Data-driven tactics don’t always pay off ... just ask Hillary
Voters are sought based on whether they get their caffeine from Dunkin’ Donuts or Starbucks
FACEBOOK and the voter-profiling company Cambridge Analytica are getting richly deserved condemnation in the growing scandal over revelations that the social network allowed millions of users’ personal information to be shared with a consultant for Donald Trump’s presidential campaign.
But perhaps we should also thank them for shining a spotlight on something that is eating at our political system.
It is far from clear whether Cambridge Analytica delivered any real results for Trump or whether it merely hyped what it could achieve with all this information. My hunch is the latter. But in some respects, this scheme represents the crowning touch for what has become a fetishisation of data in politics – the idea that if you can just identify the right niches, and microtarget those people with a diet rich in what they already think, you win.
Yes, we are all products of our demographics to some degree. But to assume that is all anyone needs to know – or ask – short changes voters and diminishes candidates.
It also accelerates the polarisation that is making the United States all but ungovernable. Voters are sought based on what they watch, and what they click, and whether they get their caffeine from Dunkin’ Donuts or Starbucks. Push the right button and deliver to them the correct message – whether it is true or false, rational or unhinged – and they are yours.
From there, it would follow that the character of the candidate, much less the quality of his or her ideas, is practically irrelevant. There is little room for creativity, no incentive to reach out or elevate. All that matters is deciding what slices of the electorate constitute your base and mobilising them in larger numbers and with greater sophistication than your opponent’s data operation does.
Ironically, given what we are learning about the Trump campaign’s efforts in this regard, it was Hillary Clinton who fell deeper into this trap during the 2016 election. She got an A in science and a D in art. Her campaign was the best we’ve seen when it comes to modern, data-driven tactics. It delivered a majority of the electorate, but it did not connect everywhere it needed to.
Clinton blames the voters, as she made clear at a recent conference in India. Trump’s “whole campaign, ‘Make America Great Again,’ was looking backward,” she said. “You know, you didn’t like black people getting rights, you don’t like women getting jobs, you don’t want to see that Indian American succeeding more than you are, whatever your problem is, I’m going to solve it.”
Her focus on those four words, however, misses the fact that enough of those same people voted for Barack Obama to make him president twice. It also overlooks the fact that Trump’s slogan was not the product of a political technocrat’s algorithm but a gut instinct he had long before anyone took him seriously as a possible president. He filed for a trademark on the phrase six days after the 2012 election and stuck with it all the way through. Clinton’s campaign, on the other hand, was at one point considering 84 different slogans, none of which stuck.
What’s lost when data starts driving politics is not only privacy – including that of 50 million Facebook users – but also an opportunity to make our politics more than the sum of its factions. (© Washington Post Syndication)
But no one should take the psychological profile stuff at face value.
No academic work exists to link personality traits, especially those gleaned from the sketchy and often false information on Facebook profiles, definitively to political choices.
THERE is, however, research showing that values or even genetic factors trump traits. It’s not even clear how traits affect political behaviour, such as the tendency to vote and donate to campaigns – some researchers, for example, have found a negative relationship between emotional stability and these measures; others have found a positive one.
This is not to say Facebook data, including data on a user’s friends, can’t be useful to campaigns.
The Obama campaign actually asked its active supporters to contact six specific friends suggested by the algorithm. So people reached a million others, and, according to data from the campaign, 20pc of the million actually did something, like registering to vote.
But did the Trump campaign need CA and the data it acquired from Kogan to do this kind of outreach in 2016? Likely not. Facebook cut off the friends functionality for app developers because it wanted to control its own offering to clients interested in microtargeting.
There’s plenty of evidence that Brad Parscale, who ran the digital side of Trump’s campaign, worked closely with Facebook.
Using the platform’s ‘Lookalike Audiences’ he could find people who resemble known Trump supporters.
Facebook also has the capacity to target ads to the friends of people who have “liked” a page – a Trump campaign page, for example.
Targeting messages to millions specific people without going directly through Facebook is messier and probably more expensive than using the social platform’s own tools. All Facebook requires for access to its data trove is a reasonable fee.
Whether CA could add anything meaningful to Facebook’s effort is unclear.
Its previous client, the unsuccessful presidential campaign of Senator Ted Cruz, has said it didn’t deliver on all its promises. Some studies have shown that Facebook ads can work quite well for businesses.
If they also worked for Trump, the CA story is a red herring – it’s Facebook’s own data collection and the tools it makes available to clients that should be the target of scrutiny and perhaps regulation, both from a privacy perspective and for the sake of political transparency. (© Washington Post Syndication)