January 28, 2008

The hunt for Jérôme Kerviel

Jérôme Kerviel is the trader at Société Générale responsible for the loss of almost 5 billion EURO or 7.2 billion USD. He is taken into custody by the French police. Journalists and bloggers have started a worldwide hunt for information on him.

A Wikipedia article describes in detail his professional career at SocGen. A photo of Jérôme Kerviel was first published in a blogpost from the Fincancial Times. His CV has been published on several news sites, including the Telegraph. According to the Guardian Jérôme Kerviel lost almost all his friends on Facebook. At this moment his Facebook account is not accessible anymore. Several fake accounts have been created. Luckely for Jérôme Kerviel, not everyone is attacking him. Different fan groups have been created on Facebook to support him.

Jérôme Kerviel does not seem to have a Twitter account. Twitter users are informing each other on their progress on finding information on him, directly in Twitter messages or indirectly by announcing posts on their blogs. His hotmail email account (listed in the CV) seems to be invalid, you cannot add this email address anymore to your MSN contacts (as found by this Twitter user). Another Twitter user says he has found Jérôme Kerviel's address and even his mobile phone number (see this tweet).

On the one hand Web 2.0 allows users to look up information on all kind op topics. The scary thing however is that in case of a sudden event, information about an individual person can be found, revealing personal information that is taken out of its context. E.g. what is the relationship between "ambitious judo teacher" and this fraud case ?

January 20, 2008

First state of the Twitosphere in Iceland

Last week check legend Bobby Fischer died in his new home country Iceland. At the same time I was wondering why Iceland had so many Twitter users. For an estimated population of 312,872 (Wikipedia article) I was able to find 84 Twitter accounts, resulting in a Twitter account for every 3,725 inhabitants.

Let's have a closer look at the twitosphere in Iceland.

History of the Twitter accounts in Iceland
This graph below shows the history of the Twitter accounts in Iceland. The graph is based on the date of the first message posted on each of the public Twitter accounts in Iceland.


Oldest Twitter accounts from Iceland
http://twitter.com/briansuda - Brian Suda - first message on 22/11/2006
http://twitter.com/gusti - Gusti - first message on 29/12/2006
http://twitter.com/rosastef - Rosastef - first message on 8/01/2007
http://twitter.com/Kiddi - Kristinn Agnarsson - first message on 27/01/2007
http://twitter.com/kjarri - Kjartan Albertsson - first message on 28/01/2007

Public of Private
13 Twitter acounts from Iceland or 15.5% have chosen to keep their updates only available to their friends.

Number of following
A Twitter user can choose to follow one or more other Twitter accounts, called "following". 23% of the Twitter accounts in Iceland do not follow other Twitter users. An average Twitter account from Iceland follows 21 Twitter users.


Top 5
http://twitter.com/bjork - bjork - 992 following
http://twitter.com/Kalli - Kalli - 102 following
http://twitter.com/briansuda - Brian Suda - 92 following
http://twitter.com/oskuskaktus - Ósk Gunnlaugsdóttir - 71 following
http://twitter.com/salvor - Salvör - 60 following

Number of followers
A follower is someone who has indicated another Twitter account as following and receives all updates of that other Twitter account. 24% of the accounts have no Twitter followers at all, meaning that no one else on Twitter has indicated this account as following. An average Twitter account from Iceland has 15 followers.



Top 5
http://twitter.com/briansuda - Brian Suda - 371 followers
http://twitter.com/bjork - bjork - 243 followers
http://twitter.com/Kalli - Kalli - 112 followers
http://twitter.com/oskuskaktus - Ósk Gunnlaugsdóttir - 57 followers
http://twitter.com/salvor - Salvör - 45 followers

Updates
4% of the accounts from Iceland are still waiting for their first update to be published. Currently an average Twitter account from Iceland has 135 updates.


Top 5
http://twitter.com/Kalli - Kalli - 4,058 updates
http://twitter.com/Fk1710 - Frida Kjartansdottir - 1,551 updates
http://twitter.com/briansuda - Brian Suda - 927 updates
http://twitter.com/hthth - Hrafn Th. Th. - 575 updates
http://twitter.com/imoxxx - imo - 497 updates

Tweet frequency
The tweet frequency is the average number of Twitter messages - also called Tweets - a Twitter users produces on a single day. In the list below I have only taken the Twitter users into account who have posted at one message in the last 30 days.

