I saw a number of mentions of this Atlantic article, How Netflix Reverse Engineered Hollywood "To understand how people look for movies, the video service created 76,897 micro-genres. We took the genre descriptions, broke them down to their key words, … and built our own new-genre generator." It's a little wordy but they wrote a script to find all the genres Netflix uses (they're numbered so they just incremented the URL). Then they talked to Netflix to get more of the story.
I found this followup by Felix Salmon much more intesting. Netflix’s dumbed-down algorithms.
Netflix’s big problem, it seems to me, is that it can’t afford the content that its subscribers most want to watch. It could try to buy streaming rights to every major Hollywood blockbuster in history — but doing so would cost hundreds of billions of dollars, and could never be recouped with $7.99 monthly fees. What’s more, the studios can watch the Netflix share price as easily as anybody else, and when they see it ending 2013 at $360 a share, valuing the company at well over $20 billion, that’s their sign to start raising rates sharply during the next round of negotiations. Which in turn helps explain why Netflix is losing so many great movies.
As a result, Netflix can’t, any longer, aspire to be the service which allows you to watch the movies you want to watch. That’s how it started off, and that’s what it still is, on its legacy DVDs-by-mail service. But if you don’t get DVDs by mail, Netflix has made a key tactical decision to kill your queue — the list of movies that you want to watch. Once upon a time, when a movie came out and garnered good reviews, you could add it to your list, long before it was available on DVD, in the knowledge that it would always become available eventually. If you’re a streaming subscriber, however, that’s not possible: if you give Netflix a list of all the movies you want to watch, the proportion available for streaming is going to be so embarrassingly low that the company decided not to even give you that option any more.
So Netflix has been forced to attempt a distant second-best: scouring its own limited library for the films it thinks you’ll like, rather than simply looking for the specific movies which it knows (because you told it) that you definitely want to watch. This, from a consumer perspective, is not an improvement.
The original Netflix prediction algorithm — the one which guessed how much you’d like a movie based on your ratings of other movies — was an amazing piece of computer technology, precisely because it managed to find things you didn’t know that you’d love. More than once I would order a movie based on a high predicted rating, and despite the fact that I would never normally think to watch it — and every time it turned out to be great. The next generation of Netflix personalization, by contrast, ratchets the sophistication down a few dozen notches: at this point, it’s just saying “well, you watched one of these Period Pieces About Royalty Based on Real Life, here’s a bunch more”.
The legal difference for Netflix between shipping DVDs and streaming is huge. They just buy the DVDs and then loan them to you, no licensing required (a nice copyright fair use concept). For streaming they have to license everything and that's expensive. I haven't streamed much from Netflix but I have a few things. Currently my DVD queue is 264 and of those just 62 are available for streaming. Usually for movies I tend to wait until they're on cable so I don't have to worry about network issues. But a lot of old or British or foreign TV series aren't going to be on cable so they tend to be what I stream from Netflix.