Prizes, pieces, and property rights
David Leonhardt’s Economix column today suggests that prizes may be a useful way of stimulating innovation. His primary example is Netflix.com’s million dolar prize for any one that can improve its movie suggestion algorithm by 10%. The current best team is at 6.75 percent. Netflix may be better positioned than most companies to be able to offer prizes that will provide the winners reputational benefits that may exceed the value of the prize itself, but Leonhardt may still be right that prizes are an often overlooked means of accomplishing corporate goals.
The Netflix.com prize is an example of “crowdsourcing.” The web site Innocentive contains many such crowdsourced offers. If, for example, you can develop a pressure sensitive adhesive for re-sealing flexible bags for salty snacks (the adhesive must not adhere to potato chips or hands reaching in to take out chips), you can make $50,000 (plus the eternal gratitude of me, a Doritos connoisseur). An advantage of this approach over conventional sourcing is that the project sponsor does not need to assess the quality of those who may work on the project.
Why do we not see more crowdsourcing via prizes? See my comments after the jump.
A significant drawback of prizes is the possibility of redundant work. There is no incentive for participants to share information about failed efforts or partial progress. Even if I am sure that I can achieve a solution in a race for a $50,000 prize, if I estimate that about four other equally competent solvers will also be racing, then the prize will only be worth about $10,000 to me in expected value terms. At least in this case, it would be cheaper for the entity needing the solution to enter into a contract with a sole solver — perhaps providing $20,000 conditional on a solution being produced.
Could crowdsourcing via prizes be improved to the point where it is generally cheaper than alternative approaches, so that even a company like Microsoft might start relying on subsidized peer production rather than on its own employees to create software? I think that is possible. We have already seen that peer production without subsidies can produce large well-integrated products, and subsidies should allow for even more impressive ones.
But there are two key obstacles, if redundant work is to be minimized. First, the sponsor of the prize must promise to divide the prize in proportion to achievement by participants, and must be able to convince potential participants that this will be done in a relatively efficient way. As long as showing that one seemingly attractive approach won’t work counts as an achievement worthy of some prize money, some redundancy can be eliminated. Moreover, dividing prizes into pieces allows for greater specialization by the sourced crowd. My own view is that it should be possible to create sufficiently independent judges of contributions that this is a surmountable obstacle.
Second, there needs to be some type of at least tentative property rights. Competitors might work on essentially parallel solutions simultaneously (for example, crafting modules of computer code that perform essentially the same function), unless there is some means of allowing one competitor exclusive rights. If the sponsor of the research can delineate property rights, some type of auction might work; alternatively, the sponsor might grant rights to the first party that proposes to work in a particular area upon submission of a property claim. I have not seen much attempt at solving this problem, but real innovation is possible here.
In the end, crowdsourcing via prizes could rival existing industrial organization structures, but it may require a fair bit of careful supervision. The ultimate question, as Ronald Coase’s Theory of the Firm suggests, is whether the costs of that supervision are less than the costs associated with the organization of hierarchical firms. Perhaps the small role of crowdsourcing in today’s economy suggests that they are not. But if information technology continues to reduce the costs of controlling crowdsourcing, that could well change.