How Netflix outsmarted everyone else in TV
recode.net
recode.net
I recently stopped using it and it was like walking out of an incredibly comfortable, yet subtly depressing cave. I found it hard to get out of the cycle of watching only films that I was likely to watch.
What humans will eventually discover after we've squeezed every last drop of value out of the monolith ML systems of today is that many facets of our culture do not benefit from traditional recommendations systems. Maybe a more democratized and transparent system for model consumption could help. If indie researchers and hackers can integrate their ML models into existing systems and offer truly diverse ML models then maybe recommendations systems have chance. But in their current state they are laughable representations of the average taste of its users, which is generally not great.
AI/ML is great at optimizing what you're currently doing. While it can also find correlations and connections that you may miss without it, it's unlikely that it will make any radical recommendations.
IMO this is ok because AI/ML can evolve into that capacity - you have to take baby steps after all.
But in the interim, it's definitely something to watch out for - you don't want to take away the element or principal of randomness out of life - it's at the core of how we operate, innovate, and (almost quite literally) how we evolve.
In other words, rather than optimizing for zero risk of disliking something, allow the user to optimize for the possibility of truly loving something, or being surprised. Especially on Netflix, where the risk is you stop part way through and switch to something else!
Also highly recommend the Norwegian “Occupied” series.
Le Chalet Suburra Dark The Break Marcella The Fall Happy Valley
However, I've also gone back to Netflix DVD. They have such a massive catalogue of movies and it seems easier to just skim through genres and find weird gems.
The non-customized (viewed in incognito) main page of Youtube that is the recommendation for the average viewer is even worse, nobody should ever touch that horror show.
Youtube is a horrific service. The recommendations are terrible, the UX sucks, too many ads, they dont know how to use the sreen real-estate.
I freaking hate youtube.
WRT to real-estate, I see that the video I am wanting is taking up ~20% of the screen, then there is a list of completely unrelated content, and I have to click to see the date of the submitted video.
The UX is different on every single device. Phone, machine, ipad, browsers differ... its just a mess.
This is a standard ML practice, otherwise ad networks would be stuck showing you the same ads (even though it feels that way sometimes). I'm willing to wager large sums of money that Netflix has tested EVERY idea mentioned anywhere in this thread and many others you couldn't guess.
This would help me to quickly sort out if I would consider watching something and I could reduce my watching list to the things I'm actually currently watching instead of abusing it for "might watch at some point".
If I have watched something I don't want it to be recommended to be and if I don't like something I don't want to see it again.
It's like if a chef based a restaurant menu on the average diet of the average person. You may find neat patterns and be able to hyperstimulate some drive in people, but that's a very different thing than carefully designing a new or engaging thing people didn't know they wanted.
It's sort of a guaranteed averaging feedback loop.
This is pretty much what fast-food chains do methinks? And I think that's an apt comparison to what people are lamenting with Netflix above.
There's definitely a DVD rental market for "every dvd that is or ever was in print."
Just to be clear/nitpicky, you're talking about an algorithm that starts with the catalog Netflix could afford to license at some moment in time, and then maximizes the viewer's enjoyment for that limited set of content.
That's probably a vastly different level of enjoyment than starting with, say, the catalog of available content listed in IMDB and maximizing for the same thing.
A user can hit those constraints when they search for something that exists but is not available on Netflix.
That point becomes noteworthy when that same content (or portions of it) are available elsewhere on the net. In my experience this is typically through:
a) downloading a torrent of the content that is missing from Netflix
b) watching all or some portion of that content on Youtube
I find a lot of my joy in watching netflix is randomly browsing through the catalogue and selecting a random film that I would never watch if I had to either schedule watching on TV or pay to watch at the movies.
The fact that Netflix brings together a wide range of different movies and television programs in one place where for no cost I can watch something you may not like has opened a world of media I would never have found before.
I get why they've moved to 'this content matches [x%] to your tastes', but I'm watching similar stuff over and over again without any insight into whether other content is also good, unless I want to spend hours pouring through IMBD and Rotten Tomatoes.
Disney is rolling out their own streaming service and I'd bet that eventually they will roll ESPN+ into the deal and maybe even ESPN entirely itself. I think once sports gets into the mix the entire streaming domain will radically change.
Still I'm rooting for the Witcher to be a big hit.
