As the publisher of USA TODAY and hundreds of local media sites across 46 states, Gannett Co., Inc. (NYSE: GCI) has a compelling proposition for advertisers — broad reach, big audiences and the ability to reach both general and targeted consumers. However, that value proposition and trust in the news organization was shaken this week when The Wall Street Journal's Patience Haggin reported that Gannett had misinformed clients of its ad exchanges during a nine-month period.
Haggin had received investigative data from Braedon Vickers, an independent ad research and self-described “Kiwi software developer.” Vickers partnered with Adalytics Research LLC, which tracks digital advertising placement and metrics. They confirmed that the digital ads placed with Gannett’s ad auction system, including for its flagship USA TODAY Network, weren’t always appearing where and under the express conditions the advertiser had been promised based on their bid-request data.
“Many brands have rules regarding which news topics they want their ads to appear next to, often to avoid being associated with controversial content,” Haggin explained in The Wall Street Journal article. “The discrepancies in the Gannett auctions meant that many ad auctions couldn’t adhere to those rules, the researchers said.”
Vickers told The Wall Street Journal that they’d found international brands like Nike, Ford, State Farm, Starbucks and Marriott were among those impacted.
This morning, Gannett released a statement in response to The Wall Street Journal article, affirming the reporting: “Gannett sincerely regrets inadvertently passing along the incorrect data parameter. This human error was immediately rectified when [Gannett] independently discovered the issue. The data parameter issue was caused due to a caching error when the Company implemented changes to how data is passed from the publisher to the ad exchanges.”
Despite some rather heavy-hitter brands being impacted, Gannett sought to temper the news by adding, “It is important to note that the revenue associated with third-party programmatic advertising changes that potentially used the incorrect data parameter in question was less than $10 million in total over the impacted period. Also, none of Gannett's direct-sold digital advertising or direct-sold programmatic advertising were affected."
In its statement, Gannett said it “has fully evaluated the quality assurance program relating to product releases and is implementing procedures to ensure that an error such as this does not occur again in the future.”
E&P reached out to Krzysztof Franaszek, Ph.D., the founder of Adalytics Research, to better understand how an error of this kind might occur and whether it was the result of malintent. In a chat with E&P, he explained, “I think it’s likely a simple ad ops error. We tried to analyze this phenomenon from a number of different angles to determine if there was some kind of material benefit to Gannett of doing this, and after an exhaustive enumeration of possibilities, we found no rational explanation of how this would benefit Gannett.
“For example, in some cases, a very ‘safe’ article about high school football was swapped out with an ‘unsafe’ article about war in Syria,” he continued. “From a brand safety or ad revenue aspect, this substitution doesn’t make sense.”
Franaszek expressed empathy for the news publisher and said he was more surprised that the problem went undetected by auditing software for so long. “Brands and marketers spend billions of dollars, hoping to reach their target audiences at the right time and place,” he noted. “In many cases, these brands rely on their vendors to implement quality controls. This Gannett situation offers an ideal natural experiment to verify the verifiers.”
For Gannett, the next question will be how to “make good” with the advertisers newly alerted to the problem.
Gretchen A. Peck is a contributing editor to Editor & Publisher. She’s reported for E&P since 2010 and welcomes comments at email@example.com.
No comments on this item Please log in to comment by clicking here