Duplicate Text and Generic Accounts: The Rarest (and Most Telling) Fake-Review Signals

Duplicate text and generic accounts are our two rarest fake-review signals, and also our most direct evidence of manufactured reviews. As of September 2026, they account for just 283 and 526 of the 92,359 flagged reviews in our dataset, a fraction of what low-activity accounts or bursts catch, but each one is close to unambiguous: a review whose wording matches another review of the same place almost word for word, or an account with no name, no photo, and no history beyond the single review it left.

#What is the duplicate-text signal?

Duplicate text means a review's wording is nearly identical to another review posted about the same place, often within days of it. A scorer that finds this treats it as strong evidence the review wasn't written independently, since two customers spontaneously describing an unrelated visit in matching phrasing is not something that happens by chance at any real frequency.

The signal is deliberately narrow. It doesn't fire on reviews that share a sentiment ("great food, friendly staff") or a common phrase pattern that real people happen to use often. It requires substantial overlap in the actual sentence structure and wording, the kind that shows up when a template gets handed to reviewers, or when the same author (or a small group coordinating) posts under different names. That narrowness is why the count is so low relative to other signals: it's built to avoid false positives, not to catch every low-effort review.

Academic work on fake-review detection has approached the same problem from the text-similarity side. One widely cited method treats review-to-review similarity as a feature and tests term-weighting schemes like BM25 to measure how much two reviews of the same subject overlap beyond what topic alone would explain.¹ The mechanics differ across systems, but the underlying logic is the same one duplicate text uses: independent, honest reviews of a real experience don't converge on the same sentence twice.

#What is the generic-account signal?

A generic account is a reviewer profile stripped of everything that would make it identifiable: no real name pattern, no profile photo, minimal or no review history beyond the flagged post. It's the account equivalent of a burner phone. On its own, a single generic account isn't proof of anything; some real people genuinely keep sparse, anonymous profiles. It becomes evidence when several generic accounts post about the same place in a short window, especially alongside other signals like a burst or duplicate wording.

Generic account is more common than duplicate text (526 flagged reviews against 283) because it's cheaper to fake at scale: creating a throwaway account with no identifying detail takes seconds, while writing genuinely distinct wording for dozens of reviews takes effort most low-budget campaigns skip. That's also why it rarely appears alone in a place's evidence. It's most often one line item in a broader pattern, alongside low-activity accounts and burst timing, rather than the deciding factor by itself.

#Why are these the rarest signals we track?

Because they require the reviewer to fail in the most careless way possible: reusing the same words, or leaving no trace of identity at all. Most manipulation campaigns we see are more careful than that. Review drives and low-activity-account patterns rely on real people (staff, friends, customers offered a small reward) writing in their own words from accounts that look ordinary, which produces volume and timing anomalies without any duplicated text. Duplicate text and generic accounts catch a narrower, less sophisticated slice: campaigns run by whoever copy-pasted a script, or bought a batch of throwaway accounts, without bothering to vary either one.

A review flagged for duplicate text or a generic account is the closest our scoring gets to catching a manipulation attempt in the act, rather than inferring one from timing or account age.

#What should you do if you spot this pattern yourself?

Read a handful of the most recent five-star reviews on a listing back to back. If two use the same unusual phrase, the same typo, or describe the same specific detail in near-identical wording, that's the pattern by hand. Check the reviewer profiles next: a wave of accounts with generic names, no photo, and a single review each is the second half of the same picture. Neither check takes more than a couple of minutes, and both are things anyone can do before booking, without needing a scoring tool at all.

Timing helps too. Open the reviewer's profile and check the post date against other reviews using the same phrasing; matching wording posted within a day or two of each other is a stronger tell than the same phrasing scattered across months, since a genuine repeat customer might reuse their own turn of phrase over time in a way a coordinated batch won't.

For a business owner on the other end of this, the fix isn't to police your own customers' wording. It's to make sure any review request you send doesn't hand people a script to copy from in the first place; a request that just says "tell us what you thought" produces reviews with nothing in common to flag.

Our guide to spotting fake Google reviews covers the broader checklist this fits into, and the review-bursts explainer walks through the timing signal these two most often appear alongside. To check a specific listing rather than reading it manually, paste the Google Maps link into our free analyzer and it runs all five checks at once.

#Frequently asked questions

#Can two real customers write nearly identical reviews by coincidence?

It's possible for short, generic praise ("great service!") to overlap by chance, which is why the duplicate-text signal requires substantial sentence-level overlap, not shared sentiment. Two independent, detailed reviews matching almost word for word is not something coincidence explains at scale.

#Is a generic reviewer account always fake?

No. Some real people keep anonymous profiles with no photo or history. A single generic account proves little; several posting about the same place in a short window, especially alongside a burst or duplicate wording, is what turns it into evidence.

#Why are duplicate text and generic accounts rarer than other signals?

Because they require the reviewer or campaign to be careless in a specific way: reusing wording, or leaving zero identifying detail. Most manipulation is more careful than that, which is why review drives and low-activity accounts, run by real people in their own words, are far more common in our data.

#Does a low duplicate-text count mean fake reviews are rare?

No. It means this specific, narrow pattern is rare. Other signals (low-activity accounts, bursts, review drives) catch far more manipulation; duplicate text is the smallest but highest-confidence slice of the total.