I'd say no, because they aren't dependent on advertising. Taking ads introduces a kind of competitiveness within the seller's financial heads. It's an irreversible state change that works for some companies, but -- and I'm no econ -- Netflix might be banking on not needing to do that for as long as they can. We see them dancing with the fire in their menu behavior and intercalary promos[1], so we'll see.
1. https://www.nytimes.com/2018/08/20/business/media/netflix-co...
I think House of Cards was at least in the same ballpark as GoT and Sopranos. I think there is something to be said for brand affiliation as well, HBO was already considered a great source of TV while Netflix was only starting to get a fraction of that credibility. I could easily imagine an alternate reality where we would say:
nothing Netflix has made can touch Game of Thrones or House of Cards for cultural impact
with the implication being The Sopranos being a Netflix property.
As with homeruns, HBO has more than its share of (sometimes critically acclaimed) base runs (Veep, Boardwalk Empire, Entourage, Curb Your Enthusiasm, etc...) as is does strikeouts (Vinyl).
Good point regarding HBO's brand and their "good" shows; and I think it helps to illustrate my point: Netflix has a lot of good shows, maybe even some great ones, what I think they need, and what I think the other streaming providers will need is really big generational hits that come out every five years or so.
Both good points, too bad that HoC looks to be Spacey's swan song (and deservedly so).
Dubrovnik has been filled with tourists for decades.
But there seem to be some other locations as well?
However, if they are as gorgeous as Dubrovnik presumably they were also tourist attractions prior to GoT as well.
Sure, but that is just the difference between fantasy/scifi and everything else.
You don't get cosplays of Tony Soprano.
The worst thing is I have not found the next Game of Thrones/Harry Potter/Lord of the Rings. Lots of good books have come out in the past few years, but nothing has had the same impact yet.
And Netflix has got options for critically acclaimed tv series that traditional networks can't fully utilize. They can and do import foreign tv shows to an American audience without worrying much about taking airtime away from other shows.
Interesting Bloomberg piece on Netflix's efforts in India: https://www.bloomberg.com/news/articles/2018-06-28/can-netfl...
Oh, I don't know. Stranger Things was a pretty big deal.
A. Streaming quality is still poor, particularly audio. The bitstreams are compressed beyond normal DTS/DD compression, and it sounds like garbage in comparison to the lossless tracks on physical disks. To me the audio track is 90% of any movie.
B. It's impossible to find the vast majority of movies on streaming. If something pops up, it disappears in a month or so. Out of FIVE HUNDRED physical discs queued up, only TWELVE are available on streaming. I seriously fear the day they get rid of physical disk rental all together. Their catalog is expansive, and we have ZERO other options now.
I just bumped from 2 discs to 3 discs out at a time, because we no longer have 1 day turn around times on returning and sending out discs. It's nearly impossible to get one back the same week unless you send it out Monday or Tuesday.
A lot of this is the classic play of using investor money to buy users. Netflix has been burning billions because they don't charge enough for the content they provide. That problem is getting worse as the streaming market becomes more of a primary distribution channel driving the cost of streaming rights up. We've already seen that happen as Netflix catalog shrinks despite all of their original programming.
Basically, Netflix's product is getting steadily worse while their costs are going through the roof.
Maybe we get more granularity, but that comes with the overhead of managing more subscriptions.
Who will laugh when we end up paying for a bundle of streaming services?
There are so many amazing films from those two decades. Where are they?
As the AI gets better for restoring old film, it might become possible to find old laserdiscs or even VHS to use as a master, but right not, that's not really possible.
Not in most cases, no. Usually everything is owned by someone else -- even the audio and the video are usually owned by separate companies.
Also, the originals that those blu-rays were digitized from are most likely the film masters themselves, which would have to be found again, and possibly restored again. You'd be surprised how easily studios lose their film masters.
You can see what they have here: https://www.filmstruck.com/us/watch/browse-all/
Netflix basically needed to bootstrap its streaming service with content, so they paid a bunch for syndication. They then sold an all-you-can-eat plan at a low price. Their success butted against pay-per-view providers, other syndicators, and content owners: Netflix blazed the trail and proved that people sign up for video-on-demand in droves if it carries content people want to watch, and is cheap enough to where one doesn't have to think about every view. But this devalued the content, so studios yanked their newer seasons and put them on knock-off platforms for a separate subscription.
This development meant Netflix had to start making deals with less controlling content providers, or produce content of their own. On the other side of the house, Hulu was able to justify its price point by offering current seasons of popular shows, and being a joint venture of influential content owners, but single-owner offerings struggled for customers, unless the IP was exceptional. This is why HBO hasn't yet died in this space, and Disney will likely succeed on their own, but also-rans no longer offer enough in a crowded market.