Top 5
http://twitter.com/Kalli - Kalli - 13,18 tweets/day
http://twitter.com/imoxxx - imo - 8,15 tweets/day
http://twitter.com/hthth - Hrafn Th. Th. - 4,83 tweets/day
http://twitter.com/briansuda - Brian Suda - 2,19 tweets/day
http://twitter.com/oskuskaktus - Ósk Gunnlaugsdóttir - 1,76 tweets/day

Degree of activity
58% of all Twitter accounts from Iceland with a public feed did not post an update in the last 30 days. They can perhaps be considered as inactive accounts.

30% of all Twitter accounts from Iceland with a public feed did post at least one message in the last 7 days, indicating that these accounts are still alive.

Conclusion
This state of the Twitosphere in Iceland is only a snapshot, meaning that the figures mentioned in this post have already changed because the Twitosphere is a very dynamic environment with several new accounts being created every day, new connections being made between accounts (following/followers) and of course new messages being written every hour of the day.

Iceland is an country complete surrounded by water. At the same time Iceland is the most developed country (Wikipedia article). Users from Iceland were rather late to discover Twitter (from November 2006). From February 2007 the twitosphere in Iceland grew in a linear way. This results in a very high penetration - compared to other (larger) countries.

The usage patterns shown above are similar to the usage patterns for other countries for which I analyzed the twitosphere (see the sidebar for a list of these countries). The Twitter users from Iceland are strongly linked to Twitter users in other countries such as the USA, the UK and Canada. The most followed Twitter account from Iceland is http://twitter.com/GunnarHafdal, someone from Denmark.

It will remain a question where the Twitter community in Iceland with currently around 30 active Twitter users is large enough to survive.

January 17, 2008

Twitter Technology Blog

The guys at Twitter started their blog more than a year ago (more precisely in August 2008 August 2006). Now they just launched another blog, the Twitter Technology Blog (dev.twitter.com). The developers at Twitter wanted a place to talk in more detail about what's going on under the hood at Twitter. If you are writing an application that talks to the Twitter API or if you are just curious about code, tools, and practices used by the Twitter developers, it might be a good idea to add this blog to your list of blogs in your favorite feedreader.

January 10, 2008

What's the cost of Twitter ?

According to this post from Pat Phelan and Florian Seroussi from Roam4free Twitter would cost the economy around $13,5 billion in 2008. Let's have a look at their assumptions.

750.000 registered users
Today Twitdir reported 753.629 users. These are only Twitter users with a public profile. Twitter users with a private profile (estimated at 15%) have to be added, giving 866.673 users.

450.000 active users
Based on a large number of Twitter users analyzed, I did find out that 45% did not post a single message over the last 30 days, 40% did post at least one message in the last 7 days. 40% of 866.673 gives 346.669 active users.

As I sometimes use Twitdir (and other sources) to look for Twitter users for a geographical location, I regularly encounter Twitter users that do not exist anymore. Twitdir seems to catch (hopefully all) new Twitter users. Twitdir however does not seem to catch all abandoners. So the figure of 346.669 active Twitter users is too high. If the number of dead accounts is 15%, we finally end up with about 300.000 active users at this moment.

60.000 new users each month
In a recent previous post I showed a graph with the growth of the number of Twitter users. Currently there are daily over 3300 new Twitter users according tot Twitdir, resulting in 100.000 new Twitter users each month.

As said before, the number of deleted Twitter accounts is unknown. I guess that Twitter users who are not interested anymore in using Twitter do not always make the effort to actually delete their accounts. They rather leave their accounts alive, but do not post new messages nor read messages from other Twitter users. They belong to the large group of sleeping accounts. Let's suppose that the number of sleeping accounts outnumbers the number of deleted accounts.

Number of messages
The author of Tweetscan announced a while ago on his blog that Tweetscan is capturing almost 100% of all public Twitter messages. Unfortunately he did not mention how many messages Tweetscan is capturing.

In June 2007 I wrote a post to announce that Twitterment (a Twitter search engine that is currently not available anymore) detected 25.000 messages per day. As the number of Twitter users has doubled since June 2007, the number of messages will probably be higher. A safe bet is at least two to four times : 50.000 - 100.000 messages per day.

Another way of looking at Twitter activity is looking at the average number of messages per Twitter account. The average number of messages for the data I collected is 253. Only 5% of the Twitter users have produced more than 800 messages. It is also interesting to have a look at the tweet frequency, the number of Twitter messages written per day. On average the tweet frequency is 2,15 (messages per day). 5% of the Twitter users write on average more than 6 messages per day. Several of the Twitter accounts with an high tweet frequency are non-human, the messages are automatically generated (e.g. news feeds).

