I work in this industry and actually we’ve found from online experimentation customers generally don’t want to tell you they dislike something. The expectation is customers get what they like (relevant) and sometimes they want to express stronger affinity for specific things. The converse hasn’t shown to be true.
Generally customers don’t have a clear picture of how their explicit signals impact recommendations. Though some customers have shown desire to want to manage these signals.
Edit: one other consideration is disliking can usually be inferred from other activity. Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
I had to keep deleting my Spotify account because my recommendations kept getting contaminated with music genres I didn't like that were impossible to get rid of. Listening to the wrong song two times could irrevocably ruin my recommendations. I would have very much liked a dislike button, or at least a "forget everything you think you know about me"-button.
Now it ended up with me getting rid of Spotify, because of this and other frustrations.
Without specifics I cannot guess as to your specific case but in general customer's just do not use dislike buttons. So exposing them isn't a value add and usually a UI clutter that most customer's have no use for and won't use.
In general, you had to be doing something to be getting recommendations so there is some action that when most customer's do demonstrates a high probability that you will like some genre and it just so happens you don't.
> in general customer's just do not use dislike buttons. So exposing them isn't a value add and usually a UI clutter that most customer's have no use for and won't use.
A feature does not need to be used often to be useful.
Take a fire extinguisher for example. Going by this logic, since most people will never even touch a fire extinguisher and most of them are never actually used, we should get rid of them because they're a cause of clutter.
I must say your dismissal of this is rather annoying. Multiple users are telling you they are facing a real -- and completely obvious and predictable -- problem due to the feature not being available. In a comment section regarding research showing it would be an improvement to recommendations.
Increased UI clutter due to a single extra icon is not a good nor sufficient argument for not including it.
My dismissal is based on a many-million digit customer base and online real world data. Sorry if a few hackernews users feel they want a dislike button, most users don't and if you give to them most users don't use it. I don't mind disagreement but dismissing real world data from this industry is sort of silly to counter with 'because I want it you're wrong'.
I'm not even saying you're wrong, I'm saying your argument for not including it for (at the very least) a sizeable minority is weak and somewhat silly.
A minority is anything that is not a majority (this is the definition I'm using). So a < 50% group can hopefully still be viewed as non-negligible, even though it is not the majority.
I disagree that your reasoning is entirely sound. It's a bit of an elitist point of view I suppose; but fundamentally I don't think all customers are created equally in terms of importance. One could even in this day and age call some of the more important ones "influencer" or something similar.
While the general customer base might not be served by this feature, what if a subset of the customer base enjoys this feature to an extraordinary extent? Maybe it fixes a lot of problems they have with the service; and negates a lot of complaints they otherwise would have posted.
Would it still not be the correct choice to "clutter the UI" with this?
I don't think it's as easy of a choice like you're trying to make it.
But I could also be jaded by the fact that Spotify for years has trying to pull me into listening to all kinds of odd European folk music that I have no interest in what-so-ever.
For a long time we called this sort of user a "power user". People who are not necessarily technical but who love your application and use it extensively will appreciate power user functionality.
It's probably fine to not have anything to appeal to power users, but those power users probably will be the sorts to be influencers to talk up your application to everyone they know.
Because the masses are now large enough that you don't need to have an actually useful application. You just need to make it kind-of work and for it to become popular.
I think for technically minded or abnormal users a dislike button would be valuable, but that's a minority of the population and I understand why a company would choose to focus on the larger market to the detriment of the enthusiasts. Props for taking the heat and trying to present this perspective.
I personally had these same complaints and acknowledged that I'm unusual and my needs are best served by my own hosted solution. I'd encourage people who are unhappy with major streaming services to roll their own if they truly feel strongly about it- I did and I'm happier for it.
How could people dismiss real world data that you didn't share? You only shared your interpretation of it and people pointed to lots of ways that this interpretation could be wrong. What data is this? Did you conduct a survey asking people if they “want” a dislike button? Or did you just measure usage frequency of the dislike button and decided that people didn't want it because it wasn't used as frequently as the like button? How could low usage frequency mean users don't “want” it?
This totally ignores effects from the long tail, black swans, power users, and criticality dynamics.
Downvotes are way more "potent" than upvotes. If I'm listening to music that I like, I'd say I probably log over 100x upvotes than downvotes. But when I downvote, I usually want to nuke the site from orbit.
What is the ratio? What ratio of users use downvote? What ratio is that compared to users that interact in any way? How are you controlling for sampling bias? Berkson's paradox? How much utility are the users that downvote obtaining, vs the disutility of being unable to?
I absolutely love it when people working on a product tell their customers that they don't need a feature they're all asking for and that would clearly improve the experience, just because not everyone uses it all the time.
> absolutely love it when people working on a product tell their customers that they don't need a feature they're all asking for and that would clearly improve the experience, just because not everyone uses it all the time.
We need a name for this, personally I call it GNOME mentality but probably is bad name since most developers are not familiar with GNOME (though the file picker issue get on top HN a lot )
The data does not speak at all. You are drawing inferences based on the data, inferences of a type that are pretty easy to poke holes in by looking at the pretty large class of things rarely used but situationaly useful (often life saving).
This type of bizarre logic makes me feel like I have to habitually use features even when they are not useful in order to protect them from UX designers in case I'm subject to some misguided A/B test.
What data are you referencing? I'm drawing conclusions based on data I have but I don't think you work for the same company as me so I'm not sure how you can speak to confidently about how wrong I am.
If x number of people are asking for something, all of those people are asking for it. I think it was pretty obvious I wasn't claiming all spotify users are asking for this.
I'm a PO so feel sufficiently qualified to state that user analytics data is completely redundant for looking at small use high value features & undervaluing direct user feedback from your 'power users' is generally a bad move.
Again, just because not everyone needs/wants a feature does not make it useless or not worth implementing.
Data doesn't tell you how useful a feature is, it tells you how frequently it's used & how many people use it. They are not the same things.
> Data doesn't tell you how useful a feature is, it tells you how frequently it's used & how many people use it. They are not the same things.
That entirely depends on the the data set. On one hand you are claiming to be an expert and on the other making generalizations about what data can be used for. In reality you know that data is contextual and depending on the data set is what determines what conclusions and how confident you can be in those conclusions.
Each of the comments that mirror yours illustrate a major problem we have in the tech sphere—assuming _we_ know better what the end user needs therefor we don’t listen to the actual user. In our defense, we’re very disconnected from the end users.
I’ve done some work with a couple disaster organizations in the past and we continually get massive praise and thanks from the victims of these disasters. The actual victims have the same praise over and over again: Unlike the red cross, salvation army, and FEMA we actually ask them what they (the victims) need. We don’t assume they need water and and a tarp, we assume they know whether or not they already have those things. Often the victims need diapers, help with flood mold remidiation, a chainsaw to cut limbs from their street so they can get their cars out so they can go help the neighbors a few streets over etc…. They know what is prohibiting them from getting out to help others much better than we.
When people are telling us what they need, we should listen to them and we should actively try to avoid our tendencies towards “I know better what you need.”
I have a couple friends who work on recommendation algorithms full time and in their defense, they’re fully aware how terrible the recommendations are—I still get my music recommendations from friends, from the local music store employees, and friends who work in music venues—their recommendations are infinitely better.
> Listening to the wrong song two times could irrevocably ruin my recommendations.
Similarly any app or website that pulls “top artists” from Spotify tends to include stuff I listens to a bit several years ago, but never stuff I’ve listened to nonstop for months.
Yes that's very frustrating. I'm scared to even listen to new music on it because I'd try to use it for discovery but it would mistake what I listen to for what I like. Besides that it hardly recommends me any new music anyway. I now just use youtube for discovery.
You could even do this maliciously. Send someone a Spotify link to a Steely Dan song and a Skrillex song, and now yacht rock and dubstep will follow them around until they delete their account (I'm assuming the intersection between those fan-bases is vanishingly small).
High school me would have been very offended to be labeled “vanishingly small.”
But you’re not wrong.
By the way, I believe Spotify does have a private mode, which at least blocks your plays from showing up in the social pane and on your profile. But I’m not sure if it cuts them out of your recommendations.
Please, dear God, Spotify: just because I listened to the Hamilton soundtrack once does not mean that I want to hear it every week in Discover Weekly and multiple times per Duo playlist, because Lin-Manuel Miranda is apparently a genre of himself.
I keep hearing from people how good is Spotify "Discovery" playlist. I've been a paid customer for 5+ years and their recommendations have a hit rate of 2%. I listen to a lot of different genres, artists, mostly dictated by my mood. Today it was some easy Dire Straits song, yesterday QOTSA, the other day some psytrance. So Spotify recommends all genres in the same playlist, creating an awful cacophony of songs that don't flow into one another.
Spotify recommendation system is very bad in cases like mine apparently. It's so bad I keep listening to the same old songs instead of discovering something new, which is why I'm no longer a customer until they improve their AI.
I listen to a LOT of different genres of music and their Discover reccomendations (not the playlists, the actual you like X artist you might like X) are absolutely brilliant. Got to be approx 50-60% hit rate.
It seems to differ wildly between people - I've heard plenty of people making the same case as you, but many that have the same as me. It never appears to be 'its ok', its either 'brilliant' or 'terrible'. I've been using Spotify for 10+ years at this point, so maybe that makes a difference?
Same here. I like Christopher Tin (e.g. Baba Yetu) and therefore for 6 years I've had no way to tell Spotify I don't like 8-bit video game music.
No amount of skipping or "I don't like this artist" seems to convince Spotify to stop offering me variations of 8-bit Tetris themes or MULTIPLE 8-bit Cowboy Bebop Tank! theme variations per every single Discover Weekly. I'd be thrilled to even get a 2% hit rate on music I like.
Seriously Spotify: two separate "Tank!" versions per Discover Weekly is too many.
Man, the spotify algo makes no sense. The other week I was in a really specific mood, starting with Sonic 1/2 OST and moving into 8bit/chiptune/keygen. I probably listened to every rendition of Chemical Zone out there.
Rest of the week, much the same. Nothing in Discover Weekly.
I think the relative "power" of a music genre is in relationship to its popularity compared to the popularity of what you usually listen to. Stands to reason that if you listen to rarer music (i.e. stuff the algorithm has less information about), then stuff that is more common (and the algorithm knows more about) will have a disproportionate effect on your recommendations.
This seems to be a problem with this general type of recommendation algorithms in general. YouTube suffers from it as well.
> one other consideration is disliking can usually be inferred from other activity. Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
This is nonsense. The fact that I skip a song does not mean I don't like it in general, it means I don't want to hear it right now. Music taste changes with mood and situation. If you treat a skip as "I don't ever want to hear this again" you will collect absolute garbage data, and if I know you are doing this, it will make your app unusable because I can't do normal everyday actions without ruining my recommendations.
It depends right. If you skip a song everytime its played in a sequence is a stronger indication of dislike compared to skip it once but listening to it a few times. The point is there is more to context and recommenders than a adding a dislike button and having a magically awesome recommender.
Have you read about illusion of control (search for elevator door button for one well known example)? If we take the idea that recommendations are really hard and dislike button doesn't help at face value, wouldn't it still improve the user experience to give them the illusion that they can directly affect the recommendations?
I'm not really convinced about recommendations being hard, though. Pandora could do them a long time ago, Netflix had good recommendations at one point but not since they dropped the star ratings and at some point Amazon wasn't that bad either.
Is this within a global context or a local one? Liking a song in Spotify is global - Spotify then adds it to a list of liked songs.
But from my experience with Pandora, radio stations were localized. Liking a song did add it to a list somewhere, yes, but disliking a song just meant that a particular radio station on Pandora wouldn't have a song in that particular style, and other stations did not seem to be affected by this.
In other words, what you want to listen to, for me, seemed to be very contextual. And since I felt my dislikes were captured in a context (radio station), I didn't have any qualms disliking frequently using Pandora.
And my experience with Pandora for song discovery has always been much better.
A problem I have with the inferred dislike is that when I skip a song it's only something I dislike half the time. The other half I'm simply not in the mood to hear it, maybe for another 10 minutes or another 10 months. The result has been that I've become more willing to listen to certain genres on another platform (say Youtube) because it won't contaminate the prevailing mood which I want the main service to recommend.
That's a good point and the next evolution recommenders try to do is to become contextual. There may be songs you like at dinner or on weekends but not vice versa. These types of things like time of day, device and other types of data points can make a recommender more precise.
> Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
How do you know I dislike the song? Instead, maybe I just heard it (or a similar song) on your platform or elsewhere outside of your app and don’t want to listen to it again right now?
I guess you could try to build a more complex model that determines what songs or music I like at what times of the day or what day of the week, or based on what I just listened to.
IMO this is rampant in FAANG companies. We create complex solutions to problems that didn’t really exist and pat ourselves on the back when our experiment show positive results.
The experiment might show you some metric (total time in app, play time at whatever percentile, etc) increased. Great! Now we launch the model in production for all customers and pat ourselves on the back. Meanwhile, the actual experience became shittier. A few promotions come out of this, and we continue living in the bubble.
The paper is making an obvious claim. If I have data showing positive and negative affinity, that will perform better than a model which just has positive day. I don't find this very surprising but my point is in practice exposing dislike buttons to customers are a CX that is hardly ever used. So the point that a model will perform better with data that customer's rarely offer up to your platform makes their claim moot in my opinion.
I downvoted your comment to demonstrate that I disagree with you.
I've heard some good arguments against allowing users to show dislikes/downvotes as well as likes, but I've not heard your argument before and it doesn't strike me as accurate. Less used than the positive direction, sure, but still significantly used when available.
Netflix, Youtube and reddit aren't the same medium as short lived music streaming is. Streaming music is a show lived precision game of guessing what's relevant. There is a lot of it depends scenarios, are we showing recommendations on a landing page, are we sequencing songs in a station or are we ordering results on a search page. The CX context is a big part of how you can use signals to benefit the customer.
