24% of Tweets Created by Bots

Compared to email that's not such a big percentage, but if you take into account how long twitter exists and how small it really still is the number is pretty scary.

Maybe it should read 'already 24% of tweets created by bots'.

A bit of a trend would be useful information.

Would be interesting to have a breakdown on 'useful' bot tweets and 'spam' bot tweets as well, the article suggests that there is a bit discrepancy but does nothing to document how big.

Combined with http://news.ycombinator.com/item?id=639266, it seems like Twitter is less about inviting millions of strangers into the minutiae of your life than it seems from a naive view.

I seem to remember danah boyd talking about something to that effect at one point; that a lot of the Twitter activity she'd observed among young people was about talking amongst a small circle of friends. The only people really broadcasting on Twitter are people who were already famous for something besides their 140 character brilliance.

And another 40% of tweets created by bot-like - although not technically bot - human marketing droids.

I think this goes well with the whining story posted a couple months ago about how the higher ups at twitter weren't using the service right, by which, they weren't using the service as useless "social media expert" follower whores want to use it.

One of the problems I've had to address in developing a social media analysis product is figuring out whose tweets get read by humans, and which ones are just followed by other bots. It's hard to approximate without access to lots of data which isn't available over the API. On the other hand, it will be easy for twitter to figure this out.

A few days ago, some users stopped showing up in search results. Are they starting to tighten the noose around spammers? The beauty of twitter is that I've already whitelisted those who I want to hear from. It can completely eliminate the spam problem with email.

about 25% - 35% of posts on delicious were due to spammers/bots/crap as well...

Bots are allowed on Twitter - that's fine with me (@beijingair is a very cool twitter bot)

But Twitter should make applications mark themselves as such to let tools filter out auto tweets (imo twitterfeed should count as a bot)

I expect this number to be 95%+ at some point, because it's so much easier to scale bots than humans, but that's not necessarily a bad thing at all. I have 3 twitter accounts, one personal and 2 bot-controlled, with one of the bots having 10 times as many followers as my own. And deservedly so, he puts out a lot more interesting content than I do.

A better metric would be something like tweets*followers, or some other way to measure consumption of tweets instead of just production. Spam-bots will do poorly on that metric, but useful bots should do fine.

It doesn't matter how many tweets are spam, all that matters is how easy they are to filter out. And the answer is, pretty damn easy.

What Twitter needs is a way to filter the drivel that comes from humans.

How many of those bots are RSS headline bots?