Report: YouTube, Yahoo Relax Standards to Inflate Views
In the last year, both YouTube and Yahoo, two of the three biggest online video sites, have become more lenient about how they measure video views, according to TubeMogul, a video analytics startup.
Last June, we asked “Hey Guys, What Constitutes a View?” after Emeryville, Calif.-based TubeMogul released the first edition of this report. We were surprised to find out how much counting methodology differed between sites. This time around, we’re surprised to find that counting methodologies seem to have changed. It appears that YouTube and Yahoo now register a view every single time a visitor clicks play.
Last year TubeMogul tested eight sites by uploading a video, making it private when possible, and viewing it in different situations. The results showed Yahoo was the most stringent about how it counted views and YouTube the third-most stringent. At the time, Yahoo did not count more than one view per computer, and it did not count videos that were viewed as embeds on other sites. YouTube would only count multiple views from a single IP address if they went all the way through to the end, and would only count embeds once per IP address.
This year, TubeMogul tested a total of 14 different online video hosts and found YouTube and Yahoo at the other end of the list. Each of them ticks off a video view every single time a video is started, no matter if it’s halfway through, if the page is refreshed, or if the video’s played from an outside embed.
TubeMogul suggested YouTube and Yahoo tweaking their formulas could be influencing “the ever-increasing numbers for videos viewed online,” which wouldn’t make sense if we’re talking external measurements from comScore and Nielsen, but certainly makes sense on a video-by-video basis.
When reached for comment, a YouTube spokesperson told NewTeeVee, “We don’t get into specifics on what’s considered a view; we’re less focused on counting what makes a view and more focused on providing the best possible viewing experience.”
 |
So who are the sticklers left in the game? blip.tv and Metacafe, which only register one view from an IP address. We should note that both of these sites share revenues with their users based on number of views. blip earned the title of most stringent of all by not counting views from embeds (which seems odd, considering they don’t emphasize their site as a destination).Update: TubeMogul retested blip and found it does count embedded videos. See the comments below.
TubeMogul’s conclusion is almost word-for-word the same as last year:
“[T]he lack of standardization presents complexity to content producers and advertisers in understanding the relative popularity of videos across video sites. To fully realize the potential of advertising models in the online video medium, increased standardization and transparency is required.”
True that.
TubeMogul plans to release the full report on Wednesday on the research section of its website.
Sponsor Gallery
Recent
GigaOM Network
- MacBook and Apple’s Black Friday Sale: “Almost,” But Not Quite Enough [TheAppleBlog]
- 10 Ways To Trick Out Your Netbook for Free [GigaOM]
- James Cameron: The World Should Be Shot in 3-D [NewTeeVee]
- Better Opportunities Than Ever in USB Thumb Drives [WebWorkerDaily]
- Hawaii Says Aloha to Better Place [Earth2Tech]
© 2008 The GigaOM Network. Powered by WordPress.com. Marketing consulting by ACS.
Hey,
Just wanted to note that blip doesn’t share revenues with content creators based on the number of views — but rather based on the actual revenue from advertising impressions. The number of views we count really has nothing to do with the revenue share. Rather we count the number of ads that are served and the CPM (cost per thousand) of those ads.
When we traffic ads we share exactly 50% of that revenue with the content creator. So we have absolutely no incentive to make view counts lower — all that matters in terms of money paid to content creators is how much money we can receive for the video views.
Also… we do count embedded video plays :)
Yours,
Mike Hudack
Co-founder & CEO, blip.tv
Mike Hudack on May 19th, 2008 at 9:46 pm - Permalink
Video is a fundamentally different medium to pages. We need to change our language and move beyond measuring “impressions” - since the analogy clearly is insufficient to measure what really matters; engament.
To do this we need not only a new lexicon, but brand new video-specific analytics tools to provide metrics on every facet of the view, from play to pause to % viewed and completion. Until we provide publishers with tools that measure all these metrics and analyze it, then this type of confusion will only increase.
At LiveRail we’re working hard to build tools that provide these types of metrics and analytics to advertisers, so even if publishers are somewhat mystified about how many views their video really got, at least the people writing the checks won’t be!
Mark Trefgarne on May 19th, 2008 at 11:16 pm - Permalink
Video is a fundamentally different medium to pages. We need to change our language and move beyond measuring “impressions” - since the analogy clearly is insufficient to measure what really matters; engagement.
To do this we need not only a new lexicon, but brand new video-specific analytics tools to provide metrics on every facet of the view, from play to pause to % viewed and completion. Until we provide publishers with tools that measure all these metrics and analyze them, then this type of confusion will only increase.
At LiveRail we’re working hard to build tools that provide these types of metrics and analytics to advertisers, so even if publishers are somewhat mystified about how many views their video really got, at least the people writing the checks won’t be!
Mark Trefgarne on May 19th, 2008 at 11:19 pm - Permalink
Hi Liz,
We disagree with the statement that YouTube and Yahoo relaxed standards to inflate views. In our June 2007 study, we concluded that standards were needed, and while greater standardization and transparency are still needed, this research indicates that there has been greater alignment among sites as to what constitutes a view.
Sincerely,
Mark Rotblat
TubeMogul
Mark Rotblat on May 19th, 2008 at 11:55 pm - Permalink
Hi Mark, I don’t think those interpretations are mutually exclusive, and both were expressed in the materials you guys sent to me.
