Executive Summary
Google's AI-powered enforcement is reshaping how healthcare practices build and maintain their online reputation. In 2025 alone, the company removed or blocked 292 million reviews, including legitimate patient feedback that many practices depend on to attract new patients.
This report draws on three sources: Google's own published Transparency Report and Trust and Safety data, an independent analysis of 335,520 deleted reviews, and RepuGen's 2026 Healthcare Behavioral Study of 647 patients. Together, they explain what is happening, why healthcare is disproportionately affected, and what practices can do about it.
Four findings anchor this report. Enforcement is accelerating year over year, with no year on record in which Google removed fewer reviews than the year before. 89.1% of deleted reviews carry five-star ratings, meaning the reviews practices work hardest to earn are the ones most exposed to removal. Healthcare practices face a structurally higher moderation bar than other industries because Google classifies regulated, consumer-safety-sensitive categories as such. And the practices that remain unaffected are not doing anything extraordinary — they are running a consistent, compliant review program.
This report gives the reader a clear understanding of the six primary removal triggers behind this pattern and a practical framework for building a review profile that withstands Google's enforcement.
Key Findings
- 292 million reviews blocked or removed by Google in 2025, up from 170 million in 2023.
- 221 million of those removals came from already-live Business Profiles, as confirmed by Google's Transparency Report.
- Google's enforcement now runs on Gemini AI, which evaluates reviews before publication and retroactively re-evaluates previously approved reviews.
- Medical services are explicitly identified as a restricted, higher scrutiny category in Google's content moderation documentation.
- Google's April 2026 policy update introduced new enforcement mechanisms affecting solicitation, shared devices, and extortion detection.
Sources: Google Maps Trust and Safety Report 2025 (published April 2026); Google Transparency Report, Maps Content Protections; independent analysis of 335,520 deleted reviews (Localo, 2026).
The Problem, Scope, and Context
Online reviews are the primary trust signal patients use when choosing a healthcare provider. RepuGen's 2026 Healthcare Behavioral Study found that approximately 89% of patients check online reviews before choosing a provider. When those reviews disappear without explanation, a practice's ability to attract new patients is directly and measurably affected.
Google's own data establishes the scale of the problem. Google's April 2026 blog post reports 292 million reviews blocked or removed in 2025. Google's Transparency Report confirms that 221 million of those were removed from already live profiles. The distinction matters: of the 292 million reviews, 71 million were blocked before publication, while 221 million went live and were then taken down. Both categories matter because one represents reviews a business never received, and the other represents reviews a business lost.
This is not a one-year event. Google removed or blocked approximately 170 million reviews in 2023, 240 million in 2024, and 292 million in 2025. Enforcement is not stabilizing — it is accelerating year over year, according to coverage from Search Engine Roundtable.
The mechanism behind this acceleration is Gemini. Google's enforcement now runs on Gemini AI, which evaluates reviews before they are published and re-evaluates previously approved reviews on an ongoing basis. There is no point at which a review is permanently safe from removal, according to reporting from Search Engine Journal.
| Year | Reviews Removed or Blocked | Year-over-Year Change |
|---|---|---|
| 2023 | ~170 million | — |
| 2024 | ~240 million | +41% |
| 2025 | 292 million | +22% |
Why Legitimate Reviews Are Being Removed, the False Positive Problem
False positives are not a fringe issue. They have been documented during fluctuations in system-wide enforcement. In 2023, Google acknowledged that its automated protections incorrectly removed a subset of policy-abiding reviews from Local Guides and indicated that affected reviews would be restored following system corrections, an event covered by Search Engine Land. If this occurred at a visible scale once, similar system-level effects can occur in less visible forms continuously.
The clearest evidence of the problem is the distribution of ratings for what gets removed. An independent analysis of 335,520 deleted Google reviews across 22,292 Business Profiles found that 89.1% of all deleted reviews carried 5-star ratings. This is not a minor statistical artifact. It reflects a structural bias in Google's moderation toward removing the highest-rated content, the content that a practice worked hardest to earn. It is also the content businesses have the strongest incentive to fake: five-star reviews inflate a rating the most, so they make up the largest share of both genuine enthusiasm and coordinated manipulation. That overlap, not a deliberate targeting of honest practices, is what drives the pattern below.
