Most Popular Twitter Clients at Chirp, by Users and by Tweets

  • strict warning: Non-static method view::load() should not be called statically in /home/sigpboo7/public_html/modules/views/views.module on line 842.
  • strict warning: Declaration of views_handler_field_comment::init() should be compatible with views_handler_field::init(&$view, $options) in /home/sigpboo7/public_html/modules/views/modules/comment/views_handler_field_comment.inc on line 50.
  • strict warning: Declaration of views_plugin_row::options_validate() should be compatible with views_plugin::options_validate(&$form, &$form_state) in /home/sigpboo7/public_html/modules/views/plugins/views_plugin_row.inc on line 135.
  • strict warning: Declaration of views_plugin_row::options_submit() should be compatible with views_plugin::options_submit(&$form, &$form_state) in /home/sigpboo7/public_html/modules/views/plugins/views_plugin_row.inc on line 135.

Last night, I posted a data set about which Twitter clients were the most popular on Day 1 of Chirp. Since then, I’ve gotten a few requests for data about which Twitter clients are popular here at Chirp normalized to users instead of to tweets.

Ask, and ye shall receive.

Here are two new data sets: one with clients counted by users, and a refresh of clients counted by tweets collected about 5 minutes after the first. Check them out:

Before we talk about the data, let’s have a look at it in tabular format:

Source Night 1Percent Source Morning 2 (Tweets)Share Source Morning 2 (Clients)Share
web19.78%web18.30%web23.70%
Tweetie15.11%API14.94%Tweetie17.99%
TweetDeck13.79%TweetDeck14.35%TweetDeck10.92%
API9.08%Tweetie14.18%foursquare6.82%
Seesmic4.16%HootSuite4.03%Echofon4.71%
Echofon3.95%Echofon3.69%Seesmic3.35%
HootSuite3.40%foursquare3.52%Twitterrific3.23%
Twitterrific3.14%Twitterrific3.02%txt2.73%
foursquare3.10%Seesmic2.85%mobile web1.74%
txt1.91%txt1.72%HootSuite1.49%
Twitter for Blackberry1.83%TTYter1.43%twidroid1.36%
twidroid1.61%twidroid1.22%Twitter for Blackberry1.36%
mobile web1.27%Brizzly1.22%Gowalla1.24%
Brizzly1.15%twitterfeed1.22%Brizzly1.12%
TTYtter1.06%mobile web0.92%UberTwitter0.99%
Other15.66%Other13.39%Others17.25%

At a glance, we notice a few things: the data didn’t change much for the tweet set from last night to this morning, which gives us confidence that our measurements are reliable; the15 most popular clients account for about 85% of volume, no matter how you measure; while there’s a little jockeying for position among the top 15 clients, the top 15 clients are pretty clearly the top 15 clients, no matter how you slice it; and the good old web interface (i.e., twitter.com) is the most used interface by far across all categories. Also, the shape of the different distributions is about the same, at least as far as client volume per position goes:

So more is the same for these data sets than is different. But one interesting observation about the differences is how chatty the users on different clients are:

Interestingly, Hootsuite is the chattiest by far, with an average of 8 tweets per user. (Now, Hootsuite only had 12 users, remember, so this looks kind of like an outlier to me.) TweetDeck, Brizzly, and Twitterrific are the 3 next chattiest clients, with about 3.8, 3.2, and 2.8 tweets per user each, respectively. Mobile web and Foursquare are the least chatty, with about 1.5 tweets each. The mean number of tweets per user is about 2.9 and the mode for the data set appears to be about 2.2, so the data set is pretty clearly skewed to the right. (Note that the only clients that appear in this analysis are the ones that appear in both the second tweet-based data set and the user-based data set.)

So, Chirp folks, there you have it. Your most popular clients are Web, Tweetie, and TweetDeck (no matter how you slice it), and the chattiest clients are HootSuite, TweetDeck, and Brizzly. If you’re unhappy with your current client, these are great places to start, and if anyone has something to say about these clients, let me know in the comments. Per usual, the code and the data for this article are attached. Enjoy!

EDIT: Thanks to Mike in the comments for pointing out that one of HootSuite’s big features is the ability to tweet to multiple accounts at once, and that people who use HootSuite tend to do this a lot. The “user-based” collection methodology for this analysis actually counts tweets by account, not by user, so that is an important grain of salt to take when interpreting this data. Excellent point, Mike, and thanks again for sharing.

EDIT: This post just got retweeted by @rsarver, Twitter’s Director of Platform. Wow. Thanks, Ryan!

AttachmentSize
clients.xlsx50.75 KB
clients.rb1.23 KB