nScreenMedia OTT multiscreen media analysis

Comscore and Nielsen Head-2-Head: And the winner is…

These are confusing times in the delivery of video entertainment. The influence of the Internet is not only impacting the viewer, it is also changing the business of video delivery. However, tracking these changes is a major headache for all concerned. Last week two companies provided two very different views of how these changes are unfolding. One claims OTT viewing is up, the other down. One says the amount of time spent viewing is 3 times the other.

So, who is to be believed?

Nielsen reports in its latest three screen report that those that watch OTT video increased the amount they watch by 23%. According to the company December 2011 OTT video viewers were watching 4 hours and 26 minutes a month. This increased to 5 hours and 51 minutes in December 2012. The biggest OTT video consumers are, unsurprisingly, the 18 to 24 year olds who watch almost double the average, at 10 hours and 41 minutes. The age groups watching the least are the 2 to 11 year olds and the over 65 set.

One thing to note about these numbers is how low they are. According to Nielsen, the average online video viewer is watching just 11 minutes a day. According to data released by Netflix last year its 25 million customers are watching video on the service, on average, for almost an hour a day. If we average this consumption out across the number of online video viewers (162 million according to Nielsen) Netflix alone accounts for 8 minutes of daily viewing. Put another way, non-Netflix online video viewing accounts for just 2 minutes a day for the average video viewer. Barely enough for one YouTube cat video!

Comscore has a different view of how much online video viewers are watching. The company reported last week that the average OTT video viewer was watching 19 hours and 10 minutes of online video in December 2012, over 3 times as much as Nielsen. Interestingly, this is a 17% decline from December 2011. Comscore data seems to show that the growth in the number of people watching has stalled as well. In December 2011 there were 181.7 million online video viewers and the same number in December 2012.

So, which of these companies is right?

Firstly let’s deal with whether the amount of online video being consumed is increasing or decreasing. In August, Comscore said they were changing their methodology slightly and cautioned about comparing previous data with new data. In other words, comparing December 2012 with December 2011 is not necessarily an apples-to-apples comparison. If we can’t rely on a year-over-year comparison from Comscore should we believe the Nielsen increase of 23%? There again we are stymied. Lurking at the bottom of the recent 3 screen report, the company mentions that:

“As of June 2012, YouTube Partner reporting became available through Nielsen VideoCensus featuring May 2012 data.”


Yikes! With YouTube being the premiere video streaming site online this sounds like a huge error and could account for the entire increase reported in December 2012 over 2011.

The only thing to conclude from all of this is that both Comscore and Nielsen have given us excellent reasons not to believe either company’s implied growth numbers. Let’s assess if there’s a good reason to believe their data for the amount of video consumed.

There is a large disparity between Nielsen and Comscore in the quantity of video consumed. Could the reason for this be due to a difference in methodology? Both companies employ two methods to measure online video viewing: opt-in panels of people reporting their usage and the tagging of video assets. Video tagging is relatively new and requires video providers to include a tag in each of their videos. When a tagged video is played the viewing is reported back to the measurement company (in this case Nielsen or Comscore.) If this technique works as advertised it should give a very accurate measurement of video consumption. But it is early days for video tagging and the majority of video delivered does not have a tag. That means the less accurate panel approach is still dominant at both companies. So, at least on the face of it, there doesn’t appear to be a big difference in the methodologies employed between the two (although I’m sure both companies will strongly disagree with this statement!)

Since they both have similar processes for counting are there other differences? Comscore claims to count both streaming and progressively downloaded video while Nielsen appears to only count streaming. That certainly might account for a somewhat higher number for Comscore over Nielsen. But I doubt it covers the large difference they are reporting.

Perhaps the best way to assess which consumption number is closest to the truth is to look to other reported values in the market. I have already mentioned the Netflix streaming number which seems to indicate higher consumption than reported by Nielsen. But there are plenty of other data available from companies such as CiscoAkamai and TDG Research. For example, Burst Media reported last year that 39% of web users watch between 1 and 5 hours a week of online video.  Taking this additional data into account leads me to believe that the Comscore number is closer to the truth.

However, by now it is probably clear that there is a great deal of uncertainty around online video measurement. Unfortunately, all of this uncertainty is exerting a drag on the business of online video delivery. In this environment, something as mundane as selling an ad becomes difficult to say the least.

So as you review the latest online streaming data keep in mind that it is far from an exact science. You should treat it with the healthy dose of doubt that it deserves.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.