On a related note, why is Netflix's interface so locked down? Why would it be bad for me to turn off autoplay for trailers? Why can't I say "never show me this show again, I'm never going to watch it"? And why, in this age of AI, and they still putting white text over a white background? That last one, especially, made me say "do people at Netflix eat their own dogfood"?
>On a related note, why is Netflix's interface so locked down?
Partly this the fault of Apple. They championed the idea that UI is an artistic expression and therefore 'configurability' and 'customizability' is akin to heresy. This view has now infected many existing products.
Part of the reason why many companies jumped on this bandwagon is also because 'customizability' is hard(er) to build in, and certainly more expensive to maintain.
> Part of the reason why many companies jumped on this bandwagon is also because 'customizability' is hard(er) to build in, and certainly more expensive to maintain.
It's also harder to A-B test with so many variables. If tests aren't statistics significant, the value of user analytics and UX experimentation decreases from a "% lift" perspective. It's harder to know if a feature change or a user defined config had a causal relationship to some other metric.
That's a good insight. A bit like unit tests, A-B tests drive the design of the website to make them testable. Not sure that is ultimately the best choice for UX, but here we are. I suppose it makes it a lot easier to justify one's perf rating.
Maybe on the perf rating, but on a bigger scale it could make it way more complex for analytics departments to function. Complexity adds real costs when that department exists to increase revenue and retention, and iterate quickly.
Those types of tests serve two very different purposes. UI is also unit testable.
Unit tests are more of a binary pass/fail. A-B tests are looking for cause-and-effect relationships by comparing some metric between a control group and a variant group.
Neither of those is likely the reason. This is about A/B testing your way to success, growth and engagement. Allowing too much disabling of engagement driving features or too much customizations hurts metrics. It's the same reason FB forces you from the date sorted timeline back to "magic" sorting every few days.
Ironically, Apple customization has become really good. Not in terms of tweaking the look, but the functionality tweaks like mentioned above? iPhone is super customizable.
I contacted Netflix support, and they basically said, "Sorry. That doesn't necessarily actually stop them. We can't share anything about whether we have plans to change that."
It has been like this for years. I don't know if it's a glitch with my account or if they are doing it on purpose. Neither possibility makes me a happy customer.
Call customer service, tell them you'd like to be removed from all A/B tests, because it's messing up your experience. They should still have a button to do that.
Most people are in the "On all other devices" category. Having to pull out a computer for what should be an in-app setting is ridiculous when you can sign up on some of those devices directly.
Are there third party frontends? Why can't you just pay them, authenticate, and use whatever frontend or player you want?
I suspect that the answer is ads. The Netflix app is an ad billboard on your phone and TV. Also, copyright and DRM.
Edit: I wonder if draconian copyright laws are ultimately to blame here. Nobody is allowed to provide such a service. This is getting a bit off topic though since this article is specifically about content aesthetics.
I feel the same way about Spotify. I'm fine paying an honest price for a useful service, but the upsells and dark patterns have gotten so aggressive I can't stand the client any more. I wish they'd at least allow third party clients to access basic functionality... but decisionmakers don't understand the attraction of a third party client in the first place, let alone feel like unlocking that functionality when it could negatively impact engagement with their dark patterns.
Aggressive podcast placement is the big problem, but they also allow artists to "promote" their music by accepting lower royalty payments. And those artists show up much more often than other artists in radios and suggested playlists.
As someone who doesn't like the podcast walled garden Spotify is trying to build (I prefer my podcast market with competitive open standards and choice of clients, thank you very much), the podcasts really bothered me. As a paying user, I should be able to turn them off entirely, but they kept overhauling the UI to make them more and more prominent. I stopped using Spotify a year ago today, and I've been very happy managing my own music library with Jellyfin since then.
Because third party front ends will show competitors shows next to Netflix shows and none of them want that.
That said, we are sort of there. Google can tell you where a show is streaming and I think Apple TV displays shows from different streaming services. Just not to the point anyone that wants to can build one.
But while there’s zero chance I’d willingly watch an ad without good reason, I might actually enjoy that promoted content because it just so happens to be of interest to me. That’s the difference between the two for me.
