Mindfuck

by Christopher Wylie

  I’d been drawn to politics because it seemed like a way to make a difference in the world, but after more than a year of banging my head against the wall, what was the point? Then came an intervention. In the Liberal Party, many of the secretarial staff came from a gaggle of older Québécois ladies, who had been around long enough to see how politics can transform someone. They took me to lunch in Gatineau, the French part of town just across the Ottawa River. After lighting up their cigarettes, they said, in their raspy, accented voices, “Listen, don’t become like us.” They told me they’d given their lives to the party, that the party had given them nothing in return except for “expanded waistlines and several divorces.” “Go, be young,” they said. “Get the hell out of here before it traps you.” And I realized they were right. At just twenty years old, I was already a midlife crisis waiting to happen.

  I decided on the law school at the London School of Economics, as London was probably far enough from Ottawa—3,300 miles and five time zones across the Atlantic. I later learned that some of Canada’s party leadership had conflicted motives. The party still awarded many of its advertising, consulting, and printing contracts to firms owned and operated by senior party members or their friends. A new, data-centric approach might mean that “friends and family” in the party would lose out. In 2011, one year after I left Ottawa, the LPC was devastated in the federal election by the Conservative Party of Canada, which had invested in sophisticated data systems at the behest of its imported Republican advisers. For the first time ever, the Liberal Party was relegated to third place, with only thirty-four seats in Parliament. It was an historic defeat.

  * * *

  —

  WHEN I BEGAN WORKING for the Liberal Democrats in London, it was for only a few hours a week, in between my classes at LSE. But almost right away I realized that, compared with the Obama campaign or even the LPC, the Lib Dems were a complete train wreck of a party. The office operated more like a stale curiosity shop than the heart of a political machine. The LDHQ staff were mostly bearded men in suits and sandals who spent more time chitchatting about the old Whigs than doing anything to mobilize their campaign. I asked to see their data systems, and someone told me about EARS, which stood for Electoral Agents Record System. “Wow, okay. This looks…old-school,” I said. “Was this made in the eighties?” It was like asking for a graphics demo and being shown one of those old Pong games. Someone told me that one of the systems had been designed during the Vietnam War.

  It soon became clear to many in the party how superior VAN was to anything else available, and the party finally approved a contract with VAN to set up the data infrastructure. But now we needed data—the fuel to run the Ferrari. This was the step where the project had gone wrong in Canada, and the process went no more smoothly in the U.K. There is no national electoral register in the U.K.—it’s all handled by town councils—so we had to approach hundreds of different councils all across Britain to get their voter data. I’d be on the phone to Agnes in West Somerset, who would be like 105 years old and had probably been managing the voter rolls since women got the vote, asking her, “Do you have a digital copy of the register?” No, she’d say, because she kept the records as they’d always been kept, on paper, but I could see a copy in a bound book in the local town hall. Sometimes the local officials agreed to give us the data, sometimes not. Sometimes it was in electronic form, sometimes it was a PDF file, and sometimes it was just reams of paper that we had to feed into an optical scanner. Excel files would usually be emailed without a password—because why would anyone want to steal voter data?

  The British electoral system was stuck in the 1850s, and, as I soon found out, so were the Lib Dems’ tactics. It wasn’t hard to understand why the party and its old Liberal Party predecessor had been on a losing streak since World War II. Leaders had lost touch with how to win and were utterly obsessed with handing out leaflets. These leaflets were called “Focus” and usually complained about parochial “local issues” like potholes or rubbish collection. The Lib Dems thought this was a clever way of “slipping in” their messaging to something that looked like a local newspaper. But there was a problem with the Lib Dems’ dollar-store Pravda: no one actually read it. Their idea of a voter was someone who spent their weekends flipping through mail order catalogs and political literature—political staffers are often so socially clueless, they forget that regular people have lives. Despite being the smallest of the three main parties, the Lib Dems had the most volunteers, because they were absolutely militant about shoving leaflets through doors, rain or shine. They would decide how many leaflets to deliver before even deciding what they should say in them.

