Imagine: A high-level political consultant admits he mined Facebook data to target likely voters in swing states. He says he helped “build this thing called targeted sharing” that “allowed us to use Facebook to persuade people.” Cambridge Analytica? No, that was Democratic strategist Jim Messina, Barack Obama’s campaign manager in 2012, discussing the ways in which his campaign employed “big data” to target likely supporters in key states.
It turns out a similar use was put to Facebook data in 2016, but the users intended to get Donald Trump elected. Why the latter is scandalous, but not the former, we are not sure.
Cambridge Analytica, or CA, a U.K.-based data-mining agency, harvested information from 50 million Facebook profiles and used it to target voters in the 2016 election. The firm received this data from Aleksandr Kogan, a researcher at Cambridge University, who developed an app that acquired personal data through a Facebook-integrated third-party “quiz.” The quiz allowed Kogan to enter profiles and snatch information such as birthdate, residence, and “likes” from quiz-takers and their friends. After Kogan obtained all this data, he shared it with CA—evidently against Facebook’s policy, which by that time had changed its rules. CA then set up psychographic models based on the Facebook users whose data they had grabbed—and used it to micro-target specified profiles.
Among those who contracted CA’s services was the campaign of Donald J. Trump.
Thus the latest theory put forward by for how Trump won the presidency. If you begin with the assumption that there’s no way Trump won the election fairly, then the Cambridge Analytica theory works as well as any of the others.
There are two problems with the theory. First, there’s very little difference between the way Team Obama used Facebook data in 2012 and the way Team Trump used it in 2016. The chief difference is that whereas Facebook was happy to give the data up in the first instance, in the latter it was shared by Kogan against the tech giant’s will.
Second, the theory is that it can’t be verified. There’s no way to gauge the accuracy of CA’s targeting. In fact, the information collected in this and similar instances of data harvesting is not high-quality and may have had no discernable impact on voting patterns. Whether or not someone likes Justin Bieber or 50 Cent on Facebook is functionally useless for political-targeting and messaging. That information does not and cannot predict an account-user’s political preferences or even whether he or she votes.
This latest theory purporting to explain Trump’s unlikely victory (we stop short of calling it a conspiracy theory, but it’s close to that) gives far too much credit to the sophisticated data-mavens who’ve supposedly mastered the science of finding and motivating voters. Surely the lesson of 2016 is precisely the opposite. It was Hillary Clinton’s campaign, not Trump’s, that employed the best and brightest data-engineers, the tech-savvy geniuses who understood voting patterns and demographic data. All they lacked was a likable candidate.
In any case, it’s a little hard to be outraged by the idea of a company improperly taking data from a social-media behemoth to which billions of user willing give their personal information. Facebook has always been—and always will be—an advertising platform. Users are Facebook’s products; advertisers are its consumers. If you hand over your personal data to an advertising company, you may expect it to be sold. On Wednesday Facebook CEO Mark Zuckerberg issued a long explanatory statement about the Cambridge Analytica controversy, but he might have just said that.
The controversy has produced one remarkable result, however: People are deleting their Facebook accounts. Brian Acton—the founder of WhatsApp who sold his company to Facebook for $19 billion a few years ago—tweeted: “it is time. #deletefacebook.” Cher, too, has deserted Facebook. The WEEKLY STANDARD’s Matt Labash is no doubt wondering what took them so long.
