Tasselfoot

Tag: Correlation

Correlation Between Views And Subscribers

by Tass on Mar.18, 2010, under Random Stuff, Site Stuff, Statistics, youtube

This post stems out of my getting 10,000 subscribers while in NYC last week. It made me think about the relationship between total views and total subscribers, especially in comparing conversion rates from a year ago and now. It also just so happens that I’ve kept track of milestone dates of both views and subscribers, so I have some nice data points to work with. All information in this post will be based on a 1000:1 view to subscriber ratio.

Basic analysis: View and Subbie growth remained relatively similar through the 5000 mark, at which point Views started to completely dominate. I could go back into my spreadsheets and pull more data points of small areas between 5000 and 10,000… but that just seems like too much work.

The interesting question to all of this is why. I have a few theories, the combination of all likely forms the real reason.

-The 5000 mark is hit in roughly July-Aug 2009, which is around when I started to get a lot of in-game features. The in-game features increases views, but the views are not on YouTube, which almost certainly is the dominant factor as to why views increase dramatically from this point onward and subscribers lag way behind.
-There are only so many people interested in flash game walkthroughs. After those 5000+ people subscribed, it took much longer to get more because they just aren’t as interested in the content.
-Aug or Sept 09 is around when I stopped doing videos for random games, focusing only on games that I can monetize (the vast majority of which are in-game). This has resulted in less videos being released overall, so less exposure.

So what this tells me is that when the views are coming from YouTube, there is about a .1% conversion rate of view to subscriber… but when the views are from embeds on other sites, the conversion rate is closer to .02% (about 1 in 5000). This .02% is more from looking at data from the past month or two specifically, not from the difference in 10,000 values in the chart above.

Leave a Comment :, , , , , , , , more...