27 minutes per day
This amount of time seems reasonable for a heavy Twitter user, with an high tweet frequency. Not all Twitter users are heavy Twitter users, implying that 27 minutes per day is too high.

$20/hour
Although this may seem a fair productivity cost, it is probably only valid for developed countries such as the USA, Canada, the UK or Germany. Wages and labour cost in other countries are much lower. What is the average labour cost for Africa - where there is also a (very small) Twitter user base ? What is the average labour cost for South American countries such as Mexico, Brazil and Argentina with also thousands of Twitter users ?

100%
I do not agree that 100% of all Twitter activity has to be considered as a (complete) waste of time.
- Not everyone is using Twitter during work hours. Already a lot of companies are blocking Twitter forcing their employees to postpone their Twitter activities until they are at home.
- There is practical usage of Twitter, even within a business context, e.g. Remember the milk.
- Monitoring Twitter can be part of someone's job. I personally know several persons who make their living as a web developer, web consultant, social media consultant, online marketing consultant, ... This kind of job implies that you have to keep in touch with recent trends, for instance while using Twitter.


If the cost of Twitter usage to the economy would really reach $13,5 billion in a single year, which company would be interested in buying Twitter ? Especially if the companies who incur these productivity losses would try to sue Twitter for compensations.

January 6, 2008

Number of Twitter users

One reliable source for the number of Twitter users is Twitdir. Twitdir reports the number of Twitter it already discovered on its homepage. As Twitdir is only capable of detecting Twitter users with a public profile, this number is a underestimation of the total number of Twitter users. I have been following the number of Twitter users reported by Twitdir for a couple of months by now. I made a graph showing the number of Twitter users according to Twitdir over the last seven months.


I want to make one prediction for 2007, more specifically regarding the number of Twitter users reported by Twitdir. I expect this number to reach 1 million before April 1, 2008.



By the way, as the percentage of Twitter users with a private profile is sometimes estimated at 10% to 15% Twitter will reach 1 million users even sooner, probably around March 1, 2008.

January 5, 2008

Twitter stats

If you are curious to have a look at some statistics about your own Twitter activity, have a look at the Twitter Stats package developed by Damon Cortesi. The original script only works on Mac, but there are several other versions available now, including a web version. The full code is also available on a Google Code page.


If you happen to use a Mac, concise instructions to obtain your stats are available here.

December 31, 2007

Top 10 posts

As a last post on this blog in 2007, I want to share the 10 most viewed pages on this blog of the past year.

  1. Number of Twitter users, July 22, 2007
  2. 4 Types of Twitter Users, May 23, 2007
  3. Second state of the DutchTwitosphere, August 21, 2007
  4. First State of the Dutch Twitosphere, April 23, 2007
  5. Virtual travel agent on Twitter, August 22, 2007
  6. First State of the Italian Twitosphere, May 7, 2007
  7. Private Twitter accounts visible on Facebook, September 8, 2007
  8. First state of the Twitosphere in Canada, December 18, 2007
  9. First state of the Twitosphere in Brazil, August 15, 2007
  10. First state of the Twitosphere in Argentina, October 4, 2007
I also would like to share the 10 least popular posts.
  1. Twitter : Buzz statistics from public sources, June 28, 2007
  2. Tweeterboard, December 19, 2007
  3. Twitterposter, November 7, 2007
  4. Twitter Down, June 22, 2007
  5. Followers, Friends, Fans, Admirers, Connections, Subscribers, Visitors, ..., September 17, 2007
  6. TwitterWhere, October 25, 2007
  7. The case for Twittter - The case against Twitter, May 2, 2007
  8. Sitevolume, August 10, 2007
  9. Twitter shares, August 14, 2007
  10. Club140, December 6, 2007
I hope that you will keep on reading the posts on this blog in 2008.

Source : Google analytics stats

December 29, 2007

Most popular Twitter bookmarks on del.icio.us

Del.icio.us is a social bookmarking site where users can share their bookmarks. Six months ago I already posted an overview of the number of del.icio.us bookmarks for Twitter related URLs (Twitter itself, Twitter tools, Twitter gadgets, Twitter blogs, ...).

I just had a look at the number of bookmarks for these Twitter related URLs. The list below shows the 10 most popular Twitter related URLs according to the number of del.icio.us bookmarks.