And Mark, do you mind responding to Mike from blip? He says they do count embeds.
Also, Mike, since the ads are based on a CPM — that is, an impression, not an action/click — doesn’t it fundamentally come back to how many people watch the video?
Thanks for the comments, guys.
Liz Gannes on May 20th, 2008 at 7:53 am - Permalink
Hi, Michelle from Metacafe chiming in here. We agree that standardization and transparency is important.
As additional background on how we handle this, a view on Metacafe (whether on our site or through the embeddable player) is based on a viewer watching 50% of a video or the first 30 seconds of it — whichever comes first.
Video is a fundamentally different medium, and we believe it doesn’t make sense — for our content creators, advertisers or our business — to count as a view a video that is initiated but immediately aborted.
Michelle Cox on May 20th, 2008 at 9:29 am - Permalink
Liz’s sage comment and Mike’s insistence led us to re-test whether blip is counting embedded views–and sure enough, they are! The error on our end had to do with the timing of when the view was counted. We are confident, however, in the rest of the results. We tested rigorously, including corroborating earlier results on YouTube, for instance, from another IP address and attempting to see if the site had a maximum (i.e. 100) number of views by IP address.
David Burch on May 20th, 2008 at 11:35 am - Permalink
Liz,
We count video views and ad impressions separately. We have an incentive to count video views accurately for the benefit of our content creators. We also have an incentive to count advertising impressions accurately for the benefit of our advertisers. These are two different metrics which often but do not always coincide.
For example, a video may be viewed one thousand times but have two thousand advertising impressions because it carries both an overlay advertisement and a postroll advertisement. Or perhaps a different video may be viewed one thousand times but have zero advertising impressions because its content is objectionable to advertisers (maybe it’s of a sexual nature).
In each and every case we try to be as accurate and transparent in our calculations as possible. We have no incentive whatsoever to deflate view counts even though we do believe it’s in our interest — and the interest of the industry at large — to be conservative and accurate in our viewership accounting.
Yours,
Mike
Mike Hudack on May 20th, 2008 at 12:18 pm - Permalink
Thanks for the thoughtful response, Mike. You guys always do the right thing for your users.
Liz Gannes on May 20th, 2008 at 2:30 pm - Permalink
From my testing, it appears youtube is doing one view count per IP address.
Aaron on May 20th, 2008 at 2:46 pm - Permalink
Liz:
Your comment about YouTube testing surprised me, so I went back to the drawing board with yet another YouTube scenario, only to have YouTube count all 15 refreshes I executed. We are confident in the original results. In the case of YouTube, we tested four different scenarios. We later did some corroboration from a different IP address, computer, OS and browser.
In the case of “Refresh,” we let the player start before clicking refresh again, so maybe that explains the discrepancy you report. Also, YouTube sometimes can take hours to log views, and other times count them almost immediately.
David Burch on May 21st, 2008 at 10:32 am - Permalink
Sorry. My comment directly above should be addressed to Aaron, not Liz. The name of the commenter appears directly below the comment (duh!) , but I was looking above.
David Burch on May 21st, 2008 at 11:07 am - Permalink
Thanks David, I was wrong, I ended up coming to the same conclusion after finding out the views took a few hours to log.
Aaron on May 21st, 2008 at 12:53 pm - Permalink
[...] I posted a comment in reply to Report: YouTube, Yahoo relax standards to inflate views: [...]
On Stats at the blip.tv blog on May 23rd, 2008 at 8:38 am - Permalink
[...] study was picked up by the Los Angeles Times, Silicon Alley Insider, NewTeeVee and others. Here is a summary of the [...]
TubeMogul Blog » “View” Equivocation Across Video Sites on May 27th, 2008 at 12:48 pm - Permalink
Is there a limit to YouTube views per computer? We are trying to boost the views for a contest and can’t seem to add more counts even by watching the video all the way through.
Doug Pripps on May 27th, 2008 at 1:07 pm - Permalink
[...] study was picked up by the Los Angeles Times, Silicon Alley Insider, NewTeeVee and others. | subscribe var disqus_url = ‘http://howardlindzon.com/?p=3623 ‘; var disqus_title [...]
Howard Lindzon » TubeMogul: What Counts As a "View?" Depends On Who You Ask on May 28th, 2008 at 3:11 pm - Permalink
People locators, general investigations, bug detectors! D&C private investigators can & will investigate anyone. The Director Josh says, “At D&C we provide a great range of specialist services, & fields such as Locating Individuals and Surveillance , are what we do best.” “We do everything from locating run-away teens, deadbeat parents, and even pets, to electronically sweeping an office or residence for eavesdropping devices, or putting one of our officers / investigator teams to work doing surveillance on a cheating spouse. We’re good at what we do. Our success rate in locating people is an amazing 85%. Our private investigators are also well-seasoned in the art of surveillance. If your phone is tapped, or your place is bugged, there’s a 95% chance that we will find it! We also welcome the challenge of unusual cases..”
D&C Private Investigators on June 14th, 2008 at 7:52 pm - Permalink
[...] that this data does undervalue Metcafe’s playcount, as Metacafe is the only site that counts one view per IP address. TubeMogul did not provide an adjustment for this [...]
YouTube Most Effective Vid Distributor « NewTeeVee on July 15th, 2008 at 11:06 am - Permalink