Micro-Targeting, or Why We Unfriended So Many People in 2016
So far, I’ve focused on the means by which this propaganda introduced hostile memes into the American population, but I haven’t provided much detail about how ads, fake news, and trumped up hashtags successfully manipulated millions of Americans into voting for Trump (or at least the 70,000 it took to swing Wisconsin, Michigan, and Pennsylvania). The story behind that is pretty interesting, and I’d be impressed with the cleverness of Putin’s tactics except for the part that they were (and still are) aimed at making my friends and family hate each other.
Promote a Candidate or Political Party
Sometimes an ad is just an ad. Facebook collects tons of data about its users. Their platform makes this pretty easy, really. We’re there to notify our loved ones about momentous events in our lives, to rant or rave about some topic dear to our hearts, and to watch other people doing the same (and possibly to respond to them). Facebook’s advertisers can target their ads, promoted news stories, and promoted groups using that data. We’re well beyond “white, married woman with no kids who likes to read,” here. Facebook knows when you’re in the market for a new car because you just asked your Facebook friends for recommendations. It knows you’ll be interested in knowing when your favorite band releases their next album because you added them to your Facebook profile or Liked their fan page. As with all such targeted advertising, it can be useful – even convenient – for most of the ads you see to be for products and services that might actually be of interest to you, but it holds opportunities for abuse, too.
At the basic level, Facebook advertisers can promote content by state, city, or down to zip code, if need be. This great for local businesses trying to reach their neighbors’ eyes, but also very useful for political campaigns attempting to encourage some people with the “right” political leanings to vote while encouraging those with the “wrong” political views to stay home – especially if those advertisers already have access to hacked voter information in those voting districts (which is a very important piece of this puzzle, but beyond the scope of this essay).
It turns out that it was also possible for them to identify, say, the “Jew haters” who used their platform, and Facebook was more than happy to help advertisers identify that tiny slice of users so that they could be targeted for advertising. Facebook’s claims that it didn’t know that its advertising platform could be used this way are rich coming from a company that made nearly 700,000 of its users unwitting test subjects in a 2012 experiment meant to control their emotions. Get out the vote is usually a nonpartisan exercise, but micro-targeting meant that a political party could target their get out the vote push solely at citizens who were almost certain to vote for candidates of that political party. This was a powerful weapon in swing districts and states where a comparative handful of votes could change the result.
As a closer fit to the definition of provokatsiya, Russia used trolls (compensated agents) or bot networks (allowing a single such agent to amplify messages via thousands of sock puppet accounts) to encourage extreme elements within movements that could otherwise participate in civil dialogue with outsiders. The goal here is threefold:
First, it encourages and normalizes lawless behavior, subverting legitimate activist movements unless the group routinely kicks out extremists and suspected agents provocateurs. Most groups can’t or won’t do this because authority in a grassroots activist organization is nearly always decentralized, so who has the power to tell someone they’re not allow to show up at next Saturday’s march? Those extreme voices can provoke other members to embrace extremism in turn. This extremism need not be violence. It can instead be, say, a refusal to vote for an imperfect candidate in favor of staying home or voting for a third party even though “standing on principle” means a much, much more troubling candidate will win the election.
For example of this in action, a woman who goes by the Twitter handle @CassandraRules started out as a vocal Bernie-or-buster, but now she is an enthusiastic Trump supporter. How could anyone who ostensibly supports single-payer healthcare and breaking up the big banks for all throw her lot in with a man who just signed an executive order stripping healthcare from 8 million people and has pledged to lower taxes for the big banks? Here’s a hint: She works for Sputnik, a Russian propaganda news outlet.
I think we’ll ultimately find that Moscow used mico-targeting to find people who would be sympathetic to progressive causes and messages and attempted to radicalize them, first. As someone who caucused for Bernie, I watched with horror as people I knew held largely the same political views as I do chose not to vote in 2016 or cast a protest vote for Jill Stein (Why was she with Michael Flynn at the Russian dinner, again?). Democrats and progressives alike have long bemoaned the tendency of working class whites to vote against their own interests by putting Republicans in office who will only make it easier for the rich to exploit them, but 2016 proved that those of us on the left are capable of putting Republicans in power, too.
Second, the presence of extremists in any organization tends to undermine the credibility of its message in the public eye and makes it more difficult for it to form coalitions with groups that would otherwise be its allies. That Russia was amplifying these extreme views on social media did not do those organizations any favors in the court of public opinion.
After the 2016 convention, a lot of Democrats had a very low opinion of Bernie Sanders supporters. The primary race was hard-fought and contentious, to be sure, but so were the 2008 primaries. I remember getting continually lumped in with sexist Bernie Bros and worthless Bernie-or-Busters, and yes, it was frustrating and I can see why other Sanders supporters may have been turned off by it. This was the harvest of the previous year of provokatsiya among progressives. People who should have been able to agree that Donald Trump would be a disastrous president and who should have been willing to fight tooth and nail to keep him out of the Oval Office wasted a lot of energy sniping at each other as Kremlin trolls cheered on both sides from the rafters of the tavern’s common room.
Third, extreme voices attributed to an organization make a marvelous recruiting tool for people who oppose that organization. Russian agents then use further provokatsiya to amplify and radicalize the voices within the organization that forms as a reaction against the first organization. And thanks to micro-targeting, the Kremlin can build counter-organizations and counter-counter-organizations ad nauseum.
In some cases, provokatsiya is hard to detect. This is the internet we’re talking about, here, and we’re all familiar with the formula “normal person + anonymity = complete asshole.” Even when looking at non-assholes, the web in general – and social media in particular – tends to bring out people’s propensity for hyperbole, particularly when we’re expressing our opinions about emotionally charged events and figures. I won’t deny that I’m guilty of this myself. Moreover, we all have a few friends who hold some opinions that make us roll our eyes whenever they bring them up, and allowing everyone a few such eccentricities has become a means of keeping the peace to avoid burning bridges.
Several examples of this have already come out, and I’m sure we’ll see many more, but let’s create an example of this process out of the largely defunct Occupy Wall Street movement (which to my knowledge was not a target of provokatsiya):