Thinking Out Loud About YouTube Metrics
BY Micah L. Sifry | Thursday, August 16 2007
Todd Ziegler of the Bivings Group, who I greatly respect, recently dinged us (gently) with a post called "The Danger of Using YouTube Views as a Metric." He argued that our daily tracking of how many times each candidate's videos are viewed on YouTube was "deceptive" due to the "vastly different ways candidates are using YouTube." And he has a point. John Edwards' numbers are somewhat higher than the other leading Democratic candidates because his campaign is using YouTube as the player for videos on his own site, while Obama uses Brightcove and Clinton uses an in-house tool.
But we've never said anywhere that counting views on YouTube is somehow equal to counting every video view a candidate receives on every platform possible. (Though we can clearly start picking up the numbers of viewers the candidates may be getting on other widely used channels like Brightcove, and will report on that soon.) What we have said is since online video is a pull, not a push, medium, online video stats are clearly a sign of interest in a candidate, and probably a much stronger sign of moving an impactful message than, say, TV advertising points.
I also think it still makes sense to keep track of the viewership numbers that YouTube makes available for each candidate's videos, even if this means we are inevitably undercounting, in some cases, at least, total viewership. Like it or not, YouTube is by far the leading hub for online video sharing and viewing, so if a candidate is doing well there, that's an important metric.
Speaking of which, I'd love some input on the following question: Is there a relationship between YouTube views and YouTube subscriptions? People who subscribe to a channel on YouTube get alerted by email whenever the publisher uploads a new video, so this is actually an important number to follow. If you look at our old YouTube charts (ignore the total channel view number, which we don't trust), and instead focus on the total subscribers, you'll discover that Ron Paul is way ahead of the other candidates with more than 25,500 subscribers. Barack Obama, by comparison, has just over 10,000 subscribers to his YouTube channel. John Edwards, who has the third highest total number of YouTube views (2.1 million, compared to Paul's 3.2 million and Obama's 2.9 million) has just 3,700 subscribers. Clinton, who has 1.9 million total YouTube views, has nearly 5,300 subscribers. Clearly, a lot of people are interested in Ron Paul, and as the mainstream media isn't satisfying their thirst for information for him, they're soaking up information they can get directly off the web. But is there a relationship between # of subscribers and viewership?
What are we to make of the fact that for every one of Paul's subscribers, he's gotten 125 views on average, while Obams's gotten 284 views, Clinton has gotten 365, and Edwards has gotten 568 views? I'm not saying that each YouTube subscriber watches X number of views; rather that because each time a video is published on YouTube, subscribers to that channel get an email alert and you undoubtedly see a ripple effect. But I'm not sure what the ratio means, if anything. Any ideas?