Pete Hanlon likes to save a few bucks when he can. So when the 44-year-old Phoenix resident walks down the dairy aisle, he keeps an eye out for the yellow tags advertising a discount, in return for scanning his plastic grocery discount card.
On one recent trip, he saved $10 on a $90 purchase.
And while Hanlon said he figured the store was keeping track of what he buys, he didn’t realize that — for a hefty price — that sort of detailed consumer data is being leveraged by political campaigns trying to decide whether Hanlon is someone they should target as a persuadable voter or even to hit up for a contribution.
“I remember as a kid hearing that when we grew up we would have bar-codes on our arms that would be scanned before we could buy anything,” Hanlon said. “It feels invasive. It feels a little ‘big brother.’”
The technique is called “microtargeting,” and firms that specialize in it fetch big money to obtain all sorts of data, like Hanlon’s purchases, and use it to help campaigns influence the outcomes of elections.
There’s practically unanimous agreement, however, that microtargeting has made an impact in political campaigning and that it’s only going to become more prevalent.
Broadly speaking, microtargeting is an emerging, highly detailed application of old campaign tactics. It’s all about finding the right voters to target and serving them a tailored campaign message that will win their support.
The techniques have been borrowed from what the consumer marketing industry has known for a long time, said Blaise Hazelwood, president of Alexandria, Va.-based Republican microtargeting firm Grassroots Targeting.
“Microtargeting is simply finding the right voter or consumer and the right message to either motivate them to vote or purchase your product,” Hazelwood said.
To do that most effectively, a microtargeting firm will want to know as much as possible about the voters in the election area they’re targeting.
And the key to accessing that knowledge, microtargeting practitioners say, lies in the faint trails of data people leave behind every day.
“There are usually about 300 to 500 variables,” Hazelwood said. “It’s everything from what kind of car you drive to the magazines you subscribe to, organizations you’ve donated to, whether you have a swimming pool, the size of your house, the size of your mortgage, or whether you take the train to work.”
Ken Strasma, president of Strategic Telemetry, the Democratic microtargeting firm that is also serving as the mapping consultant for the Arizona Independent Redistricting Commission, said that although they may rarely consider it, consumers willingly participate in data collection, since data tracking is practically always an opt-in system.
“If you’re using a store discount card, the store isn’t giving you that discount out of the kindness of their hearts,” Strasma said.
In order to get their hands on rich consumer data, microtargeting firms need to go through what microtargeting professionals refer to as “data houses.” These large companies, such as Experian, Acxiom and InfoUSA, pay for huge swaths of consumer data from an array of sources, then crunch it into usable data for sale to microtargeting firms.
Representatives from those three data houses either could not be reached or declined to comment, but Strasma estimated that the money being spent by these companies to purchase raw consumer data to be “tens of millions, if not hundreds of millions of dollars.”
Purchased personal data talks to strategists
After purchasing data, microtargeting firms begin building a model of the potential voter. Across the industry, microtargeting firm representatives find political indicators in just about every personal preference. Golf versus tennis or health food versus soda pop — each choice indicates something.
A preference for football indicates Republican leanings. Basketball indicates Democratic leanings. Religious purchases can say a lot about social-issue stances. Professional licenses can help paint a picture of a voter. Owning a motorcycle strongly suggests libertarian leanings and gun-rights advocacy. A subscription to Good Housekeeping says someone will favor female candidates.
A subscription to Mother Jones says a person probably has liberal leanings.
While many preferences can be gleaned from consumer data, it is only one part of the equation. The most accurate model of a voter will emerge from the combination of consumer data with personal voting history and census data.
Data from the decennial census and the American Community Survey, a detailed survey administered every three years in a similar fashion to the census, provides insights into ethnicity, income, family size, marital status and a number of other indicators. A voting history won’t say for whom any person voted in an election, but it will say how frequently a person votes and in which elections, indicating the likelihood a person will vote in the future — a key piece of knowledge for someone looking for a reliable vote.
CEOs of the larger microtargeting firms say the most effective microtargeting plan will also include a very large sample size survey, where voters within the campaign area are surveyed about their preferences.
So if Hanlon gets a survey call from a microtargeting firm that already has a near complete voter model built of him, they might ask him how he feels about an issue or a candidate. His answers would then be associated with the voter model built of him.
Firms repeat that process 10,000 times, and end up with specific messages that will work for different voter types within the election area.
John Aristotle Phillips, CEO of Washington D.C.-based, non-partisan microtargeting firm Aristotle, Inc., said the result of the survey will tell a candidate, who might have five main campaign messages, which of those should be delivered to which voters the candidate is seeking to win over.
“Maybe message number four and five will be effective with voters of some particular age and some particular income level, and maybe message number three will be in sync with a different set or a subset of those voters,” Phillips said.
Hanlon, along with thousands of other hypothetical grocery discount card users and survey respondents, will have helped the microtargeting firm establish which messages to push to which portions of the larger electorate.
Republicans build the Voter Vault
The history of political microtargeting methods goes back only about a decade.
Republican strategist Karl Rove is widely lauded as having helped foster these concepts that were developed and refined during the George W. Bush presidential campaigns. The increased ease in digitally archiving, analyzing and connecting disparate data solidified the method as a tenable campaign tool. The result of those efforts was a national database used by Republicans called the Voter Vault.
Hazelwood, of Grassroots Targeting, helped develop the Voter Vault before starting her microtargeting firm.