  1. Twittervision : 2866 bookmarks, first bookmark on March 17, 2007


  2. Twitterific : 899 bookmarks, first bookmark on January 15, 2007


  3. Twitter Fan Wiki : 893 bookmarks, first bookmark on November 20, 2006


  4. Twitterfeed : 778 bookmarks, first bookmark on March 22, 2007


  5. Twitbin : 563 bookmarks, first bookmark on May 4, 2007


  6. Foamee : 401 bookmarks, first bookmark on November 9, 2007


  7. Twitthis : 396 bookmarks, first bookmark on March 22, 2007


  8. Snitter : 372 bookmarks, first bookmark on September 4, 2007


  9. Terraminds micro search : 341 bookmarks, first bookmark on Augustus 20, 2007


  10. Twitterposter : 327 bookmarks, first bookmark on October 31, 2007

Do you agree that this list is a true representation of the most popular Twitter tools ?

December 26, 2007

First State of the Twitosphere in India

I published already several blogposts on this blog analyzing the Twitosphere is several European and South American countries (see the sidebar). Last week I had a close look at the twitosphere in Canada. The next country analyzed - the first country in Asia - is India.

Using the same approach as for the previous countries, I managed to find 1631 active Twitter accounts in India. A Twitter account was considered to be held by someone from India if the location explicitely refers to India or if the Twitter user was located in India according to one of the Twitter mapping mashups. The actual number of Twitter users in India is probably higher as only 50% of the Twitter users specify their location in their profile (see this previous post).

For an estimated population of 1.12 billion people (according to Wikipedia), India has 1631 Twitter accounts, meaning that there is a Twitter account in India for every 687,000 inhabitants. Currently for Iceland and the Netherlands there is one Twitter account for less then 10,000 inhabitants. For Canada there is a Twitter account for every 10,014 inhabitants.

History of the Twitter accounts in India
This graph below shows the history of the Twitter accounts in India. The graph is based on the date of the first message posted on each of the public Twitter accounts in India.


Oldest Twitter accounts from India
http://twitter.com/pajama - Pajama Pockets - first message on 28/07/2006
http://twitter.com/harshadsharma - Harshad Sharma - first message on 2/08/2006
http://twitter.com/Kalyan - Kalyan Sagar - first message on 14/08/2006
http://twitter.com/hardik - Hardik Shah - first message on 27/08/2006
http://twitter.com/Dolly - Dolly - first message on 31/08/2006

Public of Private
257 Twitter acounts from India or 15.8% have chosen to keep their updates only available to their friends. In the analyzed European countries this percentage varied from 3.5% (Italy) to 22% (Belgium). In Canada 7.1% of the Twitter accounts were private accounts.

Number of following
A Twitter user can choose to follow one or more other Twitter accounts, called "following". 36% of the Twitter accounts in India do not follow other Twitter users. An average Twitter account from India follows 10 Twitter users. An average Canadian Twitter account follows 23 Twitters accounts. In the analyzed European countries this average varied from 9.6 (Belgium) to 32 (Italy).


Top 5
http://twitter.com/vijay - Vijay - 1,435 following
http://twitter.com/Gauravonomics - Gaurav Mishra - 972 following
http://twitter.com/GharExpert - GharExpert - 799 following
http://twitter.com/regalfinance - regalfinance - 635 following
http://twitter.com/RaviKarandeekar - Ravi Karandeekar - 521 following

Number of followers
A follower is someone who has indicated another Twitter account as following and receives all updates of that other Twitter account. 25% of the accounts have no Twitter followers at all, meaning that no one else on Twitter has indicated this account as following. An average Twitter account from India has 8 followers. This average is far below the average for Canada (25 followers).


Top 5
http://twitter.com/Gauravonomics - Gaurav Mishra - 276 followers
http://twitter.com/drmani - Dr Mani - 247 followers
http://twitter.com/labnol - Amit Agarwal - 213 followers
http://twitter.com/kiruba - Kiruba Shankar - 208 followers
http://twitter.com/abdulqabiz - Abdul Qabiz - 176 followers

Updates
10% of the accounts from India are still waiting for their first update to be published. Currently an average Twitter account from India has 88 updates. This is again significantly lower than the average for the Canadian twitosphere. The top 2 Twitter accounts are non human, meaning that the messages are automatically generated. In other countries I discovered more automated Twitter accounts.


