The first thing to know about VoterMapping.com is that if you get the chance to play around with it you will love doing so. Bruce Willsie is the Princeton-based president of Labels & Lists, a company based in Bellvue, Washington, thet operates VoterMapping.com, and even he’s not immune to noodling around with random bits of information and settings just to see what comes up.
In the upstairs study at his Mercer Road home Willsie plugs in data bits (ethnicity, political leanings, voter turnout records, and so on) like some mad virtual scientist with a multi-colored chemistry set. And when you say to him “It looks like you’re having fun with that,” he smiles broadly and answers, “Everybody does.”
The catch is that to get into the system, you need to be a legitimate political entity, such as a candidate or pollster.
What is so important about the information on VoterMapping.com is that, as its name implies, it shows detailed data of American voters. And that means all 165 million-plus American voters (4.6 million of whom live in New Jersey). Every registered voter in the United States is in the system, with as much information as is available about the voting habits within that household.
Labels & Lists does not know how people vote in an election — that’s protected by law. But they can find out who did vote. “The fact that you cast a ballot in a particular election is part of the public record so it is possible to use public data to determine how frequently any particular voter votes,” says Willsie.
The voter data can then be matched with other publicly available sources, such as census information, voter registrations, and driver’s license applications, to create a complex portrait of each voter and household.
If you were to look up yourself on VoterMapping.com, you would find your address, age, party/political affiliation, sex, race, and voting frequency, among other information.
All this data, multiplied by a factor of 165 million, creates a cumbersome and heretofore unwieldy database of golden eggs for pollsters and politicians. What allows the system to physically work as you compare and contrast areas of white Republicans to those of Latino Democrats to those of active-voter senior citizens is a software platform developed by Moonshadow Mobile, an Oregon-based developer of computer technology designed to retrieve geospatial data at extremely high speeds.
Moonshadow’s software enables users to retrieve information from immense databases far faster than most other competitors. Think of it in terms of a search engine — Google can take your request for, say, lasagna recipes, and within seconds show you hundreds, maybe thousands, of matches from its database. This is possible because Google’s software can filter and retrieve information at the scale of 1 million records per 60-90 seconds.
By contrast, there is Moonshadow Mobile, which, Willsie says, has “just recently broken a technological barrier and can now visualize over 100 million records per second per core processor, utilizing ordinary servers in the cloud.
This, he says, allows his company to visualize data points for all 230 million-plus adults in the U.S. over a single map that updates in only a couple of seconds — and offers the ability to zoom down to a single household.
And for the complexity and scale of VoterMapping.com’s files, this is vital speed. Without that kind of speed, the colored maps that show up immediately when Willsie sifts through various voter criteria would take 15 to 20 minutes to resolve.
If the overlays took that long to update, they would be of little use to anyone trying to see how voting patterns or constituencies actually lay out
“This makes selections from millions of voters in less than a second,” Willsie says. “You can create spreadsheets, make maps, and make very layered data sets.”
And the thing about this all is, you don’t need a mainframe to actually use the data. Data sets the size of VoterMapping.com’s and speeds as fast as Moonshadow Mobile’s historically have only been put together in monstrous mainframes. But the partnership has found a way to get immense, multi-layered bundles of data to the average web user — or even a smartphone user with web browsing capabilities — in as much time as it takes to sneeze.
The speed is made possible by a dramatic rethinking of database architecture, Willsie says. Labels & Lists knew its main enemy was slow resolution time, and Willsie says the company found its solution in the way data was stored and retrieved. The technology is patent-pending, so Willsie does not disclose much about it.
With a subscription to VoterMapping.com you can search political information by location, district, and cable television zones, the last of which allows you to see where concentrations are located of whatever subsector of the voting public you’re researching.
In his upstairs study, Willsie browses through Latino voter concentrations in the cable districts of southern California (chosen to highlight the speed of Moonshadow Mobile’s technology as it sifts through 15 million records in about a second).
With the information returned he could, were he a political candidate looking to address Latino voters in Los Angeles County, pinpoint the cable television market on which he could advertise and get the best reach.
