A winning idea for innovation

One encouraging piece of news on the procurement front in the past few years has been the use of contests by some agencies, such as NASA and the Air Force, to buy solutions to problems that require innovation. In a traditional procurement process, the government chooses a supplier to solve a problem and, often — especially if the problem is difficult — pays them for their work regardless of whether they actually solve the problem.

Contests are different in two ways. First, the government broadcasts its needs to the public and allows anybody to submit a solution, rather than choosing one contractor to work on the problem. Second, the government pays the winner a prize for a successful solution, not simply for effort — though the contest prize often must be larger than what a contractor would have been paid for effort alone.

Whenever contests are suggested as a procurement technique, advocates (including me) repeat the obligatory statement that “the technique isn’t suitable for all procurements,” but the phrase often stops there.

Recently, though, I read a brilliant paper in the July 2012 issue of the Academy of Management Review that provides some helpful insight into determining when contests are a good alternative to contracting. (Readers of past columns referencing articles in academic journals will note that it is not unusual for me to read an article months after publication; academics like to think the content is timeless.)

In “Crowdsourcing as a Solution to Distant Search,” Allan Afuah of the University of Michigan’s Stephen M. Ross School of Business and Christopher Tucci of the Swiss Federal Institute of Technology in Lausanne offer the following guidelines:

•  The more the customer believes there is a significant likelihood that the solution to the problem might involve knowledge and skills outside what the researchers call the customer’s “knowledge neighborhood,” the more it makes sense to use a contest. One reason is that the bigger the gap between the knowledge needed to solve the problem and the kinds of knowledge the customer already has, the harder it is to select the appropriate contractor to solve the problem and the greater the likelihood of selecting a contractor that will fail. (Incidentally, this theoretical observation is supported by empirical evidence that winners in private-sector contests are often geographically distant entities or individuals unknown to the customer.)
•  The easier it is to describe what the customer needs, the more it makes sense to use a contest. If a potential solver needs to understand a lot of hard-to-describe details of the customer’s environment and culture in order to solve the problem, it is unlikely that the far-flung small businesses and innovative individuals who typically tackle such challenges will be able to solve the problem correctly because there will be something they don’t know about the situation that makes their solutions inappropriate.
•  The less contact the solvers need with the customer while they are working on the problem, the more it makes sense to use a contest. Otherwise, the communication demands on the customer will become overwhelming.
•  The cooler the problem, the more it makes sense to use a contest. Cool problems attract more solvers, and they are often willing to work on the problem for lower prizes. That is one of the reasons for the success of various open-source software innovations.

The researchers’ guidelines are sensible and practical. Federal program managers should keep this checklist in their toolkit.