Top 5
http://twitter.com/indianews - India News - 39,831 updates
http://twitter.com/PHP_programmer - onlinejobsforall.com - 10,053 updates
http://twitter.com/bluepotion - Manasa Malipeddi - 2,224 updates
http://twitter.com/aalaap - Aalaap Ghag - 1,827 updates
http://twitter.com/prsn - Prasoon - 1,756 updates

Degree of activity
60% of all Twitter accounts from India with a public feed did not post an update in the last 30 days. They can perhaps be considered as inactive accounts. In Canada the percentage of inactive Twitter accounts was 46%.

27% of all Twitter accounts from India with a public feed did post at least one message in the last 7 days, indicating that these accounts are still alive. In Canada the percentage of active Twitter accounts was 38%.

Conclusion
This state of the Twitosphere in India is only a snapshot, meaning that the figures mentioned in this post have already changed because the Twitosphere is a very dynamic environment with several new accounts being created every day, new connections being made between accounts (following/followers) and of course new messages being written every hour of the day.

Users from India were among the first to start using Twitter (from July 2006). From March 2007 the growth in the twitosphere in India accelerated. The growth of new Twitter user accounts in India remained at the same speed from March 2007 until now.

The degree of activity in the twitosphere in India is lower than in the other analyzed countries. The percentage of active Twitter users, the number of Twitters users followed and being followed, the number of updates (Twitter messages published), the number of non human Twitter accounts are all significantly lower compared to Canada and several European countries.

For a country with a huge population (estimated at 1.12 billion), the number of Twitter users is surprisingly low. Especially because India is known for its IT workforce. There are a lot of IT specialists from India working in the USA and in Europe. I have the impression that the IT workforce within India does not use Twitter to communicate intensively with their colleagues working abroad.

December 19, 2007

Tweeterboard

Tweeterboard is a way of looking at who is influential on Twitter based on their conversations with other Twitter users. Services showing how many followers a Twitter user has and how many Twitter users a Twitter user is following have been around a long time. Examples of these services are Twitterholic, Twitterposter, Twitdir and Terraminds Twitter Search.

The basic idea of Tweeterboard is that conversations - which Twitter user talks to other Twitter users - is according to the people behind Tweeterboard a better measure of influence and engagement than followers, friends and update frequency. The reputation of a Twitter user is calculated based on who talks (links) to this user. Tweeterboard is currently aware of about 2600 users, of which 1500 or so are active Twitter users with public updates.

You can have a look at the top 100 list published by Tweeterboard. This list is still pretty volatile.



Tweeterboard has implemented a very interesting concept, although it is unclear how the metrics are derived. As soon as there is more clarity in how the metrics are calculated, it is quite likely that some Twitter users will try to score high in the ranking (using SEO-like tactics).



2600 users is only a small subset of the complete twitosphere. Twitdir reported today 685,000 Twitters users with a public profile. As shown in a previous post on this blog a long tail effect can be observed in the twitosphere. Only a small subset of users have high values for the traditional metrics (following, followers, updates).

Twitter users have sometimes good reasons to keep their messages private. Twitter offers the possibilty to its users to declare their account private. Only accepted users are able to read the messages of private Twitter accounts. The private nature of a specific Twitter account does not exclude that this Twitter user can be very influential within his community of followers. Tweeterboard is only able to trace the public influence of Twitter users. Replies from public Twitter accounts to messages from private Twitter accounts are captured. The original messages however remain inaccessible to Tweeterboard.

Twitter is a global community. At the same time there are numerous subcommunities within the global twitosphere. Each of these subcommunities has its own dynamics. You can get an idea of this dynamics by browsing thru the different states of the twitosphere by country I already published on this blog (see the side bar). The subcommunities - often by topic, by language or by country - have members who are also connected to other subcommunities, resulting in subcommunities that are often interlinked. Examples of subcommunities are the Twitter users from the Netherlands (see which Twitter users TwitterNL is following), the Twitter users from Argentina (see Tuitiar), the Spanish speaking Twitter users (see TuitHispana), the Twitter users from Madrid (see Twittmad).

It would be interesting to see the influence rankings not only on a global level - as presented on the current top 100 page - but also by subcommunity. Companies only active in a specific geographical location (a continent, a country or even on a lower level) are not always interested in who's influential on Twitter on a global scale, but rather on who's influential with their geographical location and perhaps also domain of activity. It will be a real challenge to draw clear boundaries around subcommunities.

One final remark, the developers of Tweeterboard still have to fix some loose ends. While trying to find out the metrics for specific users (within my subcommunity), I sometimes got error messages.