January 2, 2006 By Chad Vander Veen
Who are these people and how do they know so much about you? Is it some shadowy government operative exploiting your personal information? Nope. Instead, it is most likely a campaign manager for one of your district's congressional candidates. And he didn't get all your information through some sinister network of spy satellites and black helicopters -- he took advantage of something called microtargeting.
The term microtargeting sounds straightforward enough, and in many ways it is. It is both brand-new and decades old. Its effectiveness in campaigns is almost impossible to measure, but it is widely regarded as the only way future campaigns will survive.
Soon, your city council candidates may deliver text messages with information specific to you directly to your mobile phone. Some are even attempting to create a way to deliver TV commercials produced for individual viewers. Such tools would dramatically alter the way political campaigns are organized.
The vision among politicos is a subtle yet significant change in the way future campaigns are run. For example, with microtargeting, one campaign might focus on spurring higher turnouts while another campaign might try to embolden its stalwart base. But for all its seeming complexity, the singular goal of microtargeting is to modify voter behavior to a candidate's benefit.
Will the strategy boost voter engagement in the election process by personalizing campaign issues? Or will it leave them feeling as if candidates are merely calculated, distant salespeople? Only time will tell.
In general terms, microtargeting is sending a message to a highly specific portion of an audience based on particular information. In most cases, this method is used to target people as precisely as possible by creating messages tailored to individuals. Though it may sound impossible, political campaigns are currently devising ways to bring their message directly to you -- and only you -- by correlating consumer microtargeting data with voting behavior.
Of course, like many practices in government, microtargeting is an extension of what's occurred in the private sector for years. In fact, microtargeting's origins date back to the late 1970s and early 1980s.
"Commercial direct marketing has been using microtargeting for 25 years," said Hal Malchow, president of MSHC Partners, a Democratic consulting firm. "It's been out there a long time, and [direct marketing companies] have built enormous databases about people."
Prior to that time, microtargeting was irrelevant because of the limited number of available media outlets -- broadcast television, telephone, radio, print and mail. Back then, all that was needed to reach the intended audience was a commercial on the broadcast television networks -- viewers had little choice but to watch it.
Now, to get a message to consumers, companies can no longer count on running their ads only on NBC. Broadcast networks, and in turn advertisers, fight for an ever-smaller viewing audience that has been wooed by an expansive number of cable channels, video games, mobile phones and billions of Web sites.
These emerging technologies forced advertisers to be more creative and get far more specific. Commercial microtargeting is the result of consumers having so many media through which to receive advertisements -- and an equal number of new ways to avoid those ads.
As it is wont to do, private industry immediately responded to each new technology that emerged over the years. With the arrival of modern database-management and data-mining capabilities, advertisers gained ground against the ever-expanding number of channels and Web sites that compete for consumers' attention.
Massive databases have been created over the last few decades. These databases hold dozens, sometimes hundreds of pieces of information on individual consumers. Powerful data-mining software and sophisticated algorithms allow advertisers to gain an almost personal understanding of potential customers' likes and dislikes, and ads are then tailored to address individual wants and needs.
Numerous companies are scrambling to develop microtargeting that taps into cell phones and customer-specific television commercials. These are the avenues currently being explored and exploited by political campaigns trying to get their message out or motivate their base.
What's new about microtargeting is just how much information campaigns can squeeze out of data sets, said Alan Gerber, professor of political science at Yale University.
"The basic idea of microtargeting is that in the past, you might target a fairly aggregated level like a precinct, whereas now people are merging lots of data that's aggregated not at the precinct level, but down at the individual level," he said. "So you are able to find pockets of voters within some aggregate that are especially attractive to persuade or mobilize.
"That's kind of new," Gerber continued. "That isn't how the Democrats or Republicans have traditionally done their targeting, though now they're starting to do it more and more. The Republicans, almost all their targeting, the Democrats are certainly moving in that direction."
Campaigns have begun taking advantage of voter files, which are databases containing collected data of voters, likely voters, and people who don't vote-- aka everyone. Both major parties have gigantic and mysterious national voter files. The Republicans generally are considered to lead the technology and microtargeting races with their national voter file, called the Voter Vault.
Malchow said Republicans have been much more aggressive in using microtargeting techniques, whereas other observers credit the GOP's massive war chest for allowing the party to experiment with extensive microtargeting projects. The Democrats are not far behind, however, with the voter file known as Demzilla.
Each voter file is estimated to contain more than 150 million names, and each connects to multiple pieces of specific information. Few people, however, are willing to divulge any detailed information about these voter file databases.
Howard Simkowitz is the director of Government Services at Caliper Corp., based in Newton, Mass. Caliper developed microtargeting software, known as Political Maptitude, which takes advantage of the vast amounts of data in party voter files. The software is a spin-off of a commercial application used by private industry advertisers, and works very similarly. Data from publicly available sources is combined with other demographic data collected over the years. The data is then mapped into GIS programs to produce dynamic and informative precinct maps.
Caliper counts among its clients the Republican National Committee, the Democratic National Committee, and numerous other state parties and public interest groups.
