I found this Netflix article just fascinating - Netflix has found a way to extract value from knowledge experts thousands of miles away who are giving the company value before ever getting paid.
We at PC-Doctor have our own software development challenges and we sure should look closely at the way Netflix is innovating. To sum up the problem. Netflix is trying to improve upon Cinematch, its movie recommending software, in an effort to increase and retain subscribers. To do so, it didn't hire more developers, it outsourced the work . . . sort of . . . or at least their model seems to work like outsourcing. Yet, their model is many times better.
Can you imagine outsourced entities delivering a product without knowing if they would get paid? Well, that is essentially what is happening with their contest. In its first two weeks Netflix had received 169 submissions and as of Feb 25th when Jordan Ellenberg of wired.com published this story, 1000 teams had entered submissions and four or five had become top contenders.
And Netflix is still holding the booty - the $1MM and all the code entries. Further, I can only imagine there are lots of good ideas in the 995 other submissions providing lots of work for their internal team of developers to sift through.
Jordan in his writing mentions that Netflix offered up its database of movies watched and the data-diving scientists and mathematicians of the world gave their brain power for a chance to roll around in it and make sense of it all. Again, the beauty is that both parties seem happy.
I can't help but think of the financial side of this situation - All this brainpower on an hourly basis would have been very expensive for Netflix to pay for outright and running rough numbers it appears they have already gained more value than they will pay out. Roughly, 1MM divided by 1000 entrants is $1000 each and I imagine that there are at least 10 hours of collective effort and research and synthesis of ideas into each of those submissions. That is $100/hour for some of the world's best intellectual talent and that hourly rate is only going to go down as time moves on.
Here at PC-Doctor or at any software company, we could surely benefit from the same sort of open competition, don't you think?
Imagine if we got users' machines to upload reliability information in some safe, anonymous way to a database that we create. If we had that installed on a few million PCs, we'd get an impressive database on PC reliability.
It could well be as much fun to give away the data as to play with it ourselves.