From Netflix to HBO, the terrible design of streaming is ruining TV | Hacker News Reader

From Netflix to HBO, the terrible design of streaming is ruining TV

fastcompany.com

46 pointsandsoitis40 comments

In general, separating data from UI would solve many problems across many domains.

Paying should be for the right to access video content alone, with many UI providers to choose from. Would increase competition within the market greatly. Centralized (or decentralized) database of all video content, with UI vendors caching from that for perf reasons.

UI would have to compete on its own merits rather than relying on data lock-in. Better products in the end.

While I’d love that, unfortunately the business and unit economics can differ wildly. Some shows may be cheaper to acquire if there’s promotion in the interface (due to tie-ins or newer content like sequels in cinemas), subscribers may churn more regularly if not shown specific things, making marketing less effective and therefore more expensive.

It’s like how a TV with no smart features would have to cost more, because there’s no post-sales revenue. We’d love them. But they’d have to cost more, so most people wouldn’t buy them.

> Some shows may be cheaper to acquire if there’s promotion in the interface

Something I would gladly pay to live without.

A Netflix subscription is ~$20. How much would you pay for that without the promotion/etc?

It might need to be $40/m to be profitable on the marketing. Now add in the extra resources to sell direct content access, price for the fact that the market is vanishingly small, and it could easily be 10x the price. Would you pay $500/m? Now it's so expensive you're way more likely to churn, and actually they need to make back money much quicker to sustain it. Would you pay $2000/m for Netflix? What sort of experience and support would you expect for $2000/m, now Netflix needs a new sales and support org, and it costs a huge amount to keep them around. Would you pay $5k/m for Netflix?

Obviously these numbers are all entirely made up, but Netflix is built on a large subscriber base and very low touch sales and support, with expensive engineering amortised across hundreds of millions of users. All that works in the opposite way, so I don't think these numbers are the wrong ballpark.

There aren't many people who would pay $5k/m for effectively the same service as they can get for $20/m.

In this way it's all quite different to my TV example. There you're paying $500 for a $700 TV because they'll make $300 from you on average. Just not including the smart TV functionality might only cost a little more, and then the TV might sell for $1k and a fair number of people would buy it.

> There aren't many people who would pay $5k/m for effectively the same service as they can get for $20/m.

But as you said, you completely made up that number. You could just as easily have made up the number $20.01.

Let's face it: that amount would be low. The costs are around $10/month for Netflix, and less for HBO and Prime. People already complain about rising costs of streaming. Paying for another graphic shell over the same content won't be worth much more than 1$/month. Of that 1$, some, possibly quite a lot, will have to be paid to the streaming service. You need to offer considerable benefits for mass adoption. Suppressing promotions is not enough to achieve that, IMO. Perhaps a better recommendation system would be a more attractive option.

Note that the provider of this UI would be bound completely to the terms dictated by the streaming service(s). E.g., it might look interesting to have all your streaming services in one app, but Disney or Netflix or Amazon or HBO can simply forbid that option.

How else will I know which movies to avoid because they're passionless trash?

So much this, Hulu might be one of the worst burying the "Continue Watching" section 3 pages down.

They all just want to slam you with what they want you to watch but make it intentionally difficult to just pick up the last thing you were watching. At the same time they roll new shows off that banner so fast you forget you were watching episodic weekly content.

UX across the board has been ruined by marketing majors.

Netflix's UI is a shell game to trick you into thinking they have more programs than they actually do.

I think the Netflix app is very performant and well build, especially compared to the ones like prime. Button presses just work even when using low end devices like a Nintendo Wii. The switching around of the different things like continue watching is very annoying though. I hardly use it anymore due to the algorithm bubble of only showing (very) similar things to what I have already watched and no tools to influence this. Also when they started out they had way more access to work from other publishers. I usually am looking for something novel and found more interesting things in the beginning when their algorithm was less "refined", they more had other's titles and it was less geared towards pushing their own content.

I find it very frustrating that the 'continue watching'/'watch again' sections are often in a different spot. And my last show will be in one or the other section depending on if my last viewing included the last episode.

I'm confused why Amazon Prime is the only streaming service with something like X-Ray. I know Amazon owns IMDB, but at the very least, you should be able to see a cast/character list in all the services...even if it's not contextual to the specific scene.

My guess is potential licensing issues and/or cost reductions. I assume someone has done a competitor analysis and decided that the lack of such a feature will not cause significant churn and is not worth the time and effort (legal, financial, development time etc.).

I was assuming when you gain rights to a TV show or movie, that it comes with basic metadata and rights to publish that. Maybe not?

I'm not familiar with how licensing works at that level, but I know there are a lot of licensing deals involved regarding music in the series and subtitles for example. It wouldn't surprise me if also meta data such as the movie cover and credits may need to be licensed separately for some markets.

I mean, you don't even automatically get rights to music used in the show, I wouldn't assume the rights to metadata are automatically granted.

Music is a creative work and people may legitimately want to consume it independently. Cast lists (and data in general) don't have copyright protection and serves little purpose except to augment the original file - both the means and the motivation aren't there.

Even if cast lists don't have copyright, they still need to source the data from somewhere. If I was responsible for this at Netflix, I would prefer to buy that data from a good quality source instead of scraping the data from external websites or running OCR on the credits screen.

