How online ad targeting weaponizes political misinformation
Photo Illustration: Sarah Grillo/Axios. Photo: Chip Somodevilla/Getty Images
Ad targeting is how Facebook, Google and other online giants won the internet. It's also key to understanding why these companies are being held responsible for warping elections and undermining democracy.
The big picture: Critics and tech companies are increasingly considering whether limiting targeting of political ads might be one way out of the misinformation maze.
- Under this approach, giant platforms would still allow campaigns and candidates to purchase political ads — but the companies would stop (either voluntarily or by law) selling messages aimed only at narrow segments of the electorate.
How ad targeting works
Facebook and Google have somewhat different systems for targeting ads, but both allow advertisers to bid on narrowly defined demographic groups or keywords.
- For instance, you can tell Facebook to show your message only to Southern men who don't have a college degree and earn less than $75,000 — or ask for married suburban moms in three zip codes outside Indianapolis who own SUVs and play tennis.
- Both platforms have some restrictions on what you can and can't target based on local laws.
In the political ad world, these tools give candidates and groups a chance to reach narrow slivers of the population at affordable rates. They also allow candidates and groups to exploit those populations' anxieties and resentments efficiently.
The link between ad targeting and misinformation
Tech platforms stand accused of multiple sins, including:
- Improperly collecting users' data to build massive databases of profiles.
- Allowing politicians and their campaigns to spread lies.
- Creating partisan "echo chambers" and "filter bubbles" that segment reality by ideology.
Facebook and Google didn't invent these phenomena — they existed pre-internet. But by tying them together, ad targeting can kick misinformation into overdrive.
- Data collection and profile building is what makes ad targeting possible. It's also what keeps getting tech platforms in trouble with users and governments.
- Campaigns have always shaded the truth and even lobbed false accusations. But in a broadcast world, it was easy for opponents and neutral third parties to witness and call out such behavior.
- In the world of micro-targeted ads, it's almost impossible — despite transparency efforts like Facebook's ad library.
- Misleading ads can fuel frenzies in the closed-loop worlds of partisan echo chambers long before platforms can step in to bar them — even if they wanted to.
As Yael Eisenstat, formerly Facebook's head of elections integrity and a past national security adviser to former Vice President Joe Biden, put it: "[Facebook's] business model exploits our data to let advertisers aim at us, showing each of us a different version of the truth and manipulating us with hyper-customized ads."
Banning all political ads vs. banning targeting
Many critics have urged social media platforms to bar political advertising altogether — a move both Google and Facebook have resisted, even as their smaller but politically high-profile competitor Twitter said it would embrace it.
- Facebook argues that such a ban would harm outsider candidates and causes.
- Twitter's ban is significant, but its ad market share and targeting capabilities are minuscule compared to Facebook's and Google's.
The idea of a targeting ban has gained momentum recently, with figures like Bill Gates and Federal Election Commission chair Ellen Weintraub endorsing it.
- "Just because microtargeted ads can be a good way to sell deodorant does not make them a safe way to sell candidates," Weintraub wrote in the Washington Post.
- "It is easy to single out susceptible groups and direct political misinformation to them with little accountability, because the public at large never sees the ad."
Yes, but: Defenders of the status quo argue that online targeting isn't fundamentally different from longtime campaign practices like zip-code targeting of postal flyers, and they maintain that it's a free-speech issue.
Our thought bubble: Politicians have always shaded their messages based on their audiences, and it's often how they get into trouble.
- Mitt Romney in 2012 said that 47% of Americans wouldn't support him because they live on government aid and don't pay taxes.
- In 2016, Hillary Clinton said half of Trump's supporters belonged in a "basket of deplorables" and were "racist, sexist, homophobic, xenophobic, Islamophobic — you name it."
- Both were talking to crowds of like-minded donors at fundraisers.
The bottom line: Targeted ad platforms online have given a turboboost to the practice of saying different things to different audiences — and made it harder than ever to counter misinformation with truth.