Being Kept Awake by a Lack of Twitter Stats?

3rd June 2009

I’ve been reading a report by the bods at Harvard Business Publishing which talks about some new research they’ve done about people who use Twitter

The results are quite interesting as they provide a good demographical picture of the type of people who use the social networking website. So if you like trawling through statistics and appreciate a good graph now and and again, read on.

300,000 Twitter accounts were analysed in the study during May 2009.

The survey said…

More women than men use Twitter: 55% compared to 45%
On average a man is twice as likely to ‘follow’ another man rather than a woman.
The average woman is 25% more likely to follow a man rather than another woman.
An average man is 40% more likely to be followed by another man than by a woman.

The study also found that 80% of the studied Twitter accounts did have at least one follower. This means that the remaining 20% of Twitter accounts were completely inactive. This is a higher rate than other social networking sites the study claims.

The report continues to say that over 50% of Tweeters tweet less than once every 74 days whilst there is only a small number of people, typically the top 10% of Tweeters, who are very active and account for over 90% of tweets. These top 10% render Twitter activity very similar to Wikipedia in its nature rather than your typical social networking site such as Facebook. Similarly, unlike Facebook which is used for connecting with people, Twitter is used more for “broadcasting information” rather than chatting with other people as you would on an instant messaging tool like MSN Messenger or Google Talk.

So what can be gained from all this data? On the face of it very little I suspect, except to say that Twitter is different to any other social networking site (no surprises there then), it’s dominated by blokes who like to tell people stuff rather than interact and with most people, once they’ve set-up their Twitter accounts up, fail to keep them up to date.