My argument isn't an argument so much as a from real world data in this industry at a major industry competitor that customer's don't use the dislike button enough to justify it taking over your CX. Using co-occurrences of playback across a customer segment and combining that with likes or other signals is good enough.
Form should follow function. CX used to mean customer experience, and when your customers are telling you they want this feature back then you should reinterperate your data, not tell them they're wrong.
It amazes me how tech savvy users think they know what the entire population of users want. I'm not telling anyone they are wrong, I'm simply sharing real world data that counters the idea that users actually want a downvote button for music streaming. The fact that most people in this thread have really strong opinions but don't work in music streaming baffles my brain.
Maybe the button’s utility doesn’t come from its use, but from its presence. The user may enjoy having the option to dislike a song, even if they never exercise that option. Taking away their choice robs them of that agency.
I know that these implicit signals affect how I use the app. I might force myself to listen to a song I like, but would rather listen to something else at the time because I don't want my choice to skip to impact my recommendations in opaque ways.
Skipping doesn't imply dislike, though it does correlate (because the reverse is generally true, people skip songs they dislike). But I routinely see skip mentioned as non-downvote sources of dislike information.
Easily half of the times I skip it's because I don't want it right now. Some of my favorite songs reproduce absolutely horribly in a car on a highway for example, because the surrounding noise utterly destroys subtleties. So maybe you add "but it's not a downvote if it's hearted or otherwise positively ranked"... but many other times I'm just not in the mood for [a song I had never encountered until now].
You are exhibiting google's mistake, caused by over educated people with zero real experience thinking they know better. Or, I should correct that - some code they wrote knows better than the complex brain of the actual person. This results in "AI" (big if/then loop) trying to figure it out, and completely blocking any input from the customer.
Here's the issue with "inferred from other activity." You know how you can find out? Give the customer a way to tell you. And no - you cannot "infer from other activity."
>Skipping a song after less then 20 seconds for example can train a model...
..to be wrong. You know when I usually skip a song after 20 seconds? When I've heard it too much, or just heard it on youtube on my laptop an hour ago. In a couple of days I'll want to hear it again. The reason I heard it too much, is I like the song. Your AI though - if I skip this favorite song of mine a couple of times, will now block it and all other songs like it. The literal opposite of what I want.
What this results in, is you have an inferior product, your customers leave, then you leave and go screw up another company's product without ever learning your lesson.
That last part is the most important point - the main point. Your comment shows the exact issue as the "you don't need a dislike button because our AI knows better." You read my post, you read all the comments on here with users either jumping through hoops or ditching the product, you read the comment like what I replied to saying "we know what you need better than you do."
That last bit is a person thinking some crap code they wrote can figure out why someone doesn't want to listen to a song at a particular time. They ignore all the inputs into that decision - how tired they are, what day they had, are they angry at something, have they heard it on another medium recently. Then ignore all that data they don't have, and have the AI make the decision: did you skip after 20 seconds? If you did, let's cut this genre out of your playlist.
This little thing, destroyed spotify's recommendation product. The people in charge of this, are going to destroy everything they touch, because they think their little piece of code, is smarter than the user, about what the user wants. You know, because the user can't code.
People making these decisions - well all I can say is I've met them. I've been personal friends with some. They have been passed up for jobs when I get called for a reference check. Because once they're on your team, they'll destroy your product, so they can feel personally superior.
So that last part is very necessary, because it's an extremely serious issue, exactly as I put it. But in your opinion it's not, so let's cut that part out because it's not necessary for what I was saying. Because you know better than me what I was trying to say.
Some people get very, um, engaged with the universe of sound they wish to live in and have much stronger opinions about that than, say, world hunger. It directly impacts their experienced quality of life, and is one of the few cultural constants they retain over the course of their entire life. Lovers, employment, friends come and go but the music remains. I may be one of them myself. Profanity generally results when it is mucked around with. Count yourself lucky the opinion was expressed as mildly as it was.
I will not permit a company/algorithm to directly determine what I want to hear. I’m old enough to know about the Payola scams with radio DJs to trust anyone with that authority.
My solution: buy CDs or vinyl, find a group of like-minded friends to share music recommendations (email newsletters would work, perhaps even…physical interaction), read music journalism, buy music online and build your personal library. Just the music on my hard drive, all directly and legally purchased, would play continuously half a year without repeats.
I use/used the Apple iTunes others-have-bought recommendations for tracks related to tracks I’d purchase. Invaluable for discovering alternative artists by someone who is nowhere near the target age demographic for that sort of music. And you buy the flippin’ track and you know the artist/rights holder is directly getting a reasonable number of pennies from that transaction.
Worst case — buy CDs and get a CD changer. New ones are still being sold. You can still get the 300-400 CD units used. No fans are more rabid than classical music fans, and they are still ecstatically buying huge boxed sets of CDs. They are expecting to listen to these tracks for decades, and there is wisdom there.
> What this results in, is you have an inferior product, your customers leave, then you leave and go screw up another company's product without ever learning your lesson.
I'm not sure what you're suggesting and I don't need to qualify my experience in the software industry or within the domain of recommenders and personalization.
I did not suggest anything. I very clearly stated what I stated, in plain English that is very easy to understand. Your qualifications are irrelevant. Your point of view on the issue we are discussing is wrong for long-term product success, and destroys the hard work others have put into the idea, in order to feel personally superior to your users. I don't know how more plainly I can put it for you if you had trouble understanding that. I do believe however that not understanding something, is not something that has ever prevented you from action or strong and wrong opinions on a subject matter.
You’re speaking as if you work beside me and have a clue about anything in the music industry. There’s a reason Spotify doesn’t but uses to have dislike buttons more accessible. I’m not sure why you are disagreeing with me but there’s no need to speak down to someone because of different perspective. I have real world data and work in recommenders I’m not sure what your background is.
It is a no-brainer to use the collected data points instead of user feedback in this case. Only a small set of users would down vote enough to make it relevant. Also hard to measure the quality of the votes and how they age.
Well, no surprise this was down voted since most people in this thread never seems to have worked in this field. I have worked with recommendations on a service with millions of users. We do not guess. We look at things like conversion rates to tell us what works. If one think down votes would play a significant part in training a model you simple does not understand the amount of metrics a modern service collects and what impact different variables have.
I am speaking as if I have no idea about your work and what you do. I am speaking as if I am addressing the specific incredibly stupid thing you said, from a user point of view. You know who does not care about your music "industry" or your "work?" The users of spotify. They care about listening to music they want to listen to. My background is I listen to music using things that play it. Spotify is absolute crap because the people that make it do exactly how you said. The fact that you think you know better than the user is why I put that last paragraph in. You still don't get it, and you never will, and that's the whole point. People who think like you shouldn't be allowed to have any input on features or product, because you make inferior products, and destroy good ones. It happens a lot, with a lot of stuff. Because of you.
Based on assuming I talk at the office to my coworkers the way I talk to some stranger with a really bad opinion online, you have deduced that coworkers at my presales job in data storage are not treated well.
Based on how you talk here, you sound like you paint a fake little world around yourself full of straw men to make you feel better. Which goes well in line with your opinion you know better than the user of a product, what that user wants. Living in a fake little world where you da man, courtesy of your brain.
This is called deep seeded insecurity - where your brain makes subconscious coping mechanisms to avoid depression. Bullied a lot in high school?
Not OP, but I think they're suggesting to spend a bit more time listening to customers, and a bit less time thinking that you know better what they want then they do themselves, even if you believe that your experience and expertise would make it that you do.
I'm using the impersonal "you" here. Don't know if you do that specifically, but OP seems to complain about software workers who do have this tendency, and seems to think that eventually leads to losing the customer.
> Though some customers have shown desire to want to manage these signals
Let me be one of those. I want to be able to control the algorithm in all of its parameters, but even the ability to control the values for new/familiar and how far to stray from the current genre would be very welcome.
>>>one other consideration is disliking can usually be inferred from other activity. Skipping a song after less then 20 seconds for example can train a model
GOD NOOOOOO
that would be terrible, there are all kinds of time where I will skip a song because I do not want to Listen to it RIGHT NOW, not because I do not like it or never want to listen to it ever again...
Music is very emotional, some times you just do not want to listen to X at the moment, 30 mins later you might.
Yesterday I tried to book a plane ticket on eDreams and no matter which card/bank I used I got a bank authorization error.
So I tried to contact eDreams, which requires a trip reference number.
Since I couldn't pay them for a plane ticket, I didn't have a trip reference number. So all their ticketing and phone tree systems would hang up on me.
eDreams customers, evidently, don't want to contact them without having a trip reference number. And they know this for sure because they've never had a customer contact them without one.
The dislike button is if nothing else a sanity check for someone to have control if your model goes haywire. Realistically it will go haywire in a non-trivial number of cases. If every month your model goes haywire for 1% of users then you may see dislike used 1% of the time. That's tiny, right? However without it those users may leave and then next month another 1% will be hit and so on. Users remember the bad things a lot more than the good things so they will remember your AI f-ups and the lack of a way to fix it.
The problem with data driven UX is that the data is often short term (ab tests are expensive) and doesn't cover long term complex user impact.
Edit: For example, I’ve seen a lot of Etsy users complain about their front page being filled with NSFW recommendations and no way to get rid of them. I doubt they’ll be using Etsy much in the future.
This and your responses below remind can be applied to handicapped parking or handicapped ramps (ie, they aren't used by most users therefore we shouldn't have them).
I prefer and use the dislike button when it exists. I don't care that 51% or 99% of users don't use it. Further, I want to send a concrete signal. Skipping a song doesn't mean I don't like it. It might just be that I didn't want to hear that particular song right now. Even if I skip it 10 times in a row that doesn't mean I don't like it.
It's honestly really stressful using systems which do this. I have to worry about the signals I'm sending to the system all the time. Did I hover over that clickbait too long? Did I alt-tab while a song I liked was playing? Did I accidentally click like on something? Did I misclick onto the wrong song?
Sure I might enjoy some clickbaity content every once in a while, but that doesn't mean I want the system to show me more of it, even if that will lead to me watching more things.
> Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
So, basically if I dislike a song but want to give it a chance and then decide that you want to never hear it again, your industry thinks that I liked it. That explains a lot.
> I work in this industry and actually we’ve found from online experimentation customers generally don’t want to tell you they dislike something.
> Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
Really? Cause I constantly wish for granularity in my ability to dislike something. Insta-skip might mean I'm just not in the mood. Or it might be "never play goddamn clownstep in my neurofunk mix". Or it might be "I actually love that song but I reserve it for special occasions to get into a specific headspace and if I listen to it too much it loses spark". Same signal, three radically different valences.
I just wish services didn't treat all their customers like simps. I'd easily pay double for the ability to like, specify "Tuesday is flying microtonal banana day" or "start switching to prog house around 16:00 because that's when I start to get really deep into code but also easily frazzled." I'd pay gobs of money for Pandora's ability to discover new music with the library of Youtube+Spotify.
I'll never use Spotify again. My account was hacked by a latino and now my recommendations is 100% latin music. They told me there is nothing they can do to fix this. My recommendations are broken and my only option is to create another account, cancel my subscription, buy it again... Not gonna happen.
Yeah me too, a couple of years ago. Initially I was in denial that my account might have been compromised, but eventually I changed my password and haven't had that issue since.
I found out you can break your recommendations without losing control of your account.
I listened to one podcast episode and it's been stuck on my home screen for two months. I've tried listening to it, ignoring it, unsubscribing, deleting (you can't!). It's just there forever now.
And of course, that means it's always recommending related podcasts all the time so most of my homepage is now filled with this stuff and I have no way to remove it.
I’ve never listened to a podcast ever on Spotify. Refuse to, given their approach of trying to wall off podcasts in their own app.
I’ve had an entire top-level list of podcasts recommended to me for over a YEAR now. I will never use it, but I have to constantly scroll past it and dodge podcast recommendations in new podcast + music playlists (that aren’t clearly marked) as well.
I used to love Spotify. I got so many friends to sign up, I discovered so much music in my Discover Weekly, I convinced friends to switch from all kinds of other platforms, I built collaborative playlists with friends…
but the recommendations seem stuck in a rut for me, probably because they can’t handle the size of my library (hundreds of artists and thousands of albums that I listen to regularly). The product doesn’t respect me. I’m actively researching ways to host my own music library and use something else for discovery. Recommendations welcome so this frog can escape this slowly boiling pot.
Argh, I’ve been having this exact issue too. Worse still, it’s a thumbnail and title I would really rather not be plastered front and centre of my home screen all the time! Absolutely frustrating.
Listening to metal gets a lot of recommendations for me in languages I don't understand much of. You would think Spotify would have a way to not change languages between songs at least.
Metal fans who listen to things beyond the mainstream hits tend to listen to a lot of music in languages they don't speak due to how much of the best metal out there isn't in English. If you listen to a lot of bands whose fans listen to music in other languages then I'd expect you to get a lot of recommendations in other languages.
I understand why it happens, but it doesn't mean I want it to happen. Unfortunately, this is a case where a recommender system runs into issues. Outside of these songs, I likely overlap a lot with the people who are okay with non English recommendations.
Hmm, I had a similar situation with an ex-partner contaminating my recommendations. I emailed support and it was pretty easy to reset my recs to a blank slate.
It happened to me with someone that loves Simple Plan. I only noticed in those end-of-the-year lists that Spotify prepares. A bunch of Simple Plan stuff on most played tracks.
I changed my password and it took me almost an year of listening to my songs to Spotify stop recommending me Simple Plan and similars.
