Netflix and the Napoleon Dynamite Problem (2008)

From what I understand, they never implemented any of the algorithms, because their existing one was good enough, and the premise has changed. It used to be, you'd make suggestions so people could get DVDs sent to them, and you'd know they would enjoy it. Now with streaming, the cost to send something you don't like is cheap. So the recommendation engine just has to come up with something you'd try to watch and perhaps enjoy. If you start watching something, it's not your taste, you switch streams. So there's not a lot to gain from improving recommendations 10% unless they were pretty low quality ratings to begin with.

"If you use a computerized system based on ratings, you will tend to get very relevant but safe answers"... I call this my "Radiohead metric": for a given music recommendation system, how long do I have to use it before it tries to recommend me Radiohead.

One of the problems I find in music algorithms is they can't know how I've changed - suggesting music based on what I play or input - but that's stuff I liked in mid-2000s when I was in peak music consumption mode. Most suggestions then get picked from the same pool.

I'd love these algorithms to take my age into consideration: "If you liked X when you were 25, you're going to love Y now that you're 35!"

All those hours and millions of dollars into perfecting their recommendation system, and Popcorn Time still has a better UI for viewing movies than Netflix does. Just show me all the top movies and let me pick one, why partition them into drama/comedy and filter out the one's you don't think I'll like? it makes the experience worse.

I've been getting increasingly frustrated at Netflix, because it keeps recommended things I have already seen, or definitely do not want to see. Having to keep scrolling through unwanted content like that really reinforces just how dumb it is, despite all these claims about their intelligence.

Eventually I did figure out that on the website you can say you aren't interested in particular items, but Android/Roku etc do not have that option. Of course those Netflix employees who work on this stuff deal with the PC/web based interface all the time, but apparently don't realise that not everyone does that.

Android/Roku etc do let you set ratings on items, but to me there is a big difference between "I am not interested in this item" and "I have watched it and give it one star".

I'm not sure there's a reasonable way past the "Napoleon Dynamite Problem". They mentioned, in the article, other movies where it's difficult to predict if someone would like them or not.

The list was “I Heart Huckabees,” “Lost in Translation,” “Fahrenheit 9/11,” “The Life Aquatic With Steve Zissou,” “Kill Bill: Volume 1” and “Sideways.”

Personally, the common thread I see across those is that the viewer would be more likely to enjoy the movie if they had the right context.

Kill Bill, for example, is far easier to enjoy if you've seen (not necessarily liked...but have seen) cheesy old Kung Fu / Karate movies.

Napoleon Dynamite is easier to appreciate if you were socially awkward yourself in high school. There's also some humor in there that's more relevant if you were a kid in the 80's, as well as some humor that might make more sense if you lived somewhere rural, or even more specifically, rural Idaho/Utah/Wisconsin.

All that to say that Netflix may not have enough data at their disposal to know if you have the right context to enjoy movies that almost require it.

To be fair, even people aren't able to judge a movie properly. People often change their opinion of movies on second viewing.

Just watching a movie at different times of day can significantly alter your opinion of a movie.

A controversial observation might be they maximize profit by having the most people sign up and then stream the least possible, correct? And they have a semi-monopoly, correct, where theres no real competition? I mean, I have and use Prime but I don't seriously watch videos and I'm in it for the shipping...

Anyway there's this assumption that they'd want to present results that encourage you to watch as much as possible, yet they have a financial incentive to present results that are just barely good enough that you won't cancel and/or sign up for a minor competitor.

This is NOT like a public library card catalog where next years funding depends on increasing circulation counts.

In some ways I think my proposed search result criteria are more difficult than presenting result I'd actually want.

Netflix, having crushed the video store industry, no longer needs this. What's the alternative? Redbox?

"Even though Net­flix has a good deal of demographic information about its users, the company does not currently use it much to generate movie recommendations; merely knowing who people are, paradoxically, isn’t very predictive of their movie tastes."

"Targeted" advertising suffers from that model. Demographics matter only a little. What you've bought (which Amazon knows) matters far more.

