Tech Trend 247

Count down three seconds in your head. It goes fast, right?  I bet you did it the "one-hippopotamus way." That's actually considerably more than 3 seconds. Ask Siri to set a 3-second timer and then try to read something on the web.

Exactly, it's barely enough time for a sentence, including this one.

A couple of years ago, Facebook apparently set 3 seconds or less as the benchmark for actual video ad engagement. A choice that, on its face, seems rational, even preferable for media planners and advertisers. 

Anyone who spends under 3 seconds watching a Facebook video ad was, for the last couple of years, not factored into the social media giant's average viewing time. 

Of course, an average that doesn't account for the total number of viewers, whether they're engaged or not, is probably not much of an average — a fact Facebook belatedly discovered and confessed to in a well-hidden blog post. Now, though, as ad partners stepped forward and called foul, Facebook made a more public apology, but insisted that "it did not impact billing," which is another way of saying that everyone got what they paid for.

The measure of a view

I bet advertisers and their partners miss the good old days, when a media platform (newspapers, magazines, TVs) simply handed them audience numbers. The size of the audience dictated how much a media company could charge for inches, pages and minutes. No one could prove if people skipped the ads in print and, before the DVR, everyone ignored the fact that most of us used commercials as snack and bathroom break times.

The problem is that the explosion of social media and mobile video has led to the rise of engagement as a measure of success. YouTube does a good job of measuring video engagement. It knows when a video is playing, for how long, and if it stopped playing. They still don't know if you left it playing and walked out of the room.

I am a little surprised that advertisers are trying to equate YouTube and Facebook video ad views. 

I am a little surprised that advertisers are trying to equate YouTube and Facebook video ad views. The interfaces and consumption styles are so different.

On YouTube, you find video and play it on a page basically devoted to that video. Facebook's newsfeed puts everything on a sort of user-controlled assembly line. Users scroll up and down, only occasionally pausing to actively engage with a piece of content or maybe an ad. Much of what they see in the newsfeed hits their retinas for a second or two. Even if a video autoplays, they might scroll right by it. This is why so many Facebook videos feature text overlays — you can ingest a sentence much more quickly that you can understand what's going on in a moving image.

I honestly don't think it's easy to know when people truly engage with video and if engagement leads to brand recognition or even action (buying a product or service), but you would think that measuring average video views would be a solved problem.

It's clearly not. In fact, there is no system or any kind of agreement about how to measure this stuff, which basically makes Facebook the poster boy for this entire mess.

Help me understand

It bothered me that I sort of sympathized with Facebook. If no one can agree on how to measure video ad views, how can Facebook be blamed for trying to do it their own way? Another part of me, though, knew that Facebook chose a measurement system that favored their platform and its performance. Once they do factor in all those Facebook users seeing a video ad for 3 seconds or less, the Facebook video add engagement numbers are going to drop, by a lot.

I decided to turn to an expert to help me work through some of my conflicting concerns and feelings.

In my inbox was an email from Brian Shin. He's the founder and CEO of Visible Measures, a social video measurement and activation company.  Visible Measures recently put out a report on engaged viewing times that shows, according to the company, that actual consumer attention is more important than view counts. 

Shin, who's been in the industry for 15 years and has a business degree from MIT's Sloan School of Management, seemed to have the answers. I wanted to clear up my own confusion and see if my feelings of sympathy for Facebook were ill-founded.

I was curious if you thought Facebook was trying to be fair by not counting any views below 3 seconds. After all, what can you glean in that period of time? Even on auto-play, you might see a second of a video ad.

Shin: Video viewership is like a going down a funnel. At the top of the funnel, you have viewers scrolling through their feed and the video player loads. On Facebook, before 3 seconds, those are considered 'ad impressions.' Progressing down the 'funnel' — if viewers stay long enough — that's when the video view event is triggered. So much money now is being spent on buying video views (which is an improvement over buying just 'ad impressions' or buying based on CPMs - cost per thousand impressions), but there is no standardization on what constitutes a video view. Hence the need to know your Engaged Viewing Time across platforms. 

How might Facebook have better measured engagement? Should they have factored in every potential view — anyone the ad is served to whether it played or not? What would that have done to their engagement numbers?

Shin: Properties [like Facebook] do not show you the whole funnel — 

video player loads ->

ad impressions ->

video views ->

viewing time ->

video completions

— for lots of reasons including not wanting to enable dissection of their economics or traffic patterns. Facebook should have shown, at a minimum, the total video views and combined it with an objective measure of Engaged Viewing Time. The inaccuracy of 'self-reported' stats is why an objective measure is needed. The lack of standardization in terms of how video views are calculated is why an objective measure of time is needed that goes across YouTube, Facebook and other places selling cost-per-view (and other forms of) video advertising.  

So much money now is being spent on buying video views... but there is no standardization on what constitutes a video view.

Facebook including ad impression data (the stuff before 3 seconds) would not help. Marketers are paying for views (post-3 seconds) and they need to know the "value" of what they bought — demonstrable by viewing time. 

How much of all this is media buyers and advertisers still not understating the nature of social media video engagement — especially video advertising on these platforms? It is fundamentally different than the video on YouTube, after all, which isn't even designed as a scroll. If you view a video on YT, you will likely see a pre-roll ad.

Shin: Much of the growth in video today is being driven by these 'social' experiences and we would include YouTube in that category. If you look at places like nbc.com, espn.com, etc., that's really repurposing and monetizing TV content or 'longer form' video. 

But places like YT, FB, Twitter, etc., are pushing advertising that offers more user choice (option to skip or scroll, or share or repost, comment, etc.) and that is presenting both opportunities and challenges for marketers.

Marketers are embracing this idea of more interactivity and choice in video advertising (hence the now billions of dollars going into cost per view and other types of video advertising). But the constantly evolving landscape requires new measures of consumer attention (vs. just ad impressions). Attention in video would include finding all the views that you are generating across all properties and copies and sharing (what we call True Reach) and the actual time that consumers are spending with your video content (what we call Engaged Viewing Time). Provide total views + viewing time (and you know the spend) — you can better assess the value of your ad campaign.

Here's what I glean from what Shin said. 

As long as companies like Facebook make up their own measurement strategies, there will be a fundamental inequity in video ad measurement on social media platforms and other services that include video ads. 

Standards would help, but so would more transparency from Facebook and others. With hundreds of billions of dollars at stake, someone needs to step in and level the playing field for everyone. 

We all agree that video is the future, so it's time to get on board with a future that's equitable and fair. Facebook's mea culpa is a step in the right direction, but no matter what it might cost them in the short term, an agreement on video ad metrics will probably help them and everyone else make billions more in the future. 

In the end, I realize this was never about 3 seconds of video. It's about math, averages, the measure of engagement and standards. Facebook simply couldn't run away from those realities anymore — certainly not in three seconds or less.