Skip to main content

Cambridge Analytica isn’t worried about Facebook ad restrictions

Cambridge Analytica isn’t worried about Facebook ad restrictions


‘Facebook is largely the Wild West in terms of regulations,’ targeting firm says

Share this story

2016 Concordia Summit Convenes World Leaders To Discuss The Power Of Partnerships - Day 1
Photo by Bryan Bedder/Getty Images for Concordia Summit

Last week, thousands of marketers, buyers, and media figures gathered in New York City for Advertising Week, a yearly industry gathering. Most of the attention was on television and print — where most of the money in the industry still goes — but Facebook was hovering ominously in the background. At one panel, aggressively titled “Will Data and Digital Resurrect TV?,” even Oprah favorite Dr. Oz got in on the action.

“I can’t say things on my show that they can say on Facebook,” Oz told the crowd. “I can’t sell children slaves on my show, but you can on the web, legally, in America. So when you see that, people start to get upset.” (In fact, placing ads involving child slavery remains illegal.)

Oz’s concerns aside, the broader anxiety about Facebook’s influence has become hard to miss. Over the last two weeks, the company has run into a string of damaging scandals that have called its entire ad-targeting product into question. It started with an admission that the company had accepted $100,000 for campaign-related ads from entities linked to the Russian government, which the company recently admitted had reached roughly 10 million users. A separate ProPublica investigation found targeting categories like “jew hater” in Facebook’s advertising system, something the company put down to unmoderated user inputs. (The categories were removed the following day.) The company is still working through its response and cooperating with a congressional investigation of the Russian ads, but the result has been a crisis of confidence in Facebook’s advertising system, and the looming threat of stronger regulation from Washington.

A crisis of confidence in Facebook’s advertising system, and the looming threat of stronger regulation from Washington

Ironically, one of the first people to address that threat onstage during Ad Week was Cambridge Analytica’s Molly Schweickert. Cambridge is best known for taking advantage of many of the same tactics as part of the firm’s work for the Trump campaign, and specializing in a kind of targeting the company calls “psychographics.” The Guardian has also connected the firm to a Brexit campaign, although Cambridge denies doing any work for the campaign and has brought a legal complaint against the newspaper in response.

According to Schweickert, a federal crackdown on Facebook ad targeting would mostly shift the power between platforms, and even that regulation is far from a done deal. “Facebook is largely the Wild West in terms of regulations,” she said. “We’re already starting to see some changes come in. It could be that those changes reduce Facebook’s impact. It may not happen, but it’s certainly a possibility.”

There’s reason for marketers to feel comfortable. For all the talk of tighter ad restrictions, most proposals circulated so far have been limited to election ads, just a fraction of the billions that are spent on the platform each month. For broader targeting, the recent restrictions are limited to explicitly named audience categories, which wouldn’t affect the more sophisticated methods used by firms like Cambridge. The result is that, even as critics push for regulation, much of the industry expects business to continue as usual.

“What’s proprietary is the research and model that was made offline and then gets uploaded onto the platform.”

Facebook’s clearest changes have to do with dark ads, paid content that can pop into users’ feeds without ever showing up outside the targeted demographic. During the 2016 US presidential election, that system let firms like Cambridge serve different ads to thousands of different micro-targeted groups. Since there was no way to see all the ads at once, there was no way for observers to know how the ads varied from group to group, and no way to hold the campaign accountable for the message of the ads. Facebook has announced planned changes to that system that would make all ads from a single client viewable in a single place, starting with campaign ads but eventually extending to commercial advertising too.

You might expect a targeting company like Cambridge Analytica to have a problem with that, but Schweickert isn’t worried about the new disclosures giving away the company’s special tactics. “What’s proprietary is the research and model that was made offline, and then gets uploaded onto the platform,” Schweickert told The Verge. According to Schweickert, just showing the individual ads won’t necessarily expose that model, even if it reveals the segments the model produces. Revealing the model itself would be far harder, she says, “unless they’re going to put regulation on advertisers where you have to talk about whether you first ran quantitative surveys and what your sample size was.”

In part, that’s because firms are increasingly doing their targeting before Facebook gets involved. In other words, Cambridge isn’t coming to Facebook and asking who likes Coca-Cola; instead, it knows who likes Coca-Cola already, and is seeking for a way to reach them. Most sophisticated micro-targeting campaigns will begin with a detailed list of customers, typically gleaned from customer lists for conventional commercials. Those lists can be uploaded directly to Facebook’s ad platform, provided they don’t come from a third-party data broker. If you narrowed that list to focus on customers that met a particular profile, Facebook would have no easy way to know, even if it violated the company’s targeting policies.

Then there’s the targeting Facebook does in delivering the ad. Within a given target audience, marketers will often let Facebook optimize on the fly, delivering the ad to whatever users within that audience are most expected to respond. But the nature of the algorithm makes it difficult to unpack exactly who the delivery is focused on. Facebook’s current targeting disclosure system (accessible through Ads preferences and the “why did I see this ad” tab) would only display the initial target set (say, people in the U.S. between 18 and 35), rather than the algorithmic delivery taking place within that field, which might end up skewing delivery towards specific regions or more abstract behavior patterns.

“These platforms are going to have to make a decision about where they stand”

So far, Facebook’s immediate plan has been to add more human oversight in an effort to tease out transparently offensive targeting categories or foreign interference in elections. A bill proposed by Sens. Warner (D-VA) and Klobuchar (D-MN) would seek to regulate Facebook ad buys under the same FEC rules as TV or print ads. But neither measure would have much effect on more sophisticated types of targeting.

Even within stronger rules around political campaigns, it’s unclear how much transparency efforts will change the basic terms of political campaigning, which tend to prioritize noise over signal. “Everyone has an interest. We’re here to elect people. Media companies are there to sell advertising. Even social media, which was supposedly the area where you could engage one-on-one with the people you trust, is increasingly under assault,” said Brent McGoldrick, whose firm Deep Root Analytics managed data for the Republican National Committee. “None of the stakeholders involved are serious enough about responsible communication to educate an electorate.”

Now, Facebook is caught in between those forces, and there’s still real pressure from marketers to fight restrictions on targeting. “These platforms are going to have to make a decision about where they stand,” says Schweickert. “If they’re going to position themselves as a tool for advertisers to get their message out there, they cannot subjectively exert that much control over what you can and cannot target.”