A year ago, Yahoo became the first company to live stream a regular-season NFL game all around the world. The broadcast netted 15.2 million unique viewers worldwide. With most Sunday NFL games in the U.S. averaging 10 million to 20 million viewers, Yahoo seemed to have hit a TV-sized NFL audience.
Except it didn’t. If you were to measure Yahoo’s live stream the same way TV is measured, the viewership was far smaller: an average of nearly 2.4 million viewers across the 195-minute live stream.
This, in a nutshell, is the biggest difference between how viewership is defined on TV versus the web: Whereas TV looks for the average number of viewers across the entire program, the web prioritizes the cumulative number of people who have watched a video.
Here’s why audiences are calculated differently across TV and web and why that matters as digital publishers and platforms devote more resources toward producing and making money off video:
OK, so Yahoo had 15.2 million viewers but also only 2.4 million viewers. Which is it?
TV ratings are measured by Nielsen, which calculates the live viewership of a particular broadcast based on a metric called “average minute audience.” It’s exactly what it sounds like: the average number of people that watched a TV broadcast during any 60-second portion of that broadcast.
So while 15.2 million people watched Yahoo’s live stream, that metric includes someone who might have caught the first few minutes but did not watch the rest of the game, as well as someone who only caught the tail-end. It’s an aggregate where TV is measuring the average.
Why is this important?
TV networks have to guarantee advertisers that a certain number of viewers are going to watch (or have watched) a certain program. With this Nielsen number, advertisers can feel confident that an ad they bought during the middle of a TV broadcast that averaged 3 million viewers was watched by 3 million people. (There are some flaws here that we’ll address below.)
“[Average minute audience] provides a much more accurate picture of how well a show held the audience’s attention, versus people just stumbling upon it,” said Alan Wolk, TV analyst and consultant at Toad Stool Consultants. “TV networks want to tell advertisers that their programs get a high degree of engagement — no matter where the ad runs, people will watch.”
What about a video view, though?
That’s just a measure of how many times a video played. Facebook counts views at three seconds, Snapchat counts views as soon as they start, and YouTube counts them around 30 seconds. One person can watch the same video five different times — that’s still measured as one unique viewer even if it’s counted as five total views.
Live video’s becoming a big deal on Facebook. Does it provide this “average minute” metric?
No. Facebook offers two live video metrics: peak concurrent viewers (which counts the highest number of viewers a live broadcast received) and a visual graph that shows the number of live viewers at any point during the broadcast. Ostensibly, a Facebook live broadcaster could write the numbers down and calculate the average minute audience, but that’s a lot of work.
Do people care that Facebook doesn’t offer this?
Not at the moment. With Facebook counting views at three seconds, it’s much easier for publishers to boast about the total views their streams get. Once Facebook opens up monetization for live videos — especially if mid-roll video ads are involved — expect advertisers to ask for it. Whether Facebook opens up its platform to more third-party measurement is an entirely different question.
So social platforms are opposed to being measured like TV metric because it makes their audiences look smaller?
Maybe. But not all social platforms are against the metric.
Twitter, for instance, is offering the average minute audience for its live NFL broadcasts. A recent Thursday night game between the Patriots and Texans drew an average minute audience of 327,000 viewers on Twitter. On CBS and NFL Network, the average minute audience was 17.5 million viewers.
Oh wow, that’s small.
Yep, but some advertisers say they appreciate that Twitter and the NFL are offering this stat, which makes it easier for them to make an apples-to-apples comparison.
Media Storm is already talking to clients about putting more money in Twitter live streaming. It’s an “area that we’re comfortable taking a risk in,” said the ad-buying firm’s chief digital officer Charlie Fiordalis. “If there is a publisher who can come out there and establish a new conversation in the OTT space, it’s Twitter. I expect them to put up much bigger numbers because they have something that’s actually social TV.”
You hinted earlier that Nielsen’s measurement has some flaws.
Yes. Nielsen’s measurements are based on “diaries” and “people meters.” It’s estimated that Nielsen has meters in about 20,000 households in the U.S., which are used as a representative sample of the entire U.S. TV population (more than a 118.4 million households, according to Nielsen). Nielsen’s TV ratings also don’t account for viewing in bars, hotels and other places outside of the home. TV viewers can also skip ads if they catch up on programs using the DVR.
“The attitude toward Nielsen ratings is simply: ‘We’ve been working off of this for years and we all agree to it, so let’s not mess with it,’” said Wolk.
Images via Netflix and Fotolia