Fake retreat reviews are easier to spot than you think. Here's what the FTC's 2024 rule bans, the patterns that signal manipulation, and what to trust.

A retreat page with 412 reviews and a 4.9 average is not automatically telling you the truth. It might be. It might also be the result of four years of quiet editing, in which every guest who mentioned the mould in the shared bathroom got held back before publication.

Learning to spot a fake retreat review has become a practical travel skill, partly because there is serious money in this category now. The Global Wellness Institute put wellness tourism at $894 billion in 2024, roughly double the $439 billion it first measured in 2012. Money on that scale attracts people who write reviews for a living.

Here’s the useful part: since October 2024, most of what those people do is explicitly illegal in the United States.

What does the FTC rule on fake reviews actually ban?

The Federal Trade Commission’s Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, was published in the Federal Register on 22 August 2024 and took effect on 21 October 2024. It is short. It is also worth knowing, because its categories map almost exactly onto the tricks you will see on retreat booking pages.

Section 465.2 covers fake reviews. It bans writing, selling, buying or distributing a review that misrepresents that the reviewer exists, that the reviewer “used or otherwise had experience with the product, service, or business,” or the nature of that experience. A testimonial invented by a marketing agency for a Bali yoga school is covered. So is one produced by an AI tool on behalf of somebody who never went.

Section 465.4 bans conditional incentives. A business may not provide “compensation or other incentives in exchange for, or conditioned expressly or by implication on, the writing or creation of consumer reviews expressing a particular sentiment.” Read that wording carefully. Emailing every past guest and asking them to leave an honest review is fine. Offering ten percent off the next booking for a five-star review is the violation.

Section 465.6 stops a company misrepresenting that a website it controls “provides independent reviews or opinions” about the category it competes in. If a retreat group quietly owns the comparison site ranking it first, that is the provision it runs into.

Insider reviews without disclosure

Section 465.5 is the one small retreat operators break most often, usually without meaning to. Officers and managers cannot post reviews about their own business without “a clear and conspicuous disclosure” of their material relationship to it. They also cannot solicit reviews from employees or immediate relatives without making sure that relationship gets disclosed.

Most retreat centres are family businesses. The founder’s sister really did attend the September silent week and really did love it. She still has to say who she is.

Review suppression

Section 465.7 matters most to you as a reader. Paragraph (a) prohibits using an “unfounded or groundless legal threat, a physical threat, intimidation, or a public false accusation” to stop a review being written or to get one taken down. Paragraph (b) prohibits a business from creating the impression that the reviews it displays “represent most or all the reviews submitted” while it holds back the negative ones.

The carve-outs are sensible. A business can still take down reviews that are defamatory, contain personal information or trade secrets, are discriminatory, are demonstrably fake, or have nothing to do with the business at all.

Why is a page with no criticism a warning sign?

Because it is the exact pattern a regulator has already caught someone running.

In January 2022, two years before the rule existed, Fashion Nova agreed to pay $4.2 million to settle FTC allegations that it blocked negative reviews from its site. The complaint said the company’s review tool automatically posted four- and five-star reviews while holding anything lower for manual approval. From late 2015 through November 2019, the lower-rated ones were never approved. Hundreds of thousands of them sat unpublished while the site displayed an unbroken wall of praise. The FTC called it the first challenge of its kind.

Remember the mechanism, because it is the cheap one. Nobody had to write a single fake review. Somebody built a filter and let it run.

Now apply that to retreats. If a centre has hosted sixty weeks over three years and its testimonial page carries two hundred quotes, none of which mention a bad night’s sleep, cold water in the outdoor shower, a group member who dominated every sharing circle, or a facilitator having an off day, that page has been edited. Real weeks contain friction. Pages that contain none have had it removed.

Does a high review count mean a retreat is good?

Volume matters, but not in the way most people assume. It is a sample size question, not a quality signal.

Michael Luca of Harvard Business School and Georgios Zervas of Boston University studied this on Yelp, in work later published in Management Science in 2016. They found that roughly 16 percent of Yelp restaurant reviews were caught by Yelp’s own filter as suspicious, that the filtered ones tended to be more extreme in both directions, and that businesses were likeliest to game reviews when their own reputation was weak.

