Monday, May 21, 2007

Ways of ranking

In comments on the previous post, AC Douglas and Roger Bourland recalculate the list using Google links. I have done the same using the unpublished list of 116 blogs I have collected. Here are the blogs that didn't make the top 53 using Technorati, but would be in the top using Google (shoehorn them into Roger's list) (and some numbers are probably different, since I checked the statistics today):
#20: Vilaine Fille 854 (Crit)
#33: Daily Observations, 465 Charles Noble (viola)
#33: Fredösphere, 465 Fred Himebaugh (C)
#37: Café Aman, 421 Anastasia Tsioulcas (Crit)
#42: Saint Louis Symphony Orchestra Blog 368
#43: Brian Dickie 366 (artistic director)
#47: In the Wings, 346 Heather Heise (piano)
#48: Today's Opera News, 338 Alan Faust (O)
#49: Of Music and Men, 324 Ilkka Talvi (violin)
#50: Sound and Mind, 323 Kris Shaffer (A)
#51: Coloratur...aaah 322 (voice)

I won't mention which blogs would be cut to make room for these worthies. As I look at the different lists, I think Technorati has a better picture of the more current buzz, as it would with the emphasis on only the last six months. Google is a more historical approach, so even if a blog doesn't get new links in the past six months, it could still rank well based on past performance. Matthew Guerreri's blog is getting linked a lot lately, as the Technorati ranking shows, but by Google standards he would be down at #90 (oops, I just mentioned one). So I trust Technorati more for this use. Here is an old study comparing Google with Technorati, though some of his interpretations of the statistics are incorrect, particularly when comparing Google and Yahoo. Technorati doesn't record nearly as many links, the question is which links are important, and how old the links are.

Update: Chris Foley has generated his own list, using subscription rates on Bloglines, an RSS reader. He also makes good arguments for the validity of subscription ranking, not just because I go up considerably (to #11). However, I would have more faith with using statistics from one RSS reader if the numbers were considerably higher. There are 67 people who choose Bloglines to read this blog, but there could be many others that choose Google Reader, Kinja, deli.cio.us, or other aggregators. And the proportion of readers choosing one service over another could change radically from blog to blog, augmented by these low numbers. If we were talking in the hundreds or thousands, I would have more confidence. But I also understand the limitations Chris worked with. You go to list with the statistics you have, not the statistics you want to have.

2 comments:

Anonymous said...

My own blog is relatively new and the majority of traffic comes from other bloggers who decide to link to it and by word of mouth. There has been comparatively little traffic from Google to date.

The theory is that Google "sandboxes" young blogs until they reach a certain maturity after which they are freed to climb in ranking.

Technorati is notoriously inaccurate, but seems to nevertheless be the most accurate statistic around. Here in Switzerland all blogs are ranked using Technorati stats.

Anyway, thanks for providing the links. I have connected with several other bloggers already and am looking forward to the reading.

Chris Foley said...

Unfortunately, Google Reader only lists feed numbers for a publisher's own blog and only if they have access to their own server logs. Hopefully Google will add subscription numbers for all feeds on Google Reader in the near future...