Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is something so many big companies do, and I hate it. Every 6 months it seems Spotify has to change their UI and make it harder to find what I want. Just make a decent UI and stick to it. I don't understand why there are redesigns over and over and over. My assumption is that it keeps people employed, even if they aren't necessarily needed. It at least takes away from more important issues being taken care of.


> why there are redesigns over and over and over

Because they aren't building UIs to be useful for pleasant for users anymore. Once these data-driven companies have enough users locked in, they begin optimizing the user interface to manipulate user behavior for profit. They shift from helping you do what you want to manipulating you into doing what they want you to do.

It's actually gotten so bad that open source software with it's notoriously unpolished interfaces is actually starting to be the better, more useful and more aesthetic option, without having improved much at all.


Spotify is an ever-evolving trash fire. It's become very difficult or impossible to get it to display a list of albums created by a given artist, and then play album X. "Oh no" Spotify says. "I think what you really want to do is play <random popular song Y by that artist> and then a whole bunch of random songs by other artists you've never heard of, right? That's what I'm going to do for you."

I realize it's still actually possible to get Spotify to do what I want, but that stuff is increasingly buried beneath dark patterns.


Yeah it’s weird to me - spotify has basically all of the world’s music. Why does it insist on playing the same 12 songs on repeat when I listen to anything? I enjoyed those songs the first hundred times they were played. How do I get out of this recommendation engine jail and get some variety? Here’s a feature I’d love in spotify: no repeat workday. Once activated, unless I explicitly play a song again, spotify is banned from repeating any song within a 24 hour - 2 week period. You know what has this feature? The radio.

Is Apple Music any better?

YouTube has the same problem. It seems to insist on recommending the same 3 rabbit holes every time I visit the site. I know there’s more stuff out there I’d love that I’m not seeing, that I don’t think to search for. But I have no idea how to get YouTube to show me any of it.


Spotify pays a discounted royalty rate when they can kick you over to random plays. Business over user experience, for sure, although they might say you wouldn't pay them a higher monthly subscription to compensate if they didn't do this.

Can't say why or how they arrive at such a tight list of random options; perhaps it's an attempt to give you something predictable while fulfilling royalty obligations. Maybe it's just broken.


My guess is that attempting to diversify what they play for you leads to much worse results, at least for the modal user. I don't think that music recommendation services actually work that well except by identifying very popular stuff.


"Jeannie Becomes a Mom" by Caroline Rose plays first every single time Spotify tries to generate radio for my Indie Mix playlist. This has been happening for months. You can click "Don't play this again", but since the contents of the *Mix playlists change every day, the don't play list is wiped every day.

I wish I could universally blacklist a song (at least for a couple of months), but that only works for entire artists. I've seen the exact same complaint for the exact same song on reddit, which makes me wonder if it has something to do with their program that lets artists jump ahead in the algorithm (https://newsroom.spotify.com/2020-11-02/amplifying-artist-in...).


My wife has been complaining about the random feature playing the same songs over and over again. It's been about ten years now. I stopped using the trash fire because of it.


Reading your comment reminded me of the blog by Jeff Atwood https://blog.codinghorror.com/our-programs-are-fun-to-use/

Surely if some company focussed on making software that puts the user at the centre they'd be able to carve out a pretty decent niche for them.

Sadly if they we're a public company their investors would probably be unhappy because they'd be sacrificing the all important "growth", but a private company might be able to get away with it.


GNOME begs to differ though.


Is GNOME a data-driven, for-profit company?


Apparently all the decisions that everyone hates much on GNOME design decisions are based on data driven analysis.

Not everyone uses feature X? It gets cut.


> Just make a decent UI and stick to it. I don't understand why there are redesigns over and over and over.

It gets people (in your example, designers, but the same organizational disease affects engineers and product managers too) promoted. Perhaps somebody got promoted this cycle for "making a decent UI", but you're not gonna get promoted next cycle for "sticking to it".

And managers get promoted by "growing" teams to build stuff.

When performance and promotion criteria incentivize "having impact", which is understood to mean "launching stuff", this is what results. It's a analog of "teaching to the test" [0], or a special case of "gaming the metrics" [1] or "you are what you measure" [2]. I don't know if there's a term for the general phenomenon.

I agree it sucks, but while I'm invoking cliches, I should remind myself: don't hate the player(s), hate the game.

[0] https://en.wikipedia.org/wiki/Teaching_to_the_test

[1] https://mitpress.mit.edu/books/gaming-metrics

[2] https://hbr.org/2010/06/column-you-are-what-you-measure


Compensation driven development


> My assumption is that it keeps people employed, even if they aren't necessarily needed.

I'd say more that they keep people relevant, particularly architects and CTOs. You can't go in front of the board and say "Everything is great and we're keeping it just the same!" You can't say that in front of the CEO, you can't say that in front of the investors. You have to keep selling the idea that Big Changes Are Coming and Our Userbase Will Increase.

So you add bloat that nobody wanted and credit anything good that happens to it.

I work for a company based in a foreign land and while there are downsides, one of my biggest reliefs was to find that there is very little interest in 'change for the sake of change'.


Twitter has been so bad with this lately. It feels like the app is different in some irritating and pointless way every time I open it.


When I did use twitter the only way that was even remotely tolerable was with a third party client (tweetbot was my client of choice; twitteriffic is quite good too).

The biggest benefit is you get a literal timeline that's unmanipulated by promoted tweets. I have no idea how anyone does anything remotely useful with the native twitter UI.


Yet every time the UI is still horrific. It’s like they’re trying the Edison way and trying to find the 99 ways not to do it.


This is why I only browse twitter from my notifications panel.

Twitter is great when it serves its purpose of being a 140 char log of people that matter.


They're running A/B tests... You like variant A, the majority picked variant B. Otherwise, someone new was hired and needed a project.


Nobody liked A or B, just some weird engagement metric chose B.


Maybe making it harder to find what you want makes you more likely to find what they want.


I used to enjoy using Spotify before they added podcasts; now it’s kinda a UX/UI mess.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: