Drive-by Web, Aggregate Contributions, and Ad Hoc Communities

I was listening to Clay Shirky on TechNation regarding his new book (Here Comes Everybody), and one of the topics he discussed was Wikipedia and the power law distribution of contributors. Clay has been down with power laws and the web for a while, although in his discussion he regarded the controlling contributors as having a special interest, and those in the long tail as drive-by contributors. He also commented on the importance of both groups, and the business perspective of only keeping the top 20% contributors and discarding the lower 80% to maximize a cost benefit equation.

Clay's comments got me thinking about what you may refer to as a drive-by web, where contributions are made by users without a special interest, that are likely content to passively graze information for most of the time. A first pass at the consideration suggests at Q&A forums and comment systems. The former case highlights that the power-law effect is likely just as applicable to mediums with an ad hoc context like someone asking a question, and the latter highlights the pervasiveness of the approach regarding web artifacts (new articles, photos, videos, bookmarks, etcetera).

In the Wikipedia, a wiki page provides a static context that promotes the formation of special interest (highly motivated contributors that take ownership). The example also highlights that with useful tools (like review and reversion), drive-by users can be integrated with special interest users toward a democratic consensus. Importantly, this second point highlights the value of the ongoing consensus process, which is the resultant aggregation of contributions: a wiki article. The examples of a forum and a comments thread for a web-artifact may or may not reach a consensus, although the information is not provided in an easily consumable form (it is linear and elaborated). I am sure that a lot of interesting and useful information is accumulated in Usenet, forums, and comment threads (Slashdot was great for this), although the aggregated value of such information is typically realised. For example, you have to reach a consensus yourself using groups search (all manner of google) or reading a moderated perspective of comments based on popularity and or reputation.

I like those websites that cater for the drive-by web as they provide excellent procrastination tools. For example, raw drive-by such as grouphug and artifact drive-by like social news. This "mostly read, occasional participation" behavior used to be called lurking in the old web days and was discouraged, in fact I suspect that it is a social-human thing used when we are outside our interest or ownership zone (area of expertise). These examples further highlight that good tools can exploit the value in the aggregation of the drive-by user contributions (such as the occasional submission or up-modding of an artifact) as a service or resource.

The pattern may be described as the capture of anonymous or semi-anonymous (user context means little) drive-by contributions such that a services can be provided that exploits the value of the aggregated (emergent) information. For example, in the case of the IWish and AskTheMob web application proposals directly exploit this effect.

Another example I have been rolling around today is called "I did, I want" which basically provides an advice asking/providing service that exploits this pattern. The important difference (coming from the highlighted use of bottom-up algorithms in collective intelligence) is the mapping between the two, likely using natural language and keywords. Obviously the matching does not have to be one-to-one, the result of which creates a real-time micro-community focused on questions and experiences that directly relate to the user (advice and adviser match maker, think dating sites again). It notion is not limited to advice, it could be skills/jobs like sourceforge or linkedin, problem solving/solvers, etc.

One may extrapolate this further, and consider any search query a drive-by contribution where the results represent Google's efforts for creating an ad hoc community for you in the form of pointers to web content. Reminds me of a service (I forget the name) where you could comment on (I think it was ad hoc IM communities) for any web page through a frame (or something). The nature of participation (live or delayed) in the community does not matter so much, although what does matter is its permanence. The longer it persists, the more likely experts are to locate and participate, like a wiki page or special interest website. Search results have no permanence, although such an artifact could be created for high-frequency searches (I guess Wikipedia addresses this for nouns).

In the 'I did, I want' example, cases of high-frequency/high-probability-of-match advice could be automatically identified and abstracted as an independent artifact to permit ongoing evolution such as further divergence or dissipation. Most importantly, these ideas integrate notions of drive-by users implicitly improving the 'adaptation of fit' of the application (data consumption and participation), and the exploitation of aggregate contributions toward the formation of ad hoc (and potentially persistent) communities.