I have been noticing this over the last couple of weeks, but is the Pubsub LinkRank jaded against WordPress blogs somehow??
It looks as if somewhere around a week and a half ago, something was changed in the algorithm that completely trounced the rank of WordPress blogs, even the really popular ones while others continue to flourish.
Photomatt’s LinkRank
BinaryBonsai’s LinkRank
Dougal’s LinkRank
WordLog’s LinkRank
However, Scoble’s LinkRank continues to grow.
All of these ranks, including the one for this blog, are either askewed or are borked. Is this a result of some sort of a WordPress bug, vendetta against WordPress or a bug on PubSub’s side? Or maybe it was due to an attempt to fix the algorithm that went horribly wrong for WordPress users?
(All of this is really not as serious as I would like to make it sound, so be gentle! 🙂 ) What do you think?
EDIT: Bob Wyman provides an explanation. Look at comment #3
On link rank changes … from Bob Wyman, PubSub man himself.
After thinking about it a bit, I’m convinced that some other change besides the one mentioned in Bob Wyman’s post has taken place. The change he describes (recognizing sub-blogs on spaces.msn.com, livejournal.com, and journals.aol.com) doesn’t account for these drastic drops. At least not in any way that I can think of.
I’m wondering if they’ve implemented some sort of filtering (a la Technorati) to discount “default links” from certain blogware like WordPress. I used to rank pretty high in the Technorati Top 100 before they made that change. I’ll grant that my ranking was “artificially” inflated due to the linkage from default WP installs. But it’s still disappointing to see those links being ignored in such a wholesale fashion.
Who’s to say which of those sites are not worth counting, and which ones kept my link (or Matt’s, or Alex’s, or…) on purpose, because they wanted to link to us? We may have had artificially inflated linkage stats, but artificially deflating them is a worse muddying of the statistical waters, as it only serves to increase the margin of error.
As a curious stat to examine more closely, check out the numbers for Kottke vs Zeldman. On Technorati, they are ranked 26 and 41, respectively (at the time of this writing). Kottke is linked from 3626 sources, Zeldman is linked from 2743.
But in PubSub’s rankings, Kottke is ranked 61, while Zeldman is waaaaay down at 35,958 (almost as low as my 36,425). What’s up with that? Okay, I realize that PubSub is only counting links from posts, not from blogrolls, and I also realize that Zeldman hasn’t been posting as frequently since he became a parent… But is Kottke really getting mentioned that much more frequently?
When you look at LinkRanks, you have to be asking the question: “What have you done for me lately?”… The number of total historical links for Kottke and Zeldman may be relatively similar, however, LinkRanks only care about recent links and then only links from entries in feeds — not blogrolls, etc. Check out the following two urls:
http://www.pubsub.com/linkrank.....kottke.org
http://www.pubsub.com/linkrank.....eldman.com
You’ll see that Kottke has had many more *recent* links then Zeldman has.
Please note that the only significant change that we have recently made in LinkRanks is that we now track individual blogs on LiveJournal, Spaces, and a number of other services. Any change in rank in the last few weeks is strictly attributable to that change — or to an actual change in the way the community links to you.
bob wyman
Ah, I had considered the ephemeral nature of the data, but I didn’t follow that thought through before I posted previously. And I also should have thought better about how many new data points you added into the equations when you broke out the sub-blogs from those services. I’m still surprised at some of the results, but I don’t doubt them anymore 🙂
That makes sense. I will add the explanation to the post.
Ooohhh, thanks for clearing that up. I was wondering why mines was so low.