By the end of the article, they do descend to reality and identify the magic formula: Netflix makes money from the subscription, and a constant flow of offerings both strong and mediocre will tide the user over to make them not consider cancelling. And in time, the ever-expanding catalog is insurance for a change in market headwinds.
The key, though, is will Netflix be able to convince the consumer to part with their money when other platforms with the same business model offer a more compelling catalog? In the long run, that remains to be seen.
They downside of showing only your own content is that you need a pretty high percentage of winners (diversity of viewers increases your odds, but still). A "drought" can kill you. And as a result, Paramount, 20th century Fox, MGM etc are irrelevant to the viewers' choices.
The only time I can think of being forced to pirate is when legacy licensing crap keeps you from being able to watch it where you want, when you want it. But if it's available to buy on iTunes or Amazon or wherever, and all you have to do is pay, then you're either choosing to pay what the owner is asking, or steal it.
When I subscribed for Netflix, the motivation was a large catalog of old movies, cool stuff from other countries (I remember watching TV series from Iceland that were great), no ads. But the greatest selling point: it was more convenient than pirating.
Now, it's all about pushing Netflix produced content down my throat using whatever trick possible, and pirating is again more convenient (large catalog, no ads) even if less user-friendly.
So yeah, "outsmarted" is really pushing it. I'm not the only one getting bored by the current Netflix direction.
The current Netflix direction is the only one possible; one big catalog only worked before they had proven the viability of mass streaming—once that happened competitors bidding for exclusives (and content owners reserving material for their own services) was inevitable.
> I sold my company to MTV in March 2000
2000 is not "early" or even "earlish" at MTV, youngun.
A lot of people like to have something playing on the TV that they passively watch while also doing something else. Maybe they put on a show while they browse Instagram, or do house work/cook dinner, or while hanging with friends.
Seemingly as a result of this, there are a lot of shows/movies where you don't really need to pay close attention to get the plot because it has been dumbed down, reduced to the most basic strokes. There's more filler in between plot developments. It's like a lot of B-movie schlock with high production value.
These shows/movies don't interest me. When I put something on to watch, I sit there and watch it. I want deep plots and fleshed out characters and a lot of Netflix originals seem to lack this.
no it really did.
"Shipping" is a classic trope where 2 characters have a will they won't they relationship for several seasons. It is always a good B story to fill otherwise empty space with when writers run out of material.
The name literally implies what it is, you are "shipping" a script. quality is slightly irreverent
As for original programming seemed dumbed down, I'd argue that for whatever reason (general technical advancements, maybe), it's far easier to make dumb shows, with higher production values. That's how you get wannabe prestige dramas that ape the style of The Sopranos, Mad Men, even costume epics like Game of Shows, but end up being far more vapid and worse in writing or plot. It's sort of how even box office stinkers these days look much better than bad movies used to do.
For sure there is still some value there, but it's more and more distributed across platforms, with worse and worse usability. Often I find I'm paying for multiple expensive services and can't find a way to stream what I'm looking for. Or (Amazon Prime!) I can rent it from a service I already pay monthly for.
TV had lower expectations. On-demand but sort-of-not-really-and-you-have-to-pay-again-and-watch-more-ads is frustrating.
Autoplay though was a profoundly good idea. It's literally the reason binge watching exist, which from a business is genius. For what reason do you dislike autoplay?
I see nothing wrong with search either. I mostly use the PS4 app in which in works fine. It's admittedly not the best but certainly no worse than any other streaming service and worlds better than what my cable provider has.
>For what reason do you dislike autoplay?
Because it doesn't bring anything but distraction. If I want it to play, I can click, if I want to browse, I don't need a preview. Most of their movies is shit anyway so why would I need to preview them?
> It's admittedly not the best but certainly no worse than any other streaming service
Amazon is way better, both on their app and website.
EDIT: amazon (prime) is better for the search, content wise, it's the same poor selection of violent movies and idiotic documentaries.
With delivery platforms all now funding their own content, the streaming entertainment landscape starts to look ... exactly like the old TV networks. Just the names have changed - replace 'ABC', 'NBC', 'CBS' with 'Netflix', 'HBO', 'Amazon'. It convergent evolution, and all of the things that drove the dreck that was on broadcast TV in the old days are there to drive it now, only with even better and realtime ratings data and targeted ads.