  In the world of hacking, the term “brute force” refers to randomly trying every possible option until you hit on the correct one. It doesn’t involve strategizing—it’s simply throwing everything at the wall to see what sticks. That’s essentially what the Lib Dems were doing, spending tons of money on leaflets without targeting particular voters. Brute force is an unsophisticated hack, comically inefficient yet occasionally successful. There were certainly more effective ways to win elections. Yet when I tried to present alternatives to spamming voters with their upper-middle-class propaganda, complete with MS Word clip art, I would get a lecture on “how Lib Dems win” and the fabled Eastbourne by-election, in 1990—the surprise victory of the first MP to win as a Lib Dem since the 1988 SDP-Liberal merger—where they had apparently delivered a lot of leaflets. To question Eastbourne was like heresy. Fringe religions demand conformity, and the Lib Dems were no different. The party was a leafleting cult.

  By-elections are irregular special elections, such as a vote to replace a member of Parliament who’s died. The Lib Dems were obsessed with them. For some reason, whenever the party won a by-election, members acted as if they’d conquered Britain with banners that proclaimed LIB DEMS WINNING HERE! But as part of my research, I decided to catalog every election and by-election since 1990 and found that the party had lost the overwhelming majority of them. “But what you are doing doesn’t help you win. It helps you lose. Here is the data,” I told them. “These are facts.”

  Some of the party leadership listened, but most just seemed pissed off. They had their own cottage industry of being “election gurus” for the party faithful and weren’t keen on a newcomer strolling in and telling them how to fix their system. This did not bode well.

  In the meantime, I started playing around with the voter data we had been able to collect from data vendors like Experian. I experimented with different types of sandbox models, similar to what I’d done in Canada. And something strange kept happening. No matter how I designed the models, I couldn’t build one that reliably predicted Lib Dem voters. I had no trouble doing it for Tories or Labour. Posh dude in some leafy rural township? Tory. Live on a council estate in Manchester? Labour. But Lib Dems were these in-between weirdos who resisted any neat description. Some of them looked Labour-ish and some of them looked Tory-ish. Ugh, what am I missing? I wondered if there was a latent variable at work. In social sciences, a “latent variable” is an element that is influencing a result, but one you haven’t yet observed or measured—a hidden construct that’s floating just out of view. So what is the hidden construct here?

  One problem was that, at a basic level, I couldn’t visualize a Lib Dem voter. I could visualize Tories, who—in the most general sense—were either posh, rich, Downton Abbey types or working-class, anti-immigrant types. Labour voters were northerners, union members, council estate dwellers, or public-sector types. But who were the Lib Dems? I couldn’t imagine a path to victory if I couldn’t imagine who’d be marching with us on that path.

  So, in the late spring of 2011, I started traveling around Britain to find out. For several months, I’d go to my classes at LSE in the morning, and then in the afternoon I’d hop on a train to places with delightful names like Scunthorpe and West Bromwich and Stow-on-the-Wold. My intent wa
s to do voter interviews and focus groups, but not the usual kind. Instead of asking prepared, scripted questions, I’d have unstructured conversations, so people could tell me about their lives and what was important to them. I could have jumped directly to polling, but I realized that any questions I asked would be biased by what I, the questioner, thought was the most relevant thing to ask. Sure, I would get answers to the questions I added to the poll, but what if they were the wrong questions? I went to speak to people because I knew I was already biased and colored by my own experiences. I did not know what life was really like for an older British man living on a council estate in Newcastle, or a single mother of three in Bletchley. I wanted them to tell me what they wanted me to know about their lives, in their own words and on their own terms. So I got local constituency parties and polling firms to help randomly select people to speak to.

