I earn between $50,000 and $75,000 a year. I have three degrees. I teach at a small Catholic liberal arts university. I own several cameras. I have two vehicles, one four-wheel-drive, the other high-mileage. I belong to two environmental organizations. I commute more than 10 miles to work. I contribute to three non-profit organizations. I am single. I buy clothes from Lands’ End and L.L. Bean. I collect Rotring pens and pencils. I play guitar and piano. I read science and detective fiction, purchased mostly from Barnes & Noble and Amazon. I have two iPods, one computer and two TVs. I have several credit cards. I drink beer. I once owned a handgun. I have a mortgage. I have voted since 1964 as an Independent or Democrat.
To our presidential candidates, I am not Denny Wilkins, an individual human being: I am a set of data points, one of about 168 million sets of data points collected by both the Democratic and Republican national committees. They are interested in me only because I am an eligible voter. Democratic and Republican presidential candidates will base their direct-mail, push-poll, robo-call and volunteers-knocking-on-my-door messages on computerized analyses of me as data. I will be micro-targeted by candidates’ campaign organizations for the sole purpose of producing a vote for a candidate.
Candidates who decide to run for office under the aegis of the two principal national parties have access to comprehensive databases of American voters. The Republicans have Voter Vault; the Democrats have DataMart and “Demzilla.” If you are an eligible voter, you’re in each database. So how’d we all end up as little ones and zeroes?
We vote. The national political parties collect state voter files.
We need licenses to drive a vehicle, fish, hunt, cut hair and run a business. The parties collect those public records.
We get counted by the federal government. The Census Bureau also asks numerous nosey questions about our ethnicity, occupations, neighborhoods, incomes, family size, value and size of homes, rent vs. own, commuting distances, etc. The parties collect those records.
We’re suckers for direct marketing. We buy stuff from mailings that show up in our mailboxes. We buy from the various home shopping channels littering satellite and cable television. We even send in those cheesy “warranty” cards after filling out the demographic and marketing preference surveys. The political parties buy those records.
We buy lots of stuff on the Internet and everywhere else. The parties buy those records from commercial data-mining companies.
We are relentlessly surveyed. We get phones calls at dinner asking, “How do you feel about X?” We answer just to get off the phone. Someone shows up at our door, clipboard in hand: “I represent X. How do you feel about Y?” Stores send us “consumer satisfaction surveys.” Even our hospitals and doctors survey us about their services. The political parties buy all that information, and much of it clearly identifies our social, political and religious beliefs.
The parties cross-index all that information. They do this to determine the likelihood of how I will vote; they micro-target me to either reinforce a tendency or counteract it. Even if you have never voted before, the political parties can be remarkably accurate in predicting your voting tendencies.
Says Ken Strasma, president of the Washington, D.C.-based firm Strategic Telemetry, a Democratic consultant:
We’re able to take several hundred different indicators from the census and commercial marketing data, look at people who do have primary voting history and see what they look like demographically. Not just the wealthier-people-vote-Republican generalities, but what does a woman who commutes more than 50 minutes a day and has a postgraduate degree — what does her vote look like? We’re able to compile a model and apply it to people who don’t vote in primaries and get an accurate picture of their likely partisanship. None of this is a hundred percent. You’re not able to say this person is a Democrat or a Republican. But even if you’re not able to say people who drive four-door foreign cars are Republicans, period, end of paragraph, you can come up with scores that say: these people with this combination of 350 census indicators are 60 percent likely to be Democrats. You can make generalizations about groups of voters that are accurate enough to deliver campaign messages.
The content of the glossy brochures you’ll receive from the presidential candidates, the actual words used by a campaign volunteer calling on you at your home, and the precise text of the robo-calls you’ll receive on an almost daily basis as primaries near will be based on a computer analysis of hundreds of data points about you. [For an example, here's how Mitt Romney's data cruncher works.]