Netflix adds on average about 1 show per day. It would be incredibly cheap to have an intern just watch the credits and write down the names. Maybe you get a second intern to doublecheck their work. Sure buying access to an already compiled database would be preferable to making your own, but even that's pretty cheap. For IMDB a million api requests per day costs $45/month - if we assume 3600 shows, 5 seasons per show, 10 episodes per season, 300 credits per episode that works out to less than 2 months of api requests, or about $90 total (substantially less assuming you can retrieve multiple credits with a single api request) to create a database for their entire catalogue. I've got more than that in my wallet right now. Of course you probably need an engineer to write a script to make all those API requests and at that scale it will take a little time to write and some effort to maintain, less facetiously it's probably going to be comparable to the cost of a few interns. Still, this is definitely not an insurmountable barrier for a multibillion dollar company.

I'm guessing it's access to that IMDB database of metadata. Anyone else would need to create that or license it which if I was Amazon I would either not sell or charge out the nose for.

For a feature-equivalent X-ray, yes, but a basic cast list, etc, should be doable.

> Good luck watching the credits on Netflix—they automatically skip to serve up a new episode.

Lol, you've never seen them play five minutes of credits in thirty seconds after a movie or show ends on basic cable?

Has Mr. Diaz never met anyone outside Hollywood? People don't pay for streaming services to watch credits.

There are some shows like Veep that actually have great content during the credits. I couldn’t stand watching that show using HBO max because I missed so many jokes.

I like to watch the credits, even for TV shows I'm watching serially. For me it's like getting that last bite of dessert before the end of a meal, plus I get to stretch my legs, check email, go to the bathroom, etc. before the next show starts.

In the Netflix web interface there's a per profile setting to turn off that behaviour so you can watch the credits in peace.

Doubtful this is a majority use case, but I tend to search for things to watch based on those who made the works I enjoy, rather than genre or franchise. Streaming service algos are useless at being used the way I want, I have to rely on third party websites like imdb and JustWatch to reliably find something to watch and where it is.

There was a video rental place in NYC sorted by genre and director.

They would never do it, but all I want is some sort of common API thing that I can plug into a system like Kodi that would let me browse and search the libraries of any service I'm subscribed to and then let me launch directly into an episode or movie from there. Why do I have to have a billion different apps, some of which don't work correctly or even have remote friendly UIs? (Hidive for example has fast forward and rewind buttons. Other services do that thing where they show frames of things that happened before/after the current spot and it's easier to navigate to where you want.)

Not to mention if I want to watch something but I can't remember which service it was on. It'd be great if I could just search for it and then go. Or see recently played things across services.

>curation algorithms

"Curation" is happening on streaming sites only in the same sense that Amazon suggests you re-buy the same toothpaste you've purchased before. The term implies two ingredients: not just a notion of what might appeal to others, but also the considerable personal experience of the curator. Knowledgable human beings have the remarkable ability to make non-linear associations.

Early streaming experiences were more satisfying because Netflix simply didn't have much to go on. It made discovery quite a bit more interesting, and gave users more of an overall sense of what could be watched instead of the carefully concentrated homepages of today that box users into their past activities.

My pet peeve: I have only one profile on play.hbomax.com, but every time I open the page I get the screen that asks "Who Is Watching?" with a single option.

I'm thinking of writing a web browser extension just to click that automatically.

If I was a cynical product owner who was responsible for the multiple profiles feature, I would probably fight to keep that extra click to remind users of the functionality. Otherwise it may affect the metrics.

BTW: My pet peeve is that it's not possible to mark an episode as "seen". They should at least allow it for those who use multiple profiles. If a family has profiles, but sometimes watch the same content together, it kind of locks you into using that profile forever for that show. It makes it harder to catch up on missing episodes on other profiles, and also messes up the interface in general.

YouTube has the best UI and Spotify has the worst. Watching Joe Rogan on TV is painful, if you ever want to pause it - which is quite often considering how long his podcasts are - resume does not work - you have to close the App and go back in - sometimes it doesn't remember where you were and starts from the beginning - fast forward is excruciatingly slow. Spotify should sack all developers who work on video they are totally incompetent.

As they say, what's old is new. A lot of these problems I remember having with VHS. Tapes would rewind too slowly, or they would rewind too fast and I would overshoot. Early IR remotes had to be aimed carefully, so pausing took a few attempts. Some remotes had a physical cord, which could come unplugged. And if a tape got jammed that means pulling it out carefully, gently rolling it back into the cartridge, and since the tape is now back at the beginning, fast forwarding to find my place again.

Resume worked fine on VHS.

I think you just need

  - Continue Watching (1 row)
  - My list (2 - 3 rows)
  - Discover (1 row) - this can also update with the "What's new" content and algorithmic recommendations

and a very efficient search interface that I can reach into. All the genre breakdown lists are not always useful.

It's a shame that Apple basically gave up on an all-in-one streaming app that's the hub of your ten-foot content consumption life—the Apple TV app (inclusive of Apple TV+) is a rather good experience, and I'd love to consume more of my content directly in it, be it Netflix, HBO, Plex, etc.

If it was that bad it wouldn’t be in 87% of households.

I don’t think any of the services offer award winning UX, but the difficulties stated in the article are overblown. Nor is it “ruining TV” … the proof is that they keep making TV (well, Until the writers strike…)

> If it was that bad it wouldn’t be in 87% of households.

The alternatives are live TV, which is infinitely worse, or nothing at all.

Install base is not an indicator of quality when all the competitors are bad.

It’s an indicator that the quality it needs to be is not that high.

Reply on news.ycombinator.com