Donald Trump not only leads the polls, but so far, his youtube channel is also the most watched.
|By Average number of views||as of Friday 29th of January 2016|
*without the record 1.7m viewers from his announcement.
Conversely, the data is grim for Jeb Bush. While his channel gets twice as many views on average than Cruz, relatively few people have indicated an interest in keeping in touch.
|By Number of Subscribers||as of Friday 29th of January 2016|
This is probably the most worrying statistic. While averages views may reflect the interest of both supporters, detractors, or mere curiosity, subscribing to a channel is more likely to indicate a clear preference. Detractors and others, may be less likely to make this commitment and might not want their subscription to be mistaken for an endorsement.
Analysing Youtube data, also gives a glimpse on how the candidates’ teams use this particular platform.
Cruz for instance, publishes several videos a day. Indeed, Cruz’ team published more videos in the past four days than the 19 the Trump’s team published in nine months, when his first official video was published. This relative abundance may, at least in part, explain why Cruz’ number are so paradoxical; because there are so many videos published in a given day, there is less anticipation.
This GOP race has also been characterised by controversy. Unlike twitter or facebook, youtube offers a way to evaluate how controversial a message is: this is the likes/dislikes ratio.
According to this ratio, if the GOP were looking for the least controversial candidate then Rubio is the man.
|Least controversial (more likes than dislikes as a ratio)||
While Trump has just about more likes than dislikes (1.15 likes for every dislike) Jeb Bush’s message simply do not resonate with his viewers. Strikingly and probably very disappointingly for the Bush’ family, Barbara Bush’s video was disliked almost five more times than it was liked.
|Views 68,047||Likes 116||Dislikes 524|
If the campaign team were hoping for a boost by bringing a big gun, this particular shot misfired.
While the statistics above were compiled from publicly available data, a campaign manager has access to a crucial set of data: the average view duration. Just like Amazon is able to know how fast we read a book on a Kindle and at what point, if any, we quit reading a particular book, youtube offers publishers the ability to know the “Estimated average minutes watched per view for the selected content, date range, region and other filters.”
To be more specific and useful, are there images or words during a given video that make viewership drop? If so, why and should anything be done about it? One way to go about optimising messages and images, is to follow the Buzzfeed A/B testing approach. This, would require publishing near identical videos and select the one that makes people watch till the end