Touchpoints: Internet Viewing and Measurement: A Brief Story - Stewart Schley - MediaBizBloggers - Stewart Schley - MediaBizBloggers.com

Published: March 17, 2010 at 05:35 PM GMT
Last Updated: March 17, 2010 at 05:35 PM GMT

If over-the-top Internet video is going to destroy the pay TV business, it had better start working on its endurance.

Data compiled by comScore and Nielsen, along with scattered reports from online video providers, suggest online video, even in the age of Hulu and Fancast, remains a medium of brevity. Among top Internet video properties, only Netflix, which streams full-length movies and commercial-free TV shows, surpasses an average viewing session time of 30 minutes, according to comScore data. Viewing durations associated with Hulu, which delivers thousands of full-length TV shows online, are no more than six minutes on average.

And across the board, the average length of an online video viewed in the U.S. during January was barely more than four minutes, according to comScore's latest Video Metrix report.

Those aren't exactly the sort of sky-is-falling numbers that indicate online video is poised to supplant cable and satellite-delivered video services as a primary conduit's to feed America's four-hours-a-day TV habit. Instead, they suggest that despite the arrival of high-quality, long-form, episodic television over the Internet, online video remains largely the province of the fleeting – an environment where music videos, news clips and mash-ups substantially overshadow consumption of full-length programs.

A look inside the numbers suggests a more complicated online video viewing picture than a cursory glance at viewing session durations suggest. While it's true that a great deal of Internet viewing reflects the selection of brief video clips, the data on average viewing times mask some underlying methodological influences that skew the numbers.

A big contributor is the definition of exactly what constitutes a "video" online. To you and me, TNT's The Closer, available from TNT.com, is just that – a video (or as we called it in the analog era, a TV show). But if you watch a full episode online, you're actually watching four or five separate video streams, each lasting nine or 10 minutes and stitched together to constitute the full program. ComScore and Nielsen, dutifully, will record and report the viewing of those individual video segments as unique assets – and what seemed to be a 45-minute viewing session will instead, from a reporting standpoint, be chopped into bit parts.

Not only that, but it's common for advertisements to be counted as separate viewing instances, further obscuring the averages. On Hulu, full-length shows are run (and recorded) as single videos, but individual commercials constitute unique viewing sessions, explains Tania Yuki, director of comScore's Video Metrix service and a savant of the new Internet video research sector. That means if you watched last week's episode of ABC's Modern Family on Hulu (clocking in at 21 minutes and 36 seconds), you actually accounted for six videos viewed – one program stream plus five 30-second commercials. Your average viewing time per "video" wasn't 21 minutes, but fewer than four minutes (21:36 divided by six viewing instances).

To be sure, compartmentalizing video sessions by unique asset isn't the only reason Internet viewing appears to be stuck in a short-attention-span zone. A profound American appetite for amateur video clips, news shorts and music videos also drags down the average length of online video sessions. Google's clip-heavy YouTube, which accounted for 39.5% of videos served in January, is a major influence on viewing session times, far outstripping No. 2 provider Hulu, which accounted for less than 3% of videos, according to comScore.

Even so, we believe reported viewing times will lengthen going forward as a function of these factors:

Video availability: Rising awareness and usage of long-form video sites will push the numbers higher as more people select more streams from the likes of Hulu, Netflix, Sony's crackle.com and Comcast's Fancast. In January, Hulu had 38 million unique viewers, a 58% rise in 12 months.

Living-room integration: A growing population of Blu-ray players, Internet-ready TV sets and bridging devices that bring online video to larger viewing screens should encourage long-form program selection and contribute to longer stream duration averages.

Methodology improvements: comScore is preparing to deliver more granular measurement of online video viewing, accounting for distinctions between commercials and programs, among other changes. That should help paint a truer picture of viewing habits.

A final factor to consider is how Internet video consumption stacks up against traditional television. Although Nielsen reports a vast gulf in time devoted to the two mediums – 31 hours per week for traditional TV versus just 22 minutes for online video – Internet video ought not to be penalized for delivering abbreviated viewing session durations. The way we see it, the mere act of changing a TV channel or surfing multiple channels with remote control in hand constitutes a rapid-fire succession of separate viewing "sessions" by source. Not to mention that it's annoying to whoever's watching with you.

Stewart Schley is Senior Director of Industry Intelligence for One Touch Intelligence, a leading provider of competitive intelligence, business analytics and market assessment services for leading providers in the cable telecommunications, TV content and filmedentertainment industries. He can be reached at stewarts@onetouchintelligence.com.

To communicate with or to be contacted by the executives and/or companies mentioned in this column, link to JackMyers Connection Hotline.

My understanding is that the ComScore VideoMetrix and Nielsen VideoCensus video view definitions (which stipulate that a new video view must be counted after each ad break on a full episode) were largely based on feedback from the agencies, who wanted a proxy measure to show how much ad inventory is available on a given video player ( I make the distinction between player and site here because the above measurement products measure activity on a video player regardless of what domain it's embedded on).

Thus the "time spent viewing" data reported from ComScore and Nielsen for any video player with midroll ad inserts, Hulu being the most obvious example, has very little to do with how much time the player's average viewer spent watching video content.

Unfortunately this means that the data reported is subject to constant misinterpretation by bloggers and the media. Nice to see that you've reached out to these services to understand how they're defining what they're measuring.

Posted at 04:43 PM on Mar 17, 2010 by David

I make the distinction between player and site here because the above measurement products measure activity on a video player regardless of what domain it's embedded on. Agreed with statement. Anyhow nice post thanks and keep on sharing as well.

Posted at 01:20 AM on Jul 21, 2010 by Aria

i m glad i found ur blog.Not everyone can provide information with proper flow. Good post. I am going to save the URL and will definitely visit again. Keep it up.<

Posted at 04:20 AM on Jul 21, 2010 by james kerry

Posted at 03:29 AM on Jul 27, 2010 by lbm0914