Google Updates Review Policing with New Machine Learning Model
How a shift in long-term signal analysis is catching 45% more policy-violating content while increasing risk for legitimate operators.
Last updated February 13, 2024. Google is now leveraging a more sophisticated Google Maps fake review algorithm designed to scrutinize long-term contributor signals rather than just isolated incidents. According to Dan Pritchett, Principal Software Engineer at Google Maps, this shift in machine learning architecture led to the removal of 170 million policy-violating reviews in 2023—a 45% increase over the previous year.
For years, the moderation process was largely reactive, focusing on immediate triggers within a single review. However, the new model adopts a holistic view, monitoring how a single user account behaves across multiple weeks and locations. We are seeing this impact businesses globally, particularly those that experience sudden surges in customer feedback, whether organic or otherwise.
How does the Google Maps fake review algorithm identify suspicious activity?
The current detection engine moves beyond simple keyword filtering. It now prioritizes "questionable review patterns" by examining signals over several months. This includes identifying when an account leaves identical feedback for multiple businesses or when a local profile receives a statistical anomaly in 1 or 5-star ratings compared to its historical baseline.
In one specific case cited by Google, this logic was used to dismantle a widespread scam network. Scammers were promising users high-paying work in exchange for writing fake reviews. By analyzing whether an account had a history of interacting with specific "targeted features," the machine learning model was able to flag five million review attempts linked to this specific network within a few weeks.
Combatting bot patterns and malicious contributor networks
While the primary goal of the Google Maps fake review algorithm is to maintain ecosystem integrity, the increased sensitivity of these models can lead to "false positives." For a dental practice in Leeds that suddenly receives ten legitimate reviews following a local marketing drive, the algorithm may flag the activity as a bot-like spike.
Unlike previous iterations of the spam filter which focused on the content of a single text block, the current system looks at the relationship between the reviewer and the merchant. If the system detects a lack of physical proximity or a history of the reviewer participating in coordinated groups, the content is suppressed. In 2023, this proactive stance also prevented over 2 million attempts by malicious actors to hijack Business Profiles that did not belong to them.
Why business owners are seeing more 'missing' reviews
A common frustration for a 12-location HVAC operator is the disappearance of legitimate customer feedback. This is often the byproduct of the algorithm's stricter "long-term signal" requirement. If a customer has a relatively new Google account or lacks a history of contributing to Maps, their positive review of the HVAC service might be sidelined because the algorithm lacks sufficient historical data to verify the user's authenticity.
This behavior marks a significant departure from how the platform operated five years ago, when the barrier to entry for reviewers was lower. Today, Google places temporary protections on businesses specifically when they are identified as targets of high-risk activity. In 2023 alone, over 123,000 businesses were placed under such protections to prevent automated review attacks.
What this means for local businesses
The increased vigilance of the Google Maps fake review algorithm means that business owners must be more disciplined in how they solicit feedback. Organic growth is now safer than rapid, orchestrated campaigns.
- Avoid aggressive review solicitations: If you ask 50 customers to leave a review on the same afternoon, the sudden spike may trigger a spam flag. Space out your requests to mimic natural customer behavior.
- Verify reviewer quality: Encourage customers who are frequent Google Maps users to leave feedback, as their established "contributor signals" make their reviews less likely to be filtered.
- Document legitimate interactions: If your business is hit by a sudden wave of removals, keep records of customer invoices or service dates. You will need this evidence if you choose to appeal through the Google Business Profile Help Center.
- Monitor for 'Review Attack' protections: If you notice your reviews have stopped appearing entirely, check if Google has placed a temporary lock on your profile due to suspected abuse.
- Audit your profile managers: Ensure that only verified employees have access to your Business Profile. Google blocked 1 million more unauthorized claim attempts in 2023 than in 2022, suggesting that account security is a primary vector for spam.
Sources
Frequently asked questions
- Why did my legitimate Google reviews disappear?
- Reviews often disappear because the Google Maps fake review algorithm identifies a 'spike' in activity that mimics bot behavior. If several customers leave reviews simultaneously or if the reviewers have low-trust accounts with few historical contributions, the system may filter them as a precaution to prevent spam.
- Can I appeal a review removal?
- Yes, business owners can use the Google Business Profile appeal tool. You should be prepared to provide evidence that the reviews are authentic, such as service receipts or communication logs that correspond with the timing of the reviews. Google's human investigators can sometimes override the algorithmic decision.
- Does responding to reviews help prevent them from being flagged?
- While responding to reviews is a best practice for customer service, it does not directly exempt a review from being flagged by the automated system. The algorithm primarily analyzes the reviewer's account history and the relationship between the reviewer's location and the business.