The mechanism is straightforward once the pattern is isolated. Google's machine learning models are trained to identify patterns associated with coordinated fake review campaigns: short text, maximum rating, low reviewer posting history, and no verified location signal. That pattern signature is algorithmically similar to the review a genuine, satisfied patient is most likely to leave. Medical experiences are often personal, and many patients are reluctant to discuss health details in a public forum, so they default to a brief, rating-only response rather than a detailed account, the same shape a coordinated fake campaign produces. The algorithm evaluates pattern, not intent, and authenticity is simply invisible to the model.
Healthcare experiences an amplified version of this problem. Google applies heightened scrutiny to regulated categories where misleading content poses a direct risk to consumer safety, and medical services are explicitly identified as such in Google's content moderation documentation. The same short, 5-star review that clears moderation for a restaurant faces a materially harder bar for a medical practice.
Practices with only a handful of reviews feel every removal acutely, since losing even a few reviews can meaningfully shift their visible rating. Practices that run a steady, compliant review-generation program are best positioned to absorb individual losses without affecting their overall reputation.
89.1%
of deleted 5-star reviews carried over
Based on an independent analysis of 335,520 removed Google reviews across 22,292 Business Profiles.
Six Primary Removal Triggers, Analysis, and Implications
Although Google's moderation systems are designed to protect consumers from spam and manipulation, the same signals can unintentionally affect legitimate healthcare reviews. Understanding these triggers helps practices reduce unnecessary exposure.
Understanding what specifically triggers removal enables practices to audit their current program and identify exposure before enforcement occurs. The six triggers below share a common underlying logic: Google's AI systems evaluate behavioral and structural signals rather than the review's underlying intent or authenticity.
1. Review Velocity Spikes
A sudden surge of reviews within a short window is a primary indicator that Google's systems monitor for coordinated fake-review activity. In healthcare, this commonly occurs when a practice runs an unmanaged mass outreach campaign to a large historical patient list all at once. This well-intentioned effort creates an enforcement-triggering pattern. RepuGen's Past Patient Outreach feature allows practices to control outreach cadence, distributing requests gradually over time so the review pattern looks organic to Google's systems.
2. Short Generic Text with Maximum Rating
Brief, positive reviews with 5-star ratings are structurally identical to bulk fake-review patterns: short text, maximum rating, and low reviewer history. This overlap, not a flaw in how Google's systems work, is why some genuine reviews get swept up alongside coordinated, fake submissions. This is the most common review type healthcare practices receive, and the most commonly removed. Review requests framed around the patient's experience, such as how a physician made them feel at ease, generate longer, more specific responses that survive moderation and are more persuasive to prospective patients.
3. Medical Content Flagged Under Misinformation Rules
Reviews referencing treatments, outcomes, medications, or clinical results face a specific problem: Google's Maps content policy explicitly prohibits harmful content that contains deceptive or misleading health or medical information and identifies health and medical services as a restricted category. A review mentioning a specific treatment or clinical outcome can pattern-match with health misinformation signals, even when it is entirely accurate. Review requests should be framed around the patient's experience, how they were treated, and how they felt, rather than clinical specifics.
4. Location and Device Signal Mismatches
Reviews submitted from devices or locations that do not match a practice's geographic footprint face closer scrutiny, because Google uses device, GPS, and IP signals to verify genuine interaction with a business. For example, a review left for a California dental practice from an IP address in New York is more likely to be removed, even when the reviewer is a genuine patient, such as a former resident or a traveling family member. Requesting a review while the patient is still at the practice also creates a risk: the request and the resulting review could share the practice's own IP address or Wi-Fi network, which Google can flag as pressured or artificial solicitation, similar to the shared-device risk described in the next trigger. Sending the request promptly after the visit to the patient's own device, once they are off the practice's network, avoids both problems while still reaching the patient when the experience is fresh.
5. Shared Devices at Point of Care
Practices that use a single tablet or iPad for multiple patients to leave reviews at checkout are exposed here, since multiple reviews originating from the same device and IP address now count as pressured or artificial solicitation under Google's April 2026 policy update. It is a common, well-intentioned front-desk practice that now carries a direct enforcement risk. RepuGen's review requests are sent to each patient's device via text message or email after the visit, eliminating exposure to shared devices.