I feel like one thing Netflix doesn't do a good job of noticing is what movies I actually order/watch. That is, the mere fact that I ordered a movie means a lot, whether or not I rate it.

The biggest manifestation of this issue is that I tend to order new movies, but I seem to get a lot of recommendations for older movies. Netflix should see that I've ordered maybe 2 out of 200 movies that are older than the year 2000.

Netflix doesn't show user comments in the app version of its (streaming) product.

I'm sure there are reasons for this. Nonetheless, it continues to strike me that, in that context, their own, internal information on product offerings is restricted to a two sentence blurb and an unqualified single five start rating metric.

There's a related problem: if you enjoy both good movies and some so-bad-they're-good movies, how do you rate the latter?

In Notes from Underground, Dostoevsky nailed the reason computers and science and algorithms will never be able to replicate humanity or predict the actions of man, when he wrote:

“What is to be done with the millions of facts that bear witness that men, consciously, that is fully understanding their real interests, have left them in the background and have rushed headlong on another path, to meet peril and danger, compelled to this course by nobody and by nothing, but, as it were, simply disliking the beaten track, and have obstinately, wilfully, struck out another difficult, absurd way, seeking it almost in the darkness. So, I suppose, this obstinacy and perversity were pleasanter to them than any advantage...

The fact is, gentlemen, it seems there must really exist something that is dearer to almost every man than his greatest advantages, or (not to be illogical) there is a most advantageous advantage (the very one omitted of which we spoke just now) which is more important and more advantageous than all other advantages, for the sake of which a man if necessary is ready to act in opposition to all laws; that is, in opposition to reason, honour, peace, prosperity -- in fact, in opposition to all those excellent and useful things if only he can attain that fundamental, most advantageous advantage which is dearer to him than all. "Yes, but it's advantage all the same," you will retort. But excuse me, I'll make the point clear, and it is not a case of playing upon words. What matters is, that this advantage is remarkable from the very fact that it breaks down all our classifications, and continually shatters every system constructed by lovers of mankind for the benefit of mankind. In fact, it upsets everything...

One's own free unfettered choice, one's own caprice, however wild it may be, one's own fancy worked up at times to frenzy -- is that very "most advantageous advantage" which we have overlooked, which comes under no classification and against which all systems and theories are continually being shattered to atoms. And how do these wiseacres know that man wants a normal, a virtuous choice? What has made them conceive that man must want a rationally advantageous choice? What man wants is simply independent choice, whatever that independence may cost and wherever it may lead. And choice, of course, the devil only knows what choice.

Of course, this very stupid thing, this caprice of ours, may be in reality, gentlemen, more advantageous for us than anything else on earth, especially in certain cases… for in any circumstances it preserves for us what is most precious and most important -- that is, our personality, our individuality. Some, you see, maintain that this really is the most precious thing for mankind; choice can, of course, if it chooses, be in agreement with reason… It is profitable and sometimes even praiseworthy. But very often, and even most often, choice is utterly and stubbornly opposed to reason ... and ... and ... do you know that that, too, is profitable, sometimes even praiseworthy?

I believe in it, I answer for it, for the whole work of man really seems to consist in nothing but proving to himself every minute that he is a man and not a piano-key! …And this being so, can one help being tempted to rejoice that it has not yet come off, and that desire still depends on something we don't know?

You will scream at me (that is, if you condescend to do so) that no one is touching my free will, that all they are concerned with is that my will should of itself, of its own free will, coincide with my own normal interests, with the laws of nature and arithmetic. Good heavens, gentlemen, what sort of free will is left when we come to tabulation and arithmetic, when it will all be a case of twice two make four? Twice two makes four without my will. As if free will meant that!”

I think it comes down to authenticity. Humans can predict the actions of humans, so long as humans act by and large as they think they're supposed to act or like the things they think they're supposed to like. The moment an individual walks in, an individual who does what they truly want to do when they want to do it, all correlations and inferences and predictors fall to waste.