by Adam Tanner
Controversial Practices
The same year Brierley started building what became Epsilon in 1969, an Arkansas businessman founded a direct-mailing company that eventually became one of Brierley’s main rivals, Acxiom. Using phone books and a computer, the Arkansas company sent out direct-mail pieces for Democratic Party candidates. The business continued to grow, and today Acxiom enjoys annual revenue of more than $1 billion. Clients include nearly half the Fortune 100 list of America’s largest companies, eight of the ten top credit card issuers, seven of the ten top automakers, four of the five top banks, and five of the thirteen largest federal government agencies. The company has widened its political focus too: Republicans buy its data as well.12
The largest part of Acxiom’s business comes from supplementing information companies have collected on individuals. Getting more data from Acxiom helps boost marketing across sectors including the Internet, mobile, direct mail, call centers, and interactive TV. For example, if a company knows a person’s name and address, Acxiom may be able to add a phone number, email address, and demographic information about average wealth in the neighborhood. A smaller chunk of the business comes from selling lists of potential customers.
Acxiom knows a staggering amount about people. It starts with public records, then amalgamates information volunteered on warranty cards and online surveys about hobbies and lifestyle information as well as commercial data from magazine publishers, retailers, and catalog companies. Acxiom also aggregates other identifying data, such as credit header data, the nonfinancial information at the top of your credit report that includes addresses and phone numbers. It may also include sensitive data such as Social Security numbers, which, the company says, are used only for fraud prevention. Acxiom’s file does not detail that you once bought a $269 TaylorMade Men’s Burner Super-Fast 2.0 TP Driver golf club, but it may indicate that your household has an interest in the sport.
The end product combines thousands of sources on a given person into two main dossiers. One is for marketing purposes and the other is for “risk mitigation,” with only the data needed for that purpose. The company provides data from these products to companies, nonprofits, and government agencies that have passed a credentialing process.
Acxiom has data on almost all US consumers. Just reading the list of information it has can prove a little overwhelming. As consumers learn how much a company like Acxiom knows, some grow uneasy and concerned. “It is new and people don’t understand what is happening. That feels a little bit creepy, that feels a little bit uncomfortable,” said Tim Suther, Acxiom’s former chief strategy and marketing officer, who insisted that consumers should not be afraid of the company.13 He stressed that it does not make this data available to just anyone or for just any purpose. Thus Acxiom data do not determine whether you get credit, or health insurance, or a job. “We use it generally for one thing, and that’s to help advertisers be more efficient with their messaging and who they want to talk to,” he said. “If that information were used in a way that disadvantages you—‘Hey, wait a minute! I’ve noticed you’ve gained a little bit of weight here. Do you mind if I share this with your insurance company?’—well, that would be a little freaky.”
* * *
A few years after the initial incarnations of Epsilon and Acxiom, an Indian immigrant started his first direct-marketing efforts, which would eventually make him a millionaire invited to spend the night at the White House. Vinod Gupta grew up in a poor village north of the Indian capital, New Delhi, without water or electricity. He studied engineering and came to the United States in 1967 to get a master’s degree and then an MBA in Lincoln, Nebraska. After finishing university in 1971, he got a job at the Commodore Corporation, a mobile home manufacturer, as a market research analyst earning $850 a month. (That same year, Harrah’s sold shares to the public for the first time.) The following year, seeking to boost sales, Gupta asked the national telephone monopoly AT&T to deliver a copy of every local phone book to him. Because Commodore had a costly toll-free 800 number that made it an important business client, AT&T sent a massive order of about four thousand books, free of charge.14
When his boss protested the massive size of the shipment, Gupta had the phone books delivered to his garage. The piles of directories sat there for a few months. Then, over the course of a few days, he compiled a sample list of mobile home vendors in Nebraska, writing them out by hand. Others programmed the data into an IBM System/3 computer, inserting punch cards to store up to ninety-six characters of data, about as much information as a line of text in this book.
He offered to sell a master list of US mobile home vendors to his company for $9,000 to compensate for his extra labor or give it to Commodore for free—provided he could market the list to rival firms as well. Commodore chose the latter option. Within a month Gupta had received $22,000 in new orders and started compiling data from other states, charging 10 cents per listing. It took three months before the list was ready. He started his own list-building business on the side and left Commodore the next year, 1973. Over time he expanded the lists to car dealers, boat dealers, and others. By 1981 he had made close to $1 million in sales.
