Ethnic Technologies, a 19-year-old company based in South Hackensack, New Jersey, develops software that helps businesses target different ethnic groups. Its algorithms are designed to identify a person's ethnicity on the basis of his or her full name, address, and ZIP code.

The company, whose software considers 158 distinct ethnicities--including more segmentation for Hispanics and African Americans--claims it has an accuracy rate of 90 percent. This could lend a huge hand in marketing: With minorities expected to account for more than half of the population by 2044, the ability to target those customers is key to driving sales growth.

How small businesses can benefit

Traditional marketing methods have been successful in reaching a white audience, but often fail to attract minorities. Brands today are working to fix that. Twitter has marketed bilingually, and other companies have partnered with local artists and filmmakers to capture the Hispanic audience (a customer group that holds up to $1.5 trillion in spending power, especially on holidays like Cinco de Mayo and Day of the Dead).

A cosmetics startup, for instance, could use Ethnic Technologies to target African American, Hispanic, or Asian customers, who often prefer different makeup. AcquireWeb, a marketing consulting firm based in Omaha, has used Ethnic Technologies to help its own clients--ranging from local flower shops to advertising agencies--identify their customers' ethnicities, and suggest marketing strategies.

"You give us an email, a name, or a general audience that you want to reach and we onboard that into our algorithms," explains AcquireWeb's head of marketing, Cort Irish, of the service. "We turn back a model that tells you more about the individuals."

Analysts see plenty of benefits from the technology--provided its not taken as absolute. "We're all looking for filters to identify our audiences, and there is no perfect science," says Rochelle Newman-Carrasco, the chief Hispanic marketing strategist at Walton Isaacson. She warns, however, that businesses need to consider cultural heritage as only part of the picture.

"There is always a search for filtering and segmentation, but it can't be done in a vacuum," she adds. "I would be cautious about too many conclusions about things like language preferences or even cultural affinities."

The rise of predictive systems

A number of companies have created software to give insight into customers' backgrounds. In the auto industry, research firm IHS Automotive (previously, Polk) has long used "ethnic coding"--data showing who owns and buys vehicles according to the actual registered owner, and not just the neighborhood someone lives in--to help automakers target minority car buyers. More recently, the producers of Straight Outta Compton partnered with Facebook to launch three separate ad campaigns (for white, black, and Hispanic audiences). Different trailers were presented to Facebook users on their news feeds according to their presumed ethnicity, determined on the basis of behaviors the social network observed over time.

Billed as a solution for companies in a variety of industries, Ethnic Technologies is different than most, in that it looks specifically at proper nouns--both first names and last names.

Lisa Spira, the director of research and product development at Ethnic Technologies, explained in a recent New York Times Q&A that these context clues offer a more holistic understanding of a person's background. As an example, she points out that Yeimary Moran is an American first name with a Hispanic string ("eimary") and a last name that could skew as either Irish or Hispanic. To achieve greater accuracy, the software would then look at Moran's Miami address and find that she lives in a predominately Hispanic neighborhood.

The challenges and controversies

It's unclear to what extent the technology could account for linguistic anomalies. African American customers with atypically white-sounding names, Newman-Carrasco points out, might live in neighborhoods that don't skew African American, and could be misread. What's more, names that combine indicators, like Mario Kreutzberger (a combination of Latino and Jewish names, and the real name of television host Don Francisco), are also likely to trip up the machines.

Tony Thorne, a linguist who focuses on naming and business communication at London's King's College, says he's excited about the possibilities of predictive software, but agrees that anomalies could be problematic.

"They [Ethnic Technologies] are obviously fully aware of the complexity of ethnic and other identities," he says. "But there's quite a lot of eccentricity in naming. And many of these names, to me, don't have a clear ethnic label, so I don't know how they can interpret them."

The technology is likely to be controversial. Thorne, who often helps brands involved in trademark disputes, notes that names are psychologically and socially charged. "It's so intensely personal," he says, "that anybody playing with names is going to make others feel uncomfortable."

Another concern is how exactly a company like Ethnic Technologies keeps the software from getting into the wrong hands. Ethnic Technologies could not immediately be reached for comment, though Spira previously noted that language in client contracts restricts the technology for only "allowable" use cases, to avoid racial profiling.

Ultimately, the possibilities that these solutions offer are exciting for brands--even given the limitations. "I'm sure that brands will see it as something helpful to them," Thorne says. "As a social practice, it's a little more problematic."

Published on: Oct 20, 2016