Measuring Views: We Can Count It, But Can We Count ON It?

There are lots of measures available - and those being measured are only too happy to suggest them to ostensible evaluators.

But as the following article reports, the fact that the data are available does not necessarily mean they are reliable. JL


John Herman and Mike Isaac report in the New York Times:

Even within the same company’s online entities, data is hard to reconcile. You have Twitter now with live video, then you have a Periscope product, then you have Vine, and they all integrate into this feed. Then you have Twitter on TV, where you can watch live video from your couch.

It’s really hard to say what is comparable to what.

ONE OF THE great promises of online videos was the potential to measure precisely how effective they were: Audiences reported not through inexact ratings and surveys but directly, down to the person and the millisecond he or she spent watching.

For the most part, that promise has been realized. Digital audiences are nothing if not minutely monitored.

But an abundance of data, largely self-reported by internet platforms that are in direct competition with one another, presents a challenge of its own. Namely, what numbers are truly reliable?

Consider the closest thing the industry has to a basic, universal metric — the video view.

Facebook counts a video as viewed after it has played for just three seconds, as does Twitter. Vine, in which videos are mostly limited to six-second loops, counts a view once a video has been watched to completion. Snapchat, which does not publicly display the number of views, counts users’ posts as viewed after one second.

Instagram, which hosts up to one-minute videos, registers a view after three seconds. Vimeo, which hosts videos that can be much longer, counts a view instantly, as does Twitch, on which broadcasts can last hours. YouTube has never fully disclosed its methods for counting views, but industry insiders have divined that it is somewhere around 30 seconds.

For decades, TV viewership was measured primarily by one company, Nielsen, and expressed as relatively straightforward ratings — a currency upon which 30- and 60-second television ads were bought and sold. Digital video pairs more precise measurement with a vast proliferation of metrics and formats: videos viewed live, or later; on phones, computers or other devices; with obvious intention, after a search, or more incidentally, as in a feed filled with other posts; for seconds, minutes or hours.

Even within the same company’s online entities, data is hard to reconcile. “You have Twitter now with live video, then you have a Periscope product, then you have Vine, and they all integrate into this feed,” said Nick Cicero, founder of Delmondo, a social media marketing and analytics firm. “Then you have Twitter on TV, where you can watch live video from your couch.”

“It’s really hard to say what is comparable to what,” he said.

Those who post videos can afford to look at these shifting definitions of what counts as a view with at least a modicum of humor.

But for the advertising industry, audience measurements translate directly into dollar amounts, and the different methods to measure views — and the fact that the information is all self-reported by the technology companies — are problems. Then came the recent admission by Facebook that for years it had miscalculated its video viewing metric, giving its partners the impression that their videos had been viewed for longer, on average, than they actually had been.

The Association of National Advertisers, a trade group, took the opportunity to scold Facebook, which has rapidly altered the online advertising industry. “With more than $6 billion of marketers’ media being directed to Facebook, we believe that it is time for them — and other such major media players — to be audited and accredited,” the agency’s chief executive, Bob Liodice, said in a statement.

Such demands are not entirely new, and some websites, in an effort to appease and court clients and partners, have already begun to take tentative steps in that direction. In April, Facebook announced a series of deals with Integral Ad Science, comScore and Nielsen to offer third-party verification. It also made a deal with a company called Moat, which seeks to standardize some forms of measurement across websites that host video. YouTube, Twitter and Snapchat have made similar deals in recent years.

“We think it is really important to have a standard set of metrics like audience reach, total minutes, average minutes viewed,” said Jessica Hogue, senior vice president for product leadership at Nielsen. “The market doesn’t want more metrics, just a refined set applied in a consistent manner regardless of the content type or platform.”

Though it has recently been criticized for failing to keep up with TV’s digital revolution, Nielsen has for decades offered familiar ratings and is hoping to find its place — by creating simple, Nielsen-like ratings across different services — despite competition. The use of familiar measurement methodology may help persuade advertisers to move more of their spending from TV, where the bulk of it has traditionally gone, to digital.

Facebook’s stumble may represent the beginning of a bigger conflict between sites that host video and advertisers.

Consistent measurement, which could allow views on one site to be compared directly with views on another, could streamline and accelerate advertiser spending. (Reverse-engineering each company’s counting methods has actually become something of a sport among marketing professionals.)

“Every other medium has a currency,” said Jonah Goodhart, the founder of Moat. “You can measure time in the same way across platforms even if the impact of time is different on those platforms.”

But such standardization could also diminish what platforms regard as their own unique advantages that cannot always be captured in raw numbers — differences in the ways users use particular apps, and consume video, that could command a premium.

It is in a platform’s interest, in other words, to be regarded as a different medium than its competitors, measured not just in views or time spent watching.

“What you’re going to have is a whole bunch of ad sales people trying to make the case as to why their thing is better,” said Tony Haile, the founder and former chief of Chartbeat, an analytics company. “But it’s going to be very difficult to try to match the experiential differences on the platforms to the media buyer’s desire for a simple message about performance.”

A compromise, he suggested, could be a clearer distinction between public metrics — view counts displayed to influence users’ behavior or provide positive feedback to those who post — and the numbers the companies use to sell advertising.

Already on Facebook, advertisers are able to buy video impressions, which are counted as soon as someone sees a video, or 10-second views, which are more expensive and counted only if the user makes it to 10 seconds. But the three-second view, the one behind the totals the public sees, is not for sale.

“A video view which isn’t being transacted on, but which is there to make someone feel good about how many views their video received, is useful for a totally different purpose,” Mr. Haile said. “They almost should diverge.”

Mr. Goodhart, of Moat, suggested a shift in attitudes over the last year. “No individual media owner can be their own referee anymore,” he said. “You really can’t be against that.”

At a news conference during the recently concluded Advertising Week in Manhattan, Facebook executives seemed to concede as much, noting plans to focus more on metrics that they think advertisers should care about, like sales outcomes, though they did not offer a specific timeline for when that would happen.

“There is very universal agreement that the industry needs to evolve to metrics that actually matter,” said Carolyn Everson, vice president for global marketing solutions at Facebook.