Speaking of speed and reach, Labels & Lists and Moonshadow Mobile have another arrow in their quiver called Ground Game. This is a mobile canvassing application for smartphones that allows for real-time updates from the field as political volunteers go door-to-door and neighborhood-to-neighborhood to canvass. Ground Game gives canvassers campaign-generated walking lists, voter survey questionnaires, current registered voter information, and Google Maps (though the main VoterMapping.com maps are Bing maps). And all this allows canvassers to see superimposed locations of individual houses containing the targeted voters, view relevant information about those voters, conduct campaign-specific survey questionnaires at a voter’s door, call and leave a message on the spot if voters are not home, and instantly upload any responses to central servers that manage the data.
On the other end, campaign field managers can keep tabs on canvassers in real time, create new filters, and transmit up-to-the-minute information that allows canvassing routes to change or refocus if necessary.
Campaign managers have a complete visual overview of registered voters. They can look at any part of a district, state, county, or town and immediately see the distribution of hundreds of thousands or millions of voters by party affiliation, age, gender, marital status, ethnicity, voting history, and many other attributes. With just a few clicks they can zoom to street level and look up voter information for individual addresses. Staff members can create walking lists graphically using Google Maps and assign the lists to canvassers in the field.
An obvious question at this point is, who is looking at this data and what are they doing with it? On the front end, you can’t be just anyone and get access to the VoterMapping.com database. You can’t sign up online, you have to call Labels & Lists and actually converse with someone on the phone and “prove that you’re a legitimate user of the data,” Willsie says.
But once someone clears the front end, there is little Labels & Lists (or anyone else) can do with how a user uses the information he finds. “Ultimately the responsibility is on the user to use the data properly,” Willsie says. And, in truth, there is pretty much nothing on Labels & Lists’ database that isn’t already available from other public records and marketing lists.
VoterMapping.com works on a subscription basis. You can buy a three, six, or 12-month subscription for a congressional (or smaller) district, a single state, a set of states, or the entire nation. The price range is $90 to $3,995, depending on the area you want to search and the length of your subscription. The fees cover usage for the time period, not the number of times a user checks in. Once you have set up “universes,” which are the folders that contain specific search criteria and results, you can access them as long as you have a subscription.
A traipse through New Jersey shows few surprises, but some definitive demarcations between registered Democrats (by default identified in blue) and registered Republicans (red).
Viewed overall, New Jersey is fairly purple, except for a dense blue diagonal streak from Trenton to Newark, and for a sparsely populated, but almost entirely red chunk in the upper left quadrant of the state. Blue clusters also thrive in larger cities, such as Atlantic City and Camden.
But as Willsie zeroes in on smaller areas, that purple gradation becomes an increasingly segregated image of blue and red dots. This separation of political philosophies in the physical world is evidence in a study Princeton University is conducting to see if people self-segregate into areas where others share their politics. And this study is part of Princeton’s ongoing use of data as supplied (for free) by Labels & Lists.
The free use of Labels & Lists’ data has its roots in the fact that Willsie is a Princeton graduate (Class of 1986). But it formally took off in 2006 when Labels & Lists donated voting data — stripped of names, telephone numbers, and addresses so that it cannot be traced to individuals — pertaining to roughly 75 million voters in 11 states to the university’s department of politics.
According to the university, Willsie realized Labels & Lists’s data might be useful to Princeton after researchers from another university purchased information from the company. Princeton politics researchers have long been frustrated by the dearth of affordable voting data and quickly grasped the value of public voting records from a number of large states merged with demographic information.
Long before Labels & Lists was connected to Princeton it was a small, nonpartisan voter information list firm that served only Washington State. The firm was founded in 1975 by Linden and Lourene Criddle, the parents of Willsie’s wife, Marcia. Willsie says the firm was originally intended as a retirement business for his future in-laws that quickly became a lot of work.
In 1975 computers able to hold large amounts of information were rare. And they were mainframes. The Criddles had to rent time at the municipal mainframe in Bellevue, then physically print out labels and lists for customers. Marcia’s mother had been a political volunteer and her father a pioneer in computer technologies for business.
At the time, information on voters was gathered painstakingly. “She knew there was a better way to handle name and address information for political candidates,” Willsie says of his mother-in-law.
Before knowing Marcia and her entrepreneurial folks, Willsie was raised on Vashon Island in Puget Sound, Washington, where his mother ran a daycare business and his father worked as a machinist “at a small factory across the street from where we lived.”
