Facebook knowingly took advantage of spam ad effectiveness estimates and lawsuits

Facebook, which, along with Google, accounts for about 60% of online advertiser spend, has knowingly built some of its astonishing success on incorrect data, according to recently unsealed court documents. Incidentally, this can be a problem for a business that generates more than 90% of its income from the sale of ads.

In short, this class action, which was filed for the first time in 2018, alleges that Facebook massed the “potential reach” numbers – an estimate that Facebook gives its advertisers the number of people likely to see their ad – to entice advertisers to spend more money on the platform, all in hoping to reach the people who Facebook had promised. The documents make it clear that some of Facebook’s top executives, including COO Sheryl Sandberg, were fully aware that the company had spent years exaggerating the number of eyeballs its advertisers could reach.

As first reported According to the Financial Times, the lawsuit claims that when Facebook’s base came up with internal fixes for those inflated numbers, senior executives repeatedly dismissed them on the grounds that their solutions would reduce the company’s massive ad revenue.

Thanks to these unsealed deposits, we know how exaggerated some of these numbers were. Here’s an example: in 2018, Facebook Told its advertisers that it had a potential reach of 230 million adults across the United States, out of the 250 million adults counted by U.S. Census data that year. But according to a 2018 Pew Research Study, only around 68% (or 170 million adults) actually use the platform. Sandberg admitted in an internal email that “she had known about Potential Reach issues for years.” But she has repeatedly repeated attempts by employees to rectify those numbers, according to the filing.

Internally, employees have recognized that if the product charges
itself as an estimate of the number “people“Your ad could reach, at best, an estimate of the number of accounts, including the countless number of false and duplicates. Some employees even ran the numbers in 2018, just to see what would happen if known duplicate accounts were removed from potential reach, and saw a 10% drop in the number of advertisers. Facebook chose not to cut them. When one of the product managers on the Potential Reach team later suggested changing the way they talked about those numbers – like, for example, replacing the word “people” with the word “accounts” – this suggestion was dismissed due to concerns about the “” impact this could have on Facebook’s ad revenue. According to the complaint, the manager replied that “this is revenue we should never have generated since it is based on false data.”

In many ways, this case mirrors another high profile advertiser costume that hit the company in 2016 alleging Facebook knowingly withheld some serious issues with the settings of its video ads in an effort to make more money from these video ad partners. In 2019, Facebook settled the claim for $ 40 million which, as others have pointed out, is roughly change of bump to a winning company tens of billions dollars in advertising revenue per year.

And apparently Facebook hasn’t learned much from that slap on the wrist. Regarding the lingering issues with Potential Reach, the lawsuit points out that the numbers Facebook continues to give to its advertisers make even less sense, such as telling them it can grow to “100 million” from 18 to 34 years of age. the country. Census data shows that there are in fact only 76 million of them – and we know not all of them use Facebook.

Both in class and on his own site, the company argued that these measurements are meant to be interpreted as estimates and not as gospel. But internally, according to the new filings, the company admitted that Potential Reach was “arguably the most important number” that advertisers relied on when deciding whether to put their ad on Facebook’s platform in the first place.

We’ve reached out to Facebook for feedback and will update here when we get back to you.