Netflix Prize: Stories in Stories, Part V

In my continued following of the Netflix Prize, I popped over to the Netflix Leaderboard again tonight, and there are now 21,076 contestants competing for the prize, from 140+ different countries.   That's 50% more contestants added to the captive pool of software development talent they have gathered since November of last year (only 4 months ago).  From a talent acquisition perspective, it's fascinating that the momentum continues to build even after the media buzz has worn off. 

The group that I interviewed in a previous post (see the categories at right to review the history on this topic) has dropped to number 5.   

Many companies pay for referrals of talent through employee referral programs and also external programs such as H3, or simply by buying research (essentially lists of names) from recruitment companies.  Buying names can be expensive, and good researchers will charge $100 an hour for their work (sometimes more).  Many companies pay thousands of dollars per successful referral.  It is interesting that on a per name basis, each software developer that is working on the prize cost Netflix less than $50 apiece (assuming the $1M is the all-in cost - which it probably is when you consider net present value; ie they haven't had to pay anything yet). 

The teams are nearly 70% of the way to the 10% improvement goal of improving the Netflix collaborative filtering engine.