I once didn't properly unlink Spotify when I returned my car to the leasing company. The driver taking it back played bhangra for about half an hour using my account before I noticed. Changing the password, unlinking devices nothing worked. I eventually got him to stop by constantly switching it to the Barbie Girl by Aqua and then finally Spotify support was able to unlink it. I was then recommended bhangra for months. Four consecutive Discover Weeklies were 100% bhangra. Now I like the odd track every now and then, but playing less than an hour of a genre should not be enough to do that.
I think the main issue is that sometimes you just don’t want the song but other songs like that are fine, and other times you don’t want that song or any song like it.
- Don’t play this song again but others like this are fine
i am actually impressed to see this come out cornell. i tend to see like and dislike buttons as too simplistic. they greatly reduce why a user would dislike something.
i am annoyed at youtube only allowing me to choose from "i didn't like it" or "i have already watched it" for example. i have thousands of other reasons.
A dislike button is too simplistic, but having no way to dislike something is extremely bad user experience. How do you train a recommendation engine without a negative signal?
Sometimes I dislike something and I don't want it around anymore. Which is the reason I've migrated to Youtube Music, which is an inferior product, but has a dislike button.
> i tend to see like and dislike buttons as too simplistic.
Same here. I’m still a big fan of star ratings in iTunes, though I’m not sure if they do anything or influence playback. I have a pretty eclectic taste in music and only want to hear certain songs once in a while and I give those songs get three stars. (Sorry Donnie Iris if you are reading this.)
I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you really like rating and being a music critic https://piki.page.link/Rng5
Mi problema con Spotify es que cuando le hackeo la cuenta a un gringo y escucho mi música, a este le molestan las recomendaciones latinas y despues de un tiempo deja de pagar la cuenta.
How do you even handle this? You can't both be listening music at the same time. Are you sure the gringo doesn't stop paying because he keeps having his playback stopped?
I would use it, and dont understand why its only avaiable in Spotify Radio and not everywhere. I think my dislike of a song in an album can still be utilized to refine my suggestions.
Whats worse, there was a song that was constantly playing after my playlist ended. No matter how many times I marked the song as Dislike, it was always in first of 5 songs that play when playlist finished. Made sure the song appeared nowhere in my playlists (had to manually search in each), and neither did the artist.
After contacting support and jumping trough few "restart app" styled suggestion, I was suggested to mark the artist as "never play this artist". This surprised me, as I never seen that option before. And then I understood why: its only in mobile version. You cant solve it on PC.
I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you want to start your recommendation journey from scratch: https://piki.page.link/Rng5
So basically they’re developing a product marketed to streaming platforms to identify what’s “more attractive” to consumers. The algorithms are still optimized for increasing sales and engagement, not improving recommendations (for the user).
That’s exactly why I’m unhappy with Spotify lately. I feel I’m being sold a bunch of corporate portfolios on what’s currently in vogue, and not just artistically but politically, with playlists like this-gender-race supporting this-cause. I actually just want to listen to music.
I still use it but stumble my way through looking at related artists while avoiding playlists and recommendations.
Yeah, this is the exact reason I switched to Apple Music even though I'm on Android, since their playlists (for the most part) seem to actually be made based on what they think will be interesting for the listeners, not what they got paid to promote. I might be wrong though and just getting played.
similar experience here, after 12 years with spotify I had enough of the confusing UI and poor recommendations. They literally know all the music I like and still I don’t discover enough new music as I did just listening to the car radio. Bought into Apple One and will see how that goes, I just need to migrate some playlists.
whoa thanks for the heads-up! I've always wanted to try other platforms but I've thought it impossible since I have so much data in Spotify. Any gotchas here?
You will have to share access to your data, so expect them to sniff a lot about you and your friends from Spotify. Highly recommended to remove access once the migration is complete.
The free version transfers only 1000 tracks, about 1-2 songs per second, and some might be missing in the destination, e.g. when moving from Spotify to Apple.
Favorite artists cannot be transferred to Apple Music.
I use Apple Music because it works well with the rest of the ecosystem. But the official playlists just seem to resurface what’s already popular, at least for dance, electronic and hip-hop. Whereas Spotify introduced me to new music. Maybe it depends on what genres you listen to?
So basically they’re developing a product marketed to streaming platforms to identify what’s “more attractive” to consumers. The algorithms are still optimized for increasing sales and engagement, not improving recommendations (for the user).
It's possible that the recommendations are both good for users and good for corporate interests. They might not work for you, but for millions of Spotify users they seem to work. People listen to them a lot.
I listen to my recommended "Discover" playlists occasionally. They're decent. They include things I haven't heard and quite like. Maybe record labels are paying to be on them. Oh well.
I’ve never really understood what variations of the rejoinder “you are not the target user” are intended to accomplish, at least in conversations like this.
When discussing things like product strategy it makes some (more than a little) sense. But in a conversation about personal preference, what do you expect the reader to take away from it? “Oh okay, sorry, I didn’t realize I wasn’t meant to like this. I guess my opinion’s invalid.”
Who cares who the average user is, when someone is saying something doesn’t appeal to them? Is the sentiment some kind of scolding for not liking it? I sincerely don’t understand.
I think the remark is meant to address internet comments' tendency to jump from "this product doesn't meet my needs", to "consequently it is a bad / mismanaged product".
From the original comment: "The algorithms are still optimized for increasing sales and engagement, not improving recommendations (for the user)." That is an opinion about 'product strategy'. The answers, in this context, are confirming that 'not for the user' part. I find relevant to highlight that HN may not be the most representative crowd in this situation.
I do not use Spotify, nor I had for years. And I do not like the level of influence that all those algorithms have on the population decisions. So, it's not about protecting Spotify but an observation to try to add another point of view to the discussion.
I’ve never really understood what variations of the rejoinder “you are not the target user” are intended to accomplish, at least in conversations like this.
It's simply a reminder that when you work at scale you can't please everyone. Someone complaining that a feature doesn't work for them is not the same as saying it doesn't work.
On a site like HN the conversation is usually about the broader picture rather than individual complaints unless someone is responding directly to the CEO of a company. I think the CEO of Spotify posts here occasionally, so maybe he'll reply. The rest of us are talking about it in more general terms.
That’s just the appeal to the majority fallacy. People may just use Spotify because it’s free with ads and they have the hook of personal libraries to keep you stuck on the service as a paying member.
I want a service where I can upload my own files and easily access them from a player interface in the browser, mobile, etc.
I want the ability to rate music across multiple user-configurable dimensions. Add tags. Create smart playlists that interpolate between these.
I want to be able to pay a fee to subscribe to music discovery, then be able to mix these with my own library. If I really like a track, I'd like to buy it and add it to my collection.
I want an open API so desktop apps can be written to use it. Also, let me export my annotations and music library on demand.
Music for power users. Don't give me a single button. Give me hundreds of them.
I looked for that when Play Music got shuttered but could not find a satisfactory solution. I even paid for one iOS app but it wanted access to my entire drive, and maybe some other unnecessary permissions.. so hell no. I should be able pick one folder and that's it.
I guess I could have made a separate Google account just for music.
I bought a jelly pro (miniscule phone), rooted it, added some automation to autoload a music player on boot and activate airplane mode. I added a 256gb memory card with a lot of tunes. It's now my mp3 player. When I deactivate airplane mode it scrobbles my played tracks and 15 minutes later it automatically reactivates airplane mode unless the phone is charging, in which case it's available for wireless music sync from MusicBee. It also powers itself off after 30 minutes of being idle. Best damn mp3 player I've ever had.
- Upload and download your own stuff if you want to (marketed as iCloud Music Library)
- Works offline
- Discovery services available
- API exists but mostly used to implement web players for some reason, works fine in desktop apps too, on top of that the current state of your music library is always available as an XML file even when you don't use the API at all.
It's fingerprint matching, and yeah it's poorly implemented. Most recent complaints I've seen on this were for the fan made explicit version of Kanye's Donda. It was basically impossible to upload since it would just get matched to the clean version.
Matching is optional. If a song cannot be matched, your own copy gets uploaded instead to your private iCloud library.
Edit: well, it's a bit nuanced of course, you can not match and not sync or match and sync but you can't mix and match that configuration. So it's either sync with matching when possible or no syncing.
I have a lot of rare stuff. Demos, live albums (grateful Dead, dmb), live captures of daft punk at Coachella, underground hip hop mixtapes, CDrips of local punk and Ska bands from the 90s
I haven't had an active subscription for a while (when iTunes was the desktop app instead of Music being the desktop app) but I distinctly remember being able to set metadata in the Info panel on a song or multiple at once that prevented fingerprint matching.
Check out Navidrome [0]. It's the closest one that I've found. It supports the Subsonic API so there are plenty of mobile apps and probably some desktop ones that work with it. I use play:Sub on iOS.
I use Airsonic right now, it is basically Jellyfin for a music collection. I use both of these as reliable alternative to Netflix/.../Spotify (where ... is a plethora of other video services such as Disney+).
3. More modern webui (though I recently found a good desktop subsonic client, sonixd which somewhat negates this)
Also, I haven't used the airsonic fork, but I was using subsonic before switching to navidrome
I found subsonic to be rather... unreliable. once in a while the database would randomly become corrupted and I'd have to delete the database files and start over. I haven't had any such problems with navidrome, it's been rock solid.
Unless I'm mistaken, this doesn't fulfill the requirement:
> I want the ability to rate music across multiple user-configurable dimensions. Add tags. Create smart playlists that interpolate between these.
I'd actually pay for a 3rd party metadata service that doesn't actually provide the music at all, but just let's me tag and rate some across all music streaming services. I've even thought about building that (simply for myself to begin with). I want ratings, instrument tags, mood tags etc. Let me search for songs with 'piano + synth + dreamy + weird' instead of throwing some stupid recommendation my way. And use those tags to find similar tracks across genres and decades of time, instead of just saying "Nirvana and Pearl Jam must be the same because they're early 90s Seattle bands".
But as the name already implies, that one preferentially tries to match your files to the ones already available in Apple's music catalogue, with file upload only serving as a fallback if it can't find any matches.
And from what I've heard, the fingerprinting algorithm is fuzzy enough that it will often match songs/song versions that are very similar, but not actually identical, like censored/uncensored lyrics, different masterings, differing fade ins/outs, alternate takes and live versions, etc. etc.
Oh interesting – somehow in my mind I've mainly associated this phenomenon with Apple, but of course in a way this make sense and saving storage space that way is a tempting target for that kind of service I guess.
It makes me curious though how the accuracy vs. deduplication efficiency trade-off looks like in practice: How much storage is Apple saving with their current settings, and how much deduplication would they lose if they made their audio fingerprinting more accurate, up to a level where even audio buffs would stop complaining
funkwhale has the web interface, the subsonic api (dunno how well that works i dont use it), you can create your own "radios" based on genre and artist
My service, Astiga, does this. You can upload your files to any supported cloud storage service, pair it with Astiga, then play either via the Web, our apps or any other client that supports Subsonic.
The fact that Spotify doesn't optimise for discovery and recommendations is why you use physical media which has absolutely no mechanism for discovery and recommendation?
That's the exact reason I switched to YouTube Music.
I remember even about 10 years ago, last.fm's algorithm on bringing songs I would enjoy basing on previously listened/liked songs was far more accurate than Spotify's today.
Does the algorithm of YouTube Music behave similar to standard YouTube? YouTube (not Music) basically always recommends the same tracks in the same order and within the same genre bubble. It is hard to discover anything new which makes me not want to try YouTube Music.
My experience is the algorithm is a bit different. I have not been on the platform for long, but my recommendations have been great. I've also been listening to music on the same youtube account for 10+ years, so I assume they have alot of data.
This was my experience with YouTube music - no matter what genre station I started off with, it eventually settled back onto the same small number of currently-trending songs. I get much more diverse recommendations from Spotify.
I just switched to youtube music, its leagues better than spotify. YouTube music seems to have a pretty solid recommendation algorithm, and I frequently find music more in tune with my tastes there than on spotify. Honestly, I would not mind seeing spotify disappear.
I'm thinking of going the other way, I just dislike how YT Music doesn't have some small feature like saving my queue to a playlist. But perhaps I should hold out.
Fwiw, I was shocked how much better YouTube music was than Spotify. I turned on the trial on a whim, and now I actually miss it (it expired a few days ago).
It was so nice that I’m seriously considering just turning off Spotify. It’s sort of interesting to analyze why we don’t —- for me, it’s become unconscious habit to reach for Spotify and not anything else. Plus a lot of other stuff integrates well with it.
(What if… what if we can use both? Mind asplode, it’s not a decision.)
Weird, I switched to Spotify once Google Play Music (or whatever silly name it had when Google ate songza) became YouTube Music because I found it such a poor experience.
Perhaps I was missing something but the change def did not improve the recommendations - it made them drastically worse.
In my experience, Spotify's recommendations are good for a few months, and then stick in a rut at some point for reasons I don't understand. So you're probably in that nice honeymoon phase where it's actually allowing you to discover new things instead of surfacing the same 20 artists over and over and over again.
For me Spotify was discovering good new music for years before getting “stuck” recommending the same artists over and over again last year. I do mark them as “I dont like this song” but it still will recommend them later. I think it’s time to reset and start over.
Its funny how much better Googles side show product is. Shows that the internal technology is truly advanced and capable of swiping away another company at the drop of a hat.
I'm just trying it out now based on the recommendations in this thread. When you start it asks you to pick some favorite artists. The first few rows were clearly based on artists that I've watched recently on YouTube, so it's pulling in history (I mean...no surprise).
Then I scrolled down a bit b/c I wanted to give it a strong signal of what I liked. I found an artist, clicked it, and noticed the recommendations below changed immediately afterwards.
So then I scrolled some more (because there were still 95% misses), found one I liked, scrolled to see the next row then clicked on the artist I liked above to see if it indeed changed.
It did. But not only that. I love every artist on the following row. Then it quickly diffuses back to noise, but holy cow that was a bit of a spine tingler lol.
