Is Investing in Digital Advertising a Miscalculation? Ask Facebook

Facebook committed a serious breach of trust with its advertisers. This post explores how it did so, the implications on digital advertising--and how, based on its actions leading up to this revelation, "This is not the Facebook we have come to know."



Is Investing in Digital Advertising a Miscalculation? Ask Facebook

Sep 29, 2016

AUTHOR: Jeff Cotrupe, Industry Director, Big Data & Analytics, Stratecast | Frost & Sullivan

Facebook has been seriously overstating to advertisers the success of ads placed on its network for approximately two years now. As reported in a multitude of media outlets[1], Facebook has been artificially inflating the impact of video ads, by as much as 80%, to its advertisers.

How They Did It

Here at Stratecast we observed the cries of despair and anger in the marketplace. Then we asked a question: specifically, how did Facebook pull it off? The answer is: by skimming the cream off the top. By cherry-picking. Need more? Ok, let’s do this with a simple math pop quiz. Which result is larger: 300 divided by 40? Or 300 divided by 100?

Time’s up. If you answered 300 divided by 40: YOU ARE CORRECT! You understand how Facebook pulled it off.

Still need more? Ok, we’ll illustrate via a hypothetical example:

  • Let’s say 100 Facebook users watched a video ad: 60 of them watched it for a second or two and then “bounced,” quit watching. However, 40 of them watched it for three seconds or more.
  • Then let’s say total time spent watching the video ad by all 100 Facebook users was 300 seconds.
  • If you were reporting results to advertisers about how much time the average Facebook user who saw the video ad spent watching it, you’d divide 300 seconds by all 100 users, right? Average time spent watching the video ad: three seconds.
  • “…but, NO.” Again using our hypothetical example, but absolutely true according to the formula, here is what Facebook has been doing, instead, for two years: dividing those 300 seconds of total time spent watching the video ad by 40. That’s right: Facebook included in its user sample only those who were most heavily engaged in the video ad, the 40 who watched it for more than three seconds. So that’s 300 seconds divided by 40 users. Thus, per our hypothetical example, average time spent watching the video ad, as reported by Facebook to its advertisers for the past two years:  7.5 seconds.[2]

As reported by those multitudinous media outlets, including the one cited above, Facebook’s (current) VP of Business and Marketing Partnerships, David Fischer, issued an apology for this systematic overreporting of advertising impact to advertisers, calling it a “mistake” and a “miscalculation.” (Although I heard a rumor that Fischer actually called it “THE miscalculation,” until Justin Timberlake arrived on the scene and coyly suggested they drop “the THE.[3]”)

Irony: Facebook Has Been Championing Increased Value for Advertisers

The irony is that, as we have analyzed previously[4], Facebook has been leading the charge for a new standard designed to deliver far greater value to advertisers. The old standard for digital advertising was known as served impressions. If an ad appeared where it was supposed to—for example, on a Web page—the system registered delivery of the ad. Every time someone visited the Web page where the ad appeared, that was counted as a served impression. The problem was that the ad could appear at the top of the Web page, seen by every visitor—or the ad could be seen by few or no page visitors, for reasons including:

  • The ad appeared at the bottom of the page, or “below the fold.”
  • Visitors “bounced,” meaning they arrived at a page and left quickly, due to lack of interest or because they clicked on a page they did not intend to visit.
  • Connectivity issues kept the ad from rendering fast enough for visitors to see it.

Yet, all of these were still being counted by many online publishers as served impressions.

Facebook uses a new standard: viewed impressions. A viewed impression occurs when an ad enters the screen of a desktop browser or mobile app. Viewed impressions add an extra layer of analytical rigor to evaluating ad impressions. They more accurately define delivery, and help ensure that people have seen the ads they are supposed to see.[5]

Facebook’s standard for viewed impressions, however, is less stringent than that of the Media Rating Council (MRC), the industry group of record in such matters. The MRC, in collaboration with the Interactive Advertising Bureau (IAB) and others, has adopted a viewable impressions standard that at least 50% of the pixels comprising a banner ad must be in view for at least one second, and that a video ad must be viewed for at least two seconds.[6]

Irony of Ironies: Using Only Viewed Impressions That Eclipse the MRC Standard

From these figures it is both clear and intriguing—and third, somewhat baffling—that, while Facebook has been arguing for a less stringent viewed impressions standard than that of the MRC, when it comes to reporting levels of video engagement to advertisers—in the form of average time spent viewing a video—Facebook limited its analysis to only the pool of users whose viewing duration exceeded the MRC’s more stringent standard.

