What Share of Places Have Manipulated Reviews? Our Data, July 2026

A third of the Google Maps places in our catalog carry at least one manipulation signal. Precisely: as of July 2026, 33.5% of 242 analyzed places show one or more, and 12.4% cross the Likely Manipulated threshold. Underneath those percentages sit 241,041 individual reviews, 49,047 of which drew a flag: 20.3% of everything read.

This is the first release from our own data, and future ones will be dated the same way. Every number comes from completed analyses rather than a survey or an estimate.

#How many analyzed places show manipulation signals?

Two places in three come back Clean, which is worth saying plainly: most review histories look like what they claim to be. A third show enough clustering to reach Suspicious, and about one in eight cross into the higher band. Average score across the catalog: 32.1 out of 100.

#What does the score distribution look like?

The histogram has two humps rather than a curve. The 10-19 band holds 120 of 242 places, half the catalog, which is what an ordinary background rate of thin reviewers and coincidental clusters looks like. A second group sits in the 60-79 range, 54 places, separated from the bulk by a thin middle.

Two humps with a gap between them is the shape you get when manipulation is something a minority of businesses decide to do, rather than a drift every listing slides into over time. What feeds an individual score is laid out in the scoring explainer.

#Which manipulation signal is most common?

Reviewer profiles holding one or two lifetime reviews outnumber the other two markers combined. That matches how the trade works. Rewriting a review costs a vendor nothing; building accounts with years of believable history costs real time and money, so the accounts stay thin and the accounts give the campaign away. The field guide shows how to check each marker by hand.

#What's driving review manipulation right now?

Federal enforcement has finally started moving. Since October 2024 the Consumer Review Rule has prohibited paying for reviews in either direction, generating them with a language model, and hiding genuine criticism, at up to $53,088 a violation. Our breakdown covers the five prohibitions and the cases closed so far.

The first enforcement step came on December 22, 2025, when the agency sent warning letters to ten unnamed companies. That was 14 months after the rule took effect, and two settlements followed within seven months. Any small business weighing a review vendor's pitch should read that sequence as the risk changing.

#Which categories are represented in the dataset?

Food and hospitality dominate, because that is where analysis requests have come from. Restaurants lead with 28 places, then hamburger restaurants (15), American restaurants (14), and pizza restaurants (10), with hotels, bakeries, and bars also present. Whether the manipulation rate holds steady across other trades and other countries is a question for a larger catalog. Treat the category picture as early signal.

#Methodology note

Every figure above comes from analyses completed by July 2026: public Google review text, ratings, timestamps, and reviewer profile metadata, scored with the four-signal model. Verdicts use each place's most recent completed analysis, so a business analyzed twice counts once. The numbers will move as the catalog grows, and every future release will carry its own date so the trend stays traceable. Coverage is Google Maps only; how the three big platforms police reviews covers what happens elsewhere.

#Frequently asked questions

#Is one-third of all Google Maps listings manipulated?

No. It is one-third of this catalog, which leans heavily toward US food and hospitality. Reading it as a rate for all of Google Maps would require a representative sample nobody has.

#Has the manipulation rate changed over time?

Unknown, since this is the first release. Later ones will compare against it.

#What counts as "Likely Manipulated" versus "Suspicious"?

Sixty or higher out of 100 is Likely Manipulated; 30 to 59 is Suspicious. Both come from the same four weighted signals at different strengths.

#Why are low-activity accounts so much more common than duplicate text as a signal?

Because a review-selling operation needs accounts, and cheap accounts have no history. Varying the wording, by contrast, takes a vendor almost no effort at all.

Any Maps listing can be checked against this same model at our analyzer, and each analysis submitted grows the dataset behind these numbers.