-star reviews across the app stores | At least 20,000 fake positive reviews bought from one broker. Many listings were phony and some apartments did not exist, yet renters paid for access | FTC and six states settled in 2023. Owners permanently banned from buying or incentivizing reviews |
| Fashion Nova (fashion retail) | An inflated average built only from four- and five-star posts | Review software suppressed every rating under four stars from late 2015 to late 2019, hiding hundreds of thousands of complaints | $4.2M settlement in 2022. About $2.4M refunded to 148,351 customers. The FTC’s first case on concealing negative reviews |
| Sunday Riley (skincare) | Five-star reviews on a major retailer’s product pages | Written by company employees under fake accounts at the CEO’s direction, with a VPN used to dodge detection | FTC settlement in 2019 barring future misrepresentations about reviews |
| Rytr (AI review tool) | A service that output polished, ready-to-post reviews in seconds | Marketed to generate reviews the FTC said would almost certainly be false for anyone who published them | 2024 FTC action under Operation AI Comply. The order was later set aside in December 2025 |
A number worth sitting with The Roomster number is the one to sit with. Investigators said the company flooded app stores with at least 20,000 fake positive reviews, bought from a single seller, while charging people to reach listings it claimed were verified. For a shopper scrolling a four-star app, nothing on the surface separated it from an honest competitor. The stars were manufactured inventory, no different from the product. |
What the rules finally did, and what they left to you
In October 2024 the Federal Trade Commission’s Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, became enforceable federal law. For the first time the agency can seek civil penalties, currently up to $53,088 per violation, for a defined set of review tricks. Each fake review can count as its own violation, so the totals climb fast.
What the FTC rule bans
|
The limit is jurisdiction. A rule written in Washington does little to a storefront run from overseas that spins up a fresh domain every month. Enforcement reaches the large and the locatable. The anonymous drop-shipper with a five-star rating page and a mailing address in no country at all is still, in practice, the shopper’s problem to spot. Which is where a little method helps.
How to read a five-star wall like an investigator
The goal is not to distrust every rating. It is to stop treating the average as the answer and start reading the wall the way an auditor would.
| On the surface | What to actually check |
| A near-perfect average | The shape underneath. Real products collect a scatter of threes and fours. A wall that is almost all five stars with almost no middle is a flag, not a trophy. |
| Lots of recent praise | The timing. Dozens of glowing reviews posted inside the same few days rarely come from real buyers arriving on their own schedule. |
| Warm, enthusiastic wording | The specifics. Genuine reviews mention sizing, shipping, a flaw, a real use case. Fakes stay vague, upbeat, and oddly interchangeable. |
| Named, verified reviewers | The history. Open the profile. One post, no photo, and a generic handle is a reviewer doing no work to earn your trust. |
| Five stars on the seller’s own site | A second opinion. Cross-check the name on independent sites, the Better Business Bureau, and the complaint records of the FTC or your local regulator. |
Three checks that take two minutes
|
The verdict
After enough hours inside these rating walls, the stars stop reading as a judgment and start reading as a line item. They are the easiest part of a storefront to buy, the first thing a serious scam operation invests in, and the signal most likely to be pure fiction on exactly the sites that should worry a shopper most. Honest sellers earn five stars. Dishonest ones simply order them.
None of that makes reviews useless. Ratings still carry real information, just not in the number on the badge. The information sits in the distribution, the timing, the specifics, and whether the company survives a two-minute search on anything other than its own homepage. Treat a perfect score as the price of admission rather than proof of anything, spend your attention on the mediocre reviews and the independent record, and the fakes lose most of their power.
The gold star was meant to be a shortcut to trust. On the sites most worth avoiding, it is the shortcut the scammer took first.