We analyzed 120,000 Google reviews - here's what actually gets removed

Between January 2025 and March 2026 we tracked every Google review dispute filed across the BGR Review network. That is 120,318 individual disputes across 15,000+ business profiles in 42 countries. We did not sample. We logged everything.

For each dispute we recorded the original review text, the star rating, the reviewer's account age, the dispute path used, the jurisdiction, the response time from Google, and the binary outcome: removed or not removed. We also tagged each review by violation category using Google's own published content policy framework.

Of those 120,318 disputes, 16,844 reviews were removed. That is a 14.0 percent removal rate at the network level. Before you treat that number as low, understand what it means. It means roughly one in seven disputed reviews comes down when you dispute correctly. The problem is that most businesses and most agencies do not dispute correctly.

We ran a categorical analysis on every successfully removed review in the dataset. Five traits accounted for 87 percent of all removals. Everything else combined made up the remaining 13 percent. If your review does not match one of these five categories, your odds of removal drop to nearly zero regardless of how you file the dispute.

Understanding these five traits is not optional. It is the difference between spending three months chasing a review that will never come down and spending 20 minutes filing a dispute that succeeds on the first pass.

Off-topic reviews were the single largest category of successful removals in our dataset. These are reviews where the content has nothing to do with the customer's experience at the business. Someone complaining about a parking ticket they got on the street outside. A political rant. A review meant for a different business entirely.

What surprised us was how often off-topic reviews look legitimate at first glance. The reviewer might mention the business by name and include details that sound plausible. But when you read carefully, the complaint is about something the business does not control, has never offered, or that happened somewhere else entirely.

The dispute approach matters here. When we flagged off-topic reviews using the standard report button and selected 'This review is not relevant to this place,' the success rate was 19 percent. When we used the Business Redressal Form with a written explanation identifying the specific off-topic elements and attaching a screenshot, the rate jumped to 41 percent. Same reviews, different process, double the results.

Conflict-of-interest reviews were the fastest to be removed in our dataset. Median time from dispute to removal was 4.2 business days, compared to 8.7 days for off-topic and 14+ days for legal channel removals.

This category includes competitors leaving negative reviews on your profile, ex-employees settling scores, and current employees of competing businesses. It also includes reviews from people who have a financial relationship with a competitor, though proving that connection requires more documentation.

The key to success in this category is evidence. A dispute that says 'I think this reviewer works for my competitor' does nothing. A dispute that includes the reviewer's LinkedIn profile showing they are currently employed by a competing business, or a screenshot of their own Google reviews showing they reviewed three of your competitors with 5 stars, changes the outcome entirely.

In our dataset, conflict-of-interest disputes with documented evidence had a 58 percent removal rate. Without documentation, the rate dropped to 7 percent. The evidence is the strategy.

This is the category that trips up the most business owners. Not because they cannot identify defamation, but because they use the wrong channel to dispute it.

A defamatory factual claim is a statement presented as fact that is demonstrably false and causes harm. 'This company is a scam' is vague enough that Google treats it as opinion. 'This company stole 5,000 pounds from me and refused a refund' when no such transaction occurred is a defamatory factual claim.

In our dataset, defamatory factual claims disputed through the standard flag had a removal rate of 3 percent. The same reviews disputed through Google's legal content removal process had a removal rate of 47 percent. That is not a marginal difference. That is the difference between wasting your time and solving the problem.

The legal removal channel requires you to identify the specific false statement, explain why it is false, and provide supporting evidence. You do not need a court order at this stage. You do not need a solicitor. You need a clear, factual, documented submission through the right form.

Coordinated attacks are less common than the three categories above, but they have the highest first-pass success rate when reported correctly. A coordinated attack is when multiple negative reviews arrive from newly created accounts within a short window, typically 72 hours or less.

The critical mistake most businesses make is reporting these reviews individually. When Google's system evaluates each review in isolation, it sees a negative review from a real account. Nothing actionable. When you report the pattern as a whole, showing that 10 reviews arrived from accounts created this week with no other review history and no profile photos, the system recognises coordinated inauthentic behaviour.

We filed 847 coordinated attack reports during the study period. Pattern-based reports submitted through the Redressal Form had a 71 percent first-pass removal rate for the entire cluster. Individual flags on the same reviews had an 8 percent rate. That gap is entirely about how you present the evidence, not what the evidence is.

This was the most surprising finding in the dataset. Policy-violating language on its own, meaning profanity, hate speech, or threats, accounted for only 5 percent of successful removals. We expected this category to be much larger.

The reason is that Google's automated content moderation already catches the obvious cases before a business owner ever sees them. The reviews that slip through tend to contain language that is aggressive but not explicitly prohibited. Calling a business owner an idiot is rude but not a policy violation. Threatening physical harm is a violation, but it is rare in the reviews that make it past automated filters.

Where policy-violating language does matter is as a supporting factor. A review that is both off-topic and contains profanity has a higher removal rate than one that is just off-topic. A conflict-of-interest review that also includes threatening language gets faster action. On its own, though, language violations are the weakest of the five traits.

One of the most important findings in our dataset has nothing to do with review content. It is about the dispute channel. Google offers three primary paths for disputing a review, and the performance gap between them is massive.

