Interesting report by folks that mined the Twitter stream.

The section on publishing tools is particularly interesting. Less than half of all tweets are made through Twitter’s web interface and I bet this includes some posts made through the API. Apprently, users much prefer to use desktop or mobile applications to post to Twitter.

Strategically, Twitter will be stronger with a Web front end that people actually want to see. Twitter’s greatest strength is its greatest weakness. The most-recent-first, single column layout is very simple and easy to consume. It’s perfect for beginners. The more people you follow and the more you search, though, the more the default Twitter interface’s simplicity starts failing. For example, I follow less than 50 people but some of them tweet many times per day (a few are businesses). It is very easy to lose tweets I’d have otherwise found worth reading and I often find myself wishing for more and better organization of tweets. That’s especially true when consuming Twitter through a mobile device as I usually do–catching up with what’s happening in-between other parts of my life.

A couple of questions I’d love answers to from Sysomos or anyone else who has the data:

1. What percentage of the 45.7% of tweets contributed through “the Web channel” came through the API w/o an identified client?
2. What’s the distribution of consumption tools (since reading is much bigger than writing)? (An interesting su-question is how does one measure consumption…)