How Facebook’s measurement errors have eroded marketers’ trust
Facebook's measurement errors have led some brands to question their investments and agencies to recalibrate their content evaluations.
Facebook’s string of measurement errors disclosed since September has some marketers shook up, even if those marketers weren’t directly affected by the social network’s messed-up math.
Soon after Facebook announced in November that it had been overstating Page posts’ organic reach, digital agency Blitz was pitching a client on a Facebook campaign idea to make people more aware of the brand. That meant being able to measure how many people were exposed to the campaign, including those who organically saw the ad because a friend shared it or commented on it, even though those impressions were free. This brand had never been exposed to erroneous organic reach measurements by Facebook, according to Blitz’s director of social, Kevin Wright, but it was worried anyway.
The brand’s team “came back and said, ‘Hey, I’m very concerned. Should we continue to spend these recommended dollars on Facebook with the awareness objective given reach may be off?’” said Wright. “That was definitely a hurdle.”
The aftershocks of Facebook’s measurement errors continue to ripple across the advertising industry. While Facebook has emphasized that the flawed figures — such as average watch time, organic reach and video completion rates — did not affect how much money it charged advertisers for their campaigns, that doesn’t mean advertisers and their agencies haven’t been affected.
“Yeah, media buys are safe, but not everything has a media buy,” said Nikki Scoggins, VP and director of social media at creative agency Deutsch.
In addition to paying for distribution, brands also pay for the production of the content to be distributed, including content published to a brand’s Facebook page that may never be run as an ad or may be an initial test to see if it performs well enough to justify a paid promotion. To evaluate that content, the brands and agencies look to stats like average watch time and video completion rate. But, having learned that these stats weren’t accurate in the past and have been corrected, marketers are now wrestling with whether to trust them and how soon they can.
“We’ll have to start almost over as far as our evaluation,” said Scoggins. “For me to feel confident going to the [agency’s creative team] and saying, ‘Your video did this or that. Let’s look at changing this at this time [in the video],’ I’ll feel confident speaking to those points in another quarter. I think.”
Not surprisingly, the less a brand or agency relied on Facebook’s own measurements, the less shook up it is by Facebook’s errors.
For creative agency 180LA and its clients, Facebook’s measurement errors have been a non-story, according to the agency’s new media and communications director, Roy Dank. That’s partly because most of the measurement errors were specific to videos on Facebook, and 180LA’s clients don’t invest much in Facebook video. And the one false metric that would have affected 180LA’s clients, organic reach, has ceased to be an important metric for clients as Facebook has shifted to a pay-to-play platform for brands over the past few years.
“We weren’t dependent on these particular metrics or their metrics in general because Facebook has been de-prioritized for a bunch of our clients,” Dank said, attributing the lowered status to Facebook cutting Pages’ organic reach.
According to Blitz’s Wright, none of the agency’s strategic recommendations to its clients were “greatly affected” by Facebook’s erroneous metrics,” but it could and does hinder a client’s trust in these channels” and its willingness to invest in the producing the content to distribute on those channels. “When we’re asking for dollars to produce video content and this is in the back of their head — are the measurements my agency is telling me accurate? — that’s a problem,” he said.
A potentially bigger problem is whether these relatively minor errors preclude a major one, one that would directly impact how advertisers are spending their money on Facebook. Like Facebook’s store visits metric, which estimates how many people saw a retailer’s ad on Facebook and proceeded to stop by one of the retailer’s brick-and-mortar stores. It’s a metric that Blitz considers when recommending large Facebook ad buys to a client that operates a hundred stores, said Wright.
“Imagine if there was an error in that that came out a few months later. That is going to make a much more significant impact than having video completion rate off by 35 percent,” Wright said. “Now it’s always going to be in the back of the head: Is the data accurate? Because there are metrics that are primary metrics that will greatly affect a brand’s or agency’s investment into the channel. It’s definitely a concern. It’s not going to go away.”
The possibility of Facebook flubbing as important a measurement as store visits would have been inconceivable for some marketers before the errors were announced. But then again, the aforementioned Blitz client that questioned whether to spend the recommended budget on Facebook ads would never have raised that question before the errors’ disclosure, said Wright.