Is This a Signal to Stop Investing in Digital Ads? Hardly

Is this admittedly serious breach of trust committed by Facebook with regard to its advertisers a reason to stop placing ads in digital media? Stratecast believes it is NOT. First, while Facebook is clearly, along with Google, a leader in online ad revenue[7], it is only one network. Opportunities abound on networks from Twitter to LinkedIn and untold media and rich content sites for brands that want to advertise digitally. In an increasingly digital and mobile world, the place for advertisers to be is where buyers are. Do brands really want to go back to the realm of old (ok, “traditional”) media? “Your ad appeared in that issue/on that program/at that location, which had X subscribers/viewers/passers-by, so it received Y gross impressions (GRPs). If it didn’t do a thing for your sales figures, clearly there’s a problem with your product or service. Now, in our special Fall issue, closing date just 10 days away, we spotlight product improvements specifically in YOUR industry…and I just happen to have ONE spot left in premium position, right-hand page far forward…”

Invest in digital media.

Yes, even Facebook.

Yes, even video ads on Facebook. Especially video ads on Facebook. A now publicly-chastened Facebook may actually be a good bet, at discounted prices. “There may be noplace safer in this big ol’ bustling city tonight than the scene of the previous (digital media) crime.”

Results-Oriented Advertising—and Advertising Attribution

One last thing: while impressions are indeed important in advertising, the most important thing is the "impression" a given ad campaign is making on your bottom line. Facebook or any advertising network or medium can quote impressions or engagement figures all day long, but what matters is whether your ads are making the cash register ring. Right? Or if pure revenue growth is not it, then whatever type of conversion(s) your organization is looking to achieve, whether it is getting people to request more information about a product, sign up for an event, or other objectives. So if you haven’t done so already, start taking a results-oriented approach to advertising.

Especially since your organization is using multiple media, venues, and vehicles to get your marketing messages out, that may be sage advice, but it is also easier said than done. To go back to that old quote attributed to U.S. merchant and political figure John Wanamaker: “Half the money I spend on advertising is wasted; the trouble is, I don't know which half.” Or, fast-forward to an analogy by this author: ad attribution can be like figuring out who gets credit for a goal in ice hockey that was deflected multiple times in front of the net.

You can support a results-oriented approach to advertising, and get ad attribution right, with an advertising attribution solution. The list of providers for such solutions is a short one, but such solutions do exist. This blog post has already covered a lot of ground, and this is not the place for an extended treatise on ad attribution solutions, but be on the lookout for a future post by this author that tackles the topic.


[1] Including this one: Startup Smart, Facebook miscalculated advertising metrics – should businesses now ditch video ads?  (26 September 2016), available here

[2] They do not hire us at Stratecast | Frost & Sullivan unless we are math WHIZZES.

[3] If you correctly got the reference from the movie The Social Network, the author hereby certifies your entry into our exciting raffle; grand prize: VP of Business and Marketing Partnerships at Facebook.

[4] Stratecast, Advertisers are Getting the Wrong Impressions; Facebook and the MRC Offer New Points of View (SPIE 2015-13, 3 April 2015)

[5] Facebook, The Value of Viewed Impressions for Advertisers (18 February 2015), available here

[6] Media Ratings Council, MRC Viewable Ad Impression Measurement Guidelines (18 August 2015), available here, and MRC Mobile Viewable Ad Impression Measurement Guidelines Public Comment Draft (31 March 2016), available here

[7] Bloomberg Technology, Google and Facebook Lead Digital Ad Industry to Revenue Record (21 April 2016), available here


Please add your bio info through your member profile page, or through your dashboard.