Netflix has a well publicized contest where the winner gets $1 Million to improve the recommendation engine for movies by 10% or more. Taking advantage of the wisdom of the crowds is excellent and something that should be supported by more companies… but I don’t want something that is 10% better.

Each time I log in to Netflix, it recommends the same 10 movies. Just like an unaccepted friend request from that person in middle school, i ignore them since my decision has already been made.

What would be interesting would be rotating algorithms.  My theory is that each group working on improving the recommendation engine has a different approach. Since they each have a different approach, there should be different results. With some regularity, rotate the algorithm.

Customers would see different results and movie options which makes it feel like there is greater selection. This data in turn creates more data inputs which should increase the predictions and recommendation engine as a whole.