Feedster has answered the Calacanis Challenge and come up with the Feedster Top 500 for August, 2005. The announced:

Today, Feedster launched the FEEDSTER Top 500, a ranking of the most interesting and important blogs in the US. The ranking is achieved by taking into account factors such as the number of inbound links over time; if the blog has been recently updated; and the elimination of obvious non-blogs that have appeared on other top-blog lists. [bolding mine]

Ok, first, why in the US? Why the restrictor? Especially since it was not effective, and blogs from all over are included. Why were blogs eliminated for appearing on other lists? That biases the list. And obvious top-bloggers are included, so how was it determined who would go and who would stay? Using inbound links and recent updates in no way eliminates the bias determined from the older bloggers – it’s the same metrics.

“We left out professional news sites, aggregation systems, and some fairly static web sites that happened to have feeds but don’t ‘feel bloggy,’” says Mr. Johnson. “This sort of filtering is a different screen than what we use to categorize news versus blogs in Feedster search and is much more subjective. Our first iteration of this list solves the staleness problems and not-a-blog problems that others have faced.”

Well, not completely. The determination of what is a blog and what is news, is, as stated, arbitrary. Would I count Boing Boing as a news site or a blog? I’d be borderline on that one, but would definitely think it more news than blog based on the volume. And DeviantART doesn’t seem anything like a blog to me.

The list is indeed longer and perhaps more diverse due to this, but the metrics of the list are still the same. From this, I don’t think the list lives up to my expectations or those of others. Perhaps if a step had been added to add the description of the blog or the topic categories, it would prove useful. If diversity was a goal, as stated, then it must be described. Looking at names of blogs does not tell unfamiliar readers that the blogs are diverse in topics – the names are often arbitrary.

As is, it’s still a popularity contest based on linkage. Still invites people to look at the blogs based on the popularity index, and not the value of the content. Plus, as the list is manually refreshed only once a month, it is not even a completely valid shot of what’s out there. It needs to be live data.

I am happy that some of our friends have made the list. Jeremy Wright, Roland Tanglao, Chris Pirillo and others. It’s a validation of sorts… but it doesn’t meet the challenge, IMHO. The list has diversity based on its scale and some subjective pruning, not based on its design. It gets Feedster into the limelight for just a bit while we all comment on the new list they’ve made. It was a good tactic for press, but nothing else.

Scott Rafer addresses some of the challenges I pose here about link popularity by saying:

First, links are not the appropriate basis for relevancy in the world of blogs; links are just one of many factors to consider. To make up for the difference in the short term, we did some editorial work to shore up the basic algorithms. None of the rankings were shuffled around, but we excluded some entries based on editorial criteria…

The criteria include having your own domain name (no LiveJournal or BlogSpot blogs included), feeling ‘bloggy’, having no Flash, and having posts in the last 7 days (no matter how many links you have). I think some criteria are valid, and others need work. As Scott is looking for more algorithms, I suggest below some I think vital.

I would expect, with all of the data on hand from Technorati or Feedster, that a list could be developed with different metrics. Metrics I’d like:

- linkage with caveats

- where the links come from

- duplication of links (i.e. choose only unique link sources)

- link diversity (from other big blogs or from a wide variety of blog types)

- number of authors

- frequency of comments

- post length

Plus, I think it crucial to include descriptions of the blogs or, even better, topic specific lists.

I think the wiki is a good place for people to make comments such as these and I hope you pull some of my suggestions in there.

Via Read/Write Web

Update: Scott Rafer notes that the inclusion of “US” in the press release was his mistake and is not actually a part of the list. The wiki URL has been corrected from Scott’s post.

Technorati Tags: feedster, top 500, calacanis, feedster top5 00, lists, a lists, top blogs