Publication of Chris Anderson's best-selling book, The Long Tail, has built awareness of the explanatory power of power law curves. Recently, Jakob Nielsen suggested applying logarithmic scaling to these curves. That got me searching my "curiosity file" to see if I could put this to work on something interesting.
As a blogger, I've been of course interested in my weblog's readership and, though I don't expect it to have huge influence, I'm always interested when someone quotes something I've posted. I check Technorati every couple of weeks to see who has linked, and I find myself checking more often since meeting its CEO, David Sifry, a few weeks ago. And beyond this, we're all interested getting our things to bubble up when people search for specific terms. Actually, reviewing the search terms that people use to access this weblog is an amazing education.
So, back to "fun with power law curves." Sunday afternoon, while going through my blogroll of about 220 feeds, I decided to do an Excel spreadsheet with a column for the number of Bloglines subscribers (Bloglines has the largest share of any feed reader), and another column for its Technorati rank (which is based on the number of sites which link to it). I stopped after 42 feeds because a scattergram pattern was beginning to take shape, one that seemed unlikely to change if I did 100 feeds. To see the pattern, I used log-log scaling and added a best-fit line. The result is below:
The data points below the line tend to be quoted in more other blogs than their number of Bloglines subscribers would predict. Those above the line tend to be quoted in fewer blogs than their Bloglines subscriber number would predict. You might say that the ones below the line get more respect; those above the line, less respect than the best-fit line would predict. The farther the dot is orthogonally to the line, the more exceptional it is.
On Sunday, this weblog had 84 Bloglines subs and its Technorati rank was 68,984 (woo-hoo!). That put it just a bit above the line. Not surprisingly to me, Ars Technica, The Doc Searls Weblog, Jeff Jarvis' BuzzMachine, Om Malik's GigaOM, and O'Reilly Radar all were exceptionally well linked-to among larger readership blogs. [itvt] Bloggit and AlwaysOn Feed were that way among smaller readership blogs. I won't name those linked from disproportionately fewer blogs except one. I subscribe to three of Doc Searls' blogs. One mentioned above is significantly below the line, but another of his, Doc Searls' IT Garage, is moderately above the line. The latter is an exception, but blogs like this one that do a significant amount of linking to other sites tend to be quoted less than those blogs that are predominantly original writing. But even that's not a rule. I'm confounded by one blog that has a lot of very thoughtful original writing and I link to it a lot, yet it tends to have fewer linking sites than its subscribership would indicate.
Although it has the largest share of feed readership (about 25%), Bloglines subs are only part of the readership of any blog. FeedBurner estimates this weblog has 350-400 feed subscribers depending on the day. There are 225 or so visitors per day who come directly to the weblog (about 70% of those through search). Another 74 people subscribe to the blog via FeedBlitz email. I'd be cautious about adding those together because of some who use the feed or email and visit the site also, but the total is probably 500-600, depending on the day.
--Dennis