Anatomy of a Tweet

Across society – from business and media to politics and international affairs – we have heard calls to embrace Web 2.0 technologies and social networking behaviors.  Fueled by the fact that the Web offers distinct lessons in discovering, contextualizing, and creating information, the movement towards "2.0" is associated with transparency, efficiency, dynamic contexts, and empowerment.  It remains worthwhile, however, to maintain focus on what drives the best of the Web and makes certain concepts and capabilities successful.

As popular Web applications permeate mainstream discourses and enter into the common vernacular (remember: "unfriend" was the Oxford 2009 Word of the Year, and "hashtag" a runner-up!), policy and technology decisions regarding the next generation of information management should not be clouded by a fleeting love for the latest and greatest apps in order to take significant steps toward leveraging "the wealth of networks."

Twitter has received an amazing amount of press recently and has become a canonical example of emergent behavior on the Web.  Beyond the buzz, what makes Twitter tick? What is interesting and unique? What does Twitter enable and what does it limit?

Some observers pigeonhole Twitter as a noisy channel to peoples' streams of consciousness: from yogurt preferences, to musings on TV soaps.  More importantly though, we think Twitter offers valuable insight into the nature of information flows in online communities  and hints at some of the compromises that must be made in reifying social behaviors in digital environments.  Beyond Twitter's simple concept of sharing brief, SMS-length messages with a virtual social network, the user community has adopted a variety of practices that, in sum, provide rich context to notifications that, as snapshots, may be insignificant.

Let's consider the case of the "retweet," the practice of re-posting or referencing another participant's contribution. As we shall see, a single re-posted message is essentially a data-laden piece of content: discoverable, indexible, and placed in context by a rich social graph.  This, hints at what will distinguish the future of the Web from even the best current examples of "Web 2.0."

Sharing and resharing information is central to the Twitter experience, and a number of behaviors have taken hold across the community.  The "retweet" is perhaps the most familiar of these, but we often see "via" or "h/t" – a virtual tip o' the hat. In any case, the online courtesy of referencing sources is well established.  So, if I want to share one of your messages with my followers, I might copy it with the addition of "rt @you" at the start of the message, indicating that mine is a retweet of your message.  Similarly, if you've pointed me to a page that I find particularly worthwhile, I might post a link, followed by "via @you" or "h/t @you," indicating that you are my source.  Such online courtesy appears not only on Twitter, of course, but is often the mark of a healthy virtual community. It needn't signify agreement or endorsement, but indicates a willingness to take part in an exchange.

The retweet is so established on Twitter that the service recently made a change to the platform functionality to implement the practice as a feature.  That is to say, with a single click of a button, users can now add other's messages to their own timeline.  This decision was not without controversy, since it represents the codification of an otherwise organic behavior. Whereas users often annotate retweets with their own observations, the new feature only permits cloning the original message.  Since there is no right or wrong way to engage with the Twitter community, just as there are a number of different functions that lead in a similar direction, participants continue to use both the original and the implemented retweet. For simple forwarding, the new model works well, but the community still has the desire to edit, comment, and find its own ways of making the most of the system.

Traditional approaches to organizing, categorizing, and applying meaning to the virtual explosions of data on the Web simply cannot keep pace. The increasingly fluid nature of information does not lend itself to the kind of categorization embodied by the Dewey decimal system, and forcing a fit where there isn't one is at least counterproductive if not impossible. Similarly, complaints over "noisy data" often stem from attempts to grasp relevance a priori, out of context, in an idealized way.  On the other hand, if information is accepted in advance as more of a flow than as a static "thing," then the various contexts of the information environment (constructed and reconstructed through social filters, real-world and digital associations, shared interests, etc.) start leading in the direction of "the relevance finding the user."

There are obvious differences between public social networking platforms and the information needs of large enterprises or government agencies. Where personal interactions don't follow a script, organizations rely on a measure of process to achieve their ends. This highlights the significance of institutional contexts, however, and so the contrast is between environments rather than between free-form and automated behaviors. We can still draw upon observations of real-time exchanges and bursts of messages that services like Twitter have enabled. We attempt to conceptualize what dynamic information flows (rather than emails or document archives, for example) may look like across distributed enterprises and how these can improve shared awareness and institutional knowledge.

The model of online communication that most of us are still most familiar with is the one-to-one, or one-to-many, implementation of email.  It's simple, it's easy to reach the people we wish to engage (or: it's easy to reach out and drop something into their inbox) and it's ubiquitous. The challenges associated with "crossing the chasm" of adoption have long since been overcome.  On the other hand, anyone with an ear to the ground may have noticed that complaints over lost information in bloated email inboxes have themselves become clichés of the digital era.  Where email embodies a "push" model, services such as Twitter rely primarily on publish/subscribe, or "post and smart pull."  Thus, the ability to add content to a generally accessible stream of information, and consume from it as needed, defines Web-native communication.  Email is not equipped to support such behavior, though it is often pushed to do so simply because of the dynamic nature of information. It is a technology that is "seriously overloaded and has been co-opted to manage a variety of tasks that it was not originally meant to support."  That conclusion was published almost a decade ago.

This all may seem trivial, but if enterprises had the means to connect all of their communications – instant messages, emails, text messages, and even phone calls –  in an accessible flow of corporate knowledge and could index such interactions with formal as well as emergent taxonomies, we may be able to take significant steps toward end-running the problem of information overload and actually start making sense of our growing volume of digital information in real time.

Let's just say a little birdie taught me that.