Most tweets have a shelf of an hour
Probably not a huge surprise to anyone familiar with Twitter, but worth reinforcing, the quick fire stream of tweets people see from their followers means the vast majority of Twitter posts have a life-span of an hour or less.
This is according to Canadian research company Sysomos, which found that 92.4% of re-tweets happen within the first hour of the original tweet being posted.
Sysomos also found that 29% of tweets generate some kind of reaction, either a reply (23%) or retweet (6%). While you could turn that into a headline of ’71% of tweets get ignored’ (and here’s the story from Mashable saying exactly that), I actually think a rate 3/10 tweets eliciting some kind of response is quite high – this is especially when you consider that a lot of tweets are simply personal musings that wouldn’t warrant any kind of reaction.
According to Sysomos, 85% of replies are ‘one level’ deep, 10.7% go down two levels (ie have two ripples of responses) and 1.53% are three levels deep. Again, you can interpret this statistic in a number of different ways. Sysomos says that “only” 1.53% of tweets go down three levels. I look it like this:
There are nearly 100 million tweets a day (image from Twitter above). So, every single day 29 million tweets elicit some kind of response.
Related articles by Zemanta
- Report: 71 Percent of Tweets Do Not Get a Single Reply or Retweet (socialtimes.com)
- The Short Lifespan of a Tweet: Retweets Only Happen Within the First Hour (readwriteweb.com)








Hi there, my super cool buddy recommended this website like a really good strategy to obtain advice and it is not bad. I love that, really put up that in my small Facebook . com. I highly recommend you have a look at my personal site http://xquickloans.co.uk/
Great site. Lots of useful info here. I’m sending it to a few friends ans also sharing in delicious. And certainly, thanks for your effort!
That’s a very good point Mark – we all know that someone will occasionally @ someone back with a reply and then *separately* RT the message onto his / her followers.
Similarly, as you say ‘no response’ doesn’t mean ignored and you could still click on a link. You might also still read it and take some kind of action based on what you’ve read, without necessarily hitting RT.
Yes – I do think 3/10 (or more) is good and also what I think you would expect. Again, I’d be interested to see how Facebook does, but certainly on my stream this morning, 3/10 looks like a comparable reaction rate.
Two things puzzle me about the Sysomos figures.
First, they add up to 100% – as if no tweet both gets retweeted and an @ reply. Second, “no response” isn’t really a good description for other tweets; if I sent a tweet with a link in it and someone clicks on that, I think that counts as a “response” on any sensible definition of that word.
As you say, 3/10 tweets getting some sort of response is actually pretty good – but bearing in mind my second point, it looks as if the true figure is even higher?
Thanks Ian, and absolutely we all know the same data can always be positioned in a number of different ways, hence the title of this blog of course!
The more I think about it the more ridiculous it is though for Mashable, Read Write Web et al to take this line.
I mean is 29% low? Low compared to what? Are more than 3/10 Facebook posts liked or commented on?
Based on an (unscientific!) look at my feed, not really.
Great post. I like the way you highlight how framing the information changes it’s meaning significantly. It is a bit like the glass half full or half empty idea. I’m surprised that Mashable went with the negative frame though because, as you have ‘spun’ it, so to speak, it is actually quite good news!