Raw

David Sifry has posted some more State of the Blogosphere statistics. Great information, but I do think some of the conclusions are flawed. I posted in comments about the written language distribution figures (not all English-language bloggers live in Califormia!). But I just read Kevin Burton's take on the big number - 50 million blogs - and while I agree with some of Kevin's criticism (there almost certainly are not 50 million blogs), there are points where I differ.

For a start I'd go for a different definition for "active blog" than Kevin's one post a day. I follow several of the "Planet" aggregated sites, and the posting frequency pattern on those probably has a longtail shape. I'd guess significantly less than half the bloggers post once a day. Some folks only post once a month or less (must be a month since I last saw Tim Berners-Lee on Planet RDF, and he's far from the most infrequent blogger). Dunno, I'd probably define active as one post a month, dead something like no posts for 3 consecutive months.

One of Kevin's adjustments of David's findings is that the number of active blogs is on a linear growth scale not an exponential scale. I suspect this is probably closer to the reality and possibly explicable as follows -



Say there are two categories of folks starting blogs: b ₁ folks that post a handful of times then lose interest, and and b ₂ folks that will blog continuously once started.

Very probably the pattern of growth in N(b ₁) is exponential, because the "ooh, let me try this new app" activity for a particular application depends on the spread of acquaintance with the app, when the app is sufficiently compelling that one new person discovering leads to >1 subsequently, you've got a chain reaction.



The pattern of growth of N(b ₂) is trickier to estimate. Ok, it will depend on exposure to the blogging application (which is N(b ₁)), but will also depend on other factors. I would guess that one could be modelled as the "born to blog" factor, a predisposition to regular posting. If there is such a thing I bet it's far more to do with nurture than nature, but for discussions sake call it the blogging gene. It's probably reasonable to assume a fixed proportion of the population at large will have this gene. Then as exposure and N(b ₁) grows, the growth of N(b ₂) should be exponential too.



But this analysis depends on the assumption that there's a discrete step between blogger/non-blogger. But blogs appear to go more with a whimper. Kevin picked up on a pertinent point:

I just took a rough look at [Technorati's] data and on June 2005 their mean post/weblog ratio was 0.0833. In June 2006 it was 0.04 which means its been cut by more than 50%.

I expect their mean post/weblog ratio will continue to fall.

So what if decline in posting approximates an exponential too? Perhaps this is, in part at least, cancelling out the growth of N(b ₁), leading to the linear growth in the real figures Kevin suggested.

I doubt it's worth going any further into the maths of this without taking into account the finite population size (assuming everyone has a one-blog quota). I would guess it'll still be another few years until the net effectively reaches saturation, and the rise in the number of bloggers locks step with the number of new Web users. By which time of course approaches to personal publishing on the Web may well have changed radically from today's notion of a blog.



Blow up

Such changes are inevitable, the only question is the timescale. With Internet time that's ever shorter. The last fundamental change in content publishing was probably the shift from static home page to episodic blog. Personally I reckon the next big shift will go far deeper, with the idea of "content" seeming increasingly anacronistic, an artifact of old media (ink and paper, 35mm film, videotape, telephone lines etc) rather than anything inherent to communications, data, computers or for that matter people. From an information viewpoint, blogging is a niche application - human readable text/markup content is the exception, not the rule.



Currently there is something of a gap between the blogosphere and data-oriented applications (e.g. social networks, geo/photos, calendering etc). But there's a lot of potential for merging the two. If, when creating a blog, the new user got a lot more immediate benefit, like Planet feeds based on their personal profile, they could be a lot more engaged with the Web (think chmod 777 web). Boost the feedback. Of course there are also microformats, Atom and all the Semantic Web activity which will no doubt have impact on this space. 

I rather like Micah Dubinko's new coinage, as far as the network is concerned I think we might just heading towards a datacratic society.

Rats, that reminds me, right now I'm meant to be writing for a print-publisher, not the blogosphere....

@en