YouTube makes recommendations around engagement metrics only (1 political video gets you Fox News recommendations for years), spotify still gives me plenty of small bands - I think if you’re listening to very popular music you’re screwed either ways
I get mostly small bands in Spotify and I worry that Spotify is actually bias towards them. I assume the royalties are cheaper for smaller bands and that may factor into recommendations.
I was at a conference the other week and Spotify had a Keynote. They talked about the tradeoff between playing "diverse songs" and "consistent songs". It is a hard problem to solve.
When you start your listening session they try and predict how long you are going going listen (based on your past history and time of day). If you are probably going to listen for a while, they are more risky and might play something "different". Playing different stuff is risky (short term) because you MIGHT not like it. Playing the same stuff is safe (short term) because they know you will like it - BUT people will eventually go searching for something new, so they have to risk diversity eventually.
Exactly. There's no reason the "personlized" Daily Mixes can't be built/labeled to indicate "acaccuracy" rate. It annoys me when I'm in the mood for new or different - which is 60% of the time - and all the new Daily Mixes feel like the day / week before.
I'd love a personalized playlist titled "Curveballs" that contined things different and/or challenging.
With respect, your comment illustrates the issue - a burden to reeducate people by shaping their culture. Music can make a statement, sure, but a world where people are force-fed curated playlists designed to mold them into some Standard Issue set of beliefs, this is a dystopian vision.
I think that people bristle at it in the same way that people used to bristle at having Christian-normative culture pushed on them. Not everyone enjoys cultural imperialism.
Sarcasm and belittlement are a terrible way to engage in conversation. You seem more interested in aggrandizing yourself than to actually make a point or bring some food for thought.
On top of that, if you listen to a wide range of genres the attempt to shoehorn your history into 6 Daily Mixes yields some comical results. Like Doobie Brothers followed by Nirvana.
Give me a home screen with 12 or more Daily Mixes that segment my listening behavior at a finer resolution, and I'd be a lot happier with the service.
I understand why all the other playlists exist, but I generally have an idea of exactly what I'm trying to listen to, so these low effort curated playlists are pretty useless for my listening style.
Then you can simply find a song that matches what you're exactly wanting to listen to and start the "song radio" from there. For me the problem is exactly the contrary, I can't find enough diversity in the daily mixes, it's mostly all songs I've either explicitly "liked" or heard a couple times before.
Diversity in daily mixes is an issue for sure. I do use the song radio feature but find it's hit or miss - they usually start off strong, but seem to lose the thread at a certain point.
It's a fine balance between existing liked songs and expanding within sub-genres, but I have at least noticed the recommendations improving over the past few years.
I suspect it would help to port over my entire pre-Spotify music library into Spotify to provide a bit more data, my current library is all post-Spotify so it fails to capture the breadth of my music taste. I've just never gotten around to it.
I've considered getting a spotify family account, using separate accounts for major genres and one for browsing. I have a browsing profile for Netflix so I can freely explore without messing up the suggestions.
For me, I listen to a lot of 40s, 50 - 60s lounge / exotica, early to mid 90s hip hop, 80s metal, and the standard indie stuff... then I have a bunch of chillwave and synthwave stuff that throws another wrench in the mix. The daily mixes I get are a total mess, much like yours.
The genre-specific mixes they make are pretty decent, but discovery is low.
I have a separate account just for Amazon Echo because I don't want my kid's selections to influence my main account suggestions.
I also have my own account for the car, where explicit songs are disallowed. This allows me to listen to the music I like, but it avoid explicit songs for when my kids are in the car and that setting doesn't affect my main account.
The local classic rock radio station I listened to growing up has continued to slowly expand what they consider "classic"; whenever I go back home, it's always interesting to hear stuff like grunge and some metal (like certain Metallica songs). Hearing a Nirvana song right after the Doobie Brothers is exactly the type of thing I've probably heard before from them!
I agree. I love the wider scope of Spotify’s mixes. If I wanted a narrow scope then I’d just have created the playlist myself like the old days (or just thrown an album on). The reason I use Spotify mixes is for a variety within an approximate mood. And I’ve discovered so many good news tunes and artists through their mixes.
That all said, I do wish I could turn off their podcast recommendations. I never listen to podcasts and worse yet they keep shoving that same comedians is absolutely hate (and there aren’t many comedians I dislike; which just goes to show how far off the mark their podcast recommendations are)
That example might be OK, but my daily mixes include "Weird Al, Nerdcore Comedy, and also Radiohead" and "60s/70s/80s Rock and also Broadway Show Tunes". The transitions are very bizarre.
Well for me at least, I listen to music depending on mood. It's not generally that I dislike mixing, quite the opposite actually. But when I'm in the mood for metalcore, I'm not in the mood for happy hardcore. And when I'm looking for something quick, I don't want a slow song mixed in.
I like the idea of mixing genres, but Spotify seems to totally miss on what aspects I want mixed.
I've never used Spotify or any other recommendation service. Music is like food: it's far too important too substitute with junk. I won't eat Subway, I won't listen to auto-generated recommendations. I browse (and support) rateyourmusic.com, I use the last.fm API (to find out what my neighbors are listening to) and I listen to music for free on YouTube before I buy it. I also heavily use tags on my purchased music so I can easily put together a playlist matching my mood.
Disagree with the premise - you don’t have to sacrifice productivity to put time into finding good music. For one, you don’t have to do your music research during working hours. But even if you do, it doesn’t mean you’re trading off productivity in order to do it.
you don’t have to do your music research during working hours. But even if you do, it doesn’t mean you’re trading off productivity in order to do it.
Trying to find music when you could be focusing is by definition, trading off productivity.
I’d rather have an algorithm dictate what songs I want to listen to (in order to focus) than spend an hour wading through junk to find something i like before I start coding.
Anything that can be exploited by the user for self-exploration and discovery is being removed - for the purpose of serving content that generates more revenue, I assume.. At first it was subtle, but the gloves came off with the recent UI overhaul.
Playlist search result page is just an endless grid of images and truncated names. For the playlist duration, description, number of songs, follower count, etc. you have to actually open each individual playlist. Good luck finding what you were looking for.
More and more Spotify-fabricated content is being pushed. Most of which contains the same limited selection of songs that Spotify keeps feeding you over and over anyway.
Podcasts aren't my thing, Spotify wants me to listen to them really badly though. Majority of the ones they're suggesting I'm not at all interested in, and sometimes some of the podcasts they're advertising seem to contain some pretty disturbing content. On a sidenote: I don't know who he is, or what he does, but I hate Joe Rogan and Spotify is to blame for it.
There is a setting hidden under advanced that is supposed to make Spotify stop messing with the shuffle functionality. It is labeled "Allow smooth transitions between songs in a playlist" vague huh, it's also placed directly underneath the song crossfade slider. I'm fully convinced that this was done on purpose. Also this setting seems to do precisely nothing at all, so I'm not sure why they even went through all the effort.
Shuffle is not random. If this is so on purpose, that purpose does not involve happy users. Else perhaps their devs are afraid of touching some jank script that might be holding it together.
Edit:
Almost forgot about the new artist's pages! They used to consist of a long list of all songs grouped by album. This was far too convenient for us users, so with the redesign they simply removed the lists of songs leaving only a grid of albums, forcing you to go into each individual album to find a specific song and play it play it.
After many complaints they implemented something vaguely resembling what we had before, but with such odd UX that it must be sabotaged on purpose again. But of course, adding this overview back to the artist page was out of the question. Instead what they implemented as the only way to access this, and I kid you not, is a plain text link in the most random place ever.
No one is going to use this feature if they don't know it exists. All this just to get a reason for removing it that is spinnable. Machiavelli would have been proud.
What’s funny is that I used to love listening to the Joe Rogan podcast and hate Spotify just as much, because I tried to listen to his podcast and somehow ended up listening to about TEN MINUTES of ads and the podcast NEVER started playing. I finally uninstalled Spotify and haven’t listened to his podcast since, sorry. It’s mind boggling that Spotify could drop the ball so hard, as ostensibly they were trying to funnel people like me into their platform, and whiffed in an absolutely astounding manner while also alienating people like you who already use Spotify.
> I feel I’m being sold a bunch of corporate portfolios on what’s currently in vogue, and not just artistically but politically, with playlists like this-gender-race supporting this-cause.
I feel the same way. It's not just music either. I get this feeling every time I try to consume anything. Everything is just so fake. Like it was made just to push some silly agenda.
"Recommendations" are ads in disguise. I already block them on YouTube. Wish I could block them everywhere.
It’s fascinating to hear this feedback on Spotify because I’ve never used it this way. I don’t let it guide anything. I just use it as a music repository and I pick the albums and make my own playlists.
Do others generally find that auto play (or whatever it’s called) works well? Or does it just feel like payola radio?
Why would you let a streaming service choose what you want to listen too or a premade playlist by someone else? Then you might as well just listen to the radio. All these on any service are always rubbish. I make my own playlists and I only listen to them, sometimes I search others for some inspiration to see what songs I forgot etc, but most of them include a lot of trash so I make my own.
The only 'auto' feature I use, is my release radar playlist which I check every Friday to find new songs from Artists I like and then add those to my playlist titled for the current year.
"Discover Weekly" tends to be one of my primary sources of new content discovery. I tend to listen to a significant amount of Death Metal, Doom Metal, Folk Metal, German metal and rock, random music in languages I don't understand because the sound is cool to me, and various other never-been-pop subgenres. Discover Weekly has been pretty on point for me to find new and interesting stuff in a way that hasn't felt like them pushing a message or some corporate catalog.
I don't tend to listen to the genres that are full of artists who are trying to push politics through their music (at least that I've consciously recognized) (except maybe System of a Down). Not because I'm necessarily against their messages, it just the sets haven't overlapped much for me at this point. It could also be that music isn't much about lyrics for me, it's about if it's something that sounds good to me.
I have a bunch of playlists and saved albums, but I do listen to discover weekly in the first few days of the week. Sometimes it's meh and I just go choose what I want to listen to, sometimes it's really good and it's on loop through the week.
Their "Daily Mix" playlists seem to be 90% stuff I've clicked "like" on in the past, so I listen to those as well.
I think it is sad that there isn’t a way to recognize the development of competence in music listening— with an achievement motivation. Like, mastering genres with a music collection. Aesthetic pleasure in pleasant recommendations without any principles makes me feel vacant after a time.
oh and for streaming platforms. sometimes, it's just because that song you recommended (and started playing) should not be played after the one that just finished because it ruins the mood. this has nothing to do with the song itself.
so again, reducing it to like/dislike buttons is missing a great deal.
I have really been hoping for a spotify-API-driven recommendation engine to improve this kind of thing. Which is why I've been scrobbling to last.fm for over a decade.
One algorithm will never work for everyone. Even if it's one of the best in aggregate, that doesn't mean it's not awful for me. It's not feasible (nor does it make business sense) for Spotify to identify and build the best system for each individual user, so I really don't expect them to do more than optimize for the majority.
Me too. I still find more interesting stuff by following last.fm recommendations than anywhere else. Apple Music knows the 10k songs that are in my library and its suggestions are still meh.
I just wish they fixed the problem with the bands disambiguation. I just hate everytime I get a recommendation in the release radar for a band that happens to have the same name as the one I like but stylewise they have nothing to do with the music I usually listen.
One is a progressive blend of funk, hard rock, jazz, psychedelic from New York, active in the mid 90s to 2010[0], and the other is an anti-fascist (anti)black-metal band from London formed in 2018[1].
Both are interesting, but I only like the music of the first one. Google Play Music used to mix them together.
I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you want the opportunity to dislike Coldplay ;) https://piki.page.link/Rng5
I cancelled spotify recently when I had a guest in my car who proceeded to play all her favorite music on my account. I wanted to remove her music from my history so spotify would hopefully not recommend her music to me but there is no way to delete your history
FYI, Google gives you some decent tools for managing your YouTube watch history. You can search by date or title, and you can also turn on auto-delete.
It hardly matters now, but before you do anything with a guest, you can start a private session to make sure their history is ignored. Still a very weird feature to omit; I'd reckon that most people would use history deletion at some point
I use this a lot when playing music for my kid, but unfortunately the feature is clunky: weird path to find it, self-auto-disable much later but with the song still loaded. Hopefully this changes someday.
I have a similar problem where I play Disney music/movie music in the car when I drive my kid to school. Every time I have to go find "private mode" deep in the settings. Typically I just forget. I am primarily a discover weekly and release radar user, so it's definitely frustrating. They have basically done almost nothing positive for users in years in terms of UX. They tried to git rid of dislike, but brought it back because it ruins those suggestion playlists, when you can't hide awful songs.
I'm in a similar situation at the moment. Somehow Spotify has decided that I listened to a bunch of Hindu music a couple of months ago and now over 50% of my recommended playlists and podcasts are in that genre.
Their refusal to let you delete history is just puzzling.
I think more likely is I searched for something, hit the wrong result, then got distracted while it was autoplaying through related artists for a few hours.
It seems to be a bit of a mix. Some is definitely religious Hindu music, other bits seem to be more broadly Indian/Bollywood.
As is the curse of modern technology I don't want to interact with it at all in case the algorithm decides I'm into it and undoes my work retraining my preferences.
I just checked in to an airbnb and someone left their google account signed in to the youtube app on the TV. The temptation to spike the recommender with something is very strong. :)
If you contact their support, they can probably help you. My account got hacked a while back and I started getting all sorts of weird recommendations. I contacted their support and they were basically able to reset it to an arbitrary date in the past, thus restoring it. In UI would obviously be better, but FYI, this is an option...
I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you want next level recommendations https://piki.page.link/Rng5
YT Music works pretty well for me, using the dislike/like buttons for when I know what I think about the song and just skipping when i'm just not in the mood for that song.
