For a generation now, a campaign manager has been able to select a list of, say, 100,000 names to receive a pre-election get-out-the-vote reminder and feel confident that the reminder will reach only those 100,000 voters—and not their neighbors. The voters’ addresses can be delivered to a mail vendor (who merges them onto glossy leaflets) or placed on a walk list (which a field director hands to canvassers), or the voters’ telephone numbers can be given to a phone vendor (whose call centers will reach them with live operators or by robocall).
This year, for the first time, campaign managers in races of all sizes will have a new option: individual-level targeting of Web ads. In this scenario, there is a crucial intermediate step. The 100,000 names selected for targeted communication are then matched to Internet cookies, which allows a campaign to buy ads on only those pages visited by its targeted voters. This represents one dimension of the most important innovation of the 2012 election cycle: the ability to match an individual’s online and offline identities. It means that campaigns can now target voters wherever they are, even if they’re at their vacation home for the summer or spend most of their online time in corners of the Internet where people do not typically seek out political content. As a result, individual targeting marks a crucial step in the maturing of the Web from a media platform and forum for fundraising and activist organizing to a corridor for direct voter contact.
But even if campaigns have finally acquired the technical capacity to target Web ads with the precision of mail or a door-do-door canvass, they are not using it that way. While political advertisers can specify which 100,000 individual voters they would like to reach, only a small fraction of those voters maintain a large enough online footprint to actually be targeted. If they wanted to, the campaigns could learn which of their 100,000 targets actually saw their ads, and which didn’t. But spooked by a furor over online data collection by consumer advertisers, major political players are voluntarily refusing that information. As a result, online targeting is lagging in efficiency behind more traditional methods: A campaign manager knows the names of targets who were unreachable by a phone vendor’s call centers, and can send them an extra mailer instead. He gets a list of which doors went unknocked upon by his canvassers so the voters living there can be targeted to receive robocalls. But political advertisers have shut themselves off from the analogous online information: The Obama and Romney campaigns do not find out which voters were actually delivered their ads, according to those familiar with each campaign’s targeting practices.
Indeed, conversations with the consultants who sell these groundbreaking individual-level online targeting services are notable for the fact that they seem to devote less time in their pitches to claiming they can match the efficiency of offline methods than they do offering assurances about how respectful they are of the online advertising industry’s self-regulating standards on obscuring personally identifiable information, known as PII.
“In terms of campaign strategy, the PII rules are onerous and out-of-step with the norms of voter contact,” says Chris Talbot, who as a Google account executive in 2008 guided the Obama campaign in its online-advertising placement and now has his own consulting firm that advises Democrats on digital strategy. “Any citizen would tell you it's silly that campaigns use massive amounts of personal information—including your party and voting history—to call your house during family dinner yet they can't use that info to serve you a banner ad while you're reading NYTimes.com.”
Even though this type of we-know-who-you-are surveillance is widely accepted as part of our political culture, it has become taboo among marketers. They have imposed standards on their own industry that enshrine consumer privacy as a way of pre-empting government regulation. (Last month, the Federal Trade Commission released a report on online privacy that applauded efforts by an industry umbrella group, the Digital Advertising Alliance.) The rules, requiring that users be clearly warned or given the chance to opt out if their PII is being collected or shared, were a minor concession by consumer advertisers. After all, as long as a commercial marketer knows that ads are reaching someone in the target audience—say, a rural man between the ages of 35 and 54 for a truck ad—it doesn’t quite matter the specific identity of the person who sees it, and it’s probably irrelevant which side of a county line the viewer calls home.
But that is crucial information to political campaigns as they seek to shift their broadcast and direct-mail budgets onto the Internet. In 2010, one of the stories in the Wall Street Journal’s “What They Know” series on online data collection focused on the San Francisco-based firm RapLeaf, which helped clients—including 10 campaigns that year—deliver targeted ads to individual voters using cookies matched to email addresses (which had been scraped from social-media networks). The firm built targeting profiles resembling those long employed by direct-mail firms. For example, there was the 67-year-old New Hampshire woman whom RapLeaf’s databases had accurately presented to a Senate campaign as a Bible-reading Republican supporter of environmental causes. Many online advertisers analyze individuals’ online behavior when placing their ads, but RapLeaf drew particular attention because its cookies had real names and ages attached to them.