  For the focus groups in smaller villages, there were often no addresses. I’d show up and be told, “We’re meeting in the cottage up on the hill. Just walk past the pub, through those fields with the daffodils, and after a while you’ll see it.” Random townspeople would show up, and perhaps Clive the local barman or Lord Hillingham the gentleman farmer would amble by. Sometimes I would just go to the village pub and chat with people there. British people were whimsical, nuanced, and often fun to chat with, and the focus groups reminded me of the town halls I had loved so much back in B.C. People would talk, and I’d just listen and take notes on what they had to say.

  Through these many conversations I traveled alone, as the party was not terribly interested in what I was up to, but I started to piece together the randomness of the Liberal Democrats. What quickly became apparent was that they lived so many different lives. They were farmers in Norfolk in tartan hats. Hipsters being artsy in Shoreditch. Old Welsh ladies in the Mumbles or Llanfihangel-y-Creuddyn. Gays in Soho. Professors at Cambridge who hadn’t brushed their hair in twelve years. Lib Dem voters were an odd, eclectic mix.

  They may have all looked different, but I noticed that they did have one common trait. Labour Party voters would say, “I’m Labour.” And the Conservatives would say, “I’m Tory.” But the Liberal Democrats would almost never say, “I’m Lib Dem.” Instead they would say, “I vote Lib Dem.” This was a slight but ultimately important distinction. It took me some thinking to figure out that it might have to do with the party’s history. The party wasn’t officially formed, in its modern incarnation, until 1988, after the merger of two smaller parties, which meant that many of its current voters originally came from historically Tory or Labour families. This meant that at some point in their lives, they’d had to make an active decision to switch from an old party to this new party. For them, supporting the Lib Dems was an act, not an identity.

  * * *

  —

  ONE OF THE PEOPLE who drew me to London was Mark Gettleson, who quickly became one of my best friends. I had first met Gettleson in, of all places, Texas. Back in 2007, when I was just starting out at the LPC, they sent me to a Democratic Party event in Dallas to do some networking. I was milling about with hundreds of people in a giant ballroom, marveling at all the Stetson hats, when a clipped British voice behind me said, “Youuuu are not from here.” I turned around to see a bloke grinning like the Cheshire cat in forest-green trousers and a floral Liberty-print shirt. Between my bleached platinum-blond hair, complete with a classic mid-noughties fringe, and his dandified manner of dress, we were drawn together like two butterflies at a moth convention.

  The son of a family of Jewish antique dealers on London’s Portobello Road, Gettleson is posh, eccentric, and delightfully camp and speaks with a delivery reminiscent of the actor Stephen Fry. In another time, he would have been a dandy in the salons of eighteenth-century London. He’s a polymath of the highest order. In conversation, he can draw connections between early-1990s hip-hop and the Franco-Prussian War without taking a breath. Gettleson and I vibed that night, and over the next couple of years I’d see him at various political gatherings in America or Britain. After I decided to move to the U.K., we immediately started to hang out, sometimes in the converted crypt underneath an old church that he’d somehow turned into a fabulous flat with bizarre antique miniatures and art lying around everywhere in a chaos that somehow worked. “I don’t do minimalism, Chris. I’m a maximalist,” he would say as I looked through all his objects.

  * * *

  —

  I FLOURISHED IN LONDON and quickly gained a wide circle of friends. Although I was studying law at LSE and working at Parliament, most of my friends were a gaggle of club kids, dancers, queens, flamboyant creatives, and design students from Central Saint Martins, one of the world’s preeminent design and fashion schools, which had produced graduates like Alexander McQueen, John Galliano, and Stella McCartney. The thing that was different about Gettleson was that, like me, he seamlessly floated between all these different worlds. At the time, he was working for the London office of Penn, Schoen and Berland, the well-known Democratic polling firm that had most famously once been affiliated with the Clintons. He was the only person I knew who could join me for a formal reception on the Terrace Pavilion at Parliament, surrounded by cabinet ministers, and then end up with me later in the night decked out in makeup, glitter, and wigs, voguing among a heaving cavalcade of queens at a Sink the Pink glam ball. Gettleson was magnetic and all my friends adored him. As gentle as he was exuberant, he shepherded twinks on nights out like a collie shepherds lambs. They would find themselves completely mesmerized by his spontaneous use of Barbie dolls and character voices to explain why Neville Chamberlain’s policy of appeasement failed—at 4 A.M. at the height of a thumping house party.

