The Unintended Consequences of Ad Placement Based on User Data Rather Than Website Content

It seems so logical and efficient: define your ad strategy based on the audience you seek to influence, then let a intermediary service algorithmically place your ads where those consumers, in whatever way they've been defined, would appear to cluster.

The problem, as it turns out, is that using broad demographic frameworks (or even more specific, targeted data) can lead a brand to be - apologies to American Express - 'everywhere you don't want to be.' JL


Sapna Maheshwari reports in the New York Times:

It is increasingly rare for marketers and their agencies to place ads directly on websites. Most ads are placed based on data about the user rather than the content of the website. "When you buy so many sites there’s no way to guarantee you’re not going to be supporting something fraudulent or sensationalist." (But) how a brand ends up on a website (is) meaningless to consumers who associate the two regardless of whether they have a direct relationship

The Vanguard Group does its best to stay away from politics when advertising, going so far as to have a policy against marketing on overtly partisan websites. So it was a surprise to Vanguard, an investment management company, when it found its ads on Breitbart News, the hard-right site that has become closely tied to President-elect Donald J. Trump.

An email from a client alerted Vanguard to the issue last week, prompting it to pull its ads from the site. Breitbart had been inadvertently included on a list of preapproved websites the company and its ad agency use to try to reach people who have visited Vanguard’s site, said Emily Farrell, a spokeswoman.

Vanguard is one of at least a dozen companies, including prominent brands like Allstate, Kellogg and Warby Parker, that have said recently that they will stop advertising on Breitbart. The brands have attributed their appearances on the site to the automated nature of online advertising, in which a complicated system of third-party networks and agencies is used to place ads. That has cast a spotlight on how the zeal to capitalize on consumer data and advanced targeting technologies has resulted in some companies’ being associated with sites they want to avoid.

It is increasingly rare for marketers and their agencies to place ads directly on websites. Unlike advertising on television or in magazines, where brands often choose placements in advance based on content, marketers online deploy sophisticated technologies to target and personalize ads in real time for the people who are most likely to be interested in their products or services, wherever they may be roaming on the web.

Kris Charles, a Kellogg representative, said that while the company reviewed sites for potential ad placements “using filtering technology to assess the words and phrases that make up a site’s content,” there is also “a very large volume of websites, so occasionally something is inadvertently missed.”

All of this has resulted in the kind of publicity — and political repercussions — brands were hoping to avoid. After Kellogg said it would stop advertising on Breitbart because the site was not “aligned with our values as a company,” pointing to publicly available marketing guidelines, Breitbart started a #DumpKelloggs boycott campaign. Larry Solov, Breitbart’s chief executive, said in an editorial that the cereal maker “has shown its contempt for Breitbart’s 45 million readers and for the main street American values that they hold dear.”

A spokeswoman for Breitbart did not respond to requests for comment for this article.

Allstate responded to one consumer’s complaint about its appearance on Breitbart by saying, “Unfortunately, the nature of internet media buys is such that we are not always able to receive full disclosure with regards to all of the websites on which our advertising may run,” though it would exclude Breitbart from its ad buys in the future. The eyeglass company Warby Parker blamed “third-party ad networks or ad exchanges,” and affirmed its commitment to diversity and inclusivity. Other brands, including Nissan, have said they aim to reach as many customers as possible and do not plan to change their advertising strategy based on complaints from some consumers.

The nuances of how a brand ends up on a website, however, are often meaningless to consumers who may associate the two regardless of whether they have a direct relationship — which is being highlighted anew as tensions run high after the election.

About two weeks ago, a Twitter account called “Sleeping Giants” was created with the goal of choking off ad dollars going to Breitbart, urging people to screenshot ads on the site, then post those pictures and flag the advertisers. (The New York Times is among companies that have been criticized for advertising on the site.)

Workable, a start-up that sells recruiting software and champions diversity, was flagged by the account, which has roughly 6,000 followers, for having a banner ad on Breitbart above the headline, “There’s No Hiring Bias Against Women in Tech, They Just Suck at Interviews.” A screenshot of the ad was posted and sent to Nikos Moraitakis, Workable’s chief executive, who said he “nearly had a heart attack” when he saw it.

Workable’s ad ended up on the site through one of the Google companies that brokers web ads, and Workable has now added Breitbart to an “opt out” list, Mr. Moraitakis said in an interview. But even as it has blocked Breitbart, “there’s probably another 10 sites we haven’t excluded,” he said.

“We rely on our ad networks to keep the networks clean, and then obviously, the big ethical question is what clean means,” he said. “Unfortunately, these hate speech sites and fake news sites are a new breed, and I don’t think Google has caught on to them.”

He added, “The unfortunate side effect of this is that to the consumer, it looks like we are directing our advertising dollars to a specific media site.”

Most ads are placed based on data about the user rather than the content of the website, and limits are set after the fact — say, ensuring that brands do not show up next to pornography or neo-Nazi literature. This process is known as blacklisting, and because of the sheer number of websites, brands tend to value its efficiency over any qualms they may have about appearing in the wrong place at the wrong time.

The blacklist approach starts with 500 billion impressions a day across the globe in real time across all screens and all formats; the white list approach starts with zero,” said Joe Zawadzki, the chief executive of MediaMath, a marketing technology company. “So it’s, do you start with the universe, then filter, or start with nothing, then build up?”

To some, that illustrates the underlying problem. Joe Marchese, president of advertising products for the Fox Networks Group, said that while audience profile “matters first and foremost,” there “has to be a better standard set for the environment” where an ad will appear.

“When you buy so many sites that you don’t know them all, there’s almost no way to kind of guarantee you’re not going to be supporting something fraudulent, sensationalist or alt-right or whatever the content happens to be,” he said.

The uproar about Breitbart with the separate issue of fake news sites has sent advertisers to online ad verification companies like DoubleVerify and Integral Ad Science. On Thursday, DoubleVerify added an “Inflammatory Politics & News” category to a list of more than 75 categories that advertisers can choose to avoid. If selected, it will block ads from appearing on sites including Breitbart, Rawstory.com, WND.com, LibertyWritersNews.com and YoungCons.com, the company said.

“With the surge in these new sites that are living on unsubstantiated stories that are oftentimes very politically charged, large brands are coming to us and ensuring that their ads don’t appear on those kinds of sites,” said Wayne Gattinella, the chief executive of DoubleVerify.

Breitbart said in its editorial that Kellogg’s move would have “virtually no revenue impact” on the site. While at least one ad tech firm, AppNexus, has barred Breitbart from using its tools, citing a violation of its hate speech rules, data from Ghostery, a data governance company, shows that the site works with many ad tech companies. And while select brands have pulled away from the site, plenty of others remain.

Still, John Montgomery, executive vice president for brand safety at the advertising giant WPP’s GroupM, predicts that the price of ads on Breitbart could decline, affecting the company’s revenue, if enough “venerable brands” act.

“Since the Breitbart issue came up, we’ve obviously been talking to our people internationally and making sure they have their clients understand if they don’t want to be on political sites, we need to put them onto the blacklist,” Mr. Montgomery said in an interview. “Particularly after the publicity of the election there are many big brands who are having these discussions with agencies right now.”