How Mass Customization Won the U.S. Presidency

February 21, 2018

Note: A version of this post originally appeared in Forbes.

When I started out in advertising, no one looked to political advertising as a hotbed of innovation. Political advertising was something like a backwater — a place for the rushed and inefficient implementation of tactics pioneered elsewhere.

Nobody would say that today. Thanks to all the money flooding each election cycle, politics has become a critical proving ground for the technological cutting edge.

It’s no wonder that there’s now so much cross-pollination between political advertising and brand marketers. Political advertising is designed to employ data to prompt individuals to take action (to vote), or maybe to deter action (i.e., keep voters at home); much brand advertising today is also oriented around getting consumers to take action of some kind.

Beginning with the online-focused approach of the first Obama presidential campaign in 2008, which placed data-driven tactics at the heart of political strategy, every successful campaign has rested on its ability to message uniquely to individual voters on a mass scale — an art that found its most recent and most instructive expression in Donald Trump’s surprise victory in 2016’s presidential election. 

Trump’s Facebook-centric campaign

Because Trump was behind in the polls for his entire campaign, the media presented him as an underdog. Actually, the media helped tip the scales because it couldn’t stop covering him: while the Hillary Clinton campaign’s budget was more than 40% larger than Trump’s, he received what The New York Times estimates as $2 billion in free media attention largely because of his ability to stay in front of the news cycle. But in many respects the 2016 presidential battle was waged and won on social media, not traditional channels — meaning credit for Trump’s victory must go to Brad Parscale, the campaign’s digital director.

Parscale’s choices were masterful. As a recent 60 Minutes report revealed, the Trump campaign worked closely with Facebook (which also offered the same help to the Clinton campaign) to create 50,000-60,000 targeted ads per day. Those ads took advantage of dynamic customization to change up the wording, images and colors to determine which drew the best reactions from the individuals being targeted — and then doubled down on ads that worked. In Parscale’s words, he was looking for ads that made voters say “Poof! I’m gonna stop and look.” To achieve that, Parscale crafted ads that focused on issues like taxes, energy or childcare.

Parscale also took full advantage of Facebook’s micro-targeting abilities to identify and influence rural voters that costly television campaigns could never reach, explaining “Now, I can find 15 people in the Florida Panhandle that I would never buy a TV commercial for.”

In addition, the Trump campaign used Cambridge Analytica, a firm principally owned by Trump supporter Robert Mercer, to create psychographic profiles that could predict the political leanings of American adults and target advertising accordingly. For instance, a “neurotic” voter would see a gun-rights ad that showed burglars breaking into a home, while a more sanguine Second Amendment supporter would see something else. The company looked for signals to create its psychographic profiles: while a Lady Gaga fan is likely to be an extrovert, for instance, a fan of philosophy is apt to be an introvert.

Parscale’s use of data didn’t end there. While the deep-pocketed Democratic National Committee’s data was “mediocre or poor,” according to Clinton, the Trump campaign used fresher data to pivot late in the election and woo crucial voters in Michigan and Wisconsin. Though Clinton had better number crunchers, they relied on out-of-date data like voter files and polls rather than more real-time information from Google searches, which indicated late in the campaign that her support was slipping in the Midwest.

Such distinctions turned out to be crucial. Trump didn’t win the popular vote, but tellingly won a plurality of electoral college votes, targeting the states that mattered and ignoring the ones that didn’t. It is likely that the micro-targeting that Parscale referenced in Florida (as well as Pennsylvania, Michigan, Ohio and Wisconsin) is the reason that we now have a President Trump. 

The results don’t lie

Putting aside political leanings, every CMO can relate to Parscale’s situation — underfinanced and overwhelmed, with few advantages over the competition apart from creativity, resourcefulness and grit. But Facebook and tools like those leveraged by Cambridge Analytica herald a new era in marketing, in which targeted messages can be delivered to specific individuals at a scale once reserved for the un-targeted mediums of network TV and radio. The playing field has leveled.

Trump’s win is the death knell for mass marketing: this is the age of mass personalization, and the Trump campaign capitalized with the type of precision and efficiency that marketers aspire to across every product category. Rather than waste money over-advertising, Parscale honed in on some pivotal voters and customized his message. In other words, he trusted the data, and had the discipline and insight to know and engage the right audience.

Of course, Russian meddling (something Parscale denied) is another possible variable. But if those allegations are true, that means the Russians also skillfully used Facebook targeting to get out the vote. Which only further proves my central point: the efficacy of people-based targeting at scale is now beyond dispute.

Whatever your feelings about Donald Trump, give him his due: his team used the latest advances in data targeting to achieve a stunning victory that few predicted. If nothing else, that underscores the real-world potency of effective data-based targeting. That’s not the only lesson from the 2016 election, but it’s one that marketers would be wise to learn.

Mike Sands is CEO and Co-founder of Signal.

Subscribe for Updates