In 2008, however, national Democratic strategists came back from behind and even surged ahead of the Republican microtargeting efforts with their own database construction during the 2008 Barack Obama presidential campaign. Their national database, dubbed VoteBuilder, has generally been accepted as a superior voter database, according to multiple sources interviewed for this article.
Though the parties’ large voter information databases were developed from within national presidential campaigns, access to them is no longer reserved for candidates or parties that can spend top dollar on the tactic.
Scaling the process down to the point where statewide, congressional, municipal and even legislative candidates can get into the microtargeting game is now a reality.
The costs are highly variable and depend mostly on how large and detailed the voter survey is. Strategic Telemetry, the firm that works for the Arizona Independent Redistricting Commission, collected about $500,000 from the 2008 Barack Obama presidential campaign and about $600,000 from Michael Bloomberg’s 2009 New York City mayoral campaign.
The 2012 presidential campaigns are expected to easily top those figures in their microtargeting efforts.
A statewide or congressional microtargeting campaign could cost tens of thousands of dollars, or even reach $100,000. Most microtargeting professionals say the cost scale for an Arizona-sized legislative campaign is only now becoming a feasible option.
Strasma said there’s not much difference between the cost for a statewide or congressional microtargeting plan and the cost of a legislative microtargeting campaign, but Strategic Telemetry is currently prospecting what they call “micro-microtargeting,” which would allow smaller races to access the sort of service normally available only to campaigns with deep coffers.
One way he has helped bring down the cost for such small campaigns is by having them partner with other campaigns, so that several groups can split the cost.
In that case, perhaps Hanlon’s data and survey responses could be sold to more than one candidate seeking his vote, since he’ll be casting a vote in dozens of elections from county attorney to mine inspector.
Stanton used microtargeting in Phoenix mayoral win
One recent local campaign where microtargeting had a tangible impact was Greg Stanton’s 2011 Phoenix mayoral race.
The Maricopa County Democratic Party handled a large portion of the successful Stanton campaign. Executive Director Ken Chapman said the microtargeting effort the group deployed definitely helped contribute to Stanton’s win.
Though the campaign did not contract with a microtargeting firm, its plan involved a 6,000-respondent survey done in 2010 by the state Democratic Party for a swath of races that year, which was then combined with voter files, consumer data and the Democratic VoteBuilder.
“One of the areas with the most direct cause-and-effect relationship was the roughly 13,000 first-time municipal voters we turned out for (Stanton),” Chapman said.
The message campaign volunteers delivered to voters was also tied to the analysis of what would most likely persuade an individual, based on all the data compiled about them.
Although a campaign volunteer knocking on a voter’s door wasn’t able to see what purchases an individual made, what kind of car they drive or exactly which elections they’ve voted in, volunteers were told what kind of voter the analysis determined they were and which sort of messaging would be most effective.
“They don’t know a lot of the individual data points or the research or analytics that go into the reason to talk to this voter, but what they do understand is why there’s a reason to contact them,” Chapman said. “They’ll know this is a high-efficacy voter, or this is a persuadable voter.”
Constantin Querard, a local Republican campaign consultant, said he was reluctant to give away his “secret sauce,” but said that although Democratic campaigns have more recently been associated with the most effective microtargeting, Republicans are increasingly using it as a campaign tool also.
Querard said he’s seen some races that were won or lost by 200 votes, and that although it’s impossible to pinpoint which campaign tactic produced how many votes, he said that statistically speaking, there has to be a few races where it’s made the difference.
“The Democrats have had a tactical advantage, with their VoteBuilder program, but the Republicans are getting more into it too,” Querard said. “We’re getting better about how we treat and retain data.”
Querard said he expects both parties to use microtargeting more each election cycle, but he said that right now, the key factor that will determine whether a campaign adds it to the toolbox is the cost.
A well-funded legislative campaign in Arizona costs about $40,000, Querard said, so in order to make a microtargeting plan cost-effective at that level, it would need to cost closer to $7,000 to $10,000.
The horizon: Micro-microtargeting?
But such a small-scaled microtargeting plan is not too far off.
Because microtargeting involves both identifying the right voters to go after and determining which message to deliver, the future of microtargeting will at least in part be shaped by the level of individualization of communication channels with voters.
Television ads might be considered a shotgun approach to political messaging, while a particular advertising circular or a campaign volunteer visit can be delivered with just one person in mind.
But Internet-based media lends itself to highly individualized messaging as well. On-demand video services like Hulu often know who the viewer is because the service requires an individual account, and the company has already begun delivering individualized advertisements. The same goes for on-demand music services like Pandora Radio.
Michael Meyers, the president of Alexandria,Va.-based Republican microtargeting firm Target Point Consulting, said he fully expects to see such individualized ads before the 2012 general election.
“I’d be very surprised if country music listeners don’t get Republican messaging in their Pandora stream within the next year,” Meyers said, hinting at the correlation between Republican leanings and a preference for country music.
Hanlon said he wasn’t too worried about the use of new technology, like on-demand media, to offer individualized political marketing messages, saying that at least he would be getting advertising that actually fits his preferences.
Arizona Democratic Party Elections Director D.J. Quinlan leads most of the microtargeting efforts for his organization. He said such personalized messaging efforts, which could be coupled with solid microtargeting data, will be the next big step forward in the battle for votes.
“It’s a campaign consultant’s dream,” Quinlan said. “And we’re almost there.”