The time invested in my Pandora stations keep me on that platform. I actually really hate that they recently removed the downvote buttons in the Android auto interface.
This is the #2 feature I most want in every music streaming app
(The feature I most want is an interface for tweaking the rec algo to my liking - at minimum, for being able to crank up its adventurousness. Rec algos are depressingly conservative as a rule, I guess because most customers are the kind of people who eat at McDonald's when they travel abroad, stay at package resorts, and complain movies have gone to shit because they can't find anything they like on Netflix or Hulu)
> at minimum, for being able to crank up its adventurousness.
Absolutely. At times, it makes me wanna blast it with cosmic rays, induce a burst of mutations, mix things up. Just give me a knob. Sometimes I want to stick on one specific sound. Other times I need to explore.
I'm the creator of Piki, the Cornell Tech app mentioned in the article. The recommendations are more "risky" than your typical streaming app recommendations. https://piki.page.link/Rng5
An option to hide all podcasts would significantly improve spotify's recommendations. I don't listen to podcasts (though I have accidentally clicked on one, when a podcast interviewed a musician I like... blech). Why is my screen filled with podcasts multiple times per day? I like the music recommendations, "so and so artist radio," etc., but they make me hunt for them.
Also, while I like a broad variety of genres, I only like listening to one at a time. I don't want a rap mix to be invaded by a Bach sonata. And yet...
Seriously. I listen to podcasts, but I'm not going to do it through Spotify, who's working to build a walled garden around a traditionally open media. I'm also not interested in Joe Rogan, which they _constantly_ plaster my front page with.
I did try a Spotify podcast once. It was on my recommended list and was called daily dad jokes or something. Figured it was worth a quick laugh. Well it turned out to be a bot podcast. They set up a bot to rip jokes from Reddit, push them through TTS, and dump it into a Spotify podcast.
No thanks.
Meanwhile, Apple Podcasts, my go to, works great, doesn't spam me, and even added a feature where you can pay for a no-ads version of a podcast. Worked seemlessly.
I rarely listen to podcasts but was going to try and listen to a specific one Joe Rogan did recently. After listening for a few minutes, an ad came on. I pay for spotify, why am I listening to an ad?
Anyway, I waited like 90 seconds and he came back on. Withing about 15 minutes, I had two more ads come on and they were 90 & 120 seconds each. After the second ad played, the podcast didn't start. Fiddling with it caused the podcast to start over from the beginning. When I tried to seek to the last place I was, it played an ad immediately. Seriously unbelievable.
I gave up and will never attempt to listen to an ad supported podcast on Spotify again.
Because Spotify doesn’t pay anything to the podcast producers. So podcasts run their own ads. (Well, some do. Others live off donations or offer a way to pay them)
I have never listened to a podcast on Spotify and I still have them on my home screen, so I think that it's hardcoded, not a generated recommendation.
The probably autogenerated daily mixes have been surprisingly non-bad for me. Nothing like rap mix invaded by Bach, but the "classic rock" mix sometimes has Ghost that doesn't really sound that out of place.
My guess is that they don't have to pay rights for the podcasts, so for them it's a win/win if you spend your attention on their app listening to free content.
I listened to one episode of "Call Her Daddy" because I saw it ranked as one of the most-listened podcasts out there and Spotify will not let go putting it front and center on my home screen every day.
I'm really disappointed how Spotify is buying several big podcasts and moving to their platform exclusively.
I tried to listen to a Science Vs episode the other day, and now my Spotify is almost completely filled with podcasts recommendation. Plus, the experience is still far from specialized podcast apps.
The result: I'll probably stop following Science Vs (and any other podcast that moves exclusively to Spotify).
They could, at least, have a separate Spotify Podcast app so things wouldn't be mixed (kind of like Wealthsimple has separate apps for Investing and Trading).
I'm still sad Spotify bought Gimlet and apparently will do (if not already) this to all their podcasts.
Why Spotify plasters your home page with Podcasts:
Short answer is $$$$.
Longer answer is:
Spotify must pay royalties for each song played. Imagine if there was a completely free (for Spotify) form of content that filled users ears for hours, thus removing the need for Spotify to pay royalties. Ahem podcasts. Now imagine if Spotify started injecting ads into said media form to grow their revenue beyond subscriptions. Again, podcasts. So now, you have a very long-form content that both saves you royalty $ and drives new revenue. QED. Podcasts will continue to be plastered all over your recommendations, be top search results, etc, until the above stops being true.
Selling a burger is inhernelty profitable for McD, they sell it for more than they spend on manufacturing and shipping and serving it. Each new burger sold adds more profit.
Serving music is unprofitable for Spotify. They collect your monthly subscription fee, then every additional song play makes Spotify lose money. They don't want to have zero song plays since then nobody would subscribe, but their goal is sell the profitable thing (subscriptions) while minimizing the unprofitable thing (song plays).
Well...yes, not necessarily fundamental, but it's the same class of problem cinemas and restaurants have (to pick two). What do cinemas mainly make money from? What do restaurants mainly make money from?
Yes, that seems to be true (and is that % you've guessed maybe too low? Maybe far too low?). Even with that though, it's extraordinarily fine margins -- w/r/t the businesses that I mentioned, the majority of the income comes from those value-adds, if they weren't there it would generally be difficult for the businesses to survive financially. That's where I was coming from.
To me, Spotify seems to be using podcasts in a similar way. I assume it's because the central business model is unsustainable when combined with investor pressure; they can't just focus on core product because they're burning too much of other people's money in an attempt to outcompete everyone else
Just to add to this restaurant based allegory, Spotify is like an all you can eat buffet that tries to serve out lots of sodas so you get bloated and don’t eat the expensive stuff they have
> their goal is [...] minimizing the unprofitable thing (song plays).
If that's the case, how do you explain those two features :
- the repeat button
- the "automatically play similar songs" option
In the first case, since I don't know how it works, maybe they have to pay only one time for a song per user listening to it during a period (say a month). But I doubt it.
In the second case, it's Spotify explicitly saying : "here are songs you don't intended to play but that we picked automatically and are playing to you".
If you look at Netflix, sure they have the feature that automatically launch the next episode. Or they try weird stuff like live or play anything... But they also have the "Are you still watching ?" feature, to make sure that they don't display content to an empty room.
Spotify doesn't have that, it can just play songs indefinitely without any human interaction.
Spotify dynamically injects ads into podcasts. I was listening to an episode of the Conan O'Brien podcast from 2018 and got an advertisement to be safe from Covid by enjoying an ice cold Miller Lite at home.
It was extremely unsettling until I realized what they were doing.
They replace the ad breaks that would normally be in a podcast. (Like when the hosts say "we're going to take a quick break to talk about our sponsors for this episode"). It's not an unskippable ad like you get when listening to music using the free tier. It's inserted directly into the audio stream.
Edit: I guess it's possible the Conan podcast is doing that somehow and not Spotify, but I've never had something like that happen when using the Apple podcast app.
Are you certain it was Spotify that did that? Unless you downloaded the episode in 2018, it could have easily been the Conan podcast. Podcast episodes aren't immutable, they're just a url. I've seen podcasts that update the ads in their back catalogue to whoever is paying them at the time the episode is downloaded.
I've also noticed some location specific ads, which isn't that surprising if you think about. There's nothing stopping them from serving you a different mp3 file depending on your IP geolocation.
When I listen through the apple podcast app I get ads that are clearly from 2018 though. Like for events that are long gone and over. Ads that don't match the year of the episode have only ever happened to me in spotify.
> Spotify Podcast Ads are powered by Streaming Ad Insertion (SAI), which leverages streaming to deliver Spotify’s full digital suite of planning, reporting, and measurement capabilities. Spotify Podcast Ads offer the intimacy and quality of traditional podcast ads with the precision and transparency of modern-day digital marketing.
You’re probably right. But then again: who cares about Spotify making money? I want to listen to music and not optimize the bottom line of the Company.
I think the more likely answer is that podcasts can be produced with exclusivity deals while songs can not.
Spotify is large because they were first and because apple/google music suck. But eventually apple and google will stop sucking and users will drain from Spotify rapidly unless Spotify can create content to keep users in.
Spotify is in the position of Firefox 15 years ago right now. Eventually the built in apps will take over.
You aren't wrong, but you would think they would then create an excellent podcast player, yet they can't even get simple things like the order of the episodes correctly, and I still haven't figured out how to setup a next queue.
Not to mention that I might listen to podcasts at home and one the way to work, but at work I won't want to listen to anything spoken, but half the time I end up accidentally starting the podcast anyway. That now means I am not using Spotify for podcasts.
I'm sure it boils down to money ultimately, but it can't be the way you suggest here.
The thing is that Spotify does not pay per play! The royalties lobby love to give the impression that they do, because it makes it look like Spotify is fleecing them.
Spotify pays percentages of what it itself collects in subscription fees. If a record company owned 10% of everything that was listened to on Spotify this month, they would make _the same_, regardless of how many absolute seconds their stuff was listened to.
So, more listening to podcasts does not mean Spotify has to pay out less. How much they have to pay out in total is fixed as a share of their subscription (and ad) revenue anyway, in multi-year, industry wide contracts with copyright owners.
There are ways Spotify could circumvent this, such as promoting content they covertly owned themselves, or content whose owners gave them kickbacks of some sort. There's some indication they do such shady things. But podcasts don't change this equation much.
I think Spotify's promotion of podcasts is simply good old fashioned loss-leading monopoly building. They're hoping that Spotify will become where listeners go for podcasts since that's where all the podcasts are, and also the place podcasters go because that's where all the listeners are.
That is a great point about loss-leading a future monopoly and of course a factor.
This isn't a hill I will die on by any means, but if you read Spotify's write up on royalties[0], they pay out based on stream share of the overall platform, on a monthly basis. E.g. If Columbia Records gets 10% of streams for a month, they get 10% of the pool of $ Spotify distributes. Now, like you said they likely have some shady practices - I would imagine Spotify has an incentive to create music like study and sleep beats themselves as those playlists get hit for 8 hrs at a time and it doesn't really matter who created the beat as long as it helps you study or sleep.
They do not indicate in [0] that podcasts are excluded from stream share. Even if they aren't included in stream share now, I'd imagine long term they very much intend for podcasts and music to be lumped into the same stream share model to drive payouts down.
I was using Spotify for podcasts because they were there, but then my home feed got dominated by them, and they occasionally get queued up and giving me a mix of podcasts and music, which I don’t want.
I started migrating to Apple Podcasts so I can context switch more easily.
It has been years since I've tried Pandora: Back when I tried it, they only let you skip a number of songs in a short time and often played things I didn't care for. Lots of slightly obscure songs weren't available. It might be better for more popular music or have improved in the last decade, I don't know. I started using Spotify.
And still now as then: It is only available in the US. I no longer live in the US.
> they only let you skip a number of songs in a short time and often played things I didn't care for.
If you aren't in the US, then it won't work, but I create multiple stations and can switch station if the music isn't good. I can't remember it ever not being good though. I use the like button to train it.
> Lots of slightly obscure songs weren't available.
If I'm looking for a specific song, I usually go to YouTube. Most of what I listen to is pretty obscure, but it might depend on the genres.
I don't usually overreact too strongly to big tech making dumb decisions like these, but Spotify's podcast pushing actually did drive me away from the platform.
I beg Spotify to think Logically about it. There is a limited amount of screen real estate available on any display, but especially phones. Q.e.d: pushing podcast-related content actually hurts the music listening experience. Conversely, pushing music content hurts the podcast listening experience. This is (mostly) inarguable; any pixel dedicated to an interface element related to podcast content is a pixel which cannot display music content; its a zero sum game.
I say "mostly" because; there are people who I'm sure love Spotify Podcasts, and having both avenues within one app is a net win. I'd be willing to accept a toggle in settings which could "focus" the app Between Music <> Music + Podcasts <> Podcasts. I don't think that's unreasonable, and I also don't feel it would be unreasonable for it to default to Music + Podcasts.
Though it brings up an interesting point: Imagine the experience of someone who does primarily come to Spotify for podcasts. Its among the worst podcast apps in the history of podcast apps! Its littered with music! That's a horrible experience!
The ironic part is, Spotify is probably happy to lose me as a customer. I'm a "Music power user" if there were such a thing. They may end up paying more in royalties for the music I listen to than I pay them, and I can't say I've listened to more than a half-dozen podcasts. "Show me the incentives, and I'll show you the outcomes": Spotify is not incentivized to build an app that benefits people who listen to a lot of music. Their bread and butter is people who listen to a bit of music and podcasts; just enough to keep them paying each month. So, I don't lose too much sleep over their slow decent into mediocrity; I'm far more concerned about the unfortunate reality that there are no longer any great options in this space, for people like me.
You’re probably right but I have yet to meet anyone who loves podcasts in Spotify. The two groups I have encountered who use it are those who never listened to podcasts at all before so don’t know any other way, and people who begrudgingly use it because some podcast they like went exclusive there. Which I guess are both good for Spotify but it puts them in a weaker position than if people actually did truly like the experience.
Spotify had both thumbs up and thumbs down button a while ago. It even asked why do you thumb down something:
Wrong Rec
Do not like the artist/show less
Do not like this song
Heard it too many times/pause for a bit
etc...
It got removed... as the whole functionality (Radio) got phased out, and the new stuff didn't incorporate it.
Employees changed, and moved on to other companies, and the new that came didn't think it will improve things. etc... and the feature got forgoten.
I think a lot of people in here are asking a 'reset my recommendations' feature. And that did exist as well at some point. Not sure why it got removed/it is not in production.
You might be on to something with the idea that it was there it was removed. I might presume it was removed because data showed that customers don't use it. The rest of this thread seems to think otherwise.
The idea of resetting recommendations is a minority feature ask from customers as well and I wish companies (including my own) handled that better. Sadly its a pretty expensive project for such a minority stake.
