Once a tweet, always a tweet?

Until October 2009, if you deleted a tweet from your Twitter profile it would still remain available in Twitter search. Then, on the 24th of October, Twitter started deleting tweets from search too. Suddenly it seemed like you’d regained a little bit of control over your content! But what about the all-indexing Google? Since long before they launched real-time search, Google has been indexing twitter statuses and a lot more than 10 days worth of tweets are available.

Zombie Tweets

I have multple Twitter accounts. One of which is my professsional account, another one is my sport and music account. On the 6th February 2010 I accidentally updated my Twitter status from @beantin instead of @dr_chasm.

Screenshot of the deleted Beantin tweet still cached by Google

The tweet only existed for a few minutes before I deleted it and that status now gives a page not found on Twitter.

No rush to re-index

I suspect that Google doesn’t rush to re-index Twitter status updates – why would it? They can’t be edited. All that can change is your profile picture, profile name and real name. Well, one reason of course is to remove deleted tweets. Given the tiny number of tweets that are deleted it would be a waste of resources to keep re-checking millions of tweets regularly over and over again just in case they ever vanish.

(Admittedly, re-indexing everything is exactly what Google does do for all the other billions of web pages it has indexed on the Internet, but you could argue that historical Twitter statuses are about as static as a web page can get.)

Some tweets do vanish from Google’s index and cache. Take the offensive VodafoneUK tweet that was made on the 5th of February 2010. It’s gone from both Twitter and Google.

How often does Google re-index Twitter statuses?

So either Vodafone UK specifically requested the indexed page for that tweet to be removed using Google’s Web page removal request tool (I presume, as the page returns a 404, Google will without question recheck the page eventually and remove it from their index), Or Google does re-index Twitter statuses.

Presuming that Google does eventually re-index all statuses – which Twitter themselves imply – then it raises the question of: what is the interval? Clearly as VodafoneUK’s deleted tweet from the 5th has been removed, but my deleted tweet from the 6th hasn’t, Beantin doesn’t get re-indexed as often as VodafoneUK.

Delete from Google!

Nevertheless, everyone, including those managing corporate Twitter accounts, need to be aware that you should take action to delete tweets from Google as well as Twitter when anything unwanted ever hastily flies out from your Twitter-mouth.

Update 2010-03-11: After just over a month, my original tweet has now gone from Google. It’s quite possible that this article itself caused the twitter page to be re-indexed, the 404 found, and the content dropped from Google. The advice above though still stands…