The loneliness of the long tail: subscription growth in techblogs, February 2005

Last month's analysis of subscription growth in techblogs focused only on a small sample of the blogs I read. This time around I'm looking at all of the 270-odd blogs I read, and I'm seeing something quite peculiar. Here's the big picture for the whole set:

This suggests that:

• The decline in acceleration of growth (not in growth itself), which we saw from September through December, affected my whole population of feeds.
• The reversal of that trend, which we saw last month, was also true for my whole set.
• In February, acceleration of growth tapered off again (albeit only slightly, considering the short month).

But if you look at just the top 10 feeds -- that is, feeds in my set with most total bloglines subscribers, which in this sample also happen to be feeds with more than 1000 subscribers -- here's what you see:

Conversely if you look at just those feeds in my set with fewer than 100 subscribers, here's what you see:

Just eyeballing the first chart, it looks like the over-1000 crowd and the under-100 crowd trended together until recently, but parted ways in February. I'll be curious to see if students of the power law and the long tail can confirm, refute, and in either case explain this finding.

Update: This is exactly what I've been wondering about too:

How much of this is due to the fact the "long tail" bloggers stopped blogging and hence people started unsubscribing? I think a lot. Blogging caught on, everyone started doing it, then it got old and now only some people do it. [netzooid]

Incidentally, I meant to provide the raw data in case anyone wants to help us visualize it differently. So, here it is. Request: tag your findings with blogviz.

Former URL: http://weblog.infoworld.com/udell/2005/03/02.html#a1188