Roughly one year ago, Donald Trump's victory shocked the world. How could such an unexperienced and understaffed campaign team beat Hillary Clinton's seasoned and tech-savvy crew?
It starts with an Englishman. Impeccably dressed, he speaks the Queen's English and his name is Alexander Nix. As CEO of Cambridge Analytica (CA), he found the silver bullet for electoral success. Wherever CA goes, electoral shocks follow. CA was part of the Brexit campaign. Five months later, they were part of the campaign that made Trump president. One and one makes two.
Their technique sounds plausible enough. CA holds databases with rich information about hundreds of millions of Americans. Using their data, they categorize every citizen. One can, for instance score high on the label: 'neuroticism' or perhaps on 'extroversion'. Messages can then be crafted to appeal to different personality types. An extrovert person might be inclined to respond to messages featuring the names and opinions of their friends, while a 'neurotic' person is more likely to respond to messages appealing to the rule of law. CA says it can correctly categorize people into a 'psychometric profile', and subsequently send them a message designed to be as effective as possible.
Sounds believable, especially since CA was part of the two biggest political upsets of 2016. Vice wrote a John le Carré-ish short story, hyping Cambridge Analytica and their democracy disrupting capabilities. While I think that their technique might work, and despite lots of overhyped media coverage, pointing to CA as magic kingmakers is too much. There are three reasons:
1) Using the data properly would be a huge technical challenge, and then require a big campaign team to create a lot of tailored messages, which Trump did not have.
2) Despite their big words, CA has never provided a sliver of proof to support their claims.
3) CA has admitted it never used psychometrics in the Trump campaign.
So, what did happen? Daniel Kreiss and Shanon McGregor recently published an article in which they explain how Facebook offered both the Trump and the Clinton campaign help. Clinton's campaign refused, but Trump's campaign welcomed every help they could get. Facebook sent a team to the Trump headquarters and they, as it seems, really helped the campaign a lot.
But Facebook did not provide Trump with the silver bullet. There are many different reasons Trump has won. Sure, Trump ran a pretty sweet digital campaign. But also: remember the slim margins of Trump's victory in Michigan (11,612 votes) and Wisconsin (27,000 votes)? The third-party candidate, Jill Stein of the Green Party, picked up more than 35,000 votes in Wisconsin, and even more in Michigan! Would these votes otherwise have gone to Clinton? Maybe, maybe not. Remember the FBI statement a couple of days before the election? The lessons here, I think, are threefold:
1) Be wary of 'magic silver bullet stories'.
2) Do not look for one specific cause for something as complex as an election result.
3) Be wary of Cambridge Analytica's marketing department, for they, as we say in Dutch, are able to sell a fridge to an Eskimo.