Google Ads — Racial Profiling?

Google Ads — Racial Profiling?


Google’s perceived problems with racist content hit the net recently – this time over its biggest bread basket: advertising. Nathan Newman over at claims that Adwords, Google’s advertising division, is profiling users based on race and class.

Newman conducted a small study (which he admits was less than scientific) with nine different user names commonly associated with certain racial and ethnic groups. Using Google Mail, he typed the same subject headings in each message, then checked the right side of the screen to see what kind of ads were generated.

He found that e-mails with Anglo sounding names like “Connor Erickson” generated different advertisements than messages from “Deshawn Washington” a name often affiliated with African Americans. When “Connor” wrote “arrested: need a lawyer” in the subject line he got ads for white-collar lawyers. However when “Deshawn” wrote the same message, two listings for DUI lawyers popped up.

The same test found that “Buying a Car” with “DeShawn” and the two other Black names produced at least one ad for bad credit auto loans or products related to used-car purchases. “Jake Yoder” received advertisements to buy new cars from BMW and GMC.

An email generated from all three users with Latino sounding names inquiring about employment generated an ad for Salsa Labs. The advertisement did not appear when the other names were used.

Newman said that class profiling was also kind of obvious with differing results. He said that in the poorer area of the South Bronx, he received ads for car lifts.  Lexus generated Caucasian-sounding names – whatever that means.

The more positive sides of his study showed that predatory lending promotions, while still rampant, were aimed less specifically at minorities, a charge that lenders are still grappling with since the housing bust disproportionately caused millions of folks of color to loose their homes.

Still, Newman said in his post that there were “enough provocative results to suggest that racial profiling is likely a reality in online advertising.” He believes that this is important because, “for those at the lower-end of the economic scale or who already suffer discrimination, the Internet may be magnifying and more precisely targeting that discriminatory treatment. ”

Google representatives say that Newman’s reports are completely false, with no algorithms for race or ethnicity being a part of their or Adwords system. They admit that they target users, but not by race or ethnicity.

“This post relies on flawed methodology to draw a wildly inaccurate conclusion. If Mr. Newman had contacted us before publishing it, we would have told him the facts: we do not select ads based on sensitive information, including ethnic inferences from names.”

Google has long been a target of Newman, who says that the company made billions from advertising for sub-prime lenders; and he isn’t the only one that’s questioned the racial content on the site (for example a few years ago when one typed in Michelle Obama they were directed to a picture of her resembling a monkey. The company apologized).

However, the issue that Newman brings up is not just about Google but also about whether online racial targeting is a problem or just another method used by advertisers.

Advertisers have long used information to target consumers, but the rise of the Internet makes monitoring consumers considerably easier than before.

“Consumer racial targeting has a positive connotation for me,” said Jerome Williams a Professor at Rutgers University’s Business School. “Now for others it has a negative connotation, because when people think of targeting they tend to associate it with things like targeting alcohol and tobacco products, but there can also be a positive connotation of targeting, like cars and computers and books and all other kinds of things.”

While Newman says racial targeting techniques can be “harmful” to minority groups, Williams says he’s yet to see how the practice is actually causing damage.

“There’s nothing in my mind wrong with targeting, even racially targeting, in fact that’s what I’ve been trying to get marketers to do for 30 or 40 years … to pay more attention to racial ethnic groups because of the growth of the economic spending power they have.”

Ross Petty, a professor of Marketing Law at Babson University agrees. “Racial profiling to target online ads is not illegal for most products, although public outcry can occur,” said Petty in comments to “Racial targeting is not always bad. If I sell adhesive bandages in colors suitable for African American skin tones, both I and my customers would probably appreciate the ability to target advertising by race.”

But Jeffrey Chester, head of the Center for Digital Democracy, has another take. He’s even lobbied the Federal Trade Commission, the government agency that is responsible for protecting consumers and regulating business practices, to stop members of racial and ethnic groups from being targeted by online businesses without their awareness. Chester claims the practice could lead to downgraded rates for credit cards or being duped into bad loans.  But he believes the practice is so rampant now that one company can’t be singled out.

“The bigger problem is not Google,” he said. “It’s an industry-wide practice.” He points to evidence of junk-food advertisers courting black and Latino users disproportionately, despite studies showing rising rates of Type 2 diabetes and obesity in these communities. He also notes that predatory lenders used online ads to unfairly target minorities during the housing boom a few years ago.

Perry, who specializes in marketing law, says that if a vulnerable group is specifically targeted with a product “of their weakness” then it could be considered “unfair” under the Federal Trade Commission Act.

“Normally, media are not responsible for deceptive advertising, only advertisers are,” Perry said. “But if an agency helps create the ad, the ad agency may also be responsible for deceptive claims in the ad. Seems to me if a medium such as Google helps marketers target vulnerable audiences with products harmful to them, an argument could be make that like advertising agencies being held responsible for ad copy, targeting media should be responsible for targeting.”

Williams says that he still hadn’t seen enough evidence that the practice was widespread or that it distressed minority groups.

“I say bring me data, show me something,” Williams said. “Because you type a word and this pops up, how do you know it’s not something random? I can be persuaded if you can bring me the evidence.”

Others, say the evidence is clear and solutions must be created. One plan by Jim Edwards on BNET suggests that Google force advertisers to sign a “non-discrimination promise.”

Chester is advocating for an opt-in process to let consumers decide if they want their information released before they click on certain buttons. He says right now the only information that is considered “sensitive” is your social security number and bank account information.

“On the one hand you want to see a diverse online publishing sphere. We want to see funding for groups that represent the concern of groups that reflect America’s diversity,” Chester said. “However … this information can be used to harm and discriminate.”