Today we are approaching the answer to this question with an analysis of the demography of commenters on YouTube. For this, we have randomly picked 2,120 videos from YouTube, for which we found 36,459 comments. That’s an average of roughly 17 comments per video, but note that this number doesn’t mean much because the random 2,120 videos include ones that have no comments as well as ones with several thousands of comments. Since YouTube doesn’t actually expose more than 1,000 comments per video, we had to limit the analysis of the comments per video to 1,000.
Out of the 36,459 randomly picked comments that were analysed 21,464 were by male commenters, 9,914 by female commenters. The remaining 5,081 commenters did not expose their gender.
Even if all the commenters that did not expose their gender were female (which is unlikely) – the clear majority of commenters are male.
Now let’s look at the age distribution. Out of the 36,459 randomly picked comments 26,668 provided their age. The distribution of ages is given in the next graph.
The graph provides the exact age distribution as given by the commenters. Assuming they have all been truthful, we arrive at an average age of commenters of 27.59 years, i.e. the majority of commenters (namely 50.4%) are below 28 years of age.
Details of the distribution are found in this table:
Now, you will have noticed that out of the 36,459 randomly picked comments that were analysed, interestingly 326 declared to be over the age of 100 and the graph clearly spike for over 100 year-olds. We cannot easily make decisions about whether people have provided a misleading age or not, but for those over 100 it is a fair estimate to say that are incorrect. Since they create an unfair bias in the statistics, we also provide the analysis with ages above 100 removed.
Now the average age of comment authors has come down to 26.6 and 52.93% of commenters are less than 27, 75% less than 36 years old.
The largest number of comments has been posted by 22 year old males (6.5%). As our initial gut feeling was the most commenters on YouTube are males between 15 and 25, this analysis has confirmed the gender and roughly the age group: the majority of commenters are between 13 and 27 years old.
We have no explanation for the weird shape of the graph which has a strong dip at 18 years and an unexpected peak at 29 years. This may well just be a problem with the small data set that we used, or there may be some fact to explain this. If you have any theories for these values, please leave a comment. We intend to undertake a broader analysis over more videos and comments in the future and may even be able to test your theories.
]]>The complete stats are listed in the YouTube blog post about the YouTube Generation.
One important outcome is that people consider brands that advertise on YouTube to be more current, innovative and dynamic. And 3 in 5 users say YouTube influences their purchase decisions.
]]>A new social video ad was uploaded on 25th March by Mini. It is about a couple of guys sitting in a car on the German Autobahn and watching two Minis do some crazy moves. At some point the moves become suspiciously unreal and the guys start commenting that they obviously got drawn into a viral video ad. The end consists of typical TV ad titles and the Mini logo.
This video absolutely hits a nerve. The stats that we have collected over the last 3 days are just amazing:
It hit 30,000 views within 3 days and continues to grow. By today it has taken 9 honours on YouTube.
What is it’s secret? Maybe it’s the honesty of the ad. It starts out like a dark viral, but at the minute that it is obviously not real any more, the comments make it funny and the effects are quite cool. There is no shame in confirming it as an ad in the end.
]]>According to Pew Internet:
The desire to share a viewing experience with others has already been a powerful force in seeding the online video market. Fully 57% of online video viewers share links to the videos they find online with others. Young adults are the most “contagious carriers” in the viral spread of online video. Two in three (67%) video viewers ages 18-29 send others links to videos they find online, compared with just half of video viewers ages 30 and older.
Video viewers who actively exploit the participatory features of online video – such as rating content, posting feedback or uploading video – makeup the motivated minority of the online video audience. Again, young adults are the most active participants in this realm.
Here is where it gets interesting: What really drives the virality of videos? Is it the activities of the “motivated minority” who provide ratings, post feedback and provide the content via uploads or is the amount of feedback and the ratings an indication of the potential for a video to “go viral”?
We are still at an embryonic stage in building credible and useful metrics about social video.
One of the reasons for this is that there are a number of different classes of user generated content (UGC) that you find on sites like YouTube.
While there may be a certain amount of totally user created content that is created from the ground up by the uploader, I believe that an enormous amount of the content is really User Mediated Content (UMG). That is: content that is captured from traditional broadcast media and edited and uploaded or vision that is edited to an audio soundtrack like a popular song.
In both these latter cases we find that there are often many instances of the same or similar material being posted multiple times by multiple people.
So being able to establish some useful data on even the number of raw views of a video regardless of establishing the number of positive or negative comments and the ratings can be a major challenge.
Whether you are a consumer like me doing research on video generally or a brand manager with a specific goal, it is relatively easy to go to YouTube (or others) and run a search. But the amount of data that you glean just makes you realize that you are facing the labours of Hercules to get the information that you need.
There are some tools that are starting to emerge that are geared to the class of creator that produces and posts truly user generated content. I would class TubeMogul as one of the more useful in this category. It gives you the ability to get data on the amount of times that a video has been viewed, over what period and what the volume of comments/ratings etc have been.
I tried using it a few weeks ago to monitor Toyota videos for their Kluger model – purely as an academic exercise.
Over the last week TubeMogul has been very capably measuring the activity of the videos that I selected, and at one level it is quite interesting: I now can tell that of the videos that I selected several are getting in the region of 20 views a day on some days.
If I was the producer of these videos, this sort of information would be very interesting and quite possibly very useful, particularly the feature that allows the overlaying of multiple data sets for videos from the same producer.
But what I think that is revealed by what TubeMogul does, is that the tools for metrics reporting in the online social video space need to be very specifically targeted at a user group. And there are a lot of groups with quite divergent needs.
TubeMogul definitely fulfils the needs of the true social video creator.
No one currently provides a set of tools that are geared to professional content owners who currently are probably not making media purchasing decisions about video online because they can’t determine what the true value is.
The corporate market needs to be able to get information that delivers insights so that purchase decisions, strategy decisions, tactical decisions can all be made.
That means that high quality information has to be gathered about the kinds of videos that are User Mediated Content rather than User Created Content
Here are some of the things that would certainly interest me as a CEO or a VP of Marketing:
1. How long does it take for a video designed to “go viral” to actually build up a level of activity that will cause a determinable increase in qualified leads or sales?
2. How many instances of “my video” are there on line?
3. Do the comments and the ratings reflect the virality of the video?
4. Do videos made by consumers have more relevance than the TVC’s that migrate to the web or vice versa?
5. Are there videos online that criticize or are likely to build a negative impression of my product or brand by virtue of the content in the video?
These are some of the early questions I would like to be able to answer…. How about you?
]]>