Google deleted 292M reviews. Here is the 2026 impact on your business.

On April 16, 2026, Google published its yearly Maps protection update with the 2025 Trust and Safety figures. One number from it went around every marketing newsletter within a day: more than 292 million policy-violating reviews. Several of those newsletters called it "deleted reviews", and that is where the story starts to drift away from what Google actually said.

This page does three things. It sets out the figures from Google's own posts, with the history behind them. It explains what the number counts and what it does not. And it turns the 2026 changes into a short list of things a business owner should check this week. We handle review removal for a living, so we read these reports as practitioners, and where the data runs out we say so.

What did Google actually report for 2025?

Google's April 16, 2026 post says its systems and analysts blocked or removed more than 292 million policy-violating reviews in 2025, while people shared more than 1 billion reviews and suggested 80 million updates to business information. The same post gives the figures for edits, fake profiles and restricted accounts. Set against the 2024 figures from Google's February 2025 post, the picture looks like this:

Google rounds its figures ("more than", "over"), so treat the change column as approximate. The reviews line is the one to remember: 240 to 292 is a rise of about 22%. The restricted-accounts line is worth a second look too. Google restricted fewer accounts than the year before while blocking more reviews, which suggests more of the work now happens at the review level rather than by banning the account behind it.

Did Google really delete 292 million reviews?

No, and the difference is the whole point. Google's wording is "blocked or removed", and its 2024 post added that the vast majority of that year's total was removed before anyone saw it. The 2025 post does not give a split. The reasonable reading is that the same pattern holds, but that is an inference. Google did not publish that number.

So 292 million measures what the classifiers stopped at the door and what they pulled down later, mostly spam, bots, copy-pasted text and review farms. It does not measure a cleanup of live business profiles, and it says nothing about how many of the reviews you can see today are fair.

A rough sense of scale helps. Against more than 1 billion reviews shared in 2025, 292 million works out to about 22% of review volume (292 divided by 1,292), if the billion counts reviews that stayed live. Google publishes neither a precise denominator nor a false-positive rate, so read that as an order of magnitude: somewhere around one submitted review in five never survives.

How fast is the number growing?

The figure has roughly tripled in four years, from 95 million in 2021 to 292 million in 2025. The pace is what changed. Growth ran at +48% in 2023 and +41% in 2024, then slowed to +22% in 2025.

Two readings fit the data. Attackers may have adapted, so fewer low-effort fakes reach the filters. Or the filters may already catch most of what they can recognise, so each extra million costs more. The published numbers cannot separate the two. Our reading is that both are happening, and that what is left on the profile after the filters are the reviews built to look like real customers.

If you want a back-of-envelope projection, growth of 13% to 22% would put next year's figure between roughly 330 and 355 million. That is our arithmetic and Google has made no such forecast, so it will be wrong in an interesting way if the 2026 changes bite.

What did Google change in 2026?

The April update added four mechanisms, plus one regional change worth knowing about:

• Scam-pattern detection. Google upgraded its systems to spot patterns such as people demanding payment to remove fake one-star reviews, and to stop suspicious posts before they go live. The rollout is global and was announced as starting in the weeks after April 16.
• Review pauses during spam spikes. If a profile gets a sudden wave of spam reviews, Google says it will remove the fake content, pause new reviews on that profile, alert the owner and show consumers a banner explaining the pause.
• Gemini screening of edits. Gemini models now screen suggested edits to place information and block those that break policy. This is already rolling out on Android, iOS and desktop.
• Email alerts for owners. From April 2026, verified and active Business Profile owners get proactive emails so they can review important edits before they go live.
• Germany: a public removal count. PPC Land reported that Google began showing a notice on some German profiles with the number of reviews removed in the past 365 days after defamation complaints. Per the report, it counts only valid complaints under German law that were not reversed on appeal, and it does not affect ranking. Reports so far limit it to Germany.

Read together, these shift Google from reacting to individual flags toward intervening at the profile level. That helps most when a whole wave arrives at once, which is the situation we cover in the review bombing playbook, and when the wave comes with a ransom note, covered in the extortion guide.