6. Retroactive Enforcement on Previously Approved Reviews
Google's Gemini systems are designed to track reviews over time and apply updated moderation standards retroactively, so there is no statute of limitations on moderation. Practices that stopped actively collecting reviews and now rely on an existing bank of reviews from 2022 or 2023 face double exposure: those older reviews are subject to retroactive removal, and patients consider reviews beyond a certain age too outdated to trust. A steady pipeline of new reviews is the only reliable buffer against both risks, a finding supported directly by RepuGen's 2026 behavioral study data on review recency.
| Removal Trigger | Risk Level | Why Google's AI Flags It | Healthcare Implication | Recommended Action |
|---|---|---|---|---|
| Review Velocity Spikes | High | A sudden increase in review volume resembles coordinated fake review campaigns. | Mass outreach to large patient lists can unintentionally trigger automated enforcement. | Spread review requests over time using a controlled outreach cadence rather than sending them all at once. |
| Short, Generic 5-Star Reviews | High | Brief reviews with maximum ratings closely match common spam patterns. | Genuine patient appreciation is often mistaken for manipulated feedback. | Encourage patients to describe their experience instead of leaving only a star rating or a brief compliment. |
| Medical Content Under Misinformation Rules | Medium | Reviews mentioning treatments, medications, or outcomes may resemble unverified medical claims. | Legitimate patient experiences involving clinical results can be flagged during automated moderation. | Ask patients to focus on their care experience, communication, and service rather than clinical outcomes. |
| Location or Device Signal Mismatch | Medium | Device location, IP address, and behavioral signals do not align with the business location. | Patients frequently leave reviews from home after appointments, which creates weaker trust signals. | Send review invitations to the patient's personal device via SMS or email after the visit. |
| Shared Devices at Checkout | High | Multiple reviews originating from the same device or network appear artificially generated. | Shared tablets or kiosks at reception create a direct enforcement risk. | Eliminate shared devices and allow every patient to review from their own phone or computer. |
| Retroactive Enforcement | Medium–High | Previously approved reviews are periodically re-evaluated as Google's AI models evolve. | Older review profiles gradually shrink if practices stop generating fresh patient feedback. | Maintain a continuous flow of new reviews to offset future removals and preserve review recency. |
Risk-level ratings reflect the relative frequency and severity of each enforcement mechanism as documented in Google's public policy guidance and the source data cited throughout this report. The 89.1% five-star deletion finding directly supports the higher ratings for the review-velocity and short-text triggers; the remaining ratings are a qualitative synthesis of Google's stated enforcement priorities rather than a trigger-by-trigger breakdown of removed reviews, a metric Google does not publish.
Google's April 2026 Policy Update, What Changed
Google's April 2026 update introduced three enforcement changes. Gemini now catches suspicious Business Profile edits before they go live. Verified owners receive email notifications before suggested edits are published. And review scam extortion detection was upgraded to identify bad actors demanding payment to remove fake negative reviews.
Three compliance implications follow for healthcare. Review gating, the practice of selectively soliciting reviews only from satisfied patients, is explicitly prohibited under Google's Maps content policy and more actively enforced following the April 2026 update. Solicitation language that guides review content, even implicitly, is now classified as manipulation under the same policy. And upgraded extortion detection means unusual negative review spikes are scrutinized more aggressively, increasing false-positive pressure during legitimate periods of critical feedback.
These changes reward practices already running a compliant program. For those who were not, the enforcement gap just became more consequential.
What to Do When Reviews Are Removed, Appeal Process, and Realistic Expectations
The realistic outcome should be stated up front: reinstatement is possible but unreliable, not fast, and never guaranteed.
Google maintains an appeal mechanism for removed contributions and has demonstrated that reviews can be reinstated at scale when system-level errors are identified. In the 2023 Local Guides case, however, restoration was driven by Google's internal system correction rather than individual user appeals.
Four steps make up a practical response protocol when reviews disappear.
- Document the removal immediately. Take a screenshot of the content, record the date of disappearance, and note the reviewer's name if known.
- Use Google Business Profile's flag or report tool to contest the removal as erroneous.
- Submit a direct support request if the profile is verified. Phone support typically yields faster resolution than the online form.
- Do not ask the patient to repost identical text. The same moderation signals will produce the same outcome.
The appeal process is a last resort, not a strategy. The more important investment is building a review pipeline robust enough that individual removals become statistically insignificant to the overall profile.
A Framework for a Moderation-Resistant Review Profile
Practices that maintain stable, growing review profiles through Google's enforcement waves share one characteristic: they treat review generation as an ongoing system. Three strategic pillars define that system.
1. Automated Review Generation
The most resilient review profiles are built on automated, continuous review generation. Whitespark's 2026 Local Search Ranking Factors survey, drawing on 47 leading local SEO experts, identifies review velocity and consistency as climbing ranking signals, while sudden volume spikes remain the most common moderation trigger. A practice that generates five reviews a week consistently is structurally safer and better ranked than one that generates fifty in a weekend.
