Choice Science · established evidence
The Economics of Fake Reviews: Who Fakes, and Why Competitive Pressure Predicts It
Fake reviews are not distributed at random across the marketplace. The most rigorous study of the question, Michael Luca and Georgios Zervas’s analysis of Yelp, shows that review fraud is a strategic response to a specific situation: a business is most likely to commit it when its own reputation is weak and the competitive pressure around it is high. Using Yelp’s own review filter as a proxy for suspicious activity, they found that fake reviews cluster in businesses with few reviews or low ratings, and that the probability of fraud rises as a business faces more direct competitors. The pattern is economic, not moral. A firm with little reputation to protect and much to gain has the strongest incentive to manufacture one, and a crowded market sharpens that incentive. Read correctly, this makes review manipulation partly predictable: you can identify the conditions under which it is most likely before you ever see a single suspect review.
Fraud is a response to pressure, not background noise
The intuitive model of fake reviews treats them as ambient corruption, a low hum of dishonesty spread evenly across every category. Luca and Zervas’s central contribution is to replace that picture with a conditional one. In their 2016 Management Science paper, the incidence of suspect reviews is not uniform; it concentrates exactly where economic theory predicts a business would have the most to gain from manipulating its reputation and the least to lose from trying.
Two variables do most of the work. The first is the strength of a business’s existing reputation: establishments with few reviews, or with low or negative ratings, are markedly more likely to show signs of fraud than those with a deep, favorable history. The second is competition: as the number of nearby competitors offering a similar service rises, so does the probability that a business engages in review fraud. Manipulation, in other words, is what firms do when the reputational stakes are high and their own position is exposed.
How you measure a crime nobody confesses to
Studying fraud has an obvious problem: the people committing it do not answer surveys truthfully, and the fake reviews themselves are written to be indistinguishable from real ones. Luca and Zervas solve this by using a signal the platform already produces. Yelp runs an automated filter that classifies reviews it judges unreliable and removes them from a business’s displayed rating while keeping them accessible. The authors treat a review’s filtered status as a proxy for likely fraud, then ask which business characteristics predict a higher filtered share.
The proxy is imperfect, and the authors are careful about it. Yelp’s filter is not ground truth; it will catch some genuine reviews and miss some fabricated ones. What makes the design credible is not that the filter is perfect but that its errors are unlikely to correlate neatly with reputation strength and competitive density in a way that would manufacture the result. The finding is the direction and consistency of the pattern across the dataset. The exact fraud rate is not the claim; the structure of who fakes, and when, is.
Why businesses fake reviews: weak reputation is the precondition
Why would a weak reputation, specifically, be the thing that predicts fraud? The economics are straightforward. A review, real or fake, is only worth manufacturing if it moves outcomes, and reviews move outcomes most for businesses that lack an established reputation to fall back on. Luca’s earlier Yelp work makes the point from the other side: a one-star increase in rating produced a five to nine percent revenue increase for restaurants, and that effect was driven entirely by independent establishments. Chains, whose reputations buyers already feel they know, showed no such sensitivity.
Put those two findings together and the incentive structure is clear. The businesses for whom an extra star is worth the most revenue are precisely the ones without a reputation buffer, and those are the same businesses the fraud study finds are most likely to fake. A firm with a thousand genuine reviews gains little from ten fabricated ones and risks much; a firm with six reviews and a middling average can change its entire public face with a handful of fabrications. Fraud follows the marginal value of the lie.
Competitors fake reviews too: competition is the trigger
The second variable, competitive pressure, explains timing and intensity. Reputation is a relative good. A buyer choosing among five similar businesses is comparing them against one another, so a rival’s strong rating is a direct threat and a rival’s weak one is an opening. As competitive density rises, the return on tilting that comparison, in either direction, rises with it.
This is where fraud acquires a second face. Manipulation is not only a business inflating its own reputation; it can be a business attacking a competitor’s. The same competitive logic that makes a weak firm buy positive reviews for itself makes it rational, under enough pressure, to plant negative reviews against a stronger neighbor. And negative fraud is a sharper weapon than positive fraud, because the revenue evidence shows losses weigh more heavily than equivalent gains. Chevalier and Mayzlin found that the impact of one-star reviews is larger in magnitude than the impact of five-star reviews, so a fabricated one-star attack does more damage than a fabricated five-star boost does good.
Not always the competitor: deception from real customers
It would be tidy to conclude that all review deception is competitor sabotage. The evidence does not allow it. Anderson and Simester, studying a large private-label retailer, found that roughly five percent of reviews came from accounts with no record of ever purchasing the product being reviewed. These non-purchase reviews were systematically more negative than verified ones, less likely to describe the concrete experiential detail a genuine buyer would mention, and more likely to contain the linguistic markers researchers associate with deception.
