What Likely Manipulated places look like: small, near-perfect, new data
Table of contents
- What did we measure?
- How big are Likely Manipulated places compared with Clean ones?
- Are the ratings really that high?
- Where are these places?
- Which signals sit behind the high scores?
- What should you do with this?
- Methodology note
- Frequently asked questions
- Does a high rating with few reviews mean the reviews are fake?
- What review count is safest on Google Maps?
- What does a Likely Manipulated verdict mean?
- How many Google Maps places have manipulated reviews?
Places that earn a Likely Manipulated verdict in Revdict's catalog tend to be small and nearly perfect. Among the 36 Likely Manipulated places with public reports (as of November 2026), the median review count is 378, against 3,979 for Clean places, and 17 of the 36 (47%) are rated 4.9 or higher, against 16% of Clean places. Size and rating alone prove nothing, but together they are a useful prompt to read the reviews.
What did we measure?
We pulled every analyzed place with a public report from the admin data endpoint on 9 November 2026: 898 places, of which 602 are Clean and 36 are Likely Manipulated. The wider analysis table, which includes places without a catalog page, holds 942 analyses: 626 Clean, 260 Suspicious, 19 solicitation campaign and 37 Likely Manipulated. That makes 33.5% of analyzed places non-clean, in line with the roughly 31% we reported in our October data update.
For this post we ignore the middle verdicts and compare only the two ends. A Likely Manipulated verdict means the score reached 70 or higher. In the catalog today the 36 such places score between 70 and 87. The question is simple: what do the ends of the scale look like on the surface, before you open a single review?
How big are Likely Manipulated places compared with Clean ones?
Smaller, by an order of magnitude. The median Likely Manipulated place has 378 reviews. The median Clean place has 3,979. Nineteen of the 36 Likely Manipulated places (53%) sit under 500 reviews. None has a five-figure review count.
The reason is arithmetic. A place with 20,000 reviews accumulates them over years from thousands of different people, and a short campaign of a few hundred reviews barely moves its distribution. A place with 300 reviews is different: a burst of 60 five-star reviews is a fifth of its entire history, and our scorers see it clearly. Small review pools are easier to distort and easier to measure, so both effects push the same way.
This matches what we found when we compared rating bands and review counts. The new cut here starts from the verdict and works backward, which shows the same shape from the other direction.
Are the ratings really that high?
Yes. The mean rating of the 36 Likely Manipulated places is 4.74, while the mean for Clean places is 4.54. More telling is the top of the scale: 17 of 36 are rated 4.9 or above, and so are 96 of 602 Clean places.
| Group | Places | Median reviews | Mean rating | Rated 4.9+ |
|---|---|---|---|---|
| Likely Manipulated | 36 | 378 | 4.74 | 17 (47%) |
| Clean | 602 | 3,979 | 4.54 | 96 (16%) |
Notice what the table does not say. Most places rated 4.9 or above in our catalog are Clean, because 96 is far more than 17. A very high rating is common among honest places too, especially small ones with loyal customers. The rating raises the odds; it does not decide the case.
Where are these places?
They are spread widely. The countries with the most Likely Manipulated places are the United States (7), Poland (5), Romania (4), India (4) and Indonesia (4), followed by Spain (3). With 36 places in total, those counts are small, and they partly mirror how many places we have analyzed in each country. We would not rank countries on this, which is why our country comparison uses shares and sample-size warnings rather than a league table.
One limit is worth stating plainly. In our records, 35 of the 36 Likely Manipulated places carry no category label, so we cannot say what kind of business most of them are. We also cannot treat the missing label as a finding, because many places in the catalog lack one for reasons unrelated to review quality.
Which signals sit behind the high scores?
Across all flagged reviews in the dataset, low-activity accounts are by far the most common signal: 127,456 flags, against 21,177 for bursts, 3,715 for review drives, 677 for duplicate text and 528 for generic accounts (as of November 2026). Those totals are dominated by large places with many reviews, so they describe the whole catalog, not specifically the 36 at the top.
What drives a score above 70 is usually several signals at once. A burst of reviews from accounts with almost no other history, posted within a few days, scores far higher than any of those traits alone. If you want the mechanics, how the fake-review score works walks through the weighting and what the score cannot tell you.
What should you do with this?
Treat size and rating as a reason to look closer, not as a verdict. For a small place with a near-perfect rating, open the newest reviews and sort by lowest first. Look at whether five-star reviews arrive in clumps, whether reviewers have written anything else, and whether the wording repeats. Then paste the listing into our free checker to see the same signals laid out with evidence.
Owners should read this the other way. If your place is small and highly rated, steady, organic reviews protect you. A single push to collect many reviews in a week can look like a burst even when every review is real, so spread requests out and never offer rewards in exchange for positive wording.
Methodology note
All figures come from Revdict's admin data endpoints on 9 November 2026 and cover only analyzed places, which skew toward restaurants and toward places people chose to submit. Scores flag patterns consistent with manipulation. They are not findings that any named business bought reviews, and we do not name any in this post.
Frequently asked questions
Does a high rating with few reviews mean the reviews are fake?
No. Plenty of small, well-run places have excellent ratings. In our catalog 96 Clean places are rated 4.9 or higher. It is a reason to read the reviews, not a conclusion.
What review count is safest on Google Maps?
No count is safe by itself. Larger counts are harder to move: the median Clean place in our catalog has 3,979 reviews. Still, check timing and reviewer history, since big places can be targeted too.
What does a Likely Manipulated verdict mean?
It means the place's score reached 70 or higher because several manipulation signals appeared together. It describes review patterns in the analysis, not proof about the owner's conduct.
How many Google Maps places have manipulated reviews?
In our catalog, 33.5% of 942 analyzed places were non-clean as of November 2026, and 37 (3.9%) were Likely Manipulated. Our places skew toward restaurants, so treat this as a sample, not a platform-wide rate.
References
- We're taking legal action to stop fake review scams — Google (2023)
Tags
About the author
Written by
Andrew Keymaster · Founder, RevdictAndrew Keymaster built and runs Revdict, a free forensic checker for Google Maps reviews. Its scoring model has examined 420,533 public reviews across 443 places as of August 2026, flagging the patterns that show up in manipulated review records — same-day five-star bursts, throwaway reviewer accounts, and near-duplicate wording. He writes here about how those patterns are detected, what platforms will and won't act on, and what the data says about how common fake reviews actually are.