Twitter has an advertising problem, but despite disappointing quarterly earnings reports and frustrated investors, it’s not what you might think. Earlier this week, a handful of users were confronted with a tweet proclaiming that “Whites need to stand up for one another and defend ourselves from violence and discrimination.” It might have been possible to dismiss this message—to treat it as a mere product of Twitter’s culture of offense, to block the user, to move on—were this not a promoted tweet. Someone had paid to ensure that this message would show up in users’ feeds. And that meant Twitter was making money from its presence.
Making money, sure, but not much money, as it turned out. The offending tweet was the brainchild of self-proclaimed hacker Andrew Auernheimer, who goes by the nom de troll “weev.” As he explains in a Storify post, Auernheimer realized that it was possible to promote tweets to specific groups of users. By way of example, he notes, “You can … choose to display ads to followers of specific users, like @Jezebel or @feministing.” This system is designed to help advertisers precisely target their campaigns at those most likely to respond to their messages. In Auernheimer’s case, this meant aiming at individuals most likely to be offended by his “lulzy” “white survival” message.
Because he had narrowly restricted his missive to a relatively small percentage of Twitter’s 302 million active accounts, Auernheimer’s calculated campaign of offense cost him relatively little. On Storify, he suggests that he spent only “a few pennies,” thanks to “some CPA pricing structures” that charge by the total number of impressions advertisers hope to generate. While he doesn’t list the actual cost of his initial campaign, Auernheimer suggests that he’s willing to “troll twitter some more” if his readers send along a bit of cryptocurrency. Nothing, it turns out, is cheaper than offense.
And offend Auernheimer’s tweets did. He reports responses that range from simple bafflement to undisguised anger. Many of those he cites reached out to Twitter itself, asking, for example, why the site had sold “ads to literal Nazis.” Of course, Twitter’s largely automated advertising system meant that no one had consciously sold these ads; the platform simply failed to prevent them from going out in the first place. A Twitter representative told Ars Technica that Auernheimer’s offensive tweets “had been deleted because they violated the service’s hate/sensitive ads policy.” It’s not clear, however, that it’s done anything to prevent such efforts in the future.
More troublingly, Twitter’s promoted tweets supersede many of the protections and privacy filters that the site offers. More than one of Auernheimer’s targets had previously blocked him. Given that Twitter gives its customers considerable control over who sees their ads, this loophole could make it easier than ever to employ the site as a platform for abuse. While the company explains that it allows users to opt out of tailored tweets in their own timelines (via a check box located in the “Security and privacy” tab of an account’s Settings page), it doesn’t let them to actually avoid the service more generally. They’ll still see promoted tweets on the profile pages of others, in search results, and elsewhere. Twitter allows only that if a user dislikes one of these messages they can “dismiss it from their timeline with a single click.”
Twitter isn’t the only major site with a self-serve advertising system. In 2013, Brian Swichkow exploited Facebook’s similar interface to torment a single user with even more targeted, albeit less hateful, messages. As Swichkow showed, Facebook is even easier to game, since messages don’t have to come from a specific account. Accordingly, he was able to gaslight his roommate with carefully constructed ads that made it seem as if the site had data on him that he’d never knowingly shared online. Swichkow claims that his prank cost him a mere $1.70.