I came across PubSub's LinkRank feature for the first time today. It's an interesting new measurement... Here is the official blurb, comparing it to Google PageRank:
"Unlike Google's PageRank system, LinkRanks are not iterative. Rather, we base LinkRanks on a simple formula that only looks at local links - links which are within one or two steps of any target site. Also, it's important to note that we only look at links which are in weblog entries - we don't read any of the other links on the page, like the side bars or blogrolls."
It's also different to Google because unlike PageRank, which is a mark out of 10, PubSub's LinkRank is a ranked list. Like a Billboard chart for blogs (reminiscent of a short story I wrote earlier this year). e.g. my current LinkRank is 2,188 (yesterday it was 2,314). Not bad, when you consider they claim to track 6.7 million sources (3.6 million "active" sources). My recently closed eBook Culture blog has a LinkRank of 888,335 - which perhaps explains its demise. The highest single blogger I found was Jason Kottke at 63 (Amazon.com was number 1).
PubSub's creator, Bob Wyman, did some tracking of Zeldman's blog a couple of months ago. He calculated that Zeldman was hovering at the 600-650 mark at that time, noting that Zeldman's LinkRank had decreased due to "a period of publishing less-then-enthralling content". Bob's conclusion:
"Watching your Linkrank can give you an idea of how well you are maintaining the interest of your particular community of readers. In most cases, but not all, if people find that what you are writing is interesting, they will link to you."
On a related but less positive note, I also found out today that Technorati has not been indexing my outbound links. So for I-don't-know-how-many months nobody would have known I was linking to them, at least via Technorati. That's pretty annoying and hopefully it'll be fixed soon.
PubSub also look like they're doing something with Topic Mapping. They ask people link to a specific URN, because:
"We want to see if we can construct a conversation thread around the topic by using a common URN. For reference, the URN form is based on NewsML (more specifically the URN namespace for NewsML resources), and PSI stands for "published subject indicator" (see XML Topic Maps [topicsmaps.org] and Published Subjects [OASIS]). If this works, we want to try to use this kind of URN to bring conversations together around various topics."
That's reminiscent of various topic mapping ideas people have come up with over the past few years, including my own Microcontent Wiki (remembered more for its entertainment value than any real-life potential!). So PubSub's topic mapping idea is another thing to keep a watch on! p.s. no smart comments asking if I'm being paid to blog about PubSub please ;-)
TrackBack URL for this entry: http://www.readwriteweb.com/cgi-bin/mt/mt-tb.cgi/2399
Comments
Subscribe to comments for this post OR Subscribe to comments for all Read/WriteWeb posts
Your eBookCulture was doing better than my webcomic ( http://www.pubsub.com/linkranks.php?dom=thechroniclesofa.com ) at 911597. But, as far as my main weblog goes ( http://www.pubsub.com/linkranks.php?dom=andrewsw.com ), you're definitely 3 times more popular than it ( at 6767 ).
So while you may want to be popular, for some of us, the best way to get links is to try to get them from you.
I'm glad you're continuing to provide coverage of some of the key developments in the topic mapping community. Thanks.
Posted by: Andrew | December 2, 2004 9:05 AM
One thing to point out is that the LinkRank system appears to depend on having an independent domain name, and one which hasn't been chopped and changed much. eg if you enter a Radio Userland domain (eg Seb's), you'll get no result. Another example is Dave Pollard's domain: http://blogs.salon.com/0002007/. Also I noticed a couple of the design blogs I follow which have extended URL's (eg http://networkname.com/blogsection) - their ranking isn't as high as I would've thought they'd be.
So the system is (like every other Web stats system I know of), not *entirely* reliable. But having said that, it nevertheless uncovers some interesting trends and insights over time. Such is the nature of web analytics.
Posted by: Richard MacManus | December 2, 2004 12:00 PM
Richard MacManus is correct in pointing out that LinkRanks currently only track links between domains or sites. The algorithm at present assumes that everything on a single site is the same. This is clearly less then optimal and will be modified in the future. The difficulty is that it appears that we need to implement domain specific rules to determine what is a "sub-site." For instance, while a "sub-site" on spaces.msn.com might be indicated as: "spaces.msn.com/member/name" the same pattern does *not* hold for LiveJournal sub-sites. So, this is going to be messy... But, we'll do it anyway.
bob wyman
Posted by: Bob Wyman | December 3, 2004 11:11 AM
Thanks for your comment Bob.
Posted by: Richard MacManus | December 6, 2004 9:09 AM