I'm the creator of Piki, the Cornell Tech app mentioned in the article. It's like a Tinder for Music, so the dislike is a big part of the recommendations: https://piki.page.link/Rng5
Radio functionality is still there but it’s a shell of what it used to be. I find it absolutely useless now actually because all it ever seems to do is play music I’ve already added to my library.
Each time the dislike button is pressed, it reminds the user that they might want to unsubscribe. So they removed it ;) I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you want next level recommendations. We actually embrace the dislike button https://piki.page.link/Rng5
Yup - cancelled my Spotify account a year ago, after they changed the UI for the 15th time, making features that I wanted to use harder, and pushing things on me that I don't want (podcasts). And introducing bugs constantly in some of the non-default features that I used a lot (eg. changing playlist sort order was broken at one point). Add to that their shuffle playing the same ~20% of songs in my playlists, and never hearing the other 80%, for some unknown reason. Then being gaslit when I tried to report "bugs".
It was just constantly frustrating to battle them, so I switched to Tidal, and it's been amazing. The UI is just simple and lets me play music, and nothing has broken in the last year. The only changes have been minor improvements that make it even easier to do the one thing that I want: to play music.
I find Spotify is riddled with small annoying bugs and antagonistic UI, I'm really really sick of it.
Instead of having a discussion about how it could be improved, maybe we can have one about which alternatives exist and how good they are (tidal, is it?)
EDIT: I tried Tidal, that app has a confusing mess of an interface just like Spotify. Plus they have a "convenient" bug during sign up that may sign up students with more expensive non student accounts.
I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you want a pure discovery app, with a focus on likes/dislikes feedback https://piki.page.link/Rng5
My Spotify Weekly is infested with Swedish songs. I don´t know if this is because I once listened to a Swedish song or because Spotify is Swedish. Their support told me me there is nothing they can do
I'm deliberate about any new music I hear, that I never use Spotify's recommendations. I prefer doing my own research and digging for stuff that I might like.
Oh, what I'd give for a dislike button. I tend to use Spotify to listen to nursery rhymes in the car while traveling with our toddler. This doesn't mean I need to have a daily mix full of nursery rhymes and there is no way to tell Spotify that.
I'm the creator of Piki, the Cornell Tech app mentioned in the article. Give it a try, if you want next level recommendations that takes into account your dislikes https://piki.page.link/Rng5
659 comments
[ 5.1 ms ] story [ 442 ms ] threadGenerally customers don’t have a clear picture of how their explicit signals impact recommendations. Though some customers have shown desire to want to manage these signals.
Edit: one other consideration is disliking can usually be inferred from other activity. Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
Now it ended up with me getting rid of Spotify, because of this and other frustrations.
In general, you had to be doing something to be getting recommendations so there is some action that when most customer's do demonstrates a high probability that you will like some genre and it just so happens you don't.
A feature does not need to be used often to be useful.
Take a fire extinguisher for example. Going by this logic, since most people will never even touch a fire extinguisher and most of them are never actually used, we should get rid of them because they're a cause of clutter.
Increased UI clutter due to a single extra icon is not a good nor sufficient argument for not including it.
While the general customer base might not be served by this feature, what if a subset of the customer base enjoys this feature to an extraordinary extent? Maybe it fixes a lot of problems they have with the service; and negates a lot of complaints they otherwise would have posted.
Would it still not be the correct choice to "clutter the UI" with this?
I don't think it's as easy of a choice like you're trying to make it.
But I could also be jaded by the fact that Spotify for years has trying to pull me into listening to all kinds of odd European folk music that I have no interest in what-so-ever.
It's probably fine to not have anything to appeal to power users, but those power users probably will be the sorts to be influencers to talk up your application to everyone they know.
I personally had these same complaints and acknowledged that I'm unusual and my needs are best served by my own hosted solution. I'd encourage people who are unhappy with major streaming services to roll their own if they truly feel strongly about it- I did and I'm happier for it.
Downvotes are way more "potent" than upvotes. If I'm listening to music that I like, I'd say I probably log over 100x upvotes than downvotes. But when I downvote, I usually want to nuke the site from orbit.
What is the ratio? What ratio of users use downvote? What ratio is that compared to users that interact in any way? How are you controlling for sampling bias? Berkson's paradox? How much utility are the users that downvote obtaining, vs the disutility of being unable to?
We need a name for this, personally I call it GNOME mentality but probably is bad name since most developers are not familiar with GNOME (though the file picker issue get on top HN a lot )
All is the flaw in your comment. One person does not imply all and data speaks for itself.
This type of bizarre logic makes me feel like I have to habitually use features even when they are not useful in order to protect them from UX designers in case I'm subject to some misguided A/B test.
What data are you referencing? I'm drawing conclusions based on data I have but I don't think you work for the same company as me so I'm not sure how you can speak to confidently about how wrong I am.
I'm a PO so feel sufficiently qualified to state that user analytics data is completely redundant for looking at small use high value features & undervaluing direct user feedback from your 'power users' is generally a bad move.
Again, just because not everyone needs/wants a feature does not make it useless or not worth implementing.
Data doesn't tell you how useful a feature is, it tells you how frequently it's used & how many people use it. They are not the same things.
That entirely depends on the the data set. On one hand you are claiming to be an expert and on the other making generalizations about what data can be used for. In reality you know that data is contextual and depending on the data set is what determines what conclusions and how confident you can be in those conclusions.
There's a difference between justifying your decisions with data, and letting data make your decisions for you.
What data would you use to justify not implementing a dislike feature that many users are asking for, for example?
I’ve done some work with a couple disaster organizations in the past and we continually get massive praise and thanks from the victims of these disasters. The actual victims have the same praise over and over again: Unlike the red cross, salvation army, and FEMA we actually ask them what they (the victims) need. We don’t assume they need water and and a tarp, we assume they know whether or not they already have those things. Often the victims need diapers, help with flood mold remidiation, a chainsaw to cut limbs from their street so they can get their cars out so they can go help the neighbors a few streets over etc…. They know what is prohibiting them from getting out to help others much better than we.
When people are telling us what they need, we should listen to them and we should actively try to avoid our tendencies towards “I know better what you need.”
I have a couple friends who work on recommendation algorithms full time and in their defense, they’re fully aware how terrible the recommendations are—I still get my music recommendations from friends, from the local music store employees, and friends who work in music venues—their recommendations are infinitely better.
Similarly any app or website that pulls “top artists” from Spotify tends to include stuff I listens to a bit several years ago, but never stuff I’ve listened to nonstop for months.
You could even do this maliciously. Send someone a Spotify link to a Steely Dan song and a Skrillex song, and now yacht rock and dubstep will follow them around until they delete their account (I'm assuming the intersection between those fan-bases is vanishingly small).
Profile > Private Session
Note: Anything you listen to in a Private Session may not influence your music recommendations, e.g. Discover Weekly.
But you’re not wrong.
By the way, I believe Spotify does have a private mode, which at least blocks your plays from showing up in the social pane and on your profile. But I’m not sure if it cuts them out of your recommendations.
Please, dear God, Spotify: just because I listened to the Hamilton soundtrack once does not mean that I want to hear it every week in Discover Weekly and multiple times per Duo playlist, because Lin-Manuel Miranda is apparently a genre of himself.
Spotify recommendation system is very bad in cases like mine apparently. It's so bad I keep listening to the same old songs instead of discovering something new, which is why I'm no longer a customer until they improve their AI.
It seems to differ wildly between people - I've heard plenty of people making the same case as you, but many that have the same as me. It never appears to be 'its ok', its either 'brilliant' or 'terrible'. I've been using Spotify for 10+ years at this point, so maybe that makes a difference?
No amount of skipping or "I don't like this artist" seems to convince Spotify to stop offering me variations of 8-bit Tetris themes or MULTIPLE 8-bit Cowboy Bebop Tank! theme variations per every single Discover Weekly. I'd be thrilled to even get a 2% hit rate on music I like.
Seriously Spotify: two separate "Tank!" versions per Discover Weekly is too many.
Rest of the week, much the same. Nothing in Discover Weekly.
This seems to be a problem with this general type of recommendation algorithms in general. YouTube suffers from it as well.
This is nonsense. The fact that I skip a song does not mean I don't like it in general, it means I don't want to hear it right now. Music taste changes with mood and situation. If you treat a skip as "I don't ever want to hear this again" you will collect absolute garbage data, and if I know you are doing this, it will make your app unusable because I can't do normal everyday actions without ruining my recommendations.
I'm not really convinced about recommendations being hard, though. Pandora could do them a long time ago, Netflix had good recommendations at one point but not since they dropped the star ratings and at some point Amazon wasn't that bad either.
But from my experience with Pandora, radio stations were localized. Liking a song did add it to a list somewhere, yes, but disliking a song just meant that a particular radio station on Pandora wouldn't have a song in that particular style, and other stations did not seem to be affected by this.
In other words, what you want to listen to, for me, seemed to be very contextual. And since I felt my dislikes were captured in a context (radio station), I didn't have any qualms disliking frequently using Pandora.
And my experience with Pandora for song discovery has always been much better.
How do you know I dislike the song? Instead, maybe I just heard it (or a similar song) on your platform or elsewhere outside of your app and don’t want to listen to it again right now?
I guess you could try to build a more complex model that determines what songs or music I like at what times of the day or what day of the week, or based on what I just listened to.
IMO this is rampant in FAANG companies. We create complex solutions to problems that didn’t really exist and pat ourselves on the back when our experiment show positive results.
The experiment might show you some metric (total time in app, play time at whatever percentile, etc) increased. Great! Now we launch the model in production for all customers and pat ourselves on the back. Meanwhile, the actual experience became shittier. A few promotions come out of this, and we continue living in the bubble.
Alright, end of rant.
I've heard some good arguments against allowing users to show dislikes/downvotes as well as likes, but I've not heard your argument before and it doesn't strike me as accurate. Less used than the positive direction, sure, but still significantly used when available.
My argument isn't an argument so much as a from real world data in this industry at a major industry competitor that customer's don't use the dislike button enough to justify it taking over your CX. Using co-occurrences of playback across a customer segment and combining that with likes or other signals is good enough.
Easily half of the times I skip it's because I don't want it right now. Some of my favorite songs reproduce absolutely horribly in a car on a highway for example, because the surrounding noise utterly destroys subtleties. So maybe you add "but it's not a downvote if it's hearted or otherwise positively ranked"... but many other times I'm just not in the mood for [a song I had never encountered until now].
Here's the issue with "inferred from other activity." You know how you can find out? Give the customer a way to tell you. And no - you cannot "infer from other activity."
>Skipping a song after less then 20 seconds for example can train a model...
..to be wrong. You know when I usually skip a song after 20 seconds? When I've heard it too much, or just heard it on youtube on my laptop an hour ago. In a couple of days I'll want to hear it again. The reason I heard it too much, is I like the song. Your AI though - if I skip this favorite song of mine a couple of times, will now block it and all other songs like it. The literal opposite of what I want.
What this results in, is you have an inferior product, your customers leave, then you leave and go screw up another company's product without ever learning your lesson.
That last bit is a person thinking some crap code they wrote can figure out why someone doesn't want to listen to a song at a particular time. They ignore all the inputs into that decision - how tired they are, what day they had, are they angry at something, have they heard it on another medium recently. Then ignore all that data they don't have, and have the AI make the decision: did you skip after 20 seconds? If you did, let's cut this genre out of your playlist.
This little thing, destroyed spotify's recommendation product. The people in charge of this, are going to destroy everything they touch, because they think their little piece of code, is smarter than the user, about what the user wants. You know, because the user can't code.
People making these decisions - well all I can say is I've met them. I've been personal friends with some. They have been passed up for jobs when I get called for a reference check. Because once they're on your team, they'll destroy your product, so they can feel personally superior.
So that last part is very necessary, because it's an extremely serious issue, exactly as I put it. But in your opinion it's not, so let's cut that part out because it's not necessary for what I was saying. Because you know better than me what I was trying to say.
<dislike>
I will not permit a company/algorithm to directly determine what I want to hear. I’m old enough to know about the Payola scams with radio DJs to trust anyone with that authority.
My solution: buy CDs or vinyl, find a group of like-minded friends to share music recommendations (email newsletters would work, perhaps even…physical interaction), read music journalism, buy music online and build your personal library. Just the music on my hard drive, all directly and legally purchased, would play continuously half a year without repeats.
I use/used the Apple iTunes others-have-bought recommendations for tracks related to tracks I’d purchase. Invaluable for discovering alternative artists by someone who is nowhere near the target age demographic for that sort of music. And you buy the flippin’ track and you know the artist/rights holder is directly getting a reasonable number of pennies from that transaction.
Worst case — buy CDs and get a CD changer. New ones are still being sold. You can still get the 300-400 CD units used. No fans are more rabid than classical music fans, and they are still ecstatically buying huge boxed sets of CDs. They are expecting to listen to these tracks for decades, and there is wisdom there.
I'm not sure what you're suggesting and I don't need to qualify my experience in the software industry or within the domain of recommenders and personalization.
Based on how you talk here, you sound like you paint a fake little world around yourself full of straw men to make you feel better. Which goes well in line with your opinion you know better than the user of a product, what that user wants. Living in a fake little world where you da man, courtesy of your brain.
This is called deep seeded insecurity - where your brain makes subconscious coping mechanisms to avoid depression. Bullied a lot in high school?
I'm using the impersonal "you" here. Don't know if you do that specifically, but OP seems to complain about software workers who do have this tendency, and seems to think that eventually leads to losing the customer.
I'm curious. Can you tell what kind of experiments you ran and what results you got?
Let me be one of those. I want to be able to control the algorithm in all of its parameters, but even the ability to control the values for new/familiar and how far to stray from the current genre would be very welcome.
