The disturbing implications of increasingly narrow political ad targeting

Halloween decorations and support for U.S. President Donald Trump are seen in the front of supporter Maranda Joseph's yard in Warren, Ohio, U.S., October 2, 2020. REUTERS/Shannon Stapleton

In a world unbridled data collection and a digital economy built on its targeted use, some technology executives are beginning to question the logic of surveillance capitalism. In recent weeks, Apple CEO Tim Cook has drawn a line regarding mass data collection as his company prepares to roll out changes to its privacy policies. The update, expected to arrive this spring, requires iPhone apps to ask users for permission, through a pop-up prompt, to track them across apps or websites owned by other companies.

“If a business is built on misleading users, on data exploitation, on choices that are no choices at all, then it does not deserve our praise. It deserves reform,” Cook declared in a recent speech.

Consent (or denial) of tracking is the first step in returning control of data to those who generate it. The next step is to interrogate the motive behind the comprehensive and granular digital tracking that is the bedrock of advertising. At what point does the precision of ad-targeting transmute from helpful to manipulative? This is the question that we must ask ourselves, as the creators of data and targets of ads, in order to decide what is right for our world and demand it. Cook put it well in the same speech: “We’re here today because the path of least resistance is rarely the path of wisdom.”

Legally collected data is often abused to do things that state and federal data privacy laws never contemplated, and according to interviews with political strategists and executives from advertising and data brokerage firms, the subversion of such data is fueling polarizing trends in U.S. politics. The propaganda research team at UT Austin’s Center for Media Engagement found that political operatives are using such data to carry out increasingly granular geo-targeted advertising—sometimes so granular that it is used to target one individual. But these operatives are finding that demographic data allows for more invasive targeting than location data and are beginning to target messages at niche audiences through the coordination of small-scale social media influencers.

Political groups rarely have the same targeting capabilities as major companies, but the use of data in politics and digital political communication is important to investigate because the outcome, whether it be the passage of a law or the election of a politician, is arguably more consequential than the decision to buy a new shampoo or lampshade. The origins of using massive sets of online data in U.S. politics can be traced to the 2008 Obama campaign. Since then, the use of such data sets has evolved rapidly and broadly.

Location data, in particular, has increased in importance, emboldening what we call geopropaganda. Tracking individuals attending political rallies and other events, and targeting them with political messages based on their attendance and other variables, has become commonplace. The 2018 Beto O’Rourke Senate campaign and the 2019 Dan Bishop campaign collected mobile ID numbers from individuals’ phones without expressed consent, as did the Trump campaign in 2020.

What changed in 2020 is the tracking and targeting of individuals, donors, and lawmakers with tailored messaging at a level so precise that it warrants intervention. One executive at an advertising and public relations firm said, “We geo-target and geo-fence donors’ homes and deliver individual ads to them.” The interviewee explained how effectively they were able to target a political donor in Colorado. “We hit them with [over-the-top] ads and then he mentioned it to the clients [saying], ‘Hey, you know, I see you guys everywhere. I even see you on my TV.’” Politicians and lobbyists are also targeted. One political strategist said, “You combine geo-targeting with employer targeting. You can really direct messages almost specifically to legislative staff or Hill staff… you’re going to pick up the reporters, the lobbyists who are visiting the Capitol and nearby. So things like that could be really fun.”

This level of detailed advertising targeting only scratches the surface of what is possible. One head of political advertising at a major advertising firm described how targeting based on race, gender, income, education level, and even “behavioral traits or attributes would be way more effective” than geolocation data. In describing how to slice and dice messaging among a given electorate, the executive described how different demographics would receive different messages: “This message is only for women. Only in these ethnicities and only at this income and education level. We’re going to tell this story to them. Then we’re going to tell a completely different story to this different ethnic group, a different story to immigrants, a different story to businesspeople. And we’re going to whip everyone into a froth and no one else is going to see the other messages.”

Against the backdrop of hyper-specific targeting, the political influence industry is also turning toward micro-influences to spread their messages. The political mobilization of small-scale social media influencers (less than 50,000 followers) offers appealing features to advertisers: They are regarded as more trustworthy by their followers, have higher engagement rates, and are less expensive than large-scale influencers. Their audience members are also easier to target as they often live in similar regions or share similar interests. One executive at an influencer firm told us they “can find influencers who live in a given area, in a given city, in a given town, and find people whose audiences also live in that area. We’ve identified 24,000 micro-influencers who live in swing states.”

As it stands, determining the limits of these types of targeting is left up to political campaigns and the companies they hire. The same interviewee who explained the efficacy of demographic targeting expressed an awareness that the tactics employed by his industry could easily fall out of favor with policymakers. If a data broker, such as Foursquare, “lets a bunch of Republicans harass Nancy Pelosi at home, it won’t be long before [she] introduces legislation that makes what Foursquare does completely illegal,” the interviewee said. Unsurprisingly, any restraint stems from self-interest, and if political campaigns with limited budgets can accomplish hyper-specific targeting, imagine what large companies can do. Even some of the strategists involved in this work find it distasteful. “I would tell you, my inner libertarian, it still makes me uncomfortable in terms of how much information is shared out there that is about me,” a strategist for an advertising firm said.

The applications of targeting that emerged from the political realm in 2020 raise serious questions about the disingenuous limits companies self-impose. Apple’s moves to update its privacy policy asks users the extent to which they are comfortable being tracked, and this move might be considered the first wave of reform. It’s time to create more tools and legal protections that enable individuals, the creators of data, to decide for themselves how it is used.

Katie Joseff is a senior research associate with the Propaganda Research Lab at the Center for Media Engagement at UT Austin.
Joel Carter is a research associate with the Propaganda Research Lab at the Center for Media Engagement at UT Austin.
Samuel Woolley is the program director of the Propaganda Research Lab at the Center for Media Engagement and an Assistant Professor of Journalism and Media at UT Austin.

Apple provides financial support to the Brookings Institution, a nonprofit organization devoted to rigorous, independent, in-depth public policy research.