I've never heard this is the case, and if there was a secret deal that this was the case I think it would have been leaked years ago, but also it would be difficult for this to be the case and not have it be public knowledge because accounting details of large companies are something that is often inspected by news organizations and others on the lookout for something to complain about.
But maybe it can be that it is public knowledge and I'm one of the ten thousand who doesn't know? If so, do you have a link?
> But maybe it can be that it is public knowledge and I'm one of the ten thousand who doesn't know? If so, do you have a link?
Nope, it was just a thought. Sorry if I came off as implying shaddy things. I agree with what you wrote and it would be surprising something like that had been going on for years without anyone noticing (or revealing it).
I suppose it'd be less expensive to advertise a show outside of Netflix UI.
My interpretation
>Maybe someone is paying Netflix to suggest this movie to you ?
Answer:you
We like have control, but we tend to prefer don't have control and have good enough decision makes for us
Sort of? Netflix paid an upfront cost for its own content, but a recurring fee for third party content. So they try to push users to consume their content, which they'll own forever, as a bargaining chip and savings potential when the time comes to renegotiate those recurring fees.
So nobody's paying them, but Netflix effectively pays more (in the long run) to show you third party content and would prefer you watch content they produce for a one-time upfront cost.
This is actually backwards. First party content costs Netflix more. They license everything the same way -- paying someone else for the content. For example, they didn't make Squid Games or Wednesday. They just got an exclusive deal for a period of time. This actually costs a lot more than licensing already existing content, because they have to pay extra for the exclusive rights and they have to pay enough to cover the cost of making it.
They push you towards exclusive content because it's what differentiates them from the competition.
Good point. Is it correct to say that on a long enough timescale, self-produced content is cheaper than licensed content? Or is the production cost so high that that basically doesn't matter?
Makes a lot of sense to push the content that folks can ONLY get on your platform, since that's a reason for customers to stick around. It's too bad they overdo it to the point where I actively dismiss (and get annoyed by constant promotions of) Netflix shows unless a personal friend indicates one is actually worth watching.
> Is it correct to say that on a long enough timescale, self-produced content is cheaper than licensed content? Or is the production cost so high that that basically doesn't matter?
Depends on the production but it definitely isn't rule or trend. For example, look at this page with costs to make each episode:
Many of those are Netflix originals. What it means to be an original is that it was on Netflix first and it's the only place you can find it at first. And back in the day, that only applied to the USA (now Netflix does worldwide rights, but House of Cards for example premiered on other networks in Europe at the same time it was released on Netflix).
But once that contract expires, anyone can license it for the next go around. Generally Netflix does a long license and locks up IP rights to prevent that, but that just makes them have to pay even more for the content.
If you look at the cost per hour streamed, the originals definitely aren't the best -- stuff like a show from the 1960s that they got super cheap that ends up being super popular end up there. But the originals are the prestige items, so they're willing to be less efficient there for the unquantifiable brand boost.
I can't stand the fact that pausing Netflix to go fix up a plate of dinner inevitably leads to an obnoxious adroll for shows I've already seen or don't want to watch. And when I try to resume my show, I end up out of the video player, have to click the resume button, and I then deal with 20 seconds of garbage quality video as the bitrate settles and buffers.
1. They've decided to have a single interface for every device and Netflix supports a lot of devices. So it basically can't have any features that require more than a d-pad input.
2. They want to obscure the fact that their catalogue is really quite small. That's why are very limited manual filtering options and no advanced search. You'd very often get 0 results.
3. They're still in the "A/B testing can solve anything" and "we must optimise for engagement!" phase and haven't realised the problem with that. They probably A/B tested showing you stuff you'd already seen, found it increased engagement and said "ok it must be good".
How about Netflix stops showing me movies I’ve already watched at the screen expense of discovering new content? Just keep all those movies in the “Watch it again” banner and out from everywhere else?
I think people in "the middle" watch and rewatch the same things. I don't do that for movies, but I am guilty of doing it for some shoes like The office (still on Netflix in my region... For now)
Not just interface lockdown, but the insistence that They Know Better. Netflix, YouTube, pretty much any video service I've tried, actively prevents someone from excluding search results while simultaneously putting the offerings They Want You To See toward the top, almost without regard for the search prompt. I know that this would make the mile-wide, inch-deep problem for most catalogs readily visible to many people, but not being able to say, "DON'T GIVE ME THE RESULTS I DON'T WANT", does not endear a site to this user and many others.
