The Most Influential Celebrities on Twitter by Clickthrough Rate





Influence on Twitter can be measured in a variety of ways, but one of the most interesting for marketers is in the amount of traffic an account can send to a link it Tweets. Months ago, when I was working on clickthrough rate (CTR) data I noticed that the CTRs from highly followed celebrity accounts varied greatly, so I graphed the data.

Below are the celebrities for whom I’ve been able to analyze at least 100 links shortened through the bit.ly service. I define “clickthrough rate” as the average number of clicks on the links they Tweeted divided by the number of followers the accounts had.

Got a favorite celebrity? Want to Tweet about their clickthrough rate?

If you liked this post, don't forget to subscribe to my RSS feed or my email newsletter so you never miss the science.

{ 21 comments }

Anonymous September 3, 2010 at 3:46 pm

Interesting data Dan. It was great seeing your presentation at seomoz conference up in Seattle. Looking forward to diving deeper into your presentation material (and blog posts) after catching back up here at the office.

Derek September 3, 2010 at 12:32 pm

Weird Al beats out Mashable and Tony Robbins? Wow.

Anonymous September 3, 2010 at 5:05 pm

Great job with this! Tweeted this from oneforty and will post to our FB page! Keep it up the interesting data, we definitely gotta team up for an infographic or something soon… Give my best to the soon-to-be Mrs. D! Sure I’ll see you both at a Tweetup soon!

Alfredo Artiles September 3, 2010 at 5:37 pm

Here you have the most recommended celebrities based on the #followfriday recommendations http://www.followfriday.com/verified

Eric Andersen September 3, 2010 at 6:04 pm

Interesting work, Dan. I think what we see is that as number of followers increases, click-through rate will go down, since there will be more older abandoned accounts following and a higher percent of spam accounts or accounts otherwise not reading the tweets. So somehow I think you have to “grade on a curve”.

Anita Nelson September 3, 2010 at 8:30 pm

This is fantastic info, Dan! I plan to look at what the top celebs for clickthrough are doing and try to mirror & match it =) Thanks for the amazing stats as always~!

Anita @ModelSupplies

Richard September 6, 2010 at 8:34 pm

Were you just taking open url shortners like bit.ly and is.gd? Were you taking overall clicks on bit.ly or just that their one link generated? Who gets credit fro something like this?

http://bit.ly/aE7R4X+

Since both sisters tweeted it are both getting to take credit for this even though it could be the one with more followers pulling her up in the rankings?

As well, not to defend Ashton :)

http://bit.ly/9k6HKq+

but are you counting all 100k for him or just the 3k that his direct link brought in? obviously this is all about him but this would be hard to tell. Other times ones that are shared might have nothing to do them but they RT’d some other big news and would then get credit.

Also, looking at that data, the 166k clicks are over LONG time, and not just recent. If someone shares the same link repeatedly are you then counting it multiple times for them including the same clicking?

As well, with links changing and especially people wanting to track their own influence, they will change out the link to their own tracker of choice and you could be losing that information. This is going to hurt the people with more savvy user bases who even have this automated for them (look into Argyle Social.)

There are really just a ton of hard to automate problems with this kind of data collection. I realize this is mainly about making a story and then being able to drive traffic to it because celebs will tweet it and such BUT I think you really need to look at lay out your data collection methods. I think your previous article you linked in this one does it a little better but is still missing a lot of these factors. And really can steer things in wrong directions. I think overall trends like the higher click rate on less tweets is correct but by your standards I’d be in double digit percents with my blog posts I make because I get 40+ CT with 300 followers.

And I think the problem you get from bad usage of data like this is highlighted by Anita’s comment where there is nothing to learn from that which would benefit her like she thinks.

Dan Zarrella September 6, 2010 at 8:54 pm

If you read the post above, I answer all of your questions and have spelled out my methodology very clearly.

Richard September 6, 2010 at 9:04 pm

Okay so you are taking every bit.ly link they tweeted and taking the “clicks” (not Total Clicks) and using the formula on them.

Now doesn’t again address any of the issues or concerns I brought up with that way of doing things. Just because someone tweets a links does not make them responsible for all the clicks that were on said link, and what about links that are reused over time and have accumulated a ton of clicks? There are a ton of easily spotable cases that break this way of collecting data as being valid. If I tweet out this link.

http://bit.ly/bZ4MWM+

I now am getting credit for all 5k+ Clicks? A lot of celebs are going to share links and have collectively large click throughs from all of them, so lower follower celebs get the benefit of that over their larger follower counterparts who may have actually generated those.

Dan Zarrella September 6, 2010 at 9:19 pm

You’re right it doesn’t take into account every little issue. It is a proxy for CTR, and especially since I’m looking at at least 100 links for every account, those issues wouldn’t skew in favor of one account or another.

Richard September 6, 2010 at 9:24 pm

It’s going to skew in favor of the smaller accounts that are sharing links of the bigger ones and getting the same credit. I’ll give you the data here later tonight when I have some time to grab the last 2k tweets of the celebs you listed because I think it will become more obvious where the the issues lies when you have multiple celebs sharing links.

But, this doesn’t seem to be as much about getting the best possible data but more about getting traffic from talking about these big name accounts. :)

Dan Zarrella September 6, 2010 at 9:27 pm

One more time: I’m only calculating CTR for accounts where I have at least 100 links to gather data from, and in most cases, the number is far higher. One or two overlapped links isn’t going to skew the CTR for an entire account very much, especially when you consider that all of these accounts have well over 1M followers.

Geoff Livingston September 7, 2010 at 1:49 pm

What’s ironic about this is in your original post, you said tweet less and you’ll get more clicks. Alyssa Milano tweets ALL the time!

Dan Zarrella September 7, 2010 at 1:50 pm

Well, the original post actually says tweet fewer clicks (like one per hour)…

Anjuan Simmons September 7, 2010 at 11:28 pm

I wrote two articles for Black Web 2.0 listing the 100 Most Powerful Black Men on Twitter (http://www.blackweb20.com/2010/08/30/the-100-most-powerful-black-men-on-twitter/) and the 100 Most Powerful Black Women on Twitter (http://www.blackweb20.com/2010/08/24/the-100-most-powerful-black-women-on-twitter/) using Klout.com scores and Twittergrader.com rankings. While Klout covers clickthroughs to a certain degree, I like you list that focuses solely on this measure.

Jonathan September 8, 2010 at 4:27 am

So the two best people are both kardashians. With a lower clickthrough rate but a higher number of followers, does kim get more total clickthroughs?

Did you look at what the links were actually pointing to? For example, would a kim link with a fashion comment convert higher than if she mentioned something about basketball?

Balance Transfer September 10, 2010 at 8:47 am

Stephen Fry came second!! Great for him

Nick Stamoulis September 10, 2010 at 11:56 am

Very interesting to see that the higher the click through, the less followers the celebs have in most cases it seems to be true.

Marilyn Moran September 10, 2010 at 2:24 pm

Aww, @wilw <3

Anonymous September 10, 2010 at 8:25 pm

This is some good stuff Dan. Thanks for sharing it!

Boris Loukanov September 22, 2010 at 1:11 pm

What about Guy Kawasaki?:-)

{ 12 trackbacks }