The season of computer-assisted consumption is hard upon us and, as shoppers, we are fairly sanguine by now about the data sweeps that infest our virtual worlds, mining our Pinterest pages and haunting our LinkedIn profiles. But it’s one thing to be manipulated when you’re picking out a new power drill from Sears or a sweater from Eileen Fisher. It’s quite another when you’re choosing a national leader.
Ever since Barak Obama deployed sophisticated micro-targeting techniques to identify voters in 2008 and 2012, candidates in both parties have used market analytics to refine their messages, slicing and dicing the electorate by age, race, gender, income, education, and subtler indicators such as magazine subscriptions or the model of the family car.
This devotion to demographics is hardly new in politics, but powerful computing and robust social media have allowed for deeper dives into our cultural preferences. For at least the last decade, we have been cheerfully dividing ourselves up into affinity groups: latte liberals, NASCAR dads, chablis-sippers, church-goers, and so on. We wear lookalike clothes, choose the same restaurants, and have our biases confirmed by like-minded media. It’s irresistible shorthand, comforting and even fun, but it’s not helpful for democracy.
Advertisement
Micro-targeting puts us into silos that make it much harder to talk about our differences — or much easier to smugly dismiss them. It makes campaigns better at something voters consistently say they despise about politics: telling them only what they want to hear. The “checklist” quality of Hillary Clinton’s campaign promises — free college tuition for millennials, tax breaks for young families, thousands of ads released in Spanish — underscored a sense that she saw voters more as a series of data points than flesh-and-blood citizens. And of course, Donald Trump stoked racial, class, and gender resentments into a raging fire.
The Electoral College, with its winner-take-all distortions, exacerbates the notion that we are a country segmented into red and blue turf, when every state really is some shade of purple. No less an authority than Bill Clinton lamented this drift into opposing camps, in May, when he told a college graduating class that despite their many advantages, the Internet and social media have “laid bare the power of persistent inequalities, political and social instability, and identity politics based on the simple proposition that our differences are all that matter.” But identity politics is precisely how Hillary Clinton built a winning coalition in the primary, only to have it used against her by an opponent gleefully deriding political correctness. Of course, Trump’s campaign also hired an international firm specializing precisely in “psychographic” targeting.
Advertisement
Much ink has been spilled over how pollsters, with all their advanced computer modeling, missed the Trump surge. The campaigns also miscalculated. But the rush to embrace big data doesn’t just steer us wrong; it also can undermine democratic ideals. Last year, Facebook unveiled a new “audience network” tool, promising its advertisers greater control over who would see their online ads. Now the social media giant is facing charges that its software allows advertisers to exclude certain groups from seeing their content. Federal officials are examining whether the tool allows users to discriminate, for example, in housing and employment ads. Facebook promises to end the practice.
Of course we should all be proud of our diverse backgrounds and celebrate the national mosaic. But identifying too fiercely with our own demographic profiles deprives us of the richness in others, and cuts us off from appreciating the values we share: fairness, kindness, acceptance. At the risk of sounding like a NASCAR dad, I’d say it’s time to resist big data’s typecasting and be Americans first.
Renée Loth's column appears regularly in the Globe.
Advertisement