Power Law of Participation

Social software brings groups together to discover and create value.  The problem is, users only have so much time for social software.  The vast majority of users with not have a high level of engagement with a given group, and most tend to be free riders upon community value.  But patterns have emerged where low threshold participation amounts to collective intelligence and high engagement provides a different form of collaborative intelligence. To illustrate this, lets explore the Power Law of Participation:

Most of Chris Anderson's Long Tail examples have focused on models of consumption, not production, where intelligence is largely artificial.  Amazonian algorythms guide users down the long tail from Britney Spears to Nobodys, made available without the constraints of shelf space.  But the interesting question is will the tail wag? Can users discover their own power together to either discover something great, or even create it?

As we engage with the web, we leave behind breadcrumbs of attention.  Even when we Read, our patterns are picked up in referral logs (especially with expressly designed tools, like Measure Map), creating a feedback loop.  But reading alone isn't enough to fulfill our innate desire to remix our media, consumption is active for consumers turned users.

Digg is the archetype for low threshold participation.  Simply Favorite something you find of interest, a one click action.  You don't even have to log in to contribute value, you have Permission to Participate.  Del.icio.us taps both personal and social incentives for participation through the low threshold activity of tagging.  Remembering the URL is the hardest part, and you have to establish an identity in the system.  Commenting requires such identity for sake of spam these days and is an under-developed area.  Subscribing requires a commitement of sustained attention which greatly surpasses reading alone.  Sharing is the principal activity in these communities, but much of it occurs out of band (email still lives).  We Network not only to connect, but leverage the social network as a filter to fend off information overload.  Some of us Write, as in blog, and some of us even have conversations.  But these are all activities that can remain peripheral to community.  To Refactor, Collaborate, Moderate and Lead requires a different level of engagement -- which makes up the core of a community.

The byproduct of use is a Conucopia of the Commons -- the act of using the database adds value to it.  As users engage in low threshold participation (read, favorite, tag and link) we gain a form of collective intelligence.  But it is important to distinguish the value of collective intelligence and collaborative intelligence, as first pointed out by Mitch Kapor:

...Tons of interesting types of collaborative filtering, like Digg, is TiVo like, indicating individual preferences, with some algorythm logic.  Valid and interesting, but people are not connecting.  Different from a bunch of people focusing on creating something.  That is higher value than collaborative filtering, my thesis, if you can get people to work together.  Look at health information, broadly speaking, why are doctors not collaborating to build such a resource -- the lack of information, locked up in a database that Harvard publishes, kills people.  I can feel the opportunity... 

When users participate in high enagement activities, connecting with one another, a different kind of value is being created.  But my core point isn't just the difference between these forms of group intelligence -- but actually how the co-exist in the best communities.

In Wikipedia, 500 people, or 0.5% of users, account for 50% of the edits.  This core community is actively dedicated to maintaining an open periphery.  Part of what makes Flickr work isn't just excellence at low threshold engagement, but the ability to form groups.  Participation in communities plots along a power law with a solid core/periphery model -- provided social software supports both low threshold participation and high engagement. 

Comments

[Kira Wampler]

Again, another excellent post. I love the framework you've put together on collaboration as I was just trying to have this conversation with a co-worker! You did a much better job than I did...keep up the great work.

[Peter Childs]

Encouraging participation is the critical factor in the success of social software - especially online communities. As you’ve pointed out there is a natural progression of behaviors, and that as those behaviors reflect increased active participation the value rises – both to the individual and the community.

Designers need to encourage participation by recognizing and taking advantage of this progression and by ensuring their applications have multiple steps that are easy to use and inherently rewarding. Encouraging ranking and rating, which generally requires a single click and results in an anonymous aggregated result is a training behavior for contribution. Tagging can be similarly anonymous but requires a formal and unique contribution raising the level of engagement.

Using these techniques allow a designer to provide multiple contribution points – each of which builds the habit of contribution while not stretching users to far from their comfort zone.

[Uri Baruchin]

A very interesting move, inverting the long tail from consumption to production.
Not sure it's still "the long tail" but the power law you describe makes sense.
However, i think there's one aspect of your model that is slightly misleading, which is that nice curve it shows, while from my experience with online communities, the consumption/production scale tend to be more dichotomous for most users, i.e. a "all or nothing" issue, with reader-participant relation between 1-10 1-8 in high participation communities and a lot worst in most.

[Jon Collins]

Very good. I'd like to draw the honorable gentleman's attention to "my" (no doubt drawn from many sources) laws of connectedness:

http://www.joncollins.net/wordpress/?p=161

Feedback welcome. All the best!

P.S. Isn't that a long head?

[Edward Vielmetti]

Ross, you're missing a whole huge piece of this picture. Where in this diagram is people talking to each other face to face in person about things they see on the net? Perhaps you don't see it because you don't measure it, but not all of these interactions around ideas are out in online space.

[Dennis Howlett]

It strikes me that Ross has painted a picture that reflects what really happens in life. We form hierarchies. Which is validated in the notion of the A...Z-list blogger and by the number estimated to be truly ative in Wikipedia's development. So what exactly have we achieved that provides a benefit to the business?

Is it enough that a reader should take something away, think it through for their own situation, implement but never return to the hive and tell anyone? But there's the rub, with a flipside:

I've recently seen sites pop up that are talking about ideas that vaqguely resemble my own. Great. But they fail because you cannot replicate my thinking any more than I can yours.

If you accept that as a valid consrtuct then it is only a small step to saying that collaboration and active contribution make sense for low impact users because they need to learn just as much as others.

[Dennis Howlett]

If you add in the 'traditional' long tail diag you end up with a pyramid - just like we have now in management structures.

[Sam Rose]

Ross, I think that some of the theme of what you are describing here shows that successful peer production systems tend to be hybrids of hierarchical (top-down) and decentralized (bottom-up) social organization. A large group of decentalized contributors coupled with a smaller core community who are generally organized around some form of hierarchy. This core group tends to refine the work of larger decentralized masses. Yet, there seems to be a sustainable tension balance that emerges between the decentralized and centralized in these successful systems. There seems to emerge a general recognition of the need for the small core hierarchy. And, a general understanding seems to usually emerge that the small core hierarchy is only a tool for time saving and efficiency.

[Dawn Foster]

This reminded me of a number of studies on open source that support this idea. For example, a case study of the Apache project published in 2000 found that 80% to 90% of the submissions came from a set of 15 core developers in a community of more than 3000 people. A study of the GNOME project had similar results with 11 people contributing most of the output. Relating this back to the Power Law of Participation, the small number of core community members leads to collaborative intelligence, while the larger community provides an important collective intelligence by contributing bug reports, ideas, and comments. These two types of contributors and the resulting intelligence generated both feed off of each other and allow the community to prosper. I would be interested to see how this applies to other communities.

Sources:
A case study of open source software development: The Apache server (Mockus, Fielding, & Herbsleb, 2000).

Effort, co-operation and co-ordination in an open source software project: GNOME (Koch & Schneider, 2002).

[David Gerard]

FWIW: The curve on English Wikipedia does in fact seem to be a smooth curve from the occasional readers all the way up to the editors with 100,000 edits.

[Mike Ellsworth]

Ross,

I love the power law graph and could like to use it in presentations. Any chance you have a large size, suitable for PowerPoint?

Thanks.

Mike Ellsworth
Program Manager
CareerOneStop

[Ross Mayfield]

Try this one, Mike
http://www.flickr.com/photos/ross/171420476/