  Gettleson was also one of the few people who understood what I was trying to do with data. One afternoon, as I was complaining about my difficulties in constructing a model for Lib Dems’ voting behavior, I told him I was thinking of asking some Cambridge profs about it. He connected me with Brent Clickard, who was completing his Ph.D. in experimental psychology at Cambridge and might be able to introduce me to some professors there. Clickard turned out to be so much more than simply a conduit to Cambridge. Like Gettleson, he was a dandy, the kind of guy who dresses in tweed and always has a crisp paisley pocket square. Though he came from a wealthy Midwestern American family, he spoke with a delightfully affected mid-Atlantic accent that he somehow picked up, as if he were playing a character in Casablanca. He’d been a dancer with the Los Angeles Ballet before deciding to move to England.

  In the course of several boozy conversations, Clickard suggested I look more deeply into personality as a factor in voting behavior. Specifically, he pointed me to the five-factor model of personality, which represents personality as a set of ratings on five scales: openness, conscientiousness, extraversion, agreeableness, and neuroticism. With time and testing, the measurement of these five traits has proven to be a powerful predictor of many aspects of people’s lives. A person scoring high in conscientiousness, for example, is more likely to do well in school. A person scoring higher in neuroticism is more likely to develop depression. Artists and creative people tend to score high in openness. Those who are less open and more conscientious tend to be Republicans. This sounds simple, but the Big Five model can be an immensely useful tool in predicting voters’ behavior. In political discourse, you find that many of the phrases used to describe candidates, policies, or parties align with personality. Obama ran on change, hope, and progress—in other words, a platform of openness to new ideas. Republicans, on the other hand, tend to focus on stability, independence, and tradition—in effect, a platform of conscientiousness.

  Reading in my flat in the middle of the night, I finally realized something. Maybe the Lib Dems didn’t have a geographic or demographic base; maybe they were a product of a psychological base. I put together a pilot study and found that Lib Dems tended to score higher on “openness” and lower on “agreeableness” than Labour or Tory voters. I
realized that these Lib Dems tended to be, like me, open, curious, eccentric, stubborn, and a bit bitchy at times. This is how an artist in East London, a professor at Cambridge, and a farmer in Norfolk could all coalesce around this party in their own way, despite living very different lives.

  The five-factor model was the key that cracked the Lib Dems code—and, in the end, provided the central idea behind what became Cambridge Analytica. The five-factor model helped me understand people in a new way. Pollsters often talk about monolithic groups of voters—women voters, working-class voters, gay voters. Although certainly important factors to people’s identities and experiences, there is no such thing as a woman voter or a Latino voter or any of these other labels. Think about it: If you randomly grab a hundred women off the street, will they all be the same person? What about a hundred African Americans? Are they all the same? Can we really say that these people are clones by virtue of their skin color and vaginas? They all have different experiences, struggles, and dreams.

  Exploring the nuances of identity and personality started to help me unpack why, despite the fact that politicians do polling all the time, they still seem horrendously out of touch. This is because so many of their pollsters are out of touch. Polling firms influence politicians’ ideas of what makes up voter identity, which are usually horrendously oversimplified or just plain wrong. Identity isn’t ever a single thing; it’s made up of many different facets. Most people do not ever think of themselves as a “voter,” let alone curate an identity around how their worldview relates to tax policy. When a person goes grocery shopping, they are unlikely to stop, drop their shopping in a moment of blinding self-awareness, and suddenly realize in the middle of the store that they are, in fact, a university-educated white suburban female in a swing state. Whenever I was doing focus groups, people tended to talk about how they grew up, what they do, their families, what music they like, their pet peeves, and their personality—the kinds of things you talk about on a first date. Can you imagine how terrible a blind date would go if you were allowed to ask only the standard polling questions? Yeah, exactly.