Yelp Reviews

Not recommended is not removed

Start here, because this one distinction causes more wasted money on Yelp than everything else on this page combined. Yelp runs two separate systems. The Content Guidelines decide what is removed. Automated software, which Yelp calls its recommendation software, decides what is shown and counted. They are not the same mechanism, they do not produce the same outcome, and published advice conflates them constantly.

When a review disappears from a business page, it has almost always been moved into the not-recommended set rather than deleted. Yelp describes what that means in its own words: the remaining reviews "don't disappear; they can still be viewed by clicking 'reviews that are not currently recommended' below the recommended reviews on business pages," they "can also be seen on a reviewer's profile page," and they "don't factor into a business's overall star rating or review count."

So a not-recommended review is still published, still publicly reachable from a link on your own business page, still on the reviewer's profile, and still readable, quotable and available to screenshot by any customer, journalist, investor or opposing party who spends ten seconds looking. The only thing that changed is that it stopped counting toward the star rating.

That is not removal and it should not be sold to you as removal. An owner told "we got that review taken down," who then finds the star rating unchanged, has been sold a not-recommended outcome or has been sold nothing. Before paying anyone for a Yelp result, ask what success looks like on the page, then click the not-recommended link yourself and check.

The scale of the difference shows in Yelp's own reporting. In its 2025 Trust & Safety Report, announced 25 February 2026, Yelp said roughly 17 percent of about 22 million contributions were not recommended, while about 193,700 reviews were removed after being reported and about 46,200 proactively. The filtered set is an order of magnitude larger than the removed set.

Nobody at Yelp can move it for you

The second thing to know is who has discretion over that software. The answer, in the operator's words, is nobody.

"because our recommendation software is completely automated, no Yelp employee can manually override the software to recommend or not recommend a review. It applies the same objective rules to every business and treats reviews of advertisers and non-advertisers exactly the same."

— Yelp Support Center, Why would a review not be recommended?, read 12 August 2026

Two sales pitches die on that sentence. The first is the vendor with a contact inside Yelp who will quietly get your bad reviews filtered. No human at Yelp is described as having that power. The second is the mirror image, and more owners believe it: your good reviews were buried because you stopped advertising, and a vendor can bring them back. Yelp's denial is explicit, and that theory has already been litigated and lost.

Yelp does publish the signals the software reacts to: a review from a less active user, one that reads as an unhelpful rant, reviews originating from the same IP address, reviews suggesting an unfair bias such as one written by a friend of the owner, and users it does not know enough about to weigh. None of that is a lever you can pull. It describes a filter, not a form you can file.

Which leaves exactly one route that produces an actual removal, and it is the Content Guidelines.

What the Content Guidelines actually reach

Yelp's Content Guidelines are the document to argue from, and several of the grounds in them are decidable from the face of the review. Those are the ones worth your time.

"Reviews aren't the place for rants about political ideologies, a business's employment practices, extraordinary circumstances, or other matters that don't address the core of the consumer experience."

— Yelp, Content Guidelines, read 12 August 2026

That relevance rule is more useful than it looks. A review about how a business treats its staff — rather than how it treated the reviewer as a customer — is off topic and reportable on that ground alone, with no argument about whether the underlying complaint is true. Employment grievances land on consumer review pages constantly, and Yelp has written them out of scope in its own text. Reviews also have to be posted to the page for the specific location the reviewer visited or contacted.

The other grounds that decide most business matters:

• Personal experience. "We want to hear about your firsthand consumer experience, not what you heard from your partner or co-worker, or what you saw in the news or on social media." A review repeating a news story or a friend's account is not a review of a visit.
• Conflicts of interest. Reviewers "shouldn't write reviews of your own business or employer, your friends' or relatives' business, your peers or competitors in your industry, or businesses in your networking group." A competitor review breaks that on its face.
• Demanding payment. "You should not threaten to post or offer to remove a negative review as a way to extract payment from a business." That is the rule for a reviewer running a shakedown.
• Privacy. Full names are not allowed unless the person is commonly referred to by their full name, so a review naming a junior employee is reportable on that alone.
• Review updates. Updates "should reflect a new experience or interaction," a reviewer can edit within 30 days of posting, and repeat posts about the same experience, or the same text across multiple business pages, are outside the rules.

The rule about AI-written reviews, and why it is broader than fake ones

One provision in the current guidelines deserves separating out, because it is new enough that most published advice does not mention it and broad enough to reach reviews that are otherwise genuine.

