Press Secretary Sarah Huckabee Sanders announced on Twitter Saturday that she and her family had been asked to leave the Red Hen, a small restaurant in Lexington, Virginia. The Red Hen’s co-owner, Stephanie Wilkinson, reportedly asked Sanders to leave because of her involvement in Trump administration policies like separating migrant children from their parents. Word of the incident quickly spread across the internet and on Monday, President Donald Trump lobbed an insult at the Red Hen, alleging that the restaurant’s exterior is “dirty.”
But the majority of the backlash against the Red Hen came on a different platform: Yelp. Many of the press secretary’s supporters spent the weekend vandalizing the restaurant’s page by leaving thousands of fraudulent one-star reviews. Others who agreed with Wilkinson’s decision responded by writing retaliatory five-star reviews, turning Yelp into an unwilling platform for political speech. In essence, Yelp became a battleground—and not for the first time. For years, crowd-sourced review sites like Yelp and TripAdvisor have been manipulated by trolls, paid reviewers, and politically enraged citizens. But we rarely consider how sites like Yelp fight back—and what their tactics mean for businesses and users.
This is far from the first time that Yelp has experienced a surge in vandalism in response to the news cycle. In 2012, for example, trolls famously defaced a pizza parlor’s page after the restaurant’s owner posted a picture of himself hugging former President Barack Obama. The issue comes up enough that Vince Sollitto, Yelp’s senior vice president for corporate communications, penned a blog post in 2016 explaining the company’s strategy for when similar incidents happen.
As Sollitto explains, Yelp doesn’t display every review left for a business in chronological order; the company uses an algorithm to sort reviews according to a number of signals, including whether they may be biased or fraudulent. Reviews that get sorted out also don’t contribute toward a business’s overall rating.
In situations like the Red Hen’s, Yelp deploys what it calls an Active Cleanup Alert, a pop-up that encourages users to discuss a business on Yelp’s discussion forums rather than leaving a review; it also warns them that fake ones will be removed. While the alert is active, Yelp employees work to identify and remove what they believe are fraudulent reviews. Again, Yelp blogged about this policy in 2016 and Active Cleanup Alerts were created the year before.
So in theory, when a local business becomes part of a national controversy, Yelp has an established strategy to deal with the fallout. In reality, the Red Hen’s Yelp page remains a mess. It had more than 15,000 reviews at the time of writing and the restaurant’s overall rating is down to 1.5 stars, from nearly five stars several days ago. People have attacked unaffiliated restaurants with similar names, and some have baselessly accused Red Hen of being run or owned by pedophiles.
Part of the problem, as Motherboard points out, is that Yelp doesn’t require reviewers to verify that they’ve actually visited a business, making it easy to turn the platform into a place for protest.
To be fair, Yelp has to manage more than 155 million reviews, according to the company, and not all of its resources can realistically be dedicated to defending one restaurant. Yelp’s moderators are also up against people like conservative activist Charlie Kirk, who encouraged his more than 600,000 Twitter followers Sunday to leave 100,000 additional reviews on the Red Hen’s page. TripAdvisor, for its part, temporarily stopped allowing reviews to be posted at all.
The Red Hen’s Yelp page looks especially troublesome compared to the restaurant’s Google reviews. At the time of writing, the Red Hen had a modest 45 reviews, all of which appeared legitimate; there’s no sign that the restaurant is at the heart of a national controversy. It’s possible that trolls simply didn’t target the Red Hen’s Google page in the same way. But it’s also possible that the tech giant is more adept at moderating.
A Google representative says that the company has a dedicated team and systems in place to identify incidents like what happened to the Red Hen. “Once identified, we use both automated and manual techniques to ensure that reviews adhere to our policies and that any new edits to business information are accurate,” the spokesperson said in a statement. (That doesn’t mean Google has always been free from manipulation; the Verge found last year that for-profit substance abuse rehabs had exploited its systems.)
Online reviews have also served as legitimate venues for social and political commentary; it’s not always clear what we lose when “fraudulent” ones are deleted en masse. In 2012, then-candidate Mitt Romney made a comment about “binders full of women” during a presidential debate. In response, a number of users left satirical Amazon reviews on listings for products like three-ring binders, a phenomenon chronicled in a 2015 study in the journal Feminist Theory.
“For any of you who might be considering, like me, purchasing this binder based on the reviews, let me just point out one glaring omission: While this is a lovely, multi-purpose binder, IT DOES NOT COME WITH WOMEN,” one reviewer wrote.
More recently, a vintage store in Brooklyn, New York was accused of racially profiling a black lawyer and her daughter last month after an employee believed the pair was shoplifting and called the police. The police didn’t find any stolen merchandise, and the incident ignited a small protest. In response, Yelp set up another Active Cleanup Alert and the store’s several dozen reviews no longer reflect the incident.
Some shoppers might want to know that a store was recently the site of a protest. Plenty of consumers may also want to choose one business over another on moral grounds. That’s a reality that Yelp even acknowledges in its 2016 blog post: “Many people understandably wouldn’t want to patronize a dentist who kills lions as a hobby, and others might legitimately be inclined to choose one pizza parlor over another based on their political views,” Sollitto wrote.
Yelp nonetheless chooses not to reflect that sort of feedback in its reviews, because it ultimately thinks “the better proposition is for Yelp reviews to be driven by firsthand customer experiences.” It has a point; most people come to Yelp simply to find genuine recommendations for places to eat and shop. And businesses can’t do anything to improve reviews that aren’t based on actual experience.
“When businesses make the news, their Yelp business page can be affected. Media-fueled reviews typically violate our Content Guidelines, one of which deals with relevance. Yelp reviews are required to describe a firsthand consumer experience, not what someone read in the news,” a Yelp spokesperson said in a statement.
But by encouraging on-topic reviews, Yelp also helps itself. The platform, like other social networks, has grown rich as a result of the free labor that people have contributed to it. Without a steady supply of helpful, pleasant-to-read reviews, Yelp doesn’t have a business.