As transfixing as the #NeverAgain movement and the Stormy Daniels saga continue to be, I’m more interested in a complicated, unfolding story that’s flying below the radar of the American news media. This is the collaborative investigation by the New York Times, the Guardian and Observer, and British TV Channel 4‘s undercover team about Cambridge Analytica (CA), the shadowy Steve Bannon offshoot of a United Kingdom (UK) psychometric marketing firm, SCL Group Limited.
CA played a key role in aggregating and weaponizing personal data harvested from Facebook that may have helped swing the 2016 Brexit election to leave the EU, and to hand the American presidential election to Donald Trump.
Since we’re all on Facebook, much of what attention there is has centered on the original Facebook data release, when the company allowed a Russian-American researcher in the UK, Dr. Aleksandr Kogan, back in 2014 to scrape FB user data gleaned from a personality-test app.
The app not only compiled data from the 270,000 or so participants, it compiled data on all their FB friends, too, a total of more than 50 million FB users, according to whistleblower Christopher Wylie, CA’s former research director. According to subsequent reporting on the scheme, a second UK academic figured out a way to reverse-engineer that ancillary FB activity and build personality profiles of those users, as well. Kogan then sold that data to Cambridge Analytica, which built a personality profile database on as many as 100 million American voters.
This violation of its policies, says Facebook, was originally brought to its attention by British journalists back in 2015. In response, FB subsequently banned Kogan’s test-app and CA from its platform and ordered them to destroy the data, which they certified in writing that they did. When FB recently learned that this was apparently not done after all, FB publicly announced they were suspending Kogan’s, Wylie’s, and CA’s accounts, and posted a highly unusual personal apology from FB founder Mark Zuckerberg.
FB obviously is trying to contain the PR damage to its brand, and stem the tide—well, more like a trickle—of angry users and advertisers vowing to abandon the platform. But even if they do, they’ll be back, because FB is basically the only game in town. Too big to fail? You betcha.
But here’s the real danger, and why it should matter to Americans: our democracy is hanging by a thread. Trump was not elected by a majority; he wasn’t even elected by a plurality. For the second time in 20 years, the ballot-box loser ended up in the White House. Ponder this: Hillary Clinton beat Trump in the popular vote by almost 2.1% and 2.9 million votes yet lost in the Electoral College by 304-227; that was virtually the same margin, 303-219, that elected John F. Kennedy over Richard Nixon in 1960, when Kennedy narrowly defeated Nixon by 0.17% and 113,000 votes.
Where does Cambridge Analytica fit in?
Stay with me here: the 2016 presidential election turned on three states that had been counted safely in the Democratic column: Michigan, Wisconsin, and Pennsylvania, where Clinton lost by a combined total of only 77,759, about half of one percent of the total vote in those three states.
In December 2016, the nonpartisan Cook Political Report crunched the numbers and figured out that it came down to only three counties: Macomb County in Michigan, York County in Pennsylvania, and Waukesha County in Wisconsin. Therefore, if only half, or 39,000, of those ballots had been cast for Clinton instead of Trump, she would have picked up those 46 Electoral College votes and flipped the results, winning 273-258. In fact, that’s exactly what Cambridge Analytica CEO Alexander Nix himself boasted in Channel 4’s March 20, 2018 undercover video (at the 4:22 mark).
Then there’s the protest vote. The ratio of third-party ballots to Trump’s margin was more than 8:1; in Pennsylvania, 5:1; in Wisconsin, 23:1. Even a fraction of those votes could have saved the day.
Out of 136,669,237 votes cast, only 0.00057% of the turnout made all the difference. In local terms, it’s as if the 2014 county supervisor election for the Third District, out of nearly 227,000 votes, had been decided by a few dozen audience members on a slow night at the Theatricum.
That’s the danger we face when you combine such weaponized granular data analysis and diabolically effective targeted advertising—delivered by a hidden hand through front groups and fake personal accounts—with an uninformed, easily manipulable electorate.
The proverb, “For want of a nail, the kingdom was lost”, reminds us how seemingly innocuous things can have unforeseen, catastrophic consequences.
For want of a few votes, an election was lost; for want of an election, will our fragile American democracy be lost?