Yelp's rule is that "you shouldn't use third-party AI tools and chatbots to create reviews or other content, including using such tools to draft or revise content." Read the clause after the comma. It does not only bar a wholly fabricated review generated by a chatbot. It reaches a real customer's real complaint that was drafted or tidied up by an AI tool before posting.

For a business facing a coordinated attack, that matters. Machine-drafted text tends to share structure, rhythm and vocabulary across a set in ways unconnected customers do not. The case to make is not "these feel fake." It is a set of reviews presented together, with the timing, the account histories and the shared characteristics laid out, measured against a rule Yelp has written down.

Two cautions. Detection tools that claim to score text as AI-generated are not evidence, and I would not build a report around one. And the rule cuts at your own review drive too: content your business drafted or polished for a customer sits inside the same prohibition.

The banner a heavy-handed letter can earn

Yelp does something no other operator in this group does. It publishes a warning banner over the reviews on a business's own page when it concludes certain things about the business, and one of the six alert types is triggered by the business's own conduct in trying to deal with a reviewer.

Questionable Legal Threats alert — placed on "evidence that a business may be abusing the legal system to intimidate or silence a reviewer." Yelp adds that "reviewers have a First Amendment right to honestly describe their experiences with a business," and says consumers should know when a business tries to deny that right.

— Yelp, Consumer Alerts, read 12 August 2026

The other five are Compensated Activity, for evidence that cash or incentives were offered for a review; Suspicious Review Activity, for large numbers of reviews from a single IP address or from users connected to a group coordinating incentivized fake reviews; Unusual Activity, for a spike, where Yelp temporarily disables posting; Public Attention; and Business Accused of Racist Behavior, which Yelp says appears only where a credible media organization has reported the incident and there is clear factual evidence. Compensated Activity and Questionable Legal Threats alerts are published in the United States only.

Yelp placed six Questionable Legal Threat Alerts in 2025 by its own count, so it is uncommon. But the exposure is real, it is public, it sits on your page rather than the reviewer's, and it is the exact opposite of what the letter was sent to achieve. Yelp also states that pressuring or offering payment to change or remove reviews violates its guidelines and may violate the Federal Trade Commission's endorsement guides. Since 21 October 2024, 16 CFR § 465.7 has prohibited using an unfounded or groundless legal threat to get a review removed.

If you have already sent a demand letter about a Yelp review, read that again before you send another one.

The rule pointed at the business, not the reviewer

Yelp is stricter than the other operators here on one point, and businesses trip over it constantly because the advice they read was written for Google.

Yelp's guidelines state: "Businesses should never ask customers to write reviews." Not "never incentivize reviews" — never ask. Google and Glassdoor both permit soliciting unincentivized reviews. Yelp does not, and a review campaign that is perfectly ordinary on one platform is against the rules on this one. Yelp also tells businesses not to use direct messages or public comments to launch personal attacks or to offer an incentive to change a review.

Two consequences follow. The standard advice to fix a rating by generating more reviews does not transfer to Yelp. And a business that ran a solicitation campaign has created a record that can surface later, which is a weak place to argue from when claiming someone else's reviews are inauthentic.

What is left is not nothing: report the reviews that break a written rule, keep the reporting factual and specific, and put the effort you would have spent on removal into a public response the next customer will read.

What will not come down, and what there is no appeal for

The honest list, since nobody selling Yelp services will give you one.

• A review at the business's request. There is no request route. There is only a report against the Content Guidelines, decided by Yelp.
• A not-recommended review. It stays published and publicly linked from your own page, and the reverse job is equally impossible: a filtered positive review cannot be promoted back into the recommended set, because Yelp says no employee can override the software. Any vendor selling that is selling something the operator says cannot be done.
• A bad review from a real customer who had a bad time. Opinion, however brutal, is not a guidelines violation.
• A Consumer Alert banner, on demand. Once placed, it is Yelp's decision, and pursuing the reviewer legally is itself a documented trigger for one of them.
• A decision, on appeal. No formal named appeal process appears on any Yelp policy page I read on 12 August 2026. Removal and recommendation decisions are stated to be final, in sharp contrast with Google's published one-time appeal.

Whether the words in a review are legally actionable is a question for a defamation attorney, and it is separate from all of the above. Yelp is not deciding that and never offers to. Policy here moves — the AI provision is recent and the alert program changes — so read the current text on Yelp's own site before acting on any of this. And preserve what is on your page now: the evidence of a coordinated campaign is the pattern across accounts, and that pattern disappears when the accounts do.

Frequently Asked Questions

Published 2026-08-12