Google says it took down 2.3 billion “bad ads”—the equivalent of more than six million per day—last year for violating both new and existing advertising policies.
In its annual “bad ads report,” the company also touted its efforts to address what it called “fundamental concerns” around protecting election advertising, specifically questions about who is purchasing political ads.
Google claimed that it verified nearly 143,000 election ads ahead of the U.S. midterm elections. It also launched a political ads transparency report giving more information about who purchased election ads, and said it plans to launch similar tools ahead of elections in both the EU and India.
Google’s boasts about protecting elections comes just one week after it seemingly rebuffed attempts by the Canadian government to increase protections against questionable ads in the upcoming federal election.
Specifically, Google cited concerns around creating a searchable public registry of political advertisers. The search giant determined it could not meet Ottawa’s demands, and decided instead to ban political advertising for the duration of the campaign. According to reports, the ban does not extend to enhanced search results and platforms such as YouTube.
But one Canadian academic says that Google’s real intent in refusing to accept Canadian political ads is to send an unmistakable message to other smaller democracies that might be contemplating their own reforms around election advertising: Don’t mess with us.
“This is a rap across Canadians’ knuckles, so that they don’t become a showcase for what’s possible for other countries,” says Dwayne Winseck, a professor of journalism at Carleton University in Ottawa. “It’s a way of saying ‘Let’s see if you can make this stick.'”
Google Canada’s head of public policy, Colin McKay, called Google’s decision to not accept political ads “painful.” But Winseck says it is actually a “brazen” response to the December passage of Bill C-76—also known as the Elections Modernization Act—requiring online players such as Google and Facebook to create a public registry of all “partisan advertising and election advertising” appearing on their platform during the election period.
“[It’s] putting a gun to politicians’ head and saying ‘If you try to operationalize this new law, we’re going to close down this significant channel of communications,'” says Winseck, who is also director of the Canadian Media Concentration Research Project.
Winseck said that digital platforms like Google and Facebook have become an integral part of the marketing toolkit for political strategists over the past decade. “To knock these guys out at the kneecaps because you’re pissed at a new law just seems really offside to me,” he says. “It’s a way of undercutting the way election campaigns are taking place in this country.”
According to published estimates, Canada’s major political parties spent approximately $110 million on advertising during the 2015 federal election (that figure does not include approximately $6.7 million in third-party advertising from the likes of unions, advocacy groups etc.). Google did not respond to repeated interview requests from The Message.
As for its efforts to eliminate bad actors across the broader ad landscape, Google said that its 2.3 billion takedowns in 2018 represented a 28% decrease from 3.2 billion takedowns the previous year. Those bad ad takedowns included nearly 207,000 ads for ticket resellers, over 531,000 ads for bail bonds—a sector it found was taking advantage of “vulnerable communities—and approximately 58.8 million phishing ads.
It also said that it saw a rise in ads promoting “deceptive experiences” to users seeking addiction treatment. After consulting with experts, it restricted advertising to certified organizations.
The tech giant says it is making a “concerted effort” to go after the individuals and companies behind the bad ads, rather than just the ads themselves. Through the use of improved machine learning technology, the company says it terminated nearly one million bad advertiser accounts last year, nearly double the number it terminated in 2017.
Google also continues to use technology that allows for what it describes as more granular removal of ads from websites. It launched 330 “detection classifiers”—more than double the number it launched in 2017—to help it better detect “badness” at the page level. It terminated nearly 734,000 publishers and app developers, and removed ads completely from nearly 1.5 million apps
It also took ads off nearly 28 million pages that violated its publisher policies, using a combination of manual reviews and machine learning.
It also worked to ensure that ads are supporting “legitimate, high-quality publishers,” removing ads from 1.2 million pages, more than 22,000 apps and nearly 15,000 sites across its ad networks for what it characterized as “misrepresentative, hateful or other low-quality content.”
Google plans to launch a new Policy manager in Google Ads similar to its AdSense Policy Centre, providing tips on “common policy mistakes” aimed at helping well-meaning advertisers create and launch ads that are policy-compliant.