"thumbs down" a movie to say "I'm never going to watch it", you won't see it again (great for getting rid of dumb movies in the big promotional banner on the home screen as well)
> On a related note, why is Netflix's interface so locked down?
A/B testing. The product org has always been driven by A/B testing, which means it sometimes finds a local maxima which they then stick to. They've also been driven by keeping things simple, both for the user and the developers. Having a single interface means the user gets a consistent experience (ironically broken by having so many A/B tests) and also means that testing on the backend is easier, because you don't have a geometric explosion of combinations of settings to test.
The fewer settings there are the fewer things to test when you want to change something.
> Why would it be bad for me to turn off autoplay for trailers?
Because A/B testing has shown that overall it's better for customer retention. They're willing to give up some users in exchange for the gains they get from having it on.
> Why can't I say "never show me this show again, I'm never going to watch it"?
Because people lie to themselves. A lot of people will say, "Schindler's List is an amazing movie!" and keep it on their watch list, but yet never watch it again, and then will say "Jackass 3 is the dumbest shit I've ever seen" and never put it on their list but watch it 15 times. People's intentions and actions don't always match. Netflix used to have a way to remove stuff from being shown, and then get calls into customer service from people who had chosen that button asking how to get the movie back because they actually want to watch it.
> And why, in this age of AI, and they still putting white text over a white background?
While you could fix this with AI, it would take a heck of a lot of computer power for a relatively small fix for few people (not a lot of people use captions). You'd have to review every film with every set of captions. That's a lot of hours of AI review.
> That last one, especially, made me say "do people at Netflix eat their own dogfood"?
When I worked there, most everyone watched Netflix every day. In fact, the only thing we didn't dogfood was the billing system because we all had free Netflix, and then when it broke, they gave everyone an $8/mo raise and told us all to sign up for paid accounts so that we could test the billing system with different billing days and different payment methods.
But very few that I knew of who used captioning. Most people don't like it and find it distracting. So when there were captioning errors they were usually caught by customers.
Did you also notice that everyone wears white shoes when on that auditorium stage? I wonder how many other dress code-related tips or requirements every presenter is given?
For context: watchmakers have often used 10:10 for marketing materials because it's symmetrical and it doesn't obscure the brand logo, which usually goes below the 12 position. Apple uses 10:09 probably because they want to be "ahead" of everybody else.
Similarly, iPhones are all set to 9:41 in marketing materials because that's the approximate time Steve Jobs introduced the first iPhone back in 2007.
You realize the event wasn't live, right? And Tim's watch clearly shows 10:09 the second he shows up in the video which was 10:01 at the latest "real time".
I know it isn't live, and it shows because they've been sloppy throughout. Check out his watch at the 20 minute mark, it's way off.
In general, Apple's presenter's watches (and demo devices) show the local time for live events, and they tried to do that here but obviously didn't do a very good job.
But they didn't do that, as I just mentioned and gave a fairly exact timestamp (to within a few seconds) for for you to check that they did not in fact do this.
You have a point, now wondering if there was some some memo akin to - pretend you are landing a plane as many appear to be doing just that with their arm flail.
Second this - I think part of being a programmer is having a really hard head that stands up to being banged against that wall, over and over, until the thing you're trying to make work works:)
Love this article! I don't think the guy is playing a 'persona' at all in this piece. These are the kinds of things he thinks about on a daily basis. Heck, these are the kinds of things I think about on a daily basis too, and I'm sure most devs who have been writing code for a few years do.
In my daily work, I don't have time to understand everything. There, i've said it. I'm not lazy. I'm not stupid. I just don't have time when I'm trying to get a task/job/chore done. I appreciate good abstraction/encapsulation. A good tool does this. A bad tool fails to do this. Same with documentation, really.
I set up Confluence and used it a lot for "how to do this" and "how to do that" notes for myself. It was helpful to me, and I figure that it MIGHT be helpful to other people. It doesn't HURT to have the information searchable, and with a GUI on it for the less-UNIXy folk. Some people use it, most don't.
