The Netflix Challenge

Posted February 28, 2008 by Adam Gautsch

So, Netflix offering a million dollars to the person or team that can improve their recommendation algorithm by 10 percent is a much more serious affair than I thought. It includes interesting lessons in software development and crowd sourcing. It even includes the mystery contestant, coming out of nowhere, who could win the whole thing.


olivier ~ February 28, 2008

Is Netflix assuming that selecting movies that are more in-tune with what a customer wants to watch will actually yield more revenue?

That would be incorrect.

  1. The reality is that Movie viewing time is pretty much inelastic, so better movie recommendations aren't likely to drive up demand (or frequency... or velocity).
  2. I'm not even going to bother with #2 and #3. #1 is enough of a roadblock to make algorithms mostly irrelevant when it comes to creating a 10% increase in revenue for this sort of business model.

Netflix can grow revenue by increasing its marketshare, not by increasing the number of movies watched by existing customers.

Lovingly crafted by orangecoat with some rights reserved, and a promise not to spam you.

Back to top