The paragraph below is taken from the NYT and their article today about the winners of the Netflix contest. For those of you who do not know about this contest. Netflix created a contest in 2006 offering $1 million to any person or team who could develop an algorithm that would improve on the Netflix recommendation results by 10%. Essentially, if you can give us something that works at least 10% better than what we have, we'll pay you a million bucks.
From the NYT
The Netflix Prize contest has been hailed as prime example of “prize economics” and the crowdsourcing of innovation.
Prize economics refers to running a contest to generate a new
innovation at less cost than an in-house research and development
effort, and crowd-sourcing refers to using the proverbial wisdom of
crowds to accomplish a task. Netflix has said that $1 million would be
a bargain price for an improved recommendation engine, which would
increase customer satisfaction and generate more movie rental business.
So what does this mean?
#1 Netflix Got a Bargain
Netflix got loads of high level and professional engineers to work on their problem for FREE as a result of promising the winner $1 million.
This $1 million is likely equivalent to paying only 10-15 of these engineers their annual salaries.
If Netflix had tried to solve this problem by hiring an internal team, they might have hired the wrong people and, over a two year period the internal department would have likely cost more than $1 million without necessarily being sure that they would have succeeded in their goal.
Also, there is nothing in this deal saying that Netflix can not purchase algorithm's that improve their searches by, say, 8% for far less than $1 million. As all these engineers handed in their homework over the last two years, surely Netflix got some accidental gems handed to them on silver platters that they could pay for outside the bounds of the contest.
#2 There is a Serious Shift in HR under this model
With this model of development, it was the "workers" who identified what & who they needed to succeed. There was no HR department that culled resumes;no department head that matched engineers with statisticians. They found and selected each other. Is this an indicator that the innovation team is best qualified to see what skills are missing for a particular objective?
#3 Each member of this team is an Entrepreneur
We have all been talking forever now about the information age facilitating individual workers, globalization allowing collaboration around the planet. People exercising greater personal ownership over their skills and abilities as opposed to promising those skills to an employer and applying them as directed by others. Perhaps this is a signpost on this road towards more entreprenurailism and less corporate domination over innovation? hmmmmmm
#4 The problem required Teamwork
Before any proclamations are made about the lone wolf triumphing over corporate herds.....these guys did need to collaborate. The problem required a team to be solved and even those who take ownership of their skills and work as entrepreneurs, still can not run with scissors and still need to put the cap back on the paste when they are done. Many of these people met each other through or as a result of the contest itself. Therefore Netflix still facilitated in bringing the team together, not through employment or a skilled HR manager, but by putting out the bait and letting the best people mingle and meet each other at the top of the pyramid. So who built/assembled this team? The individuals did it themselves...but they never would have met without the contest.
#5 What works for algorithms would not work for toaster ovens.
This kind of contest is a really interesting model. But it does not quite work the same way for object or product design as it does for algorithms. The cost ratio of algorithm implementation and testing compared to the development cost is minuscule. It is absolutely and empirically possible to test an algorithm's's effectiveness without implementing it in the marketplace. It is not possible to know that you have a kick ass toaster design until you have sourced it, made it, bought it, shipped it, and put it on store shelves and tested it with consumers. Unlike computing, the costs of production & distribution for products is often greater than the costs of innovation. Netflix was able to be certain that they were buying a 10.05% improvement before they paid. It is very difficult for Black and Decker to know what kind of improvement a particular drill design will have over their current sales.
So what does this all mean? No idea....not really. I think Netflix is brilliant to have created the contest and congratulations to the winning team. As an entrepreneur, I am heartened by the way individuals came together to collaborate on the problem, managed themselves, and succeeded.
Comments