In the mid-1980s Gupta added lists of people across the United States, layering public and commercial data onto basic phone book listings. Computer storage was becoming far cheaper, and his firm could do a lot more. “Once you know the name and the address, you had the home value, you had the phone number, then you can estimate the income of the household and then you can also overlay other information anytime a person fills out some stupid warranty card,” he recalls. “Suppose it’s a warranty card and some coupon for an arthritis drug, then you know this guy has got arthritis. So we had a lot of these clearinghouses who basically are selling that information.”15
Gupta included whatever categories made sense from a commercial standpoint. For example, he says, he would not have had a problem listing that someone suffers from HIV. Yet with fewer than 1.8 million Americans in that category, he says, there is not enough marketing demand, so he does not include such information.16
In 1992, Gupta took his company, then called American Business Information, public. His wealth led to greater social prominence. He donated generously to both Bill and Hillary Clinton’s political campaigns and was invited to spend the night in the White House’s Lincoln Bedroom, a perk shared with top donors and friends. He appeared on 60 Minutes II to explain data broker lists. After Bill Clinton left office, he hired the former president to consult for his company at a cost of $3 million over several years.17 Gupta traveled on a corporate jet and enjoyed a life of luxury. Eventually he sold his company, which was later renamed InfoUSA, for $680 million in 2010.18 By then he and his family had homes in Omaha; Aspen, Colorado; Washington, DC; Miami; Yountville and Hillsborough, California; and Maui and Kauai, Hawaii.19
His high living attracted the attention of federal authorities. In 2010 the Securities and Exchange Commission charged him with milking his former company for $9.5 million in unauthorized perks between 2003 and 2007, including personal use of jets to Italy, the Virgin Islands, Cancun, and other places, as well as billing for expenses related to his yacht, a winery in Napa Valley, twenty cars (including a Jaguar, two Mercedes, and a Hummer), life insurance, credit cards, twenty-eight club memberships, and other expenses.20 “Gupta stole millions of dollars from Info shareholders by treating the company like it was his personal ATM,” said Robert Khuzami, director of the SEC’s enforcement division.21 Gupta agreed to pay millions of dollars to settle the case but did not admit or deny the allegations as part of the settlement.22
After selling his company in 2010, Gupta formed DatabaseUSA, which continues to compile data. The service lists one hundred million cell phone numbers.23
Despite so many years of work in direct marketing, or perhaps because of it, Gupta is contemptuous of many in his field. He says about half of today’s list brokers are “shysters” cutting corners or lying about the quality of the personal
data they are selling. “They are basically like traitors,” he says. “They never tell the truth, you know. They claim to be like a multi billion-dollars company and at the home they only got one employee.”
The DMA
Many of the firms arousing Gupta’s ire gather every year at a convention hosted by the Direct Marketing Association, the leading industry group, whose roots date back nearly a century. The event, held in 2012 at the Las Vegas Mandalay Bay Hotel, celebrates data gathering and capitalism. Data merchants set up booths in a vast convention center. Even as they try to find new clients for their lists of consumers, they continue gathering data on one another. Some sponsor old-fashioned raffles, encouraging visitors to plop their business cards into a bowl for a chance to win a prize. As I approached one businessman’s booth, he asked if he could scan my convention badge. The matrix code contained my name, address, phone number, and email.
Different brokers advertise different specialties. A company called T5 Healthy Living asks people to share intimate details via online surveys on subjects such as constipation, bipolar disorder, depression, cancer, and other conditions. Participants “receive information and exclusive offers with valuable coupons and savings on the latest products available.” One question asks if you or a partner have had a sexually transmitted disease. The end of the form asks users to confirm: “I authorize T5 Healthy Living and its trusted third party partners to use my personal and household information to send email, postal, and valuable information to me.”
Another broker even sold lists of the deceased. In Nikolai Gogol’s satirical novel Dead Souls, the protagonist buys identities of those who have passed on to boost his apparent wealth and standing. That was not the goal of Omaha, Nebraska–based CAS Inc. It wants to improve clients’ chances of selling to the living by purging the names of others in the household who have died. “Sending direct mail to the recently deceased adds insult to injury for the remaining members of a household,” a company brochure said. “Not only is this practice bad manners, but it is also bad for business.”
The US Postal Service has a large display stand at the convention, advertising its services to update addresses for direct marketers when people move.
One conference highlight is the DMA head’s keynote speech. In 2012, with Congress and the Federal Trade Commission stepping up scrutiny of data brokers, DMA acting CEO Linda Woolley struck a fiery tone as she rallied thousands of gathered marketers in a large ballroom.24 She said firms spent $168.5 billion on direct marketing in 2012, which generated more than $2 trillion in incremental sales—nearly 9 percent of US gross domestic product.25 “Our data-driven world is a better world,” declared Woolley, a former consultant and lobbyist. “It’s not just that we have big data, it’s that we are using it to give customers what they want.”
Then she portrayed a dark future in which the Federal Trade Commission and “privacy zealots” had gotten their wishes and persuaded Congress to bar the collection of personal consumer data without their permission. The nightmare-scenario law would bar Internet tracking and prohibit the use of public records to gain insights. “Consumers even have the right to say: ‘My marketing data is mine, and it’s private, and you can’t use it or sell it,’” she said. Woolley’s vision aimed to rile up the crowd much as talk of a new prohibition might horrify a convention of brewers. “In this picture of the future, marketers didn’t fight back hard enough, and the FTC and the privacy zealots got this bill passed,” she continued. “Is this the picture of the future that you want? Of course not!”
She closed her remarks by leading the crowd of several thousand in a chant of “We are DMA! We are DMA!”
