I presented some ideas over at developer.netflix.com and by directly messaging Netflix through their “Business Development” contact page about an idea on improving movie matches and like/dislike predictions for customers.
I’m not expecting much of a response considering the thousands of messages they deal with daily and the fact that there are far more brilliant minds at work, but it couldn’t hurt:
Human Rating
“Could also be called “Keyword Rating” or “Keyword Review” or even possibly “Mini Review”. The idea is to improve movie like/dislike predictions through a more human-layered rating system.
This would not be to replace the 5-star rating system, but to work along side it as an optional enhanced rich rating feature.
Essentially this would be a form field under the star rating that would ask:
“Describe this movie in 5 keywords:” and there would also be a 3 option checkbox feature to describe what connotation the words have: Positive | Neutral | Negative
So, someone might describe a movie like so:
“dark weird nostalgic awkward brilliant” | Positive
(there would of course need to be a check/filter against misspellings and obscene language)
Once the system was in place and used substantially, users could start to get matches based on positive words they’ve used to describe other films, they may see other films labeled under “nostalgic”. Of course it’s all a bit subjective, but any rating/voting system is, but with this system it would add more human depth to matches and suggestions, getting inside movies instead of generalizing their outer shell.
This might be used to enhance Netflix search as well.
This isn’t quite an idea for an app, more of an idea to be integrated into netflix.com. I would however still be interested in an API Key.
Thanks for listening, Bryan”
I did end up receiving an API Key. I could check out some of the data, but I’m hardly a developer or programmer myself, just endless ideas.