Several Facebook advertisers are pressing to move forward with a class-action complaint accusing the company of duping them by inflating the possible reach of their ads.
“Advertisers ... flock to Facebook’s platform because of its supposedly 'unmatched' potential reach,” a group of advertisers including Danielle Singer, who owns the Kansas-based business Therapy Thread, says in papers filed Thursday. “This foundational representation is false.”
Singer and the others are urging U.S. District Court Judge James Donato in the Northern District of California to reject Facebook's request to dismiss the lawsuit.
The dispute dates to last August, when Singer alleged that Facebook induced advertisers to purchase ads -- and pay higher prices for them -- by inflating the number of users the ads could reach. Singer alleged that she purchased $14,000 in ads on Facebook between October of 2013 and April of 2018. After she sued, three other advertisers joined the case.
The complaint draws on reports by outside groups, including the industry organization Video Advertising Bureau, as well as a survey commissioned by the plaintiffs. The Video Advertising Bureau reported in 2017 that Facebook's estimates of audience reach in every U.S. state were higher than the states' populations according to the U.S. Census Bureau.
Facebook has asked Donato to dismiss the complaint, arguing that estimates about campaigns' reach are not guarantees, and that the estimates don't affect billing.
“Advertisers are charged based on specified outcomes such as website conversions, clicks, or impressions an ad receives,” the company wrote in papers submitted to Donato in February. The company adds that pricing depends on Facebook's auction system.
Facebook also says it discloses that potential reach estimates aren't based on census data or third-party data. “It is simply not plausible that plaintiffs would not have placed ads on Facebook if only they had known the potential reach estimates did not match census and survey data about Facebook users, given the conspicuous Facebook disclosures,” the company argues.
But Singer and the others say they should be able to proceed with their claims.
“Facebook argues it should be free to inflate its core metric. The court should reject this extraordinary argument,” they argue in their new court papers.
Donato is scheduled to hold a hearing in the case on May 16.
Facebook is also facing a separate lawsuit over allegations that it inflated video metrics. That earlier matter, which dates to 2016, is pending before U.S. District Court Judge Jeffrey White in Oakland, California.
While understood by the media cognoscenti, the term "reach" may be misinterpreted in this action. "Reach" sometimes referred to by its full description, "net reach", is unequivocally used in advertising and media to mean the unduplicated number of the target group potentially exposed to the ad message over the campaign period. This is very different from the "gross impressions" delivered by the campaign over its timeframe which is a count of every potential exposure to the ad message including any and all multiple "impressions" by the same person, i.e., duplicated.
A gross impressions count can and often does exceed the population of the target group. Based on this report the inference is that the reach claim is specious and is in fact a gross impression count. A significant difference . The "reach" of a campaign cannot exceed the target group population and often falls well short of even 50% of any total target group. Social media campaigns typically struggle with achieving any meaningful level of reach notably at the 100% viewable level for an ad now the industry standard. Entities like Facebook also ONLY have their users as the basis of any campaign target metrics with non-users excluded. Consequently the real total campaign target group for a campaign across its entire geography is usually significantly larger that that used by the social media channel with those not using the channel having a "reach" of zero!!
Traditional media, including newspapers and out-of-home, have long provided accredited estimates of both reach and gross impressions over time relative to the full total campaign target group for years and are still considered the bedrock media for any campaign.
"Fakebook" sadly has a track record of misrepresenting established media metrics and best practices as well as providing and/or misleading advertisers with unaudited (and possibly inflated) data. Clear as mud? If so let me know .
Too true Tony.
My investigations to the same problem here in Australia found that the data is 'account' data.
In the case of mobile advertising to 'the Sydney market' includes any account within that IP-defined area. That includes visitors to Sydney. Overseas visitors numbered 4.2m and stayed for 80m nights, Domestic vistors (i.e. other Aussies) numbered 10.3m and stayed for 27.2m nights. You can also add in people like myself that don't live in Sydney but travel back up fror work, social events, dinners, and of course the football.
What's worse, if you have micro-defined areas (e.g. a sub-region or even a suburb) you only need to travel through the area with your mobile tracking on and you are accounted .. for every region/postcode etc. you travel through.
The 'reach estimator' here was regularly tracking around 140% of the population.
This whole issue encapsulates the difference between conducting 'research' and 'counting things'.
John: Excellent reminder of how "gross impressions" can be gross gross impressions by conveniently "counting" any and all potential target group exposures (loosely defined as far as Facebook is concerned) whatever their actual geographic or residential relevance to the advertiser actually is. As Erwin Ephron stated, it is the reach, ideally based on measured actual exposures to the target, over a specific time frame or frames that is the key media metric not impressions, especially when they are really gross, which merely drive frequency of potential exposures against the target group. Frequency, he reminded us, " is like crab grass!" To estimate unduplicated reach one needs a population for the target of course which typically requires geographic boundaries for the estimate.
Along with all the issues often associated with social and digital media metrics - bots; viewability; known exposure; duration; etc. - your reminder regarding, "How gross are your gross impressions and what about the reach?" underlines the value of quality media research, offered via those media platforms from independent rating services. These media truly want to provide accountability and comparable metrics across all platforms to their advertisers.
We rejected self measurement and metrics from media entities a long time ago. It should not have been allowed back notably from an entity that is being investigated for privacy breaches worldwide and that has already been heavily fined for such practices.
Perhaps we will be appearing in front of Judge James Donato?
Great points Tony.
The tricky part is that the media owner is the one with the big money (and most to gain - or lose) with a media agency charging around 3% of media billings as an agency fee. So, on a $10m annual campaign the media agency typically charges $300k. There is still a 'wild-west' in the digital world with lots of other players, including media agencies, clipping the ticket - but that is a different issue.
When it comes to funding industry 'currency' research it is basically an 80/20 split here - 80% by the media owners, 20% by the buyers. Fortunately here in Australia we have quite strong 'technical research committees' - where we only talk tech and not funding - and we're pretty much on equal footing (albeit with buyers outnumbered by media owners) on that basis.