29 July 2009 • 7:00 am

Competition and Collaboration from Netflix

Netflix prize web page

Netflix prize web page

The Tortoise’s fascination with Netflix was enhanced with the news yesterday that Netflix’s public competition to improve the effectiveness of its recommendation system has drawn to a close with two teams essentially tied for the prize. But each of the two teams are actually consortia; teams comprised of other teams. Netflix has harnessed the power of competition and collaboration to solve a challenging business problem.

Subscribers to Netflix are asked to rate (on a scale from one to five stars) each movie they’ve rented, and are even able to rate movies seen in theatres or elsewhere. The data base of millions of individual ratings are used to predict and recommend movies to individual subscribers. The system works pretty well; around two thirds of all rental decisions made by Netflix subscribers are the result of a computer-generated recommendation.

But Netflix wanted the system to work even better. Instead of turning to their own staff, in 2006 they captured the attention of computer scientists, mathematicians, psychologists, and other professionals from around the world with a contest to “help improve our ability to connect people to the movies they love.” They offered a $1 million prize to the first team to improve the performance of the system by 10{7d517eca6fa2b1f37358396ef304f8a78637162298d2da9398058e81473e3d6a}. Here’s a video of some of the contestants.

Thousands have entered. With standings and results visible on a public Leaderboard, it became clear that competitive advantage would come from collaboration, and the two teams who’ve exceeded the prize threshold (The Ensemble and BellKor’s Pragmatic Chaos) are themselves comprised of other teams.

What a brilliant strategy. Not only does Netflix have access to some of the world’s sharpest minds, it has done so at a tiny fraction of the cost of actually paying everyone engaged in the effort. A great solution, low cost, and incredible free publicity as the competition itself has become a news story.

Of course, not every challenge faced by an organization can be solved in a public competition, but there are some valuable lessons to be learned. First, the ability to objectively compare competing ideas will likely result in a better solution than one without competition. Second, the ability to test hypotheses, and measure and quantify performance precisely enables those objective comparisons. Third, the most valuable resources to an organization do not necessarily lie within that organization.

We’ve come from a world in which organizations make decisions and solve problems by assembling their own resources and relying on them to excel. Relying on teams of expert staff members sounds like a good idea, but is it? The expected value of a solution is judged on the basis of who the people offering the solution are rather than the quality of the solution itself. This approach naturally excludes input from those not seen as experts, and alternative solutions are simply not examined.

The higher up one goes in an organization, the fewer people are seen as qualified to propose solutions and to make decisions. Collaborative problem solving happens less at the top than at the middle and foundation of an organization. It seems to me that the high failure rate of strategy execution can be traced in part to an absence of collaboration and competition in developing strategic hypotheses. While it may be impractical to make every decision in this way, Netflix’s radical approach to the challenge of improving a key business system is causing at least this strategist to question the status quo approach to decision making and problem solving.

Does your organization use collaboration and competition to make key decisions? Please share your experience with using these techniques to improve outcomes.

Comments are closed.