Update on Twitter CTR Rates
July 9th, 2009, 6 Comments »
Last month, I wrote about what sort of clickthrough rate you could expect from links shared on Twitter. I rewrote and expanded that post over on Mashable. Thanks to those who commented on my original post–your thoughts informed my rewriting.
I hoped that the Mashable post would enable me to expand the survey I ran, so that I could gather a more consequential amount of data. I was a little surprised to only gather another 80 or so data points, but that did move the numbers a little bit.
Working with about 140 responses who collectively have 333,000 folllowers, I found that the average clickthrough rate was 1.7%. The trend of more followers equaling a lower click through rate has definitely held true, though. For those respondents with more than 5000 followers, the click through rate is a mere 0.9%. For those with less than 5000 followers, it’s 3.5%. For those with less than 1000 followers, it’s 6%. This scatter graph hopefully illustrates that trend. That’s followers on the x-axis, and CTRs on the y-axis. To make the diagram more readable, I excluded a couple of the really big follower counts (click for the largess):
A recent FriendFeed comment from Tim O’Reilly also bears this out. He’s discussing the huge boost in followers he’s received since being added to Twitter’s Suggested Users List:
I had about 60K twitter followers when I went on the SUL; my peak click through-rate has perhaps doubled now that I have 10 times as many. Organic followers are what matters, except, as I say, for the media credibility that you get from people who don’t know any better.
As I mentioned on Mashable, I’m not a statistician, so take this kind of analysis with a big salt lick. I also mentioned several other mitigating factors. I’m uncertain about what impact bots spidering Twitter and following links are having on these results. I welcome any insights you might have on that, or any aspect of my sketchy math. Another consideration is that when a URL gets retweeted, you’re adding the retweeting user’s followers to the mix. Lastly, I asked those who completed the survey to pick an ‘average’ number of clicks, so that depends on each respondent’s potentially unscientific estimate.