A Week in the Life of XBRL Twitter

Written by Bob Schneider     Posted on September 26, 2009

Opinion of the microblogging service Twitter falls between two extremes. The “complete waste of time” view is captured in the Verizon commercial where the zombie-like Dad is slumped in a chair, tweeting “I am sitting on the patio…” and every useless thought in his head. The opposing “truly innovative” vision is represented by the brave Iranian students whose tweets propelled a nation to the brink of revolution.

Both image are facets of the truth. But for those who want to participate in the XBRL conversation, it’s the picture of vigor and purpose that is relevant, even if the consequences don’t rise to the level of political insurrection. Sure, there’s the occasional tweet by a 20-year-old of “I’m still WASTED from last night and my XBRL paper is due in TWO hours. HELP!” But as a whole, the tweets that discuss XBRL represent reliable news and useful commentary on the interactive data world.

The 140-character limit of tweets defines the essence of the service, and it’s Twitter’s best and worst attribute. The required brevity concentrates the mind wonderfully and makes for quick reading. But some thoughts do require more than 140 characters, and by the time you include your username -- and quite possibly a web link, a reply address, and an RT for retweets -- there isn’t much room left over for content.

One enormous positive of Twitter that is perhaps not universally known, however, is that its database is updated continually in real-time (or thereabouts). This is in striking contrast to Google, which, even with diligent pinging, may not index your blog post for several minutes, a few hours, or even a few days.

I recently did a small study to examine XBRL tweets in some depth. I had two main questions: (1) Who tweets about XBRL, and (2) What XBRL topics do they tweet about.

Using Twitter search, I retrieved all tweets between August 23 and August 29 containing the keyword XBRL, reviewed the tweet, and classified the content by subject (eg, SEC filings, UK taxation, meeting/event). I also recorded each tweeter, reviewed their bio on their Twitter homepage, and classified them by occupation (eg, accounting, marketing) or, where more appropriate, function (news feed, XBRL organization, etc.).
 
As an attempt at a rigorous study of XBRL tweeting, the shortcomings of my work are multiple and substantial:

• Naturally, only public tweets are included, no private ones
• It's highly likely there were a few XBRL-related tweets that did not include keyword XBRL (such as a reply to another tweet, where the XBRL subject was understood)
• Twitter search (formerly Summize) has a less-than-sterling reputation and is known to miss a few tweets. I did compare my hits to those from one of the independent Twitter search engines, however, and didn’t find significant differences
• The one-week sample was taken during late August, a quiet, semi-holiday period
• Whenever the study is done, a one-week sample will be greatly influenced by what is being published on the Web at the time. For example, during the week an interview with Michelle Savage of XBRL US in which she gave advice to IR professionals was tweeted and re-tweeted several times, boosting the IR content category
• Classifying both tweet (for content) and tweeter (for profession or auspices) in single categories was naturally sometimes difficult. When I tried to classify myself as a tweeter, for example, I had trouble choosing among several categories

With all those caveats, I still think the results provide some indication of who tweets and what they tweet about. In all, I looked at 143 tweets. Thirty of the 143 were retweets; but that number is understated, since I only counted a retweet if it included the characters RT (other tweets similarly repeated tweets without saying so).  Ninety-eight of the 143 included hyperlinks, which indicates well that XBRL tweets often recommend resources on the Net.

The content of the tweets can be broken down as follows:

Tweets by Content Category
Category	Number of Tweets
Australia	7
Corporate Actions	1
Education/Course	2
Financial Reporting (not otherwise classfied)	2
Foreign (Non-English)	7
IR	10
Job Opening	3
Meeting/Event	10
Municipal Government	1
RDF	5
SBR	3
SEC (general)	1
SEC (XBRL filings)	12
Social Media	1
Software	4
Student Shout-Out	1
UK Financial Reporting	6
UK Taxation	2
US GAAP Taxonomy	4
Web Tools	3
XBRL General	45
XBRL Company	2
XBRL GL/Internal Reporting	6
XBRL-Dedicated Journal	1
XBRL.org	4

The high count of 45 for XBRL General reflects a few broad articles that were published during that time about XBRL, including CFO.com’s XBRL: The Inside Story, which was tweeted and retweeted often (and which could have easily gone in the "internal reporting" category).

XBRL tweeters represent a wide variety of backgrounds, as shown by the breakdown of the tweets by tweeter category:

Tweets by Tweeter Category
Category	Number of Tweets
Accounting	9
Education (eg, professors)	4
Employment (eg, recruiters)	4
Financial	11
Government	2
IR	14
Marketing	2
Media (periodicals, etc.)	16
Miscellaneous	3
News Feed	17
PR	12
Student	1
Technology	13
Unknown	11
XBRL Institution	6
XBRL Maven	9
XBRL Provider	9

As I said, these data should be regarded with more than a pinch of salt, but it still demonstrates the range of background of XBRL tweeters. I think the one (slight) surprise for me is that there are few, if any, security analysts tweeting about XBRL. I can think of several reasons for this – not least that brokerages have strict rules about external communications – but I still find the lack of participation disappointing.

In total, there were 74 tweeters of XBRL during the period. The categories with the highest numbers were technology and media. But the overall distribution was relatively even. (XBRL Maven is a category I used for people like Eric Cohen and Neal Hannon who are leaders in the field.)

Tweeters by Tweeter Category
Category	Number of Tweeters
Accounting	6
Education	3
Employment	3
Financial	6
Government	1
IR	5
Marketing	2
Media	9
Miscellaneous	3
News Feed	4
PR	4
Student	1
Technology	9
Unknown	8
XBRL Institution	1
XBRL Maven	3
XBRL Provider	6

It will be interesting to do this study again in a few months for a different time period to see how the data have changed. In the interim, I will write an article soon that addresses the larger topic of XBRL and social media.