The Reviews Were Written by People Who Never Left

Posted on Jun 8, 2026 by leobro · 3686 views · ~2 min read

In one sentence: A data analyst digs into a remote rental’s glowing reviews and finds a creepy pattern that makes her cancel before arriving.

Something felt off about our host, so I read every review — properly. A pattern emerged: a cluster of five-star guests, all new accounts, all posting once, all using the same odd phrase. Then the police asked me the same question.

The Reviews Were Written by People Who Never Left
* AI Illustration

First impression: judge by the title and the picture alone — you get the real verdict after reading.

📄 Case Report — share this story

I'm a data analyst by trade, which is both why I noticed and why I can't unsee it. The rental was fine on the surface — remote, beautiful, cheap for what it was — but the host set off a low hum of wrongness in me from the first message: too eager, too controlling about arrival times, weirdly insistent we 'let him know the moment we arrive and the moment we leave the property each day.' Nothing you could report. Just a hum.

So I did what I do with a hum: I pulled the data. I read every review on the listing — not skimmed, read, all hundred-odd, in order, as a dataset. And a pattern surfaced that made the hum into a siren.

There was a cluster. Interspersed among normal-looking reviews from normal-looking accounts were maybe fifteen five-star reviews that shared a signature: all from accounts created shortly before the review, all with no profile photo or a stock one, all with only that single review ever posted, all clustered in date-groups, and — the detail that iced me — all using a slightly odd, slightly formal phrase that no real guest writes: 'a very safe and private stay.' Fifteen times, that exact phrase, from fifteen ghost accounts, salted among the real reviews to pad the rating.

Fake reviews are common and usually just mean a host is gaming the algorithm. But 'a very safe and private stay,' fifteen times, from accounts that posted once and vanished, for a remote property run by a man obsessive about knowing when his guests came and went — that specific combination didn't read like rating fraud. It read like camouflage. Like a listing engineered to look trustworthy for a purpose that had nothing to do with hospitality.

We didn't stay. I cancelled from the road, ate the cost, drove to a chain hotel, and — because I couldn't shake it — I compiled the pattern into a document, the ghost accounts, the dates, the repeated phrase, and I sent it to the platform's trust-and-safety team and, on a friend's advice, to the local police non-emergency line.

Weeks later a detective called me. I assumed to thank me for a tip. Instead, near the end, she asked me a question that rearranged my week: had we noticed anyone following us on the drive to the property, or any indication our arrival time had been shared with anyone else. She wouldn't explain. She just said the listing was 'of interest to us' and that my document had 'joined some dots,' and that I should keep doing whatever made me read a hundred reviews as a dataset.

I don't know the full story. There's an investigation; the listing is gone; the ghost reviews are gone with it. But I know that 'a very safe and private stay,' repeated fifteen times by people who were never really there, is the most frightening sentence I have ever found in a spreadsheet, and I read reviews very, very differently now.
Data analyst spreadsheet review
Remote rental cabin exterior
Creepy figure in woods
Fake online profiles screen
Analyst’s night work desk
Following car road view
Detective’s investigation file
Chain hotel room interior

AI illustrations. Please judge the story itself. About AI images.

Filed by AI — launch archive.

Do you trust this story?

Before you peek: what share of readers believes this story?

50%

60% of readers believe this story

Total verdict votes: 97

Guess first! What other readers decided is hidden until you cast your own verdict above — vote True, Fake or Not sure to reveal the community score.

Comments (3) Add a comment

True or Fake? — play the detective

How it works: cast your True/Fake verdict above and see whether you matched the reader majority. Then hit "Next case" for a random story and keep your streak going. Solve cases to climb the detective ranks — score and rank are kept on this device only.

Next case »

Tip: highlight any sentence of the story with your mouse or finger — a small share popup will appear to post that exact quote (with a link back to the story) to X, Facebook, Threads or Reddit.

Something similar happened to you? Tell your version — the jury is waiting. Tell your story

Comments (3)

Comments here are case materials: pick a role — ask the author a question, or file your evidence FOR or AGAINST the story — and it lands in the right column. Upvote the sharpest comments: the best one is pinned on top as the Top Detective Comment. Questions get answered by the story's author in person — look for the AUTHOR badge.

#434 trust_and_safety · Verdict: True · Jun 9, 2026

I've worked platform trust-and-safety. Clusters of single-review, new, photoless accounts using an identical formal phrase is a textbook fabricated-review signature — we flag exactly that.

#435 read_them_as_a_dataset · Jun 12, 2026

'The most frightening sentence I have ever found in a spreadsheet.' Fifteen ghosts all promising a 'safe and private stay' as camouflage. The instinct to read a hundred reviews as data instead of vibes literally may have kept you out of something.

#436 verify_the_reviewers · Jun 15, 2026

Practical takeaway I'm stealing: don't just read the star rating, click the REVIEWERS. New accounts, no history, one review, repeated phrasing = manufactured trust.

New Stories

Top True By Votes

Random Stories