In an excellent posting titled “Building a better collective memory”, Michael Nielsen makes the point that science currently lacks the ‘trust infrastructure’ and incentives necessary for free, unrestricted trading of questions and ideas. Imagine two scientists; each has information that could benefit the other more than it benefits themselves. In an ideal world, they’d exchange this information, and both would be better off. This is the concept of ‘comparative advantage’. Unfortunately, in the real world these scientists:
-Will probably never meet in the first place
-If they should happen to meet, they won’t likely talk about the relevant gaps in their work
-Even if they discuss their needs, they don’t have any basis on which to trust each other enough to engage in collaboration
Michael envisions an ideal “collaboration market” that will enable the open (or at least productive) exchange of ideas. This engendered lots of interesting debate, mostly about why none of the existing collaboration sites, publication archives, and the like are NOT fostering this type of exchange. Since we’ve been thinking about this problem for some time at InnoCentive, I thought I’d share some perspective on what characteristics we believe a collaboration platform needs to be effective.
I propose that at least three elements need to be present to drive effective collaboration in a community:
First, there must be a well-defined artifact around which to collaborate. This defines what you’re going to collaborate about. In Science Advisor, for examples, the artifact is a scientific paper. On eBay, the artifact is the item for sale. In open source software, the artifact is the code. On InnoCentive, the artifact is the “Challenge” posted by a “Seeker” – essentially a problem statement.
To be useful as a subject for collaboration, an artifact needs to be very clearly defined. A scientific paper, a Pez dispenser, a block of C code – these are all very concrete and specific things worthy of comment, purchase, or debugging. But an “idea” or a “topic” – these are often not crisply defined enough to foster focused, useful interaction. Collaboration on ambiguous things often descends into the cacophony of random shouting and discord that accompanies most political blogs.
Second, an incentive to collaborate is essential. Why should I review your paper? Why should I debug your code? There needs to be something in it for me – reward, some form of recognition, or at least personal satisfaction. At InnoCentive, we’ve found that participation goes through the roof when you can combine all three factors together: some currency I can spend at Wal*Mart, the opportunity to be seen as an expert by my colleagues, and a chance to positively impact the world. A generic “certificate of participation” for posting comments doesn’t cut it. Real, valuable incentives are required, tied to visible and (ideally) measurable contribution to understand/improving/solving the artifact.
When artifacts meet incentives, a transaction can occur – an incentive is awarded for contribution related to the artifact. I believe strongly that Michael’s term “collaboration market” is perfectly chosen. Markets are where value is exchanged via transactions. Without discrete transactions, it’s much more difficult to understand the value of what’s been exchanged. Virtually all of the examples cited in the discussion around Michael’s ideas lack the concept of a transaction. For a market to work, value has to be judged fairly, and fairly consistently. This isn’t easy and deserves further discussion; it’s a likely topic of my next posting.
Finally, there must be a mechanism to establish trust. Fortunately for our thesis, considering our platform as a market with transactions gives us a very workable answer. People participate in collaboration, and incentives are awarded, creating a transaction. Just like on eBay, the participants in a “collaboration transaction” should be able to rate each other – but not on “intrinsic value”. The magic that allows the eBay rating system to work so powerfully is that each rating is applied to a person’s role and behavior in a transaction, not to the person more generally. On eBay, you would say “The bidder didn’t pay me for the Pez dispenser he bought” – an objective comment, versus “He’s not a good person to do business with” – possibly true, but subjective and therefore less useful.
In the scientific community, one can establish technical credibility through a history of publications, but one can’t really establish trust that way. Both are critical to successful collaboration. In an effective market, each potential collaborator should have a visible badge next to their name indicating their history of ratings by their past collaboration partners. A market with this mechanism could provide fairly clear guidance on whether a particular participant is worthy of trust. (Bootstrapping such a rating system is an acknowledged challenge.)
If we put these three elements together into a market-based community platform, we have at least a chance of getting our two hypothetical scientists together and enabling productive collaboration.
-They would come to the market and identify an artifact of common interest. So they could meet, and actually have something specific to talk about
-Because they’re both interested in the incentive, they have a reason to disclose their individual ability to contribute, and the gaps they need to fill
-Since they’ve been rated by their peers based on their previous participation in collaborations, they have at least a basis from which to begin a trusted partnership
Lacking any of the three critical elements tho, these two ships would most likely still pass in the night.
At InnoCentive, we’re developing these ideas in anticipation of enhancing the ability of our Solvers to collaborate on Challenges. Please share your input on this rich and important topic, and we’ll keep the dialog rolling.