The standard in-product flag, the one most business owners use, had a 4.1 percent overall success rate in our dataset. The Business Redressal Complaint Form had a 22.3 percent success rate. The legal content removal channel had a 38.6 percent success rate on eligible reviews. That is a nearly 10x difference between the worst and best channels.

Why does this gap exist? The in-product flag feeds into an automated triage system that evaluates the review against a set of binary rules. If the review does not trigger those rules, it is rejected within hours, often minutes. No human reviews the flag. No context is considered.

The Redressal Form puts your dispute into a queue that receives human attention. You can provide context, evidence, and a written explanation of the violation. That human review is the difference. The legal channel receives the highest level of attention because legal liability is involved, which is why it has the highest success rate for eligible reviews.

Not all industries see the same removal rates, and the reasons are structural rather than random. We segmented the dataset by industry vertical and found consistent patterns that held across geographies.

Healthcare and medical practices had the highest removal rate at 19.4 percent. This is because medical reviews are disproportionately targeted by competitors, and because patient privacy considerations create additional dispute paths. A review that discloses protected health information, even the reviewer's own, can be disputed on privacy grounds.

Legal services came second at 17.8 percent. Law firms are frequently targeted by opposing parties in litigation, which creates clear conflict-of-interest grounds for disputes. Home services (plumbers, electricians, contractors) sat at 15.2 percent, driven largely by competitor review attacks that are easier to document in local markets.

Hospitality had the lowest removal rate at 8.3 percent. Hotels and restaurants receive a high volume of legitimate negative reviews about subjective experiences, and those reviews almost never qualify for removal. The solution for hospitality is not removal. It is operational improvement and response strategy.

Retail sat at 11.7 percent and professional services (accountants, consultants, financial advisors) at 14.1 percent. The pattern is clear: industries with more competitor activity and more identifiable conflicts of interest see higher removal rates.

We measured the time between when a review was posted and when the dispute was filed. The results were clear and actionable.

Reviews disputed within the first 14 days of posting had a 16.8 percent removal rate. Reviews disputed between 15 and 90 days had a 12.1 percent rate. Reviews disputed after 90 days dropped to 9.4 percent. And reviews older than 18 months at the time of dispute had a removal rate of just 2.7 percent.

The decline is not dramatic in the first few months, but it is real. The practical takeaway is simple: dispute early. The moment you identify a review that may violate Google's policies, file the dispute. Waiting costs you nothing except success probability.

There is one exception. If Google updates its content policies, which happens roughly twice a year, reviews that were previously ineligible may become eligible. We saw a measurable spike in successful removals in April 2026 following the filter update, including reviews that had been previously rejected. A quarterly re-audit of your rejected disputes is worth the 15 minutes it takes.

This section is the one most agencies will never publish because it kills their sales pitch. But it is the most important part of this research for any business owner who wants to make honest decisions about their reputation.

Subjective complaints never come down. A review that says 'the food was bland and the waiter was rude' is an opinion. It might be unfair. It might be exaggerated. It might be completely fabricated as a subjective account. Google does not evaluate truth in opinion. They evaluate policy violations. Opinions are not violations.

Reviews about pricing, wait times, product quality, and service speed are untouchable. These are core consumer feedback categories that Google explicitly protects regardless of how you frame the dispute.

Old reviews with no policy violation are essentially permanent. If a negative review is legitimate, two years old, and does not violate any content policy, no dispute channel, escalation form, or legal strategy will remove it. The energy spent trying would be better spent generating 20 new positive reviews.

One-star reviews with no text are the most frustrating for business owners and the hardest to remove. A rating-only review violates no content policy because there is no content to evaluate. Google treats the star rating as protected expression. In our dataset, rating-only reviews had a 1.2 percent removal rate, almost entirely from cases where the reviewer's account was subsequently suspended for other violations.

One pattern we tracked specifically was re-disputes, cases where a first dispute was rejected and the business filed again through a different channel or with different evidence. This is something most businesses never do. They flag once, get rejected, and give up.

In our dataset, 23,441 reviews were disputed a second time after initial rejection. Of those, 3,892 were removed on the second attempt. That is a 16.6 percent success rate on re-disputes, which is actually higher than the overall first-attempt rate.

The reason is selection bias working in your favour. If you are re-disputing, you presumably believe the review genuinely violates a policy. And if it does, a better-documented second attempt through a higher-authority channel has a real chance of success.

The most effective re-dispute path was flag rejection followed by Redressal Form submission. Reviews that failed the in-product flag and were re-submitted through the Redressal Form with supporting evidence had a 34 percent secondary removal rate. That is one in three. For reviews you genuinely believe violate Google's policies, a second attempt is not pestering. It is process.

If you take one thing from this research, take this: removal is real, but it is narrow. It works for a specific subset of reviews through specific channels with specific evidence. It does not work for legitimate negative feedback, no matter how unfair that feedback feels.

The businesses in our network that manage their reputation most effectively use removal as one tool in a larger system. They dispute what qualifies. They respond strategically to what does not. They generate consistent positive reviews through operational excellence and smart request timing. And they monitor their profiles with the same attention they give their bank accounts.

Anyone selling you guaranteed removal of generic negative reviews is either misleading you or working outside Google's policies in ways that put your entire profile at risk. The 14 percent that comes down is the 14 percent that should come down. The other 86 percent is a different problem, solved by a different discipline.