I Sold Political Ads for Google. Banning Them Won’t Work.

Tech companies can help fix what’s wrong with digital political advertising. Instead, Twitter is declaring the problem insoluble.

In the novel and movie Thank You for Smoking, the antihero Nick Naylor is part of a group named the “Merchants of Death”—or the “MOD Squad,” as the members also call themselves. They’re shills for the most hated industries in America: firearms, alcohol, and, of course, Big Tobacco. I think they need to make room in the club for me, because I used to sell a product that has become almost as reviled: digital political ads.

While political campaigns have used digital ads for more than a decade, the practice came under intense scrutiny after the 2016 election, after Facebook revealed that Russians had bought $100,000 worth of targeted ads that 10 million Americans had seen. In recent weeks, more than 250 of Facebook’s own employees have criticized the founder Mark Zuckerberg’s decision to exempt all political ads from fact-checking. Late last week, Twitter’s Jack Dorsey went to the opposite extreme by announcing a ban on all political ads—and won widespread praise for it.

But Dorsey’s decision, I suspect, will prove myopic, counterproductive, and ultimately untenable, and I hope Zuckerberg and Google’s Sundar Pichai don’t follow his lead. Most digital advertising, including most digital political advertising, is far from nefarious, and tech platforms can take steps to limit the downside. To ban political ads outright, though, is to declare the problem insoluble—a fatalism the tech companies should reject.

I joined Google in 2011 and worked for the company’s ad-sales department through the 2014 midterm-election cycle and into the 2016 primary races. (I left before the height of the campaign, to attend graduate school.) My job was to help political campaigns, along with the digital-advertising agencies they had hired, understand how Google’s various advertising tools worked—and, ideally, why they should spend more of their marketing budgets on us rather than Facebook, Twitter, or TV and radio. To tech companies, a political campaign is just another big brand with a message to promote. From the standpoint of revenue generation, Donald Trump and Hillary Clinton were no different from Coke and Pepsi.

A cynical reading of Dorsey’s move would be that he has simply become exhausted by the recurring bad PR generated by controversy after controversy over political advertising. Political ads account for only about $3 million a year of Twitter’s revenue, according to the company’s chief financial officer. (By comparison, Facebook expects to make several hundred million dollars in the 2020 election cycle.) For Dorsey, one might conclude, the revenue from political ads apparently isn’t worth the hassle.

But let’s take Dorsey at his word. He is right to worry. Digital political advertising really is more targeted than traditional advertising, as I explained many times to political campaigns whose managers and general consultants were stuck in 1995. The ability to find and speak to groups of people based on their characteristics—age, location, and gender, plus any of your niche interests—is much more efficient on digital platforms than blasting your “Save Social Security” TV ads at the Golf Channel because you think a lot of its viewers are older than 65.

This increase in ad-spend efficiency has been a boon to normal commerce. When selling a new widget or service, a start-up that would never have been able to buy expensive TV ads can actually compete with major brands for your attention. Allbirds and Dollar Shave Club, for example, probably wouldn’t have made it in a pre-digital world. Similarly, digital ads can be the most cost-effective way for new political candidates—especially those, such as Andrew Yang, who aren’t celebrities before they run for office—to make the electorate aware they exist in the first place.

One can fairly ask whether, in a democracy, we should sell policies and political candidates the same way we sell sneakers and razors. If I’m already in the market for a toaster, I’d much rather Google knew that and served me an ad for the perfect appliance for me—preferably on sale, so that I can get the transaction over quickly and efficiently. But complex, thorny topics such as abortion and immigration are up for debate, and some friction in a user’s mind is a feature, not a bug. When digital ads tell us exactly what we want to hear about these subjects, they nudge us away from considering nuances, ambiguities, and alternative points of view.

Nevertheless, the differences between traditional and digital advertising are differences in degree, not in kind. In any form of advertising, marketers try to match their message to their target audience. To ban digital ads because they’re better at achieving that goal is to draw a distinction without a coherent line. Campaigns have used “micro-targeted” direct mail for decades without nearly as much public outcry. TV advertising gets more targeted all the time. And, of course, as traditional pay-TV declines and some of the new streaming platforms decide to drive revenue with ads rather than compete with Netflix for subscriptions, the gap between digital and traditional will continue to erode until it disappears entirely. What will we do then?

Ban targeted political ads entirely, perhaps? But here we encounter the conceptual confusion at the core of Dorsey’s decision: What, exactly, is a political ad? If a newspaper promotes an op-ed by a columnist arguing that President Donald Trump is unfit for office, is that a political ad? If a movie studio run by a left-wing billionaire pays for ads promoting a hagiographic documentary about an up-and-coming progressive candidate, is that a political ad? If an eco-friendly brand such as Patagonia runs an awareness campaign about conserving the environment and fighting climate change, is that a political ad? There is simply no reasonable way to enforce a ban like this consistently and objectively. Partisans will continue to decry ads with messages they don’t like—the only difference is that they’ll be arguing over whether the ad’s content is political, rather than whether it’s true.

Even if it proved enforceable, Dorsey’s decision would have unintended consequences. Banning digital political advertising is likely to add to the already substantial advantage incumbents have over challengers. In his announcement, he claimed that the only political messages that deserve amplification are those that “earn reach when people decide to follow an account or retweet.” But the people most capable of generating “earned media” are those who already have lots of followers. In most cases, that’s the people already holding office—and the exceptions show that Democrats especially should be careful what they wish for. No neophyte candidate has mastered “earned media” better than Donald Trump did in the 2016 campaign.

Banning digital political ads will also likely hurt candidates who have innovative ideas that are unpopular with large donors. Would an underdog such as Bernie Sanders—the Dollar Shave Club of candidates—have been able to compete against Hillary Clinton as well as he did without the small donors that his digital presence enabled?

Instead of trying to ban digital political ads, tech platforms can use their power to improve them. They can make those ads more transparent by slapping larger, clearer labeling on those purchased by verified political candidates and super PACs—a step that would also distinguish such ads from ones placed by shadowy, less accountable sources. Platforms can also make it much easier for average people to tell how they’re being targeted. Each platform already allows you to see the data it has collected on you that it uses for ads. For example, here are Google’s directions for how to access your data; the steps are too long and convoluted, and unsavvy users will hardly ever bother. To their credit, since the 2016 election, Facebook and Google both developed searchable archives, accessible to the public, of every ad from campaigns and issue groups.

What all these steps require is for tech companies to acknowledge the broader social consequences of the ads they carry and make good-faith efforts to force certain ad buyers to identify themselves. The companies should produce more detailed information about who is seeing which ads and make it more accessible to outside observers. Big Tobacco was a Merchant of Death because it didn’t care about, and refused to acknowledge, any harm it was causing. Tech platforms have a more constructive role to play.

Jack Dorsey isn’t Nick Naylor. Neither is Mark Zuckerberg. If not for Trump’s victory over Clinton in 2016, the current collective freak-out over digital ads might not be happening. Still, the conversation is long overdue. But the tech platforms are too important to our political conversation for their leaders to simply huff, puff, and walk away—instead of focusing on what they can do to help.