Facebook Metrics Show Danger of Buying in Walled Gardens

Nothing says more about the danger of buying only from walled gardens than Facebook’s recent admission that people were not watching as much video on the social network as it had  reported. The average video on Facebook was counted as “viewed” after as little as three seconds, but Facebook didn’t calculate in the number of people who don’t watch video on the site at all.

Facebook apparently made a division error of the kind any normal human could make,

Instead of dividing the total time spent watching a video by the total number of people who watched that video, Facebook’s metric reflected the total time spent watching divided by the number of views the video had generated. With Facebook counting views at three seconds, that meant anyone who had seen just a glimpse of the video was not getting represented in the metric. In fact, Facebook told advertisers that its metric was off by 60 to 80 percent, according to The [Wall Street] Journal.

Advertisers seemed not to care, revealing they they don’t buy on time watched, but on either 10-second views or completed views. In this case. Facebook’s being wrong didn’t seem to cost advertisers money. But they should care, because it turns out 80% of Facebook users don’t watch video at all. Brands looking to shift large budgets from TV to digital video can’t do that safely until the know what their return on investment will be similar. It would be more advantageous if buyers spent more with independent publishers, who are closer to their audiences. Facebook’s audience is simply too large to count properly, and too uncommitted to give good results.

Most independent publishers are forced to accept some kind of third party verification of their views, but Facebook does not use third party vendors; it does its own analytics internally. After this admission, self-attestation will not work anymore for Facebook. It must allow in the same third party vendors, ComScore, Nielsen, or someone else, that the rest of the publishing world uses to tell its story.

For publishers, this inflation of the video watching time is worrisome at best, because most publishers felt they had to pay ball with Facebook and when the platform put its emphasis on video, publishers scrambled to provide video content. But all these publisher resources would be wasted if no one were watching. Because of the way advertisers pay, they can afford to wait and see if their ads work. Content publishers have no such luxury. Once they throw money at an expensive initiative like video, they would like to be sure they’re getting paid.

It will be a while until all this sorts itself out, and we figure out whether mobile video deserves the dollars being pulled out of trusty old TV.


Metrics Will Change Again

In the beginning, there was the CTR. It seems like so long ago (actually a scant two decades), but online advertising started by measuring click through rates. Never mind that those rates fell precipitously over the years as advertising proliferated. It took a long time to figure out how, or if, we could measure the ROI on digital advertising by using something other than CTRs. In those early years, the supply side only got paid on CTRs. Some in the industry realized that they could hire people to click on ads. Thus the first online advertising scams were born.

We should have known from the start  that CTRs were a ridiculous metric.  Never mind the bots that clicked on ads, or the fraudulent “click farms.”  Even with a clean supply chain, CTRs weren’t a good metric. After all, if an agency or a brand submitted an unmemorable ad, how could that have been the fault of the publisher on whose site the ad ran? It took more than a decade to bring that conundrum to light, and to agree that CTRs were only valid for performance advertising, and digital had “graduated” to more than just performance advertising; it was also useful for brand advertising.

So we moved as an industry from one lousy metric to another: impressions. The rise of programmatic and RTB gave media planners the ability to buy “impressions” at scale. Now it became a question of how many people were “exposed” to the brand. The thirst for impressions, however, did more harm than good, as publishers redesigned their sites to handle as many impressions as they could. For a long time, buyers didn’t even know what they were buying; they bought blind, to the potential detriment of their brands, and they bought impressions that weren’t even viewable.

But it gets worse. In the thirst for impressions, we lost the sense of advertising’s true intent: to convince a customer to buy a product or service from a brand. Even viewable impressions don’t help measure this return on investment. You and I instinctively know that just because we saw an ad, we might still not buy the product. We might be counted as an impression, but we might not be the target, or we might not have the need, or the funds.

Despite all the metrics we’ve tried over the years, we still haven’t solved the problem of predicting which consumer is going to make a purchase. The newest suggestion for a metric is engagement. But how do we measure that?  Cross-channel campaigns are held out as the answer: keep hitting the same consumer with a brand message everywhere she goes.

This metric is too novel to be judged right now, but because of new EU privacy laws, the increasing use of ad blockers, and the impatience of consumers with being tracked, we hesitate to say it will work any better than the metrics before it.

Let’s quit overthinking this stuff.  It’s time we were honest with the consumer, told our story well, and went back to the store to wait. The consumer is in control.


Publishers Must Help Advertisers Buy Video

Recent tectonic shifts in the advertising ecosystem have left all parties reeling and reaching for something firm to hold on to. In the past two years, You Tube, where dancing babies and talking heads once prevailed, has become the home of major celebrities. The online content world has shifted from text to video almost overnight, and audiences have shifted with it, from TV to tablets. Advertisers know they have to play here, but they’re still not sure how.
What’s holding up the growth of online video ad dollars? Media planners don’t know how to buy, and publishers don’t know how to deliver value in familiar terms. By familiar, I mean a term like “GRPs”, the traditional way TV is bought. GRP is a difficult metric to apply to online video, but it’s the common currency of the advertising industry and a $450 billion industry doesn’t shift overnight.
In addition, most media planners are thinking of online as a “second” or “third” screen, part of an integrated campaign that begins with TV and includes online, and that raises the question of attribution modeling. How to tell which part of the campaign was the most effective? There are many questions.
So when advertising and online video people get together, as they did recently at OMMA Europe, what do they talk about?  Metrics. Everyone is trying to figure out how to measure the effectiveness of online video so they can decide what percentage of their budgets to shift.
The uncertainty means that ad dollars are indeed shifting to online video, but not nearly as quickly as audiences are. The industry, which used to be about stirring emotion (see Mad Men for details) has now been taught to rely on data. But we don’t yet know how to interpret the data generated by online video viewers.
Indeed, when we talk about the worlds of second and third screens, we don’t even know with certainty who in the family is watching on that iPad. Mobile devices don’t allow cookies, and devices are regularly passed from spouse to spouse and parent to child. So how do we track response?
This is made more complicated by another shift: from performance advertising, which characterized the old world of display, and was easily tracked, to brand lift — the specialty of TV, but a more amorphous metric.TV is about brands and uplift, but where in the marketing funnel does online video play?
Not much brand language is coming out of the digital video industry, at least not language that can help media planners learn how to buy video intelligently, The industry has to find a common currency that plays across all media.
Old school planners only recognize GRPs, the traditional TV measurement. This makes it hard for niche publishers to position their value propositions, especially as RTB takes over.
More relevant metrics might include viewability, a first step, and engagement, measured by how much of the video a viewer watched. You can’t have brand uplift without more than just viewability. This number could be a proxy for engagement. Audiences are being validated by Nielsen and comScore after the fact, but what do you buy on?
It’s time for all of us in the industry to quit talking geeky jargon and start helping advertisers feel comfortable. We know the audience is online; now let’s work together to help the advertisers get there.