Great things about Twellow:

• Categorizes twitter streams by vertical/niche, eg: automotive, movies, blogging, news
• Groups related twitterers together

It’s offset by one major flaw, which unfortunately is tied to Twitter’s current state of development – analytics don’t mean much more than a brute force “followers” number.

The higher the number of followers, the higher you’ll rank in the results, with the net effect that Robert Scoble is ranked first with 28,000 followers, followed by Jake Marsh with 12,000, in the blogging category.

The results are limited by the enrollment of your twitter feed into the system for benchmarking and indexing.

But I’m having serious doubts about using followers as the determining criteria.

Could social networks be the new playground for blackhat marketers?

In the initial stages of Yahoo MyBlogLog’s launch, I know of a number of programmers who automated the creation of new profiles and artificially inflated community sizes and went out to spread spammy marketing messages.

With the current state of Twitter development, what’s to stop marketers from either automating or farming out the creation of a horde of twitter followers, thereby artificially inflating the size of their community.

It’s not secret that a forum with 100,000 registered members with 5-10 active users is pretty much dead. What about a twitter account with 50,000 inactive followers?

I wonder if Twitter will enforce a minimum activity threshold (a number of twitters sent out within a set time interval) like some webmail providers have instituted to continually prune dead and inactive accounts.

Will Twitter be able to present useful stats in the future? Likely, but not in the near future as they’re still coping with growing pains and the resource overhead of third party Twitter API calls, to the extent that pagination (an archive of your twitter conversation) and “replies” (twitter messages that you have received) have been disabled.

–

It could be a while before we see Twitter metrics evolve to comprehensive display the time visitors are spending reading your updates (ie stickiness) and clicking through to your links (conversion?).

Till then “followers” could be a proxy for a twitter popularity contest (traffic quality is a big question mark).

Tread lightly. A conservative approach can do more to preserve your social goodwill, rather than making rash experimental stabs.