The uncomfortable twist is who was writing them. Many of the non-purchase reviewers were not competitors or hired operations but the retailer’s own most loyal, highest-spending customers, reviewing products they had not bought, often to voice a grievance or advocate for a position. Deceptive-style reviewing, then, is not a single actor with a single motive. It is a set of behaviors that emerges wherever the cost of writing a review is near zero and the writer has something to gain from the verdict, whether that writer is a struggling rival, a paid operation, or a devoted regular with an opinion and no receipt.
Are fake reviews illegal? The rules changed in 2024
For most of the review era, manipulation was an ethical problem and a platform-policy problem, not a legal one. That changed in the United States. The Federal Trade Commission’s Trade Regulation Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, took effect on October 21, 2024, and converts the practices the research describes into federal violations carrying civil penalties of up to $51,744 each.
The rule is specific. It prohibits reviews from reviewers who do not exist or who never experienced the product, reviews bought or sold, incentives conditioned on the review being positive, undisclosed insider reviews from officers or managers or their close relatives, selectively suppressing negative reviews while displaying positive ones, and buying or selling fake indicators such as followers or likes. Read against Luca and Zervas, the rule targets almost exactly the behaviors their data predicted a pressured, weak-reputation business would reach for. Reputation management stopped being a courtesy and became, in part, a compliance discipline.
What the pattern lets you predict
The practical value of this literature is not that it lets anyone prove a specific review is fake from the outside; usually it does not. Its value is that it makes the phenomenon legible in advance. Because fraud is conditional on weak reputation and competitive pressure, the conditions under which it is most likely are observable before any suspect review appears. A category crowded with similar providers, a competitor whose review count suddenly accelerates out of step with the apparent size of the business, a cluster of unusually generic five-star reviews arriving in a tight window, or a run of vague one-star reviews with no experiential detail aimed at a strong local incumbent: none of these is proof, but each is a marker the research tells you where to look for.
The same lens reads competitor behavior. When a rival’s rating moves in ways the underlying business cannot explain, the fraud literature supplies the priors: who has the incentive, under what pressure, and in which direction. This is the difference between watching reviews and interpreting them.
What is established, and what is not
The core finding is established. Luca and Zervas is peer-reviewed, published in a leading journal, and its structural result, that fraud concentrates in weak-reputation businesses and rises with competition, is consistent with independent work on the economics of reputation. The FTC rule is binding law, not interpretation. Anderson and Simester’s non-purchase finding is likewise peer-reviewed and directly measured.
What is not established is any precise, transferable fraud rate for a given local vertical. The Yelp study is grounded in restaurants, and the deception study in a single retailer’s catalog. The mechanism, the who and the when, travels well because it is driven by incentives that apply anywhere reputation is relative and cheap to fake. The magnitudes do not automatically transfer, and this piece does not claim they do. Anyone quoting a specific percentage of fake reviews for med spas or contractors, absent primary data on that category, is asserting more than the evidence supports.
The evidence
Key findings, with their sources
-
Fake reviews concentrate in businesses with weak existing reputations (few reviews, low or negative ratings) and become more likely as a business faces more direct competition; manipulation is a strategic response to competitive and reputational pressure, not random noise.
established Luca, M. & Zervas, G., "Fake It Till You Make It: Reputation, Competition, and Yelp Review Fraud", Management Science, 62(12), 2016.
-
About 5% of reviews on a large private-label retailer’s site came from accounts with no purchase record for that product; these unverified reviews were systematically more negative and carried more linguistic deception markers, and many came from the retailer’s own loyal customers.
established Anderson, E.T. & Simester, D.I., "Reviews without a Purchase: Low Ratings, Loyal Customers, and Deception", Journal of Marketing Research, 51(3), 2014.
-
Since October 21, 2024, fake, incentivized, insider, and suppressed reviews are federal violations in the US, carrying civil penalties of up to $51,744 each.
established Federal Trade Commission, "Trade Regulation Rule on the Use of Consumer Reviews and Testimonials", 16 CFR Part 465 (effective Oct 21, 2024).
-
A one-star increase in Yelp rating produced a 5 to 9 percent revenue increase for restaurants, an effect driven entirely by independent establishments, which explains why weak-reputation businesses have the strongest incentive to manufacture reviews.
established Luca, M., "Reviews, Reputation, and Revenue: The Case of Yelp.com", Harvard Business School Working Paper 12-016, 2011/2016.
