• E-mail this page
• |  Print this page
• | 
• | 
• | 

Are Bots On Twitter Hurting Overall System Performance?

If you use the "real-time short messaging service" Twitter, you know that over the last year there have been many performance issues. So many, in fact, that users have created a name for the outages, called the "fail whale." I started to poke around and began to wonder if bots running on the Twitter network could be causing some of the system performance issues on the service.

A bot basically performs a function or functions. In the case of Twitter, most of the bots either post links to content from across the Web or provide a reply service for certain phases.

To get an idea of just how many bots there are on the Twitter service, over the past couple of weeks I analyzed blog posts from more than 10 popular tech blogs. While this won't account for all of the reply service bots, it did provide good insight into the massive number of link-posting bots. For the purposes of this post, the bots listed below come from one post on the tech blog TechCrunch. Here's a list of the bots that reposted a link to the post on TechCrunch along with the number of subscribers (i.e., followers) to each one:

• top_post - 104
• tc2tw - 25
• techupdates - 171
• RSS_TechCrunch - 350
• techupdate - 31
• TechFeed - 207
• twittfeed - 116
• RossRSS - 70
• SocialMedia411 - 51
• TechGlance - 449
• top_blog - 112
• nuws - 132
• readmashcrunch - 96
• wiredgnome - 43
• techcrunchtweet - 47
• balduaashish - 28
• tinycomb - 31
• readburner - 348
• normman - 89
• crazyengineers - 25
• TwitLinksRSS - 392
• seomasterlist - 207
• simplynews - 56
• tweetmeme - 643
• memebot - 48
• foobot - 42
• readburnerRSS - 112
• techgurus - 225
• linksgoogle - 300

What's really interesting is that in the case of TechCrunch, its Twitter account has the full feed in the account, so none of these bots are even necessary for TechCrunch, although some of the bots listed also scrape feeds from other tech sites as well.

As Twitter works on implementing a revenue model, I can only imagine that bot traffic could be the first hit.

« Imation Finds Scary Data On "Recertified" Tapes | Main | IBM's Enterprise Content Management Push »