It has been widely reported this week that 24% or nearly a quarter of all twitter messages are generated by accounts controlled by automated ‘bots’ and the robots are on the rise. The research was done by social media analytics specialists Sysmos.
Some of this automated traffic is from useful sources but much of it is spam and as previously discussed here, spam is polluting the social web to an alarming extent.
There is a way of dealing with this. Don’t follow in the first place, but more importantly if you find a bot block it. If an account is blocked numerous times it will be suspended by twitter. Some of the bots and spammers are obvious, the avatars are usually a dead giveway for the porn spammers but many are becoming more sophisticated. There are still ways to weed them out. Here are some tell tale signs of bot behaviour:
- Too many tweets – bots often but not always pump out 150+ per day.
- Tweeting famous quotes – this is a technique increasingly used to make the bot look real.
- Lot of homilies for the same reason.
- Duplicate messages – some bots pump out exactly the same streams of messages.
- A high Follower to Following ratio.
- No @posts or @mentions. Bots don’t do conversation.
- Exclamation marks next to a link!
- Tweets seem disjointed and unrelated to each other
- The tweet source – this can be a dead give away, for example if the tweets are from ‘API’ be very wary.
OK, get ready to block.
PR and the Social Web 
But my goodness they’re getting better. It sometimes takes a minute or so to spot some bots. When you find one you can quite often discover more through other accounts they appear to be interacting with. If you have the time or inclination you can squash several at once.