Last week I participated in a very interesting conference focused on innovation. In particular, attention was given to what drives innovation and what policy decisions could be made to maximize it. As is clear to anyone who lives in the Bay Area, innovations cluster geographically and all sorts of correlations can be made in a post hoc way. Note, however, that opportunities for correlation/causation errors abound. For example, patent lawyers cluster around areas of high innovation but few would assert that patent lawyers cause innovation -- Ed Felten made the observation that it would be like asserting that ants cause picnics. Potentially, if you buy into Wisdom of Crowds there might be an argument that monitoring the flow of patent attorneys might be a useful leading indicator -- that collectively the set of patent attorneys are aggregating knowledge about where the clusters are emerging -- but that is a topic for a different post. I want to talk about innovation, specifically two aspects of innovation where Second Life enjoys significant advantages relative to the real world. Both of them are Long Tails.
I hope that by now everyone is familiar with Chris Anderson's concept of the "Long Tail." If not, go over and read his blog and preorder his book. Chris' famous graph is more properly thought of as "Long Tail Consumption," namely that the total value of goods and services of below average popularity bought by consumers is a meaningful percentage of the total value of the popular items. Like most great ideas, this shouldn't be a surprise since we know from Clay Shirky that choice drives power law creation and that more choice creates steeper power laws. These power laws have the non-intuitive properties that most of the items fall below the average value and in the examples that Chris sites -- books, music, and movies -- there are huge numbers of choices. Thus when Wal-Mart decides to only stock the "most popular" music, for example, they are ignoring a huge chunk of the market Rhapsody or iTunes are then able to capture.
I want to look at two additional Long Tails that factor into innovation: Long Tail Communities and Long Tail Innovation. Communities first. As anyone who has used Second Life can tell you, one of its strengths is the ability for relatively small communities -- stroke survivors, for example -- to find each other online. In this way, it is somewhat like the web where the online experience mitigates the geographic clustering that can make it hard for communities to find each other. In cities we find diverse communities but we don't find them in the countryside because population densities are lower. The web allows those communities to bypass geography. However, Second Life has the ability to look more like a city then the web, because those disparate groups still interact with each other. These are Long Tail Communities -- small population communities that exist within a framework that allows interactions between them.
When Dreams/Shockproof began building Dreams Island for stroke survivors, they were able to reach out to an amazingly broad degree of groups within Second Life for help. Members of those other groups came to Dreams and met, talked, and interacted. These interactions, these bridges built between diverse sets of knowledge, experience, and culture, form the basis of Long Tail Innovation. The rate of innovation, as North suggested, is most strongly influenced by the cost of learning. Throwing money at the problem can change the direction of innovation, but the cost of learning -- of acquiring new knowledge as diverse groups and individuals interact -- drives the rate. Thus, maximal innovation requires maximal learning between as much diversity as possible. Where Second Life excels, and certainly will exceed the real world, is both in the breadth of diversity and the cost of learning. Breadth of diversity because Second Life embraces not only individual diversity, but allows those individuals to reside in multiple Long Tail Communities. Cost of learning because bits will always be easier to manipulate than atoms.
A different way to think about these advantages is to think of innovation as an initially random walk of design space. Somewhere in design space an innovation exists that will change the design space itself, as well as all subsequent searches into that space. While it is tempting to liken this to Darwinian evolution, be careful because innovators -- unlike genes -- are able to observe the fitness landscape and to move off of local maxima during their search. Moreover, extensive communication between innovators -- remember that cost of learning -- also changes optimization strategies. However, Darwinian or not, innovation is still a search of design space, and if it is even mostly true that it is initially a random walk, maximizing agents, communications, and variety of search strategies are all good heuristics.
So, whether you think about innovation's Long Tails or random walks, Second Life residents bring huge advantages. I can't wait to see what they do next.