-
The revenue impact of one-star reviews is larger in magnitude than the impact of five-star reviews, an asymmetry that makes fabricated negative reviews a sharper weapon than fabricated positive ones.
established Chevalier, J.A. & Mayzlin, D., "The Effect of Word of Mouth on Sales: Online Book Reviews", Journal of Marketing Research, 43(3), 2006.
Calibration
What is proven, what is promising, what is unproven
| Evidence tier | Tactics | What the evidence says |
|---|---|---|
| Established | Positive self-fraud by weak-reputation businesses (few or low reviews) | Luca & Zervas 2016, Management Science |
| Established | Fraud probability rising as a business faces more direct competition | Luca & Zervas 2016, Management Science |
| Established | Negative reviews from non-purchasers, skewing more critical | Anderson & Simester 2014, Journal of Marketing Research |
| Established (binding regulation) | Buying, incentivizing, insider-posting, or suppressing reviews | FTC 16 CFR Part 465, effective 2024 |
| Emerging / not yet localized | Applying Yelp-restaurant fraud rates to other local verticals | Extrapolation only; no MSME-vertical primary data yet |
Reference
Glossary
- Review fraud
- The manufacturing, buying, incentivizing, or suppressing of reviews to misrepresent a business’s reputation. It includes both positive self-inflation and negative attacks on competitors.
- Filtered review
- A review an automated platform filter (such as Yelp’s) judges unreliable and removes from the displayed rating. Researchers use filtered status as a measurable proxy for likely fraud, not as proof.
- Reputation buffer
- A deep history of genuine reviews that makes a business’s rating resistant to individual swings. Businesses without one are both more affected by each review and more likely to manipulate.
- Proxy variable
- An observable, imperfect stand-in for something that cannot be measured directly. Yelp’s review filter is a proxy for fraud because actual fraud is unobservable.
- 16 CFR Part 465
- The US Federal Trade Commission rule, effective October 21, 2024, that makes fake, incentivized, insider, and suppressed reviews federal violations with civil penalties.
Straight answers
Frequently asked questions
Which businesses are most likely to have fake reviews?
The evidence points to businesses with weak reputations (few reviews, or low or negative ratings) operating under high competitive pressure. Luca and Zervas found fraud concentrates precisely there, because those firms have the most to gain from manufacturing a reputation and the least established history to lose.
Are fake reviews illegal?
In the United States, yes. Since October 21, 2024, the FTC rule 16 CFR Part 465 prohibits fake, bought, incentivized, undisclosed insider, and selectively suppressed reviews, with civil penalties of up to $51,744 per violation.
Do fake reviews always come from competitors?
No. Competitive sabotage is one source, but Anderson and Simester found that roughly 5% of reviews came from non-purchasers who skewed negative, and that many were the business’s own loyal customers reviewing products they had not bought. Deceptive-style reviewing has several motives, not one.
Can I tell whether a competitor is faking reviews?
You usually cannot prove any single review is fake from the outside. What the research does let you do is read the conditions and patterns: crowded categories, review counts accelerating out of step with the business, tight clusters of generic five-star reviews, or vague one-star runs against a strong incumbent. Those are markers to investigate, not verdicts.
Does a higher star average mean better quality?
Less than buyers assume. de Langhe, Fernbach and Lichtenstein found average user ratings correlate weakly with independent quality scores and are often built on too few reviews to be informative, which is part of why manipulating a thin rating is so effective.
Provenance
Sources
- Luca, M. & Zervas, G., "Fake It Till You Make It: Reputation, Competition, and Yelp Review Fraud", Management Science, 62(12), 3412-3427, 2016 (established)doi.org
- Anderson, E.T. & Simester, D.I., "Reviews without a Purchase: Low Ratings, Loyal Customers, and Deception", Journal of Marketing Research, 51(3), 249-269, 2014 (established)doi.org
- Luca, M., "Reviews, Reputation, and Revenue: The Case of Yelp.com", Harvard Business School Working Paper 12-016, 2011/2016 (established)hbs.edu
- Chevalier, J.A. & Mayzlin, D., "The Effect of Word of Mouth on Sales: Online Book Reviews", Journal of Marketing Research, 43(3), 345-354, 2006 (established)doi.org
- de Langhe, B., Fernbach, P.M. & Lichtenstein, D.R., "Navigating by the Stars: Investigating the Actual and Perceived Validity of Online User Ratings", Journal of Consumer Research, 42(6), 817-833, 2016 (established)doi.org
- Federal Trade Commission, "Trade Regulation Rule on the Use of Consumer Reviews and Testimonials", 16 CFR Part 465 (effective Oct 21, 2024); FTC press release, Aug 14, 2024 (established, binding regulation)ecfr.gov
Every figure above is attributed to a real, dated source and tagged with its evidence tier. Where a claim could not be verified to a primary source, it is not stated as fact.