Yelp Prioritizes Trust and Safety - New Updates

Yelp’s Inaugural Trust and Safety Report Released

This week, Yelp published its first-ever trust and safety report. The report details the company’s efforts to combat fraudulent and inaccurate content on its platform.

The Importance of Review Integrity

Given its focus on local business reviews, one might assume Yelp is less susceptible to misinformation than other social media sites. However, Yelp reviews carry significant weight. They can substantially influence a business’s financial performance.

A Combined Approach to Content Moderation

Yelp, like many online platforms, employs a combination of automated software and human review. Software plays a key role in categorizing reviews as either recommended or not recommended.

Sudheer Someshwara, Group Product Manager for Trust and Safety, explained that reviews may be flagged as “not recommended” if they appear to originate from individuals with potential biases. This includes reviews solicited by the business itself, or those submitted by users with limited activity.

Ensuring Fairness and Consistency

Yelp emphasizes its commitment to fairness and integrity in its moderation processes. No employee has the authority to override decisions made by the automated systems, even engineers.

The company maintains that all businesses are treated equally, regardless of their advertising status.

Review Statistics and the “Not Recommended” System

Last year, over 18.1 million reviews were posted by users. Approximately 25%, or 4.6 million reviews, were categorized as “not recommended” by the software.

It’s important to note that “not recommended” reviews are not deleted. Users can still access them, but they are located in a separate section of the platform.

Human Review and Content Removal

Content removal is handled by Yelp’s user operations team. Businesses and consumers can flag reviews for review.

Aaron Schur, Vice President of Legal, Trust & Safety, stated that every flagged piece of content is assessed by a human moderator to determine if it violates Yelp’s guidelines.

Review Removal Statistics

In the past year, around 710,000 reviews (4%) were completely removed for policy violations.

Of these, over 5,200 were removed due to breaches of Yelp’s COVID-19 guidelines. These guidelines prohibit claims of contracting COVID-19 at a business, complaints about mask mandates, or criticism of safety-related closures.

An additional 13,300 reviews were removed between May 25th and the end of the year for containing threats, hate speech, or other harmful content.

Responding to Current Events

Yelp acknowledges that current events often manifest on its platform. People frequently turn to Yelp and other social media to express their opinions.

Balancing Free Speech with Genuine Experience

Yelp’s “guiding principle” is to prioritize genuine, first-hand experiences. The company uses software to detect unusual activity and adds Consumer Alerts when it identifies attempts to manipulate ratings and reviews.

There was a 206% year-over-year increase in incidents fueled by media coverage.

The Importance of Firsthand Accounts

While political opinions are not prohibited, reviews must be based on personal experience. They should not be prompted by external sources like negative articles or social media posts.

Noorie Malik, Vice President of User Operations, noted that the team sometimes removes content even when they agree with the expressed viewpoint.

Addressing Accusations of Racism

Yelp will remove reviews that appear to be driven by media reports alleging racist behavior by a business owner or employee.

However, Yelp also labels businesses with a “Business Accused of Racism” alert when there is substantial evidence of egregious racist actions.

Proactive Measures: Sting Operations

Yelp proactively conducts “sting operations” to identify and dismantle groups engaged in posting fraudulent reviews.

The company shut down 1,200 user accounts linked to review rings and reported nearly 200 groups to other platforms. A new algorithm has been implemented to better detect and demote reviews originating from these groups.