"The software comes with census data down to the census block level of detail," explained Simkowitz. "By using tools in the GIS software, campaigns can create dynamic maps of the different areas, understand their constituency better -- understand, for example, where there are pockets of senior citizens, minorities, families with a lot of young people ... and better target their message to these areas."
Campaigns that exploit this type of microtargeting technology have a distinct advantage over those using traditional, broader targeting.
For example, one critical element to any campaign is door-to-door neighborhood canvassing. The problem is volunteers who do these neighborhood walks often don't know the neighborhood. What's more, these volunteers are typically college students with limited time to complete the task. Armed with door hangers, they are essentially following a haphazard list of addresses to be visited, which isn't very efficient.
With microtargeting software like Political Maptitude, campaigns can design intelligent maps and specific messages for individual voters in a neighborhood. Data from a voter file is entered into the fields in the program, as is additional information desired, such as subscription lists or church member names. This results in a highly detailed and organized map, not only of what kind of voter lives where, but also the most efficient way for canvassers to reach them.
"When they're actually doing the neighborhood walks and knocking on the doors, they know exactly who these people are," said Simkowitz. "They know if they voted in the last three elections, generally what party they are, what age they are, what gender. The feature in Political Maptitude that people are using will create very logical, efficient [canvassing] walks. You indicate where you want to start your walk and it gives you directions to the first household, then gives directions to the next household. The software will also estimate the walk time."
For now, microtargeting is limited to the big guns in political campaigns due to costs of purchasing lists and the local level economies of scale. The more local a campaign, the more cost-prohibitive the strategy generally becomes.
"All this costs money. It costs the same to do for a small campaign as it does for a big campaign," said Malchow. "If your budget is $1 million and you want to spend $15,000 on targeting, that's a reasonable amount to spend. But if your budget is $25,000, that's a different story."
That will all likely change if microtargeting proves itself a viable way to conduct campaigns. If deemed successful, national databases will likely open up to local candidates, making microtargeting a reality at the city and county levels. The result of microtargeting in politics, however, is difficult to measure, as is its success.
Will It Work?
For now, microtargeting is relevant only to state and national parties. And the strategy is very much in its infancy. According to Darrell West, professor of political science at Brown University, the first true test of microtargeting was only completed in 2004, with still inconclusive results.
"The only test we had was 2004, and [voter] turnout did go up, especially in competitive states where this technology was used," he said. "Both parties are using [microtargeting]. I think Republicans are a little ahead, just because they had the money that could finance this. But I think it's one of the priorities of Howard Dean to bring Democrats back up. The next big test will be ."
If voters don't come back to the polls next year, it will be hard to say microtargeting was effective. On the other hand, it might be that the enormity of the 2004 presidential race and effective microtargeting together were responsible for such turnout.
Unlike private industry, where a microtargeting campaign's success can be measured by a product sale, the trouble with measuring microtargeting in political campaigns is that the only meaningful way to tell if the campaign was successful was if the candidate won.
Paul Wilson, CEO of GOP consulting firm Wilson-Grand Communications, said the concept of microtargeting remains unproven, but has shown a lot of promise.
"It is still, I think, a concept that needs to be further shown to be true," he said. "The premise is that there are consumer behaviors, as in the purchasing of certain items, which correlate perfectly or near perfectly to voting behavior. Let's say I purchase the hunting and fishing licenses of the state of Missouri. Now a hunter's behavior being indicative of political philosophy is what microtargeting is. If I know who has a fishing license, are they more likely to be a Republican? We know that to be true -- but that is less true in Wisconsin. But now, suppose I say you own a white Lexus. Does that behavior correlate to partisan affiliation?"
Much of the consumer microtargeting data comes from clustering people. For example, an ad agency might take a large sample of people and ask them loaded questions in order to polarize them. Then, the demographic characteristics are grouped into clusters. Finally advertisers examine these clusters and find similar people in the general population to whom they can deliver a specific message that should appeal to them.
Thanks to advances in data mining technology, huge amounts of data can be analyzed and compared to find more obscure commonalities among clusters of people. Campaigns with access to extensive voter files like Voter Vault and Demzilla can then use this consumer information -- combined with information such as census data, license registration, and publication subscription lists -- to deliver a uniquely tailored message.
The means of such delivery are limited, though the Internet and e-mail are the present leading media. Telephoning people at home is becoming increasingly inefficient as caller ID and annoyance keep people from answering their phones.
While companies like Finity Technologies work to bring microtargeting to mobile phones and others are devising ways to deliver customized television commercials to voters, direct mail, door-to-door and e-mail are presently considered the most viable means to reach people and get them to the polls.
"I would say this strategy contributed [to the 2004 turnout]," said Wilson. "I would say it was interest in the presidency combined with communication techniques that reached people. Microtargeting was a factor, but just the closeness of the race was also a factor. [Does] microtargeting, where we are using characteristics that are consumer-driven, correlate to political behavior? That's where I think we need more work."
The State of Microtargeting
If the 2004 presidential election represented microtargeting's potential on the national stage, California is a testament to its viability at the state level. Of course, even there, opinions differ as to how effective and innovative microtargeting really is.
According to California Democratic Party Spokesman Bob Mulholland, microtargeting has gone on a lot longer than people think.