Willsie had served on a Navy submarine for six years and met Marcia in 1981, when she was applying for the Peace Corps. They soon married, and Bruce matriculated at Princeton when he was 25. He chose Princeton in large part because in 1982 it was the only school that accommodated them with a married student apartment. In 1986 Willsie graduated with a bachelor’s in politics and then attended law school at Harvard.
From 1989 to 1990 Willsie was an attorney at Perkins Coie, a Seattle-based firm that practices everything from product liability law to regulatory and government affairs. And while Willsie still has an affinity for law (he is fond of English common law books and has donated certain historical books to Princeton University), his interest in politics lured him to accept his mother-in-law’s offer to join Labels & Lists in 1991. “It sounded like a lot more fun than being a lawyer,” he says.
Throughout the 1990s Labels & Lists grew as technology allowed. The firm’s reach spread from Washington to Oregon and moved fairly quickly across the country. In 1999 Willsie was on a business trip to Princeton, to which he had harbored a desire to move back. At the same time Marcia, who had been a teacher, had wanted to go to cooking school. With two hours to kill before his plane ride home, Willsie thumbed through a stack of real estate leaflets and saw a 300-year-old farmstead on Mercer Road that originally was owned by Ezekiel Smith, an old Quaker bachelor who died without a will or heirs in 1766.
The house, nestled on three-and-a-half acres overlooking Stony Brook, was “the coolest house,” and Willsie bought it that day. At the time there were no cooking schools in Princeton, so Marcia found her education at Seattle Culinary Academy. But she eventually opened her own school at the Smith house, Ezekiel’s Kitchen (U.S. 1, March 26, 2008).
Willsie spends three weeks a month in Princeton and one week in Bellevue and is one of several employees who work remotely. The firm employs about 20, scattered across Virginia, Hawaii, Oregon, and California, Willsie says. Add another eight employees from Moonshadow Mobile, which operates as Labels & Lists’ tech staff.
Labels & Lists has been a part owner of Moonshadow Mobile since 2010. “The two companies are joined at the hip with a division of responsibilities,” Willsie says. “The high-tech visualization is being developed on the Moonshadow side while all other data processing takes place on the Labels & Lists side.”
Technology, having completely revolutionized the market for Labels & Lists, has also made the company the go-to source of voter information, Willsie says. And electronic delivery of all the company’s labels and lists is 98 percent of its output now. That isn’t even including electronic media, such as DVDs, which Willsie says his company hardly ever sends out anymore. There still remain a few holdouts who want printed mailing lists, which the company handles, but almost all Labels & Lists clients get their information electronically.
Once all this information is in a customer’s hands, though, what is it used for? “It’s major purpose is for targeting,” Willsie says. A candidate can do mailings based on findings, locate his constituents, and even identify clusters of people in his own party that he might not have known were there. This is especially handy when political maps are redistricted. “It’s been a surprise to many congressmen who’ve found pockets of Democrats in the heavily Republican areas,” he says.
Pollsters use the data to understand voter patterns and trends. And with VoterMapping.com’s real-time capabilities, pollsters can adjust and readjust their sampling areas and find likely voters, and make their best guesses as to who is most likely to come out to the polls.
A random skip through Princeton Borough and Township shows mostly blue (Democrat) dots and high amounts of voter frequency. It virtually mimics the party layouts of Trenton, Ewing, Lawrence, and Hopewell, though it clashes with the red concentrations of Hamilton Township, particularly Hamilton Square.
Middlesex County gets decidedly more purple on the overall map, though the closer you get to urban areas, the bluer the map becomes. Still, closer inspection does give evidence that at least in New Jersey, people do tend to cluster in politically oriented groups.
All this boils down to those “what-if” questions Willsie has mentioned — What is the relationship of income distribution to education? Are single females more likely to donate to charitable or political causes than single males? How does voting performance vary by age? Is America Balkanized into sub-communities by political leanings?
“Users have found that our technology leads them to ask questions they had not previously imagined,” he says. “The technology is also not strictly limited to geospatial visualization. We are planning soon to offer a cloud-based application of the technology to speed up the retrieval of address-based data from any traditional database.”
Whatever it is, its future promises to get here in a real hurry.
Labels & Lists, 2500 116th Avenue NE, Bellevue, WA 98004. 800-842-5478. Bruce Willsie, president. www.labelsandlists.com