[danShumway]

> It is pretty much impossible to build a music UX that optimizes for all three of these user types

This is a good argument for not tying music subscriptions to a specific client -- ie, allowing 3rd-party clients that are optimized for my specific user-type to freely hook into and stream from my Google Play Music subscription via an official API. The Google Play Music client is not optimized for users like me. I'm sure it was a lot of work to build, but I am not the target demographic of any of the interface decisions that have been made, and I curse the designers every time I open the app.

So it's weird that, say, Shuttle, a music client that is optimized for me as a user, can't hook into my account or even download my purchased music.

I don't mean to call out Google Music in specific, because while it doesn't have good public APIs, they're at least consistent enough that alternative clients are being made on desktop.[0]

But I do see this as a trend across a lot of SaaS services -- Netflix, Twitter, Facebook, Apple Music. People complain that the services aren't optimized for them, and designers will roll their eyes and say, "you can't optimize for everyone".

And that's true. But who's fault is it that you have to optimize for everyone? Who's fault is it that you have to try and figure out how to balance a bunch of wildly diverse and often contradictory needs? There used to be really good alternative Twitter clients, and if Twitter wants to complain now that building one client for everyone is hard, I just can't muster the energy to feel sorry for them.

Apple's whole shtick is trying to get everyone to use the same official apps, to have the same consistent, good experience. I'm not sympathetic to a company that deliberately puts themselves into that position and then complains, "optimizing for everyone at the same time is hard."

[0]: https://github.com/MarshallOfSound/Google-Play-Music-Desktop...

[chipotle_coyote]

There are third-party music players for iOS that hook into Apple Music -- Soor and Marvis, for instance -- and take different approaches to their UX. (Disclaimer: I haven't used them and can't speak to their particular advantages and disadvantages; while I don't love the iOS 13 music player, I can deal with it, and I routinely use features like "Radio" and "For You" that the original poster was complaining about.)

But, yeah, I'd definitely like to see more services offer full-featured APIs and actively support, if not necessarily promote, third-party clients. Watching what Twitter did to their ecosystem was supremely frustrating.

[mcpeepants]

Google Play Music happens to be absolutely perfect for my usage (including the ability to upload my own tracks AND stream them). I'm extremely annoyed they are killing it off as I've tried everything else and it just doesn't work for me. The replacement YouTube Music is... abysmal. Well IMO anyway.

Only pointing this out because it speaks to your point. I think you're spot-on that APIs and embracing 3rd-party clients are the way forward. Not just for music, but any service where the primary UI challenge is making everyone happy (hint: a lot of services). I'd instantly support and switch to a service like GPM that put APIs first, assuming all the same back end functionality and catalog was available. Even if the "official" client didn't suit my needs.

[Finnucane]

You could have different clients for different kinds of users. But then you would almost certainly have users complaining that they need more than one client to do what they want.

[danShumway]

This would still (often) be strictly better than having those same users complain that there are no clients that do what they want.

It's also an easily solveable problem if the APIs are actually open. If Apple put out three music clients, you would definitely get people complaining that they weren't unified. If Apple shipped an open iOS API that covered everything in their official client(s), somebody in the community would just build one client that did everything a complaining demographic wanted, and then sell it for $4.99 on the app store.

When you see large groups of people complaining about current tech offerings for sustained periods of time, that usually means that the legal and/or technical barriers to entry to build alternatives are too high.

[azinman2]

Apple does have a public API for Apple Music [1], as does Spotify [2]. In fact, I frequently use a 3rd party client [3] that's all about curated music leveraging the Spotify API to stream the tracks.

[1] https://developer.apple.com/documentation/applemusicapi

[2] https://developer.spotify.com/documentation/web-api/

[3] https://noonpacific.com

[danShumway]

I can't speak to Apple (I'm on Android), but I did look into 3rd-party clients for Spotify at one point and my impression was that the web API only allowed you to get information about songs, and remote-control the official client. The advice I saw online was LibSpotify was basically dead and that Spotify was probably going the same direction as Twitter: more locked down, more onerous developer TOS, fewer capabilities.

I'd be pretty happy to be wrong about that, since I'm fairly annoyed with Google Play. At the time, if I had found good enough API support and a good enough 3rd-party client, I would have switched services.

It's been a while since I looked into it, and maybe I missed something when I first did. I can see that Spotify is experimenting with a web playback API now[0], but as far as I can see it's still pretty limited.

[0]: https://developer.spotify.com/documentation/web-playback-sdk...

[Finnucane]

And then people would complain that there were 1200 apps in the app store that did the same thing, and most of them badly.

[danShumway]

And those people can just be safely ignored. I get that people get annoyed at needing to sift through a lot of options, but the way to solve that is by having more curated lists, rankings, and recommendations.

I am almost entirely unsympathetic to people who argue that having too much choice is a bad thing on an app store. That's not a real problem.

[Finnucane]

>he way to solve that is by having more curated lists, rankings, and recommendations.

Yeah, sure, then we only need a few meta-rankings to know which of those we actually need to pay attention to.

[danShumway]

Yes? I'm guessing you're being sarcastic with your comment, but lists of lists genuinely scale quite well. You can be as meta as you want, and in fact the more meta you get, the easier it becomes to filter out choices for people who are paralyzed by having too many options.

Of course, when you build recommendation engines and lists, you run the risk of filtering out quality offerings. But if the alternative you're proposing is we only have 1 or 2 choices for each type of app, then accidentally filtering out quality offerings is probably not a real concern for you.

Filters, reviews, and lists are great. They're how I select products to buy online, how I get book/movie/game recommendations, how I decide what Linux software to install on my computer. Curated content recommendations are the entire premise behind sites like HN and Reddit. This stuff really does work.

[Shebanator]

I have a rule of thumb I tell my friends when they ask me questions of the form "X is so lame, why can't GPM just do Y?". The answer, almost always, boils down to laws, regulations, and contracts. This is one of those times.