2. Compliant Solicitation Design
Three requirements are non-negotiable. Every patient must receive a review request regardless of perceived satisfaction. Requests should go to individual patient devices by text message or email, never from shared on-site hardware. The request language should ask about experience, not ratings or clinical outcomes. These three requirements satisfy both Google's April 2026 policy standards and the Federal Trade Commission's Rule on the Use of Consumer Reviews and Testimonials (16 CFR Part 465, effective October 21, 2024), which independently prohibits review suppression and incentives conditioned on sentiment, and carries civil penalties of more than $51,000 per violation.
3. Active Profile Engagement
Responding to reviews, both positive and negative, does two things at once. It signals to Google's systems that the profile is genuinely managed rather than artificially inflated, and it demonstrates to prospective patients the interpersonal quality of care they can expect. Whitespark's 2026 findings confirm that response cadence is a rising local-ranking signal.
The stakes for patient acquisition are direct. RepuGen's 2026 Healthcare Behavioral Study found that approximately 89% of patients check online reviews before choosing a provider. A review profile eroded by moderation is not a technical problem — it is a patient acquisition problem with a measurable impact on practice revenue.
Conclusion
Google's AI-powered moderation removed 292 million reviews in 2025, and the enforcement curve — 170 million in 2023, 240 million in 2024, 292 million in 2025 — points in one direction. Healthcare is disproportionately exposed because the algorithmic pattern of a genuine, satisfied patient review often overlaps with that of a fake review. The false positive reality this creates is not a flaw that practices can write their way around. It is a structural feature of the moderation system. The durable solution is not a defense against individual removals; it is an automated review-generation system that consistently produces reviews, so any single removal is statistically insignificant.
Practices that build moderation-resistant review programs create a compounding competitive advantage. Practices that do not will feel each enforcement wave more acutely as their review profiles thin and age.
As Google's AI-driven enforcement continues to evolve, healthcare organizations that invest in generating compliant, consistent reviews today will be better positioned to maintain patient trust, preserve online visibility, and compete effectively in an increasingly digital healthcare marketplace.
About RepuGen
RepuGen is a HIPAA-compliant healthcare reputation management and patient experience platform that helps medical practices, hospitals, dental groups, and multi-location healthcare organizations strengthen their online reputation, improve patient satisfaction, and increase patient acquisition. Trusted by healthcare organizations across the United States, RepuGen brings together reputation management, marketing visibility, patient satisfaction, and AI-powered engagement within a single integrated platform designed exclusively for healthcare.
The platform automates compliant post-visit review requests via SMS and email, helping practices consistently generate authentic patient reviews while reducing the risk of review-velocity spikes that can trigger platform enforcement. It also enables organizations to monitor and manage reviews across major healthcare review platforms; collect structured patient feedback through surveys and NPS reporting; analyze patient sentiment using AI; improve local search and AI visibility through listings management and testimonial tools; and gain actionable insights through comprehensive reporting and analytics.
Specialized capabilities further strengthen every stage of the patient reputation journey. ReplyWize helps practices respond to reviews quickly with AI-generated, HIPAA-compliant responses. Past Patient Outreach enables organizations to reconnect with previous patients through cadence-controlled campaigns that safely accelerate review growth. CommentWiz transforms patient feedback into actionable insights through AI-powered sentiment analysis, while Review Pulse provides centralized monitoring, alerts, and reporting across multiple review platforms. Together, these capabilities help healthcare organizations continuously collect, understand, manage, and amplify authentic patient feedback.
See Where Your Practice Stands
Healthcare organizations seeking to strengthen their review strategy can schedule a personalized demonstration of RepuGen or request a complimentary Reputation Scorecard to evaluate the health of their current online reputation.
Data Sources and Methodology Note
This report draws on the following primary sources:
- Google Maps Trust and Safety Report 2025, published April 2026
- Google Transparency Report, Maps Content Protections, 2025 data
- Independent analysis of 335,520 deleted Google reviews, Localo, 2026
- Whitespark Local Search Ranking Factors 2026, n equals 47 local SEO experts
- Federal Trade Commission, Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, effective October 21, 2024
- Search Engine Roundtable, coverage of Google review enforcement trends, 2023 to 2025
- Search Engine Journal, Google Maps Gemini upgrade to combat fake reviews, 2026
- Search Engine Land, coverage of the 2023 Local Guides review reinstatement
- Google Business Profile content policy, restricted categories, and review solicitation rules
All figures were current as of publication. This report may be cited with attribution to RepuGen.