What does this mean for your business?

Obvious fakes reach you less often. Bot waves, copy-pasted text and extortion bursts are exactly what the new detection targets. If you have been hit by one of those, the exposure window should now be shorter than it was a year ago.

The review on your profile is, almost by definition, one the filters let through. What survives looks like real customer writing from a real-looking account: an ex-employee, a competitor, someone who never bought anything, a billing dispute that turned into a verdict. Google's systems cannot check your customer records, so they cannot tell that the reviewer was never a customer. That is why "no policy violation found" is such a common reply to a flag, and why the pillar guide puts DIY flagging success at around 30%.

False positives are the part nobody publishes. Google gives no rate for legitimate reviews caught by the filters. If a genuine review of yours disappears, common triggers include links, phone numbers and off-topic text, and the usual first step is asking the customer to repost without them. Do not ask the customer to change the rating or the sentiment, which is covered below.

A pause on your profile is a containment measure. If new reviews stop appearing and a banner appears, Google has detected a spike and is containing it. Use the pause to flag and document the wave. Do not fill the gap by buying reviews.

Vendor claims need an attribution test. When Google itself removes hundreds of millions of reviews a year, a vendor that says "we removed 5,000 reviews" may simply be counting reviews Google would have removed anyway. Ask whether a review was live on the profile when the case started and confirmed gone afterwards. We invoice only after Google confirms removal, which is the only test that means anything. Our due diligence checklist covers the rest of the questions.

What are the legal stakes around reviews in 2026?

Enforcement against fake reviews has moved from Google's policy team to regulators:

• United States. The FTC's Consumer Review Rule has applied since October 21, 2024. It bans writing, buying or selling fake reviews, incentives conditioned on a particular sentiment, and review suppression through false accusations, intimidation or unfounded legal threats. Press coverage of the FTC's warning letters to ten companies cites civil penalties of up to $53,088 per violation.
• United Kingdom. The Digital Markets, Competition and Consumers Act has banned fake and concealed incentivised reviews since April 6, 2025. According to CMS, the CMA can fine up to £300,000 or 10% of global turnover, whichever is higher.

Two practical points follow. First, never buy reviews to compensate for a bad one, and never gate reviews by sentiment. Second, asking Google to remove a review that breaks its policies is a different thing from threatening the reviewer, and only the second one looks like suppression. Our method is a proprietary, direct removal process. We do not use account compromise or fake reporting networks, and we tell clients to avoid any provider that does.

What should you do this week?

1. Check your profile for a pause banner and make sure owner emails reach you. Alerts about edits go to verified, active owners, so confirm the owner email on the profile is one you read.
2. Sort reviews by newest and look for clusters. Several one-stars within hours, new accounts with no history, and vague complaints that never name a product or a staff member are the usual signature.
3. Screenshot first, flag second. Capture each review with its date and the reviewer profile, then flag it under the policy category that actually fits. A flag with the wrong category is the most common reason for a quick "no violation found".
4. If anyone asks for money to remove reviews, stop and use the extortion form. Google's merchant extortion report form needs the demand and the screenshots. The full sequence is in the extortion playbook.
5. Do not buy, trade or sentiment-gate reviews. Under the FTC rule and the UK Act, that is the expensive mistake.
6. If the flag comes back "no violation found", escalate. The denied request guide sets out the three-step path, and a professional service is the fastest route for a review the classifiers have already waved through.

What can Google still not fix?

The most expensive review is the plausible, specific one that is false in a way only you can prove: the complaint about a visit that never happened, the "scam" accusation from someone with no purchase on record, the review of the wrong branch. Google cannot see your appointment book, and a classifier reading the text sees nothing wrong with it.

That is the gap the 292 million figure does not close, and it is the gap our work sits in. One negative review costs a typical local business thousands of dollars a month, so the fact that Google is filtering more spam in aggregate is little comfort to the owner looking at one specific review.

If you know a business owner or an agency client stuck in that spot, our referral program pays 10% of what they pay us, and the terms are short and public.