Remember, politicians are fundamentally advertisers with a message to sell. Anything you can imagine an advertiser doing to promote a good or service a politician also can and will do — because politicians offer themselves as a packaged good or service for voters to buy. Micro-targeting voters allows campaigns to narrowly match specific parts of that package to resonate most effectively with the demographic and ideological propensities of individual voters. This is how campaigns operate. This is the behavior that candidates engage in to persuade us to vote for them. They precisely target us with highly specific messages.
All these computer-driven database analyses of likely voters allow campaigns to be remarkably efficient and effective in their messaging strategies. Too few volunteers to canvas a particular congressional district? Screen the database for voters likely to resonate with your message. In as little as an hour, you can blanket that district with 100,000 robo-calls for as little as $2,000.
According to USAToday, candidates generated more than 5 million robo-calls during the early caucus and primary states. Add the presidential general election contest to House, Senate and gubernatorial races, and the number of robo-calls could reach hundreds of millions by November — all generated through the highly sophisticated data-mining and -crunching techniques used by the national parties.
Political campaigns’ use of computers, databases and the Internet has far exceeded the successes and sophistication of conservative icon Richard Viguerie’s computerized direct-mail efforts of the previous century. Viguerie’s innovations allowed the GOP to bypass the mass media as principal conduit to likely voters. As early as Sen. Barry Goldwater’s 1960 presidential campaign, wrote Mr. Viguerie, “We were learning how to mobilize grassroots Americans for door-to-door campaigning as well as raising money.” [emphasis added]
All this will come into play on “Super Tuesday,” Feb. 5. Voters will choose 41 percent of Republican delegates and 52 percent of Democratic delegates.
The sheer number of states in play — indeed, the sheer number of Congressional districts in play — has presented an extraordinary tactical challenge to these candidates at a time when they are running low on resources. It is prohibitively expensive to poll in all these states and districts to determine where to spend money. It is also prohibitive to run voter identification operations or advertise everywhere a candidate might be competitive.
Aides to Mrs. Clinton and to Mr. Obama said they had tried to compensate for that by building models, based on past voting history and even consumer data, to pinpoint Congressional districts where voters would seem particularly open to their candidate. [emphasis added]
Many critics of such campaign machinations decry them as Orwellian or an invasion of privacy. Perhaps, but so much of the data campaigns use we surrendered voluntarily. We opened this Pandora’s box by being so free with details about our lives, beliefs and consumer and political decisions. Stop participating in polls. Shred the warranty cards. Demand better protection of your private data from the companies you buy stuff from. (Yes, I know … a drop in the bucket, too little, too late.)
The press rarely writes about such computerized campaign techniques. It’s too complicated, too technical, too … boring. It’s a hard story to cover because campaign aides don’t wish to talk about this seamier side of politics. Aides don’t want stories portraying their candidates as numbers-driven rather than engaged face to face with voters in a made-for-TV town hall setting.
The press wants conflict. It drives ratings. So, rather than examine data mining closely, Wolf Blitzer & Co. say, “Your opponent called you a moronic flip-flopper. How do you respond?” Then they take that response and trot back to get another rebuttal. Such is the sophistication of presidential campaign coverage in these days of media cutbacks …
Our candidates tolerate this, this rudeness of data mining practiced in as much secrecy as possible in their names. (Our candidates also tolerate behind a cloak of deniability independently funded splinter groups operating in a candidate’s behalf with the candidate chiding the group’s behaviors and demanding, “Stop it. *wink, wink*”)
In November someone will get elected president. (It will not be New York City Mayor Michael Bloomberg running as an Independent. He won’t have a big enough Voter Vault or Demzilla to compete. No third-party candidate has a prayer.)
But consider this: The eventual winner will spend tens of millions of dollars (raised from the usual suspects among retired folks and in the law, real estate, investment and securities, pharmaceuticals and finance industries) on data mining and crunching to maximize electoral votes.
What reason has any candidate given us to believe that his or her adherence to such manipulation of the electorate will cease after the election? Presumably he or she wants a second term.
And what reason has any candidate given us that data mining and crunching will not be used to determine voter acceptability of public policy proposals by his or her administration?
I wish Wolf would ask the candidates about that in the next presidential “debate.” Once again: Past behavior indicates future performance.