Worth being honest about the limits of that number. A review caught by a filter is not a review proven fake, and Yelp’s algorithm is proprietary, so the 16 percent is a proxy rather than a count. It tells you the scale of the problem is meaningful. It does not tell you precisely how big.

The travel-specific figures are firmer because the platform publishes them. In its 2025 transparency report, covering 2024, Tripadvisor said it received 31.1 million review submissions and safeguarded travellers from 2.7 million fraudulent ones. That is close to one in eleven. The single biggest category was review boosting, where an owner, a staff member or a friend posts positive reviews, at 54 percent of all fraud detected.

So the arithmetic is simple. On a listing with twelve reviews, two planted ones move the average enormously. On a listing with four hundred, they barely register. Volume dilutes manipulation. It does not prove quality, and a 4.6 across three hundred reviews should interest you far more than a 5.0 across nine.

Why does review recency matter more for retreats than hotels?

A hotel is a building with a maintenance budget and a brand standard. A retreat is usually one or two people.

The lead teacher leaves. The chef who made everyone cry with the Sunday soup goes back to Lisbon. The land changes hands and the new owners add a second cohort each week, so the group that was twelve is now twenty-six. A glowing 2021 review describing a particular facilitator’s presence is worthless if that facilitator stopped working there in 2023.

Weight the last twelve months heaviest. Treat anything older than about three years as evidence about the site and the setting rather than the experience: the food, the walk to the beach, the state of the rooms, the mosquito situation. Those things change slowly. People change fast.

And check names. If reviews from 2024 all praise a teacher who does not appear anywhere on the current retreat listing, ask the operator directly who is running your week.

What patterns signal a manipulated review page?

  • Clusters. Fourteen reviews in nine days, then nothing for seven months. Genuine retreats produce a trickle after each cohort, not bursts.
  • Reviews that echo the sales page. If three reviewers use the phrase “held space” and the homepage uses it too, somebody supplied the vocabulary.
  • Praise with no texture. People who genuinely loved something still complain about small specifics. “The 6am bell was brutal for the first two days” is a mark of authenticity that no copywriter thinks to fake.
  • A missing middle. Healthy distributions contain threes and fours. A wall of fives with a handful of furious ones often means the fives were solicited and the ones slipped through.
  • Reviewer accounts with a single review, posted the same week, all praising businesses in the same small town.
  • The operator’s tone in replies. Defensive, legalistic responses to criticism tell you plenty on their own, and an unfounded legal threat aimed at a reviewer is now unlawful under 465.7(a).

Can you rely on the FTC rule to protect you?

Not entirely, and this is where the confident version of this advice deserves pushback.

Enforcement is uneven and the agency’s own position moves. In September 2024 the FTC brought a case against Rytr, an AI writing tool that sold a testimonial and review generator, and finalised a consent order that December barring it from selling that kind of service. In December 2025 the Commission reopened and set aside its own order in a 2-0 vote, concluding that the complaint had not adequately alleged a Section 5 violation and that a categorical ban burdened AI innovation. Whatever you make of that reversal, it tells you the practical enforcement floor sits lower than the rule’s text implies.

There is a jurisdictional gap too. Plenty of retreats sit in Costa Rica, Portugal, Thailand or Peru, run by operators with no US entity. The rule reshapes the behaviour of large platforms and US-facing businesses. It does not reach a five-room plant medicine centre outside Iquitos.

Which is the argument for reading distributions rather than averages, and for using directories that publish their method instead of a number.

How we handle this at Retreat Ledger

We don’t average stars and call it a score. Ratings get weighted by recency, by how much of the guest experience a review actually describes, and by whether the reviewer’s stay can be corroborated. Retreats that show no critical feedback at all across a meaningful number of stays get flagged rather than rewarded. The full method, including what we down-weight and why, is written up at how scores work, and the checks behind a verified listing are documented separately.

Operators are welcome to disagree with a score in public. What they cannot do is have criticism removed, which is the same standard 465.7 sets, and the reason we publish the rules we hold owners to.

Next time you’re looking at a retreat page, scroll to the three-star reviews first. If there aren’t any, you have learned something.