GOD NOOOOOO
that would be terrible, there are all kinds of time where I will skip a song because I do not want to Listen to it RIGHT NOW, not because I do not like it or never want to listen to it ever again...
Music is very emotional, some times you just do not want to listen to X at the moment, 30 mins later you might.
So I tried to contact eDreams, which requires a trip reference number.
Since I couldn't pay them for a plane ticket, I didn't have a trip reference number. So all their ticketing and phone tree systems would hang up on me.
eDreams customers, evidently, don't want to contact them without having a trip reference number. And they know this for sure because they've never had a customer contact them without one.
The problem with data driven UX is that the data is often short term (ab tests are expensive) and doesn't cover long term complex user impact.
Edit: For example, I’ve seen a lot of Etsy users complain about their front page being filled with NSFW recommendations and no way to get rid of them. I doubt they’ll be using Etsy much in the future.
I prefer and use the dislike button when it exists. I don't care that 51% or 99% of users don't use it. Further, I want to send a concrete signal. Skipping a song doesn't mean I don't like it. It might just be that I didn't want to hear that particular song right now. Even if I skip it 10 times in a row that doesn't mean I don't like it.
Sure I might enjoy some clickbaity content every once in a while, but that doesn't mean I want the system to show me more of it, even if that will lead to me watching more things.
So, basically if I dislike a song but want to give it a chance and then decide that you want to never hear it again, your industry thinks that I liked it. That explains a lot.
> Skipping a song after less then 20 seconds for example can train a model as well as incorporating an explicit dislike button.
Really? Cause I constantly wish for granularity in my ability to dislike something. Insta-skip might mean I'm just not in the mood. Or it might be "never play goddamn clownstep in my neurofunk mix". Or it might be "I actually love that song but I reserve it for special occasions to get into a specific headspace and if I listen to it too much it loses spark". Same signal, three radically different valences.
I just wish services didn't treat all their customers like simps. I'd easily pay double for the ability to like, specify "Tuesday is flying microtonal banana day" or "start switching to prog house around 16:00 because that's when I start to get really deep into code but also easily frazzled." I'd pay gobs of money for Pandora's ability to discover new music with the library of Youtube+Spotify.
I listened to one podcast episode and it's been stuck on my home screen for two months. I've tried listening to it, ignoring it, unsubscribing, deleting (you can't!). It's just there forever now.
And of course, that means it's always recommending related podcasts all the time so most of my homepage is now filled with this stuff and I have no way to remove it.
I’ve had an entire top-level list of podcasts recommended to me for over a YEAR now. I will never use it, but I have to constantly scroll past it and dodge podcast recommendations in new podcast + music playlists (that aren’t clearly marked) as well.
I used to love Spotify. I got so many friends to sign up, I discovered so much music in my Discover Weekly, I convinced friends to switch from all kinds of other platforms, I built collaborative playlists with friends…
but the recommendations seem stuck in a rut for me, probably because they can’t handle the size of my library (hundreds of artists and thousands of albums that I listen to regularly). The product doesn’t respect me. I’m actively researching ways to host my own music library and use something else for discovery. Recommendations welcome so this frog can escape this slowly boiling pot.
I changed my password and it took me almost an year of listening to my songs to Spotify stop recommending me Simple Plan and similars.
- Don’t play songs like this again
i am annoyed at youtube only allowing me to choose from "i didn't like it" or "i have already watched it" for example. i have thousands of other reasons.
that's why i don't use dislike buttons.
Sometimes I dislike something and I don't want it around anymore. Which is the reason I've migrated to Youtube Music, which is an inferior product, but has a dislike button.
Same here. I’m still a big fan of star ratings in iTunes, though I’m not sure if they do anything or influence playback. I have a pretty eclectic taste in music and only want to hear certain songs once in a while and I give those songs get three stars. (Sorry Donnie Iris if you are reading this.)
My guess is they don’t. Did you use the like/dislike buttons they tested on Spotify radio? I didn’t, too much effort.
They probably get more signal from the skip button
Whats worse, there was a song that was constantly playing after my playlist ended. No matter how many times I marked the song as Dislike, it was always in first of 5 songs that play when playlist finished. Made sure the song appeared nowhere in my playlists (had to manually search in each), and neither did the artist. After contacting support and jumping trough few "restart app" styled suggestion, I was suggested to mark the artist as "never play this artist". This surprised me, as I never seen that option before. And then I understood why: its only in mobile version. You cant solve it on PC.
Bleh
A button that says (and does) "wipe my recommendation profile clean, and start learning recommendations from scratch".
That’s exactly why I’m unhappy with Spotify lately. I feel I’m being sold a bunch of corporate portfolios on what’s currently in vogue, and not just artistically but politically, with playlists like this-gender-race supporting this-cause. I actually just want to listen to music.
I still use it but stumble my way through looking at related artists while avoiding playlists and recommendations.
I remember seeing a promoted playlist about empowering women voices (this is music)
I was so confused. In literally most of the world, this isn’t even a point of contention.
https://newsroom.spotify.com/2020-11-02/amplifying-artist-in...
[1] https://www.tunemymusic.com/
It's possible that the recommendations are both good for users and good for corporate interests. They might not work for you, but for millions of Spotify users they seem to work. People listen to them a lot.
I listen to my recommended "Discover" playlists occasionally. They're decent. They include things I haven't heard and quite like. Maybe record labels are paying to be on them. Oh well.
That is the most reasonable explanation. HN users are not the average user of Spotify.
When discussing things like product strategy it makes some (more than a little) sense. But in a conversation about personal preference, what do you expect the reader to take away from it? “Oh okay, sorry, I didn’t realize I wasn’t meant to like this. I guess my opinion’s invalid.”
Who cares who the average user is, when someone is saying something doesn’t appeal to them? Is the sentiment some kind of scolding for not liking it? I sincerely don’t understand.
I do not use Spotify, nor I had for years. And I do not like the level of influence that all those algorithms have on the population decisions. So, it's not about protecting Spotify but an observation to try to add another point of view to the discussion.
It's simply a reminder that when you work at scale you can't please everyone. Someone complaining that a feature doesn't work for them is not the same as saying it doesn't work.
On a site like HN the conversation is usually about the broader picture rather than individual complaints unless someone is responding directly to the CEO of a company. I think the CEO of Spotify posts here occasionally, so maybe he'll reply. The rest of us are talking about it in more general terms.
I want the ability to rate music across multiple user-configurable dimensions. Add tags. Create smart playlists that interpolate between these.
I want to be able to pay a fee to subscribe to music discovery, then be able to mix these with my own library. If I really like a track, I'd like to buy it and add it to my collection.
I want an open API so desktop apps can be written to use it. Also, let me export my annotations and music library on demand.
Music for power users. Don't give me a single button. Give me hundreds of them.
I'd pay $30/mo or more for this.
I looked for that when Play Music got shuttered but could not find a satisfactory solution. I even paid for one iOS app but it wanted access to my entire drive, and maybe some other unnecessary permissions.. so hell no. I should be able pick one folder and that's it.
I guess I could have made a separate Google account just for music.
- Rating and tagging
- Upload and download your own stuff if you want to (marketed as iCloud Music Library)
- Works offline
- Discovery services available
- API exists but mostly used to implement web players for some reason, works fine in desktop apps too, on top of that the current state of your music library is always available as an XML file even when you don't use the API at all.
https://developer.apple.com/documentation/applemusicapi/
Most of these services aren't ever true uploads and do matching to save time, bandwidth and space.
Edit: well, it's a bit nuanced of course, you can not match and not sync or match and sync but you can't mix and match that configuration. So it's either sync with matching when possible or no syncing.
Fingerprint matching barfs on all this stuff.
0: https://github.com/navidrome/navidrome
I use Airsonic right now, it is basically Jellyfin for a music collection. I use both of these as reliable alternative to Netflix/.../Spotify (where ... is a plethora of other video services such as Disney+).
1. Doesn't use Java
2. Simpler/Quicker to configure
3. More modern webui (though I recently found a good desktop subsonic client, sonixd which somewhat negates this)
Also, I haven't used the airsonic fork, but I was using subsonic before switching to navidrome
I found subsonic to be rather... unreliable. once in a while the database would randomly become corrupted and I'd have to delete the database files and start over. I haven't had any such problems with navidrome, it's been rock solid.
> I want the ability to rate music across multiple user-configurable dimensions. Add tags. Create smart playlists that interpolate between these.
I'd actually pay for a 3rd party metadata service that doesn't actually provide the music at all, but just let's me tag and rate some across all music streaming services. I've even thought about building that (simply for myself to begin with). I want ratings, instrument tags, mood tags etc. Let me search for songs with 'piano + synth + dreamy + weird' instead of throwing some stupid recommendation my way. And use those tags to find similar tracks across genres and decades of time, instead of just saying "Nirvana and Pearl Jam must be the same because they're early 90s Seattle bands".
The software just wasn't as good of an experience of just having my trusty 400gb sdcard.
Amazon used to provide this service, but they shut it down.
And from what I've heard, the fingerprinting algorithm is fuzzy enough that it will often match songs/song versions that are very similar, but not actually identical, like censored/uncensored lyrics, different masterings, differing fade ins/outs, alternate takes and live versions, etc. etc.
It makes me curious though how the accuracy vs. deduplication efficiency trade-off looks like in practice: How much storage is Apple saving with their current settings, and how much deduplication would they lose if they made their audio fingerprinting more accurate, up to a level where even audio buffs would stop complaining
https://www.ibroadcast.com/home/
Is what you are looking for, it has been changed since the last time I've been using it.
https://asti.ga/
Spotify can't even do a Playlist larger than 50 songs on shuffle. It's a garbage web app.
It was so nice that I’m seriously considering just turning off Spotify. It’s sort of interesting to analyze why we don’t —- for me, it’s become unconscious habit to reach for Spotify and not anything else. Plus a lot of other stuff integrates well with it.
(What if… what if we can use both? Mind asplode, it’s not a decision.)
Perhaps I was missing something but the change def did not improve the recommendations - it made them drastically worse.
Then I scrolled down a bit b/c I wanted to give it a strong signal of what I liked. I found an artist, clicked it, and noticed the recommendations below changed immediately afterwards.
So then I scrolled some more (because there were still 95% misses), found one I liked, scrolled to see the next row then clicked on the artist I liked above to see if it indeed changed.
It did. But not only that. I love every artist on the following row. Then it quickly diffuses back to noise, but holy cow that was a bit of a spine tingler lol.
When you start your listening session they try and predict how long you are going going listen (based on your past history and time of day). If you are probably going to listen for a while, they are more risky and might play something "different". Playing different stuff is risky (short term) because you MIGHT not like it. Playing the same stuff is safe (short term) because they know you will like it - BUT people will eventually go searching for something new, so they have to risk diversity eventually.
I'd love a personalized playlist titled "Curveballs" that contined things different and/or challenging.
Give me a home screen with 12 or more Daily Mixes that segment my listening behavior at a finer resolution, and I'd be a lot happier with the service.
I understand why all the other playlists exist, but I generally have an idea of exactly what I'm trying to listen to, so these low effort curated playlists are pretty useless for my listening style.
It's a fine balance between existing liked songs and expanding within sub-genres, but I have at least noticed the recommendations improving over the past few years.
I suspect it would help to port over my entire pre-Spotify music library into Spotify to provide a bit more data, my current library is all post-Spotify so it fails to capture the breadth of my music taste. I've just never gotten around to it.
For me, I listen to a lot of 40s, 50 - 60s lounge / exotica, early to mid 90s hip hop, 80s metal, and the standard indie stuff... then I have a bunch of chillwave and synthwave stuff that throws another wrench in the mix. The daily mixes I get are a total mess, much like yours.
The genre-specific mixes they make are pretty decent, but discovery is low.
I have a separate account just for Amazon Echo because I don't want my kid's selections to influence my main account suggestions.
I also have my own account for the car, where explicit songs are disallowed. This allows me to listen to the music I like, but it avoid explicit songs for when my kids are in the car and that setting doesn't affect my main account.
What's inherently comical about this?
One of my least favorite experiences is having a radio station based on a song and getting nothing but songs that sound just like it.
Unless you have only played 90s grunge or 70s soft rock I'm not sure why this juxtaposition is not considered a feature.
That all said, I do wish I could turn off their podcast recommendations. I never listen to podcasts and worse yet they keep shoving that same comedians is absolutely hate (and there aren’t many comedians I dislike; which just goes to show how far off the mark their podcast recommendations are)
Not happy about being forced ads though (on some podcasts)... I thought the idea of paying monthly was the value exchange for the content.
It's like watching a movie that's half Fred Astaire and half Freddy Kreuger.
I like the idea of mixing genres, but Spotify seems to totally miss on what aspects I want mixed.
Trying to find music when you could be focusing is by definition, trading off productivity.
I’d rather have an algorithm dictate what songs I want to listen to (in order to focus) than spend an hour wading through junk to find something i like before I start coding.
Playlist search result page is just an endless grid of images and truncated names. For the playlist duration, description, number of songs, follower count, etc. you have to actually open each individual playlist. Good luck finding what you were looking for.
More and more Spotify-fabricated content is being pushed. Most of which contains the same limited selection of songs that Spotify keeps feeding you over and over anyway.
Podcasts aren't my thing, Spotify wants me to listen to them really badly though. Majority of the ones they're suggesting I'm not at all interested in, and sometimes some of the podcasts they're advertising seem to contain some pretty disturbing content. On a sidenote: I don't know who he is, or what he does, but I hate Joe Rogan and Spotify is to blame for it.
There is a setting hidden under advanced that is supposed to make Spotify stop messing with the shuffle functionality. It is labeled "Allow smooth transitions between songs in a playlist" vague huh, it's also placed directly underneath the song crossfade slider. I'm fully convinced that this was done on purpose. Also this setting seems to do precisely nothing at all, so I'm not sure why they even went through all the effort.