One developing area of marketing on display at the annual DMA conference is marketing through racial targeting. The field is attracting some interesting entrepreneurs who are making new uses of big data. After immigrating to the United States from India in 1997 at age thirteen, Ajay Gupta became fascinated by the important role minorities played in politics. He ultimately started developing software to identify the ethnic origins of millions of Americans from their names and other clues.
In high school in San Francisco, he taught himself computer coding by helping to develop a website on professional wrestling. He studied financial economics and creative writing in college and graduate school, but saw better career opportunities in big data and formed his company, Stirista, after his 2008 graduation. He bought lists of registered voters across the United States, and cross-referenced names with those on the rolls in eight southern states that also listed ethnicity. Using this and other data he could eventually predict ethnic origins and religions for most names and could discern unusual patterns. For example, of Ann Smiths, he detected that about three-quarters are white and the rest black. But 70–80 percent of Annie Smiths are black.
Gupta learned as well about the need for multiple clues such as combinations of first and last names, ZIP codes, and other details. A surname like Sen could be Bengali or Scandinavian. Das could be Indian or German. “What is the most challenging is when you have a name like Obama. There are only going to be about twenty people in the country with that,” he says.
As he built his database, big companies started buying the data and his software. Cable operators including Dish Network, Time Warner, and DirecTV promoted Russian, Hindi, Chinese, and other foreign-language channels. Phone and VoIP telephony providers appealed to emigrants with ties to the old country. For example, Vonage could directly advertise very specific offers such as three thousand minutes a month to the Philippines for one cent a minute.
Ever more detailed personal data allows well-known names in retail, finance, health care, and other fields to tailor messages to different ethnic groups. McDonald’s, T-Mobile, Diageo (distiller of brands such as Johnnie Walker whisky and Guinness beer), Wells Fargo, Blue Cross Blue Shield, and Procter and Gamble have all embraced what they call multicultural marketing.
Throwing away notions of a color-blind American society in a giant melting pot, data scientists and entrepreneurs like Gupta have stepped up efforts to use big data to segment customers into ethnic, racial, and religious categories. This ability is one of many areas in which marketers are gaining ever more intimate insights about customers who typically are unaware such information is being gathered.
“There was a sense that people came here and they were expected to take on the American culture, so you marketed to them as that. And now there may be a growing awareness that people maintain their ethnicities,” says Donna Lillian, president of the American Name Society, a group of a few hundred experts who specialize in the study of US names.
For Stirista, a San Antonio, Texas–based startup where top executives include Gupta’s wife and mother-in-law, this awareness has translated into rapidly growing revenue. It expected to bring in $3 million in 2013, up from $1.4 million in 2012.
Such targeting may be as simple as swapping ad photos to show a different ethnic group. Other times the message varies to reflect cultural sensitivities, a practice the American Heart Association embraces. “In certain cultural groups there is a certain guilt or lack of trust in working with their doctors,” says Gerald Johnson II, the group’s chief diversity officer and senior vice president. “There are the cultural nuances that say that messaging would be different.”
In other cases, marketers will only target certain ethnic groups. “If I send out twenty thousand emails just for a Hispanic grocery store, for example, the response is not necessarily that great,” says Lynwood Shackelford, director of sales for the Washington Suburban Press Network, which helps firms with marketing. If they can narrow that original mailing list to just Hispanics with the right messaging, the ad may deliver ten or twenty times better results, he says. In another campaign, he promoted a new Native American museum to people of that heritage.
Direct marketers have gathered information about ethnicity, race, and religion for decades, but advances in computing allow ever greater sophistication in modeling a person’s lik
ely origins, as well as details such as whether someone is a first-generation immigrant likely to speak a foreign language. Before the 1990s such marketing was often based on surnames or ZIP codes, with religion assigned based on broad assumptions (for example, designating those with Italian surnames as Catholic).
“There was a lack of trust and common knowledge that compiled lists through inferred information were not accurate,” says Peter Brownstein, former chief information officer at Ethnic Technologies, a leading ethnic data broker and processor. Today Ethnic Technologies uses multiple sources of information to predict 95 percent of the names it encounters. “As that marketing method evolved, they started looking at geography, neighborhood analytics,” says Karen Sinisi, director of sales.
Today, large data brokers such as Acxiom buy ethnic data and related software from companies such as Stirista and Ethnic Technologies, making the data widely available to marketers. Gupta says his program starts out by seeking matches to the five hundred thousand most common US surnames that make up about 70–80 percent of the nation’s total. Another five hundred thousand surnames account for everyone else, he estimates. For less common names, his team studies emigration and government records for clues. First names help determine if people are first- or second-generation immigrants. US-born citizens are likely to have Americanized names, with the exception of Indians, he says. Some marketers also rely on self-reported ethnic data from surveys, but Gupta says people do not always answer truthfully.
On its website, Stirista shows how it can supplement basic information about a customer. Jennifer Pham? A Protestant of Vietnamese ancestry who speaks Vietnamese. John Washington in Palmetto, Florida? A Protestant African American. Marshall Begay? A Catholic Native American Navajo. Sometimes the first name contains the decisive clue.