Median # of tweets = 1

The figure is meaningless on its own, and the post creates some false significance by trying to extrapolate the success of Twitter from it. Clearly if 40% of several million account signups use your service regularly, you're still doing alright, even if your median usage is low, and that's excluding the impact of hits rather than contributions.

This is made worse by the fact that the Harvard Business Review blog article quoted and linked (http://blogs.harvardbusiness.org/cs/2009/06/new_twitter_rese...) actually provides decent reasoning for this behaviour - that Twitter is largely used as many-to-one publishing platform, and as such should be compared to the amount of posts-to-signups that blog hosts get, and I suspect it compares favourably. The HBR article itself is pretty good and quite interesting about the real-life usage of Twitter - just don't try and make conclusions about the success of Twitter from it.

What I really want to know about Twitter is whether anyone is listening. Obviously some people are, but if 50% of the people have only tweeted once or never, I'm guessing they're not glued to their account. And within that top 10%, there's an awful lot of people who have so many thousands of followers that they can't be paying much attention. So that leaves the 40% in the middle...how many of those people are really paying attention? Is everyone on Twitter just shouting into the void?

As mentioned in the article, this is really no surprise. Even more than other social networks, Twitter is a publishing platform - and it's only really interesting if you actually have something to say.

This certainly shouldn't be read as "Twitter's days are numbered", since so many people (even if they only represent 10% of its users) have found real value it its offering.

I wonder what is the median number of sales for all products sold by Amazon?

How many of these are from one-time spam accounts? I know I get people following me from time to time who follow a bunch of people hoping to get reciprocal follows, then send out a spam tweet and move on to another account. It probably doesn't account for the majority of these, but I would guess it's non-trivial.

Techcrunch should read this.