Just because Cambridge Analytica tells its customers it can sway elections, it doesn't follow that they're any good at it

Unilever founder John Wanamaker famously said, "I know that half the money I spend on advertising is wasted. My only problem is that I don't know which half." It's an odd testament to the power of advertising, an industry whose executives are incredibly effective at selling their services to other executives, even if they can't prove they're any good at selling their customers' products to the public.

For more than a year, Cambridge Analytica has been trumpeting its incredible ability to use "psychographic targeting" and the "big five personality traits" to change the way people vote. It's an incredibly self-serving claim, marketing hype dressed up as a news story. The last election was carried by a whisker-thin margin of electoral college seats, and virtually any factor can be said to have "delivered the election," because if literally anything changed (voter suppression, propaganda, Dem establishment demoralizing Sanders voters, Jill Stein, etc) the election might have flipped.

So, as I've written before, we should take Cambridge Analytica's claims to Svengali-like mind-control with a boulder of salt, because until Sunday, they made these claims to drum up business (now they're busily declaring that they are no more persuasive than any other ad agency, of course, because they've gotten in trouble for it).

That doesn't mean that Cambridge Analytica and the Trump campaign (and their other customers) don't believe that they're right, of course. By definition, their campaigns exist in a world of secrecy, where critics and skeptics aren't allowed in the room — like any other secretive organization, they're prone to reinforcing one-another's unfounded beliefs, unchecked by the need to defend these beliefs to anyone else (see also: climate deniers, anti-vaxxers, young Earth Creationists, eugenicists, Islamic extremists and readers of Gwyneth Paltrow's Goop).

It's no coincidence that Cambridge Analytica's execs boast of their ties to ex-spies. Spies are the original self-deluders, living in a bubble of secrecy, where there were no consequences for getting it wrong. That's why the reality of spying is such a shitshow of nutty conspiracy theories, double- and triple-agents, and always, always bungling.

Psychographics is a perfect example of insider-knowledge bullshit, a go-to tool for advertising execs who are pitching potential clients. Not only are the claims of the persuasiveness of psychographic targeting unsupported by the literature of psychology (because these are all "proprietary techniques" that can't be subjected to outside scrutiny without revealing precious trade secrets), but that literature is catastrophically riddled with irreproducible results that calls much of what we claim to know about human psychology into question.

Which is not to say that there isn't a thriving and effective marketing discipline online: Amazon and other big ecommerce platforms have shown that tactics like retargeting (where you see an ad for something you shopped for earlier, but didn't buy) can bring in additional revenue — but these gains are very modest. The baseline conversion rate (the rate at which someone buys after searching) is less than one percent; retargeting can improve this to a figure that is still in the sub-one-percent range.

But notably, Amazon and the other big ecommerce platforms (who do this stuff all day long, pay better than anyone else, and aren't limited to elections every four years) don't use psychographics to sell stuff. If it worked, it's a sure bet they wouldn't scruple to eschew it.

Cambridge Analytica's psychographics claims are untested and implausible. So did they have an effect on the election? Maybe. By their own account, they make up discrediting stories about candidates and use sex workers and blackmail to influence election outcomes. They bought a hell of a lot of Facebook ads, too, and maybe some of these fit in the Wanamakerian "which half" that worked.

Take Cambridge Analytica out of the picture and what do we have left? An electorate that is so demoralized with the way things are going they'll vote to burn it all down. Victory for candidates (Trump) and ideas (Brexit) that are manifestly unfit, but at least aren't more of the same. Racism. Xenophobia. Lies that are believed.

So long as we take Cambridge Analytica's claims at face value, we can say that the problem wasn't business-as-usual — mass inequality, policies that benefited the richest and whitest among us — it was Robert Mercer's magical mind-control ray. Likewise the claims of Russian interference: if Russian interference was the problem, then all we have to do is fight off Russian propaganda — no need to purge the Democrats of advocates for the finance industry, mass surveillance, authoritarianism and endless war (Russia definitely interfered, and the interference may have moved the needle for Trump into the red-zone, but only because the needle was right on the edge to begin with).

The Cambridge Analytica scandal is a life-preserver to people who benefited from business-as-usual, a story that papers over the need for change. If Cambridge Analytica are sorcerers, then there's no problem with business as usual, and once we burn the witches, we can get back to the way things were, forever.

They had to train a predictive model that guessed what sorts of Likes or Facebook profile data their targeted political archetypes possessed. In other words, now that Cambridge had a test set of people likely to vote for Trump, and knowing their profile data, how do they turn around and create a set of profile data the Trump campaign can input to the Facebook targeting system to reach more people like them?

Note that the aspiring psychograficist (if that's even a thing) is now making two predictive leaps to arrive at a voter target: guessing about individual political inclinations based on rather metaphysical properties like "conscientiousness;" and predicting what sort of Facebook user behaviors are also common among people with that same psychological quality. It's two noisy predictors chained together, which is why psychographics have never been used much for Facebook ads targeting, though people have tried.

While these conclusions are hard to make categorically even with the data in hand (and impossible to make without), a straw poll among my friends in the industry reveal near-unanimous skepticism about the effectiveness of psychographic targeting. One of the real macro stories about this election and Facebook's involvement is how many of the direct-response advertising techniques (such as online retargeting) that are commonplace in commercial advertising are now making their way into political advertising. It seems the same products that can sell you soap and shoes can also sell you on a political candidate.

Conversely, if this psychographics business is so effective, why isn't it commonly used by smart e-commerce players like Amazon, or anyone else beyond the brand advertisers who like keeping old marketing folklore alive?

The Noisy Fallacies of Psychographic Targeting [Antonio García Martínez/Wired]