"Back in 1992, this party -- at the time chaired by Phil Angelides -- took that voter file [of] 13, 14 million, and meshed that against every possible file in the country," he said. "We were able to figure out if a person had another phone number that we did not have on the voter. We'd be able to figure out if they responded to mail, we'd be able to figure out and code as such if they sent money into some organization for some cause. So we've been doing it since '92 [in California], and each cycle gets better."
If microtargeting becomes more feasible for local campaigns, does that eliminate the personal connection a candidate must make to win voters? If campaigns can craft a message that touches on all the issues important to an individual, has that candidate taken the easy way out?
"I think it has always been the case that politicians are going to use the most intelligent, tactical strategy they can devise," said Yale's Gerber. "We understand that politicians are going to do whatever they can within the law to get elected. It's not a comment on this technology, that's just the nature of politics."
Many people believe microtargeting will help candidates deliver a better product to voters. In much the same way advertisers use microtargeting to provide a more meaningful experience to customers, campaigns and their candidates will be able to improve the voting experience.
"[Microtargeting] gives you a lot more data points to find out what people care about. In some respects, it is the old technology, but it gives you a much richer picture of who's out there and what their concerns are," said Duf Sundheim, chairman of the California Republican Party. "For most people, they can't name their assemblyperson or state senator. A lot of them would have to think who their congressman is. So this gives you the ability to, in a cost-effective manner, identify people and say, 'Hey, did you know your assemblywoman really cares about education issues? Here's her record on education issues.'"
Mulholland agreed, saying microtargeting allows candidates to better understand where voters come from, especially ones that appear very similar.
Take San Francisco versus Bakersfield, Calif., he said. A 60-year-old Democrat in San Francisco and a 60-year-old Democrat in Bakersfield are handled differently even though they appear the same. The reason is that in Bakersfield and Kern County, the Democrats typically lose, while in San Francisco they typically win. Microtargeting helps campaigns identify such differences on a much smaller scale.
It will take a few years before anyone can determine if microtargeting is a success. But judging by the secrecy surrounding Voter Vault, Demzilla and companies like Finity and Caliper, it seems a lot of people take microtargeting very seriously.
Microtargeting's future will probably be the integration of local, state and national databases, allowing candidates at all levels to understand exactly what issues are important to individual voters. In addition, this integrated microtargeting will allow state and national parties to follow voters as they move about the country, which will give local candidates a chance to connect with newly arrived constituents.
Furthermore, campaigns will develop an in-depth understanding of voting behavior, which will let them decide if it's worth sending someone a mailer, e-mail, text message, or possibly even a TV commercial. Campaigns will develop what Simkowitz calls a selection set -- a group of voters who need motivation to go to the polls. In addition, the costs of obtaining voter lists, mailing candidate information, and hiring staff to run a campaign means that intelligent use of money is a high priority. So if microtargeting allows campaigns to skip over people they know will never vote for their issue or candidate, or are a guaranteed vote, it becomes a valuable resource.
On the other hand, these same techniques can be used to find voters who really might put a candidate over the edge if a message can be tailored to bring them to the polls to vote for the candidate.
"We are constantly looking for people who are at the margins," said Wilson. "If [John] McCain were to run in 2008, wouldn't he want to know who might vote in the Republican primary if prompted ... who are the independent Republicans? So [microtargeting] will tackle those questions. It will find people in strong Republican areas who should be registered but aren't yet. And that's the Ohio example. It's going to find people who really should be participating but aren't."
Another method that will continue to grow is microtargeting the people who disseminate information to the appropriate voters. Campaigns are microtargeting people, like bloggers, who have a well informed and well defined readership. By delivering a personal message to bloggers, campaigns take a lot of the burden off themselves. Targeting bloggers has proven an effective and inexpensive way to get a specific message to a specific group.
"We start our campaigns with banner ads to reach the influentials, the information seekers who are the first people to tune in and who will shape the opinions of others down the line," said Malchow. "The Internet is a great medium for that, and it's cheap. Blogs are a steal. You are talking to a well selected audience early in the campaign. They're all partisans, but getting the partisans excited and talking is an important part of the campaign."
And while Sundheim and Mulholland might disagree on just how innovative microtargeting is, both believe microtargeting will continue to get more sophisticated and improve the experience voters have with their party.
"This is a state where both parties are light-years ahead of most other states," Mulholland said. "First of all, we have party identification on the file, and second, we have vendors that actually collect stuff for decades. Just like the federal government and the credit card companies are keeping more and more track of us, so is the voter file."
Sundheim believes aggressive leadership and new technology will continue to drive the use of microtargeting. He also cautions that the strategy is a tool and not a solution.
"I think it's the leadership shown by the RNC to understand that this technology could be used. Campaigns tend to think in terms of their campaign and not on what could be beneficial 20 years out," he said. "In the old days, 78 percent of the people watched network TV. That's now down well under 40 percent, so you need to develop new tools to reach these voters. And microtargeting is one of the tools of the new technological age. It's one tool. It's not the magic bullet. But we've been very pleased from what we've seen from it."
You may use or reference this story with attribution and a link to