Hosted by Kai Ryssdal

Get the Podcast

Get the Podcast

• Apple
• Spotify
• Amazon Music
• Stitcher
• RSS

Recently on Marketplace Tech, we interviewed Sen. John Kennedy of Louisiana, who said this: “People have been coming to politicians for a long time now saying, ‘I can win your election for you on the internet.’”

After all, politics is all about whom you reach, whether it’s with a handshake, a kiss on a baby’s head, a TV, radio, newspaper or billboard ad. Or all about, increasingly, whom you reach with a specifically targeted social media ad that’s carefully calibrated to your particular interests and desires and could have been paid for by a campaign or by one of thousands of political interest groups flooding the web with money.

By far, the most money gets spent on broadcast TV; but about a billion dollars was spent on digital ads in 2016. By contrast, when new records were set for political spending in 2008, the total amount spent across all media was $2.6 billion.

As online advertising expanded in both reach and technology, so did political ads. And they used data to get a lot more precise — much of that data from Facebook.

“Microtargeting is not something that’s new,” said Sunshine Hillygus, a professor of political science and public policy at Duke University, referring to the process of using personal information to craft personalized messages. She said the practice goes back to the late 1990s, but it was happening on postcards and through direct mail. Campaigns would maintain detailed “voter files,” but all the information was offline. Then Facebook came along.

“It has been with Facebook that you have connected offline and online identities,” Hillygus said. “Now what campaigns can do is give a list of people that they have identified within the voter files, and using consumer data, they’ve identified the subset of gun owners in Pennsylvania who appear to care about a certain given issues. And because so many people are using Facebook, they can submit that list of names, these online identities, to Facebook and target them online.”

She said that before Facebook allowed that level of individual targeting, campaigns would simply have to guess at likely social media ad success. That’s essentially what the Obama campaign was doing in 2012.

“You would find that the advertisements from Obama that were put on women’s focused websites tended to be more likely to mention women’s issues,” she said. “But it wasn’t the type of really, really individualized targeting that that you can do on Facebook.”

Cambridge Analytica notwithstanding, all kinds of campaigns are harvesting data from Facebook in order to target ads. When tech entrepreneur and investor Mitch Kapor, founder of Kapor Capital and the Kapor Center for Social Impact, downloaded his Facebook data (and quit the service, he said), he found that companies weren’t the only ones that had accessed his profile information without his knowledge.

443 advertisers uploaded a contact list from Facebook with my info BUT NOT MY INFORMED CONSENT including Amex, Amazon, Salesforce, Spotify, Samsung, Netflix, The Nature Conservancy, PBS, Consumer Reports, Wall St. Journal Elizabeth Warren, Nancy Pelosi, Ron Wyden, Joe Kennedy III

— Mitch Kapor (@mkapor) April 12, 2018

The list included several different politicians (and several media outlets, too — this is a good time to remind you that pretty much every company is participating in the data economy on some level).

Now, this is the part where it’s tempting to think that elections can be won or lost by the campaign that hops on the tech bandwagon quickest. But most elections turn on lots of different factors. Online advertising and social media and even the psychographic targeting that Cambridge Analytica was promising — the idea that you could know so much about someone that you could emotionally manipulate them with ads — are probably not a win-or-lose advantage. In fact, there’s no proof at all, say Hillygus and others, that hyper-targeted ads work to change people’s minds or even convince them to stay home and not vote.

“I think they [Cambridge Analytica] were trying to take advantage of the fact that they were pitching something new, and entirely overclaiming about the impact,” Hillygus said. Obama’s campaign was most successful at using online ads with a “donate” button on them, which helped him raise unprecedented amounts of money in small donations, and she said that’s probably still the grandest use of online ads.

Plus, Hillygus said, “every time I talk about online advertising with my undergraduates, I mean, they tend to chuckle, because they have ad blockers, and so they don’t even see the ads.”

But, she cautions, “there’s no going back. Nobody’s going to turn away and not use data.”

In fact, what we might start to see more of will be trying to not look like advertising. It might be what’s called “native content,” paid material designed to appear to be a news video or article. Such content might not, according to this research from the Southern California Law Review, even fall under disclosure rules governing political ads. Or the messages you end up seeing might be from Facebook groups or pages, like those that were created and maintained by Russian propaganda groups to spread misinformation during the 2016 election — even promoted posts that also don’t necessarily follow the few ad-related rules that exist online. 

“We need to, looking forward, really think through how this means of communication might be used by different interests,” Hillygus said, “and how it could be gamed for people’s financial benefits or political benefit.”