Shuffle is not random. If this is so on purpose, that purpose does not involve happy users. Else perhaps their devs are afraid of touching some jank script that might be holding it together.
Edit:
Almost forgot about the new artist's pages! They used to consist of a long list of all songs grouped by album. This was far too convenient for us users, so with the redesign they simply removed the lists of songs leaving only a grid of albums, forcing you to go into each individual album to find a specific song and play it play it.
After many complaints they implemented something vaguely resembling what we had before, but with such odd UX that it must be sabotaged on purpose again. But of course, adding this overview back to the artist page was out of the question. Instead what they implemented as the only way to access this, and I kid you not, is a plain text link in the most random place ever.
No one is going to use this feature if they don't know it exists. All this just to get a reason for removing it that is spinnable. Machiavelli would have been proud.
I feel the same way. It's not just music either. I get this feeling every time I try to consume anything. Everything is just so fake. Like it was made just to push some silly agenda.
"Recommendations" are ads in disguise. I already block them on YouTube. Wish I could block them everywhere.
Do others generally find that auto play (or whatever it’s called) works well? Or does it just feel like payola radio?
Why would you let a streaming service choose what you want to listen too or a premade playlist by someone else? Then you might as well just listen to the radio. All these on any service are always rubbish. I make my own playlists and I only listen to them, sometimes I search others for some inspiration to see what songs I forgot etc, but most of them include a lot of trash so I make my own.
The only 'auto' feature I use, is my release radar playlist which I check every Friday to find new songs from Artists I like and then add those to my playlist titled for the current year.
"Discover Weekly" tends to be one of my primary sources of new content discovery. I tend to listen to a significant amount of Death Metal, Doom Metal, Folk Metal, German metal and rock, random music in languages I don't understand because the sound is cool to me, and various other never-been-pop subgenres. Discover Weekly has been pretty on point for me to find new and interesting stuff in a way that hasn't felt like them pushing a message or some corporate catalog.
I don't tend to listen to the genres that are full of artists who are trying to push politics through their music (at least that I've consciously recognized) (except maybe System of a Down). Not because I'm necessarily against their messages, it just the sets haven't overlapped much for me at this point. It could also be that music isn't much about lyrics for me, it's about if it's something that sounds good to me.
I have a bunch of playlists and saved albums, but I do listen to discover weekly in the first few days of the week. Sometimes it's meh and I just go choose what I want to listen to, sometimes it's really good and it's on loop through the week.
Their "Daily Mix" playlists seem to be 90% stuff I've clicked "like" on in the past, so I listen to those as well.
I podcast elsewhere, and am never going to switch to Spotify for it, just let me turn it off!
so again, reducing it to like/dislike buttons is missing a great deal.
One algorithm will never work for everyone. Even if it's one of the best in aggregate, that doesn't mean it's not awful for me. It's not feasible (nor does it make business sense) for Spotify to identify and build the best system for each individual user, so I really don't expect them to do more than optimize for the majority.
One is a progressive blend of funk, hard rock, jazz, psychedelic from New York, active in the mid 90s to 2010[0], and the other is an anti-fascist (anti)black-metal band from London formed in 2018[1].
Both are interesting, but I only like the music of the first one. Google Play Music used to mix them together.
[0]:https://en.m.wikipedia.org/wiki/Gaylord_(band)
[1]:https://genius.com/albums/Gaylord-band/The-black-metal-scene...
Including social media.
https://community.spotify.com/t5/Your-Library/2021-how-do-we...
Account deleted. I know I'm not the norm but Spotify has never done a good job for me recommending music, ever
There should be an Incognito mode on every music streaming service for this reason.
https://support.spotify.com/us/article/private-listening/
Their refusal to let you delete history is just puzzling.
As is the curse of modern technology I don't want to interact with it at all in case the algorithm decides I'm into it and undoes my work retraining my preferences.
Last.fm has always been great for recommendations.
(The feature I most want is an interface for tweaking the rec algo to my liking - at minimum, for being able to crank up its adventurousness. Rec algos are depressingly conservative as a rule, I guess because most customers are the kind of people who eat at McDonald's when they travel abroad, stay at package resorts, and complain movies have gone to shit because they can't find anything they like on Netflix or Hulu)
Absolutely. At times, it makes me wanna blast it with cosmic rays, induce a burst of mutations, mix things up. Just give me a knob. Sometimes I want to stick on one specific sound. Other times I need to explore.
Also, while I like a broad variety of genres, I only like listening to one at a time. I don't want a rap mix to be invaded by a Bach sonata. And yet...
I did try a Spotify podcast once. It was on my recommended list and was called daily dad jokes or something. Figured it was worth a quick laugh. Well it turned out to be a bot podcast. They set up a bot to rip jokes from Reddit, push them through TTS, and dump it into a Spotify podcast.
No thanks.
Meanwhile, Apple Podcasts, my go to, works great, doesn't spam me, and even added a feature where you can pay for a no-ads version of a podcast. Worked seemlessly.
Anyway, I waited like 90 seconds and he came back on. Withing about 15 minutes, I had two more ads come on and they were 90 & 120 seconds each. After the second ad played, the podcast didn't start. Fiddling with it caused the podcast to start over from the beginning. When I tried to seek to the last place I was, it played an ad immediately. Seriously unbelievable.
I gave up and will never attempt to listen to an ad supported podcast on Spotify again.
Because Spotify doesn’t pay anything to the podcast producers. So podcasts run their own ads. (Well, some do. Others live off donations or offer a way to pay them)
The probably autogenerated daily mixes have been surprisingly non-bad for me. Nothing like rap mix invaded by Bach, but the "classic rock" mix sometimes has Ghost that doesn't really sound that out of place.
An option have a personal search expression with a variety of tags include and excluded and other stuff neutral is what I want for everything.
It's twenty years since Alt-Vista allowed search with logic expressions and "recommendation" has gotten more and more railroad-y since then.
The "don't give options, make it moron proof" paradigm literally forces us all to be morons.
I love podcasts, on my podcast app.
I'm really disappointed how Spotify is buying several big podcasts and moving to their platform exclusively.
I tried to listen to a Science Vs episode the other day, and now my Spotify is almost completely filled with podcasts recommendation. Plus, the experience is still far from specialized podcast apps.
The result: I'll probably stop following Science Vs (and any other podcast that moves exclusively to Spotify).
They could, at least, have a separate Spotify Podcast app so things wouldn't be mixed (kind of like Wealthsimple has separate apps for Investing and Trading).
I'm still sad Spotify bought Gimlet and apparently will do (if not already) this to all their podcasts.
Short answer is $$$$.
Longer answer is: Spotify must pay royalties for each song played. Imagine if there was a completely free (for Spotify) form of content that filled users ears for hours, thus removing the need for Spotify to pay royalties. Ahem podcasts. Now imagine if Spotify started injecting ads into said media form to grow their revenue beyond subscriptions. Again, podcasts. So now, you have a very long-form content that both saves you royalty $ and drives new revenue. QED. Podcasts will continue to be plastered all over your recommendations, be top search results, etc, until the above stops being true.
Wrote about this in detail recently: https://www.towardssoftware.com/spotify.txt
Now given that burgers cost us money to produce, wouldn’t you rather have a podcast?
Does this whole thing not seem insane to anyone else?
Serving music is unprofitable for Spotify. They collect your monthly subscription fee, then every additional song play makes Spotify lose money. They don't want to have zero song plays since then nobody would subscribe, but their goal is sell the profitable thing (subscriptions) while minimizing the unprofitable thing (song plays).
The buffets may be the closest and they don’t really seem to care at all, because the price differential is so high.
So Spotify is likely undercharging by 50% or more.
To me, Spotify seems to be using podcasts in a similar way. I assume it's because the central business model is unsustainable when combined with investor pressure; they can't just focus on core product because they're burning too much of other people's money in an attempt to outcompete everyone else
If that's the case, how do you explain those two features : - the repeat button - the "automatically play similar songs" option
In the first case, since I don't know how it works, maybe they have to pay only one time for a song per user listening to it during a period (say a month). But I doubt it. In the second case, it's Spotify explicitly saying : "here are songs you don't intended to play but that we picked automatically and are playing to you".
If you look at Netflix, sure they have the feature that automatically launch the next episode. Or they try weird stuff like live or play anything... But they also have the "Are you still watching ?" feature, to make sure that they don't display content to an empty room. Spotify doesn't have that, it can just play songs indefinitely without any human interaction.
So, I don't know.
It was extremely unsettling until I realized what they were doing.
They replace the ad breaks that would normally be in a podcast. (Like when the hosts say "we're going to take a quick break to talk about our sponsors for this episode"). It's not an unskippable ad like you get when listening to music using the free tier. It's inserted directly into the audio stream.
Edit: I guess it's possible the Conan podcast is doing that somehow and not Spotify, but I've never had something like that happen when using the Apple podcast app.
I've also noticed some location specific ads, which isn't that surprising if you think about. There's nothing stopping them from serving you a different mp3 file depending on your IP geolocation.
Also this is specifically a Spotify feature
https://ads.spotify.com/en-US/news-and-insights/streaming-ad...
> Spotify Podcast Ads are powered by Streaming Ad Insertion (SAI), which leverages streaming to deliver Spotify’s full digital suite of planning, reporting, and measurement capabilities. Spotify Podcast Ads offer the intimacy and quality of traditional podcast ads with the precision and transparency of modern-day digital marketing.
https://www.adexchanger.com/podcast/spotify-snaps-up-podcast...
It's a bit jarring to listen to podcasts full of Americans and Brits and and hear ads for some Australian bank or something
Spotify is large because they were first and because apple/google music suck. But eventually apple and google will stop sucking and users will drain from Spotify rapidly unless Spotify can create content to keep users in.
Spotify is in the position of Firefox 15 years ago right now. Eventually the built in apps will take over.
Not to mention that I might listen to podcasts at home and one the way to work, but at work I won't want to listen to anything spoken, but half the time I end up accidentally starting the podcast anyway. That now means I am not using Spotify for podcasts.
The thing is that Spotify does not pay per play! The royalties lobby love to give the impression that they do, because it makes it look like Spotify is fleecing them.
Spotify pays percentages of what it itself collects in subscription fees. If a record company owned 10% of everything that was listened to on Spotify this month, they would make _the same_, regardless of how many absolute seconds their stuff was listened to.
So, more listening to podcasts does not mean Spotify has to pay out less. How much they have to pay out in total is fixed as a share of their subscription (and ad) revenue anyway, in multi-year, industry wide contracts with copyright owners.
There are ways Spotify could circumvent this, such as promoting content they covertly owned themselves, or content whose owners gave them kickbacks of some sort. There's some indication they do such shady things. But podcasts don't change this equation much.
I think Spotify's promotion of podcasts is simply good old fashioned loss-leading monopoly building. They're hoping that Spotify will become where listeners go for podcasts since that's where all the podcasts are, and also the place podcasters go because that's where all the listeners are.
This isn't a hill I will die on by any means, but if you read Spotify's write up on royalties[0], they pay out based on stream share of the overall platform, on a monthly basis. E.g. If Columbia Records gets 10% of streams for a month, they get 10% of the pool of $ Spotify distributes. Now, like you said they likely have some shady practices - I would imagine Spotify has an incentive to create music like study and sleep beats themselves as those playlists get hit for 8 hrs at a time and it doesn't really matter who created the beat as long as it helps you study or sleep.
They do not indicate in [0] that podcasts are excluded from stream share. Even if they aren't included in stream share now, I'd imagine long term they very much intend for podcasts and music to be lumped into the same stream share model to drive payouts down.
[0] https://artists.spotify.com/help/article/royalties
I started migrating to Apple Podcasts so I can context switch more easily.
I've never used Spotify, so I can't compare. I don't like podcasts either. I get more out of audiobooks.
And still now as then: It is only available in the US. I no longer live in the US.
If you aren't in the US, then it won't work, but I create multiple stations and can switch station if the music isn't good. I can't remember it ever not being good though. I use the like button to train it.
> Lots of slightly obscure songs weren't available.
If I'm looking for a specific song, I usually go to YouTube. Most of what I listen to is pretty obscure, but it might depend on the genres.
I beg Spotify to think Logically about it. There is a limited amount of screen real estate available on any display, but especially phones. Q.e.d: pushing podcast-related content actually hurts the music listening experience. Conversely, pushing music content hurts the podcast listening experience. This is (mostly) inarguable; any pixel dedicated to an interface element related to podcast content is a pixel which cannot display music content; its a zero sum game.
I say "mostly" because; there are people who I'm sure love Spotify Podcasts, and having both avenues within one app is a net win. I'd be willing to accept a toggle in settings which could "focus" the app Between Music <> Music + Podcasts <> Podcasts. I don't think that's unreasonable, and I also don't feel it would be unreasonable for it to default to Music + Podcasts.
Though it brings up an interesting point: Imagine the experience of someone who does primarily come to Spotify for podcasts. Its among the worst podcast apps in the history of podcast apps! Its littered with music! That's a horrible experience!
The ironic part is, Spotify is probably happy to lose me as a customer. I'm a "Music power user" if there were such a thing. They may end up paying more in royalties for the music I listen to than I pay them, and I can't say I've listened to more than a half-dozen podcasts. "Show me the incentives, and I'll show you the outcomes": Spotify is not incentivized to build an app that benefits people who listen to a lot of music. Their bread and butter is people who listen to a bit of music and podcasts; just enough to keep them paying each month. So, I don't lose too much sleep over their slow decent into mediocrity; I'm far more concerned about the unfortunate reality that there are no longer any great options in this space, for people like me.
Wrong Rec Do not like the artist/show less Do not like this song Heard it too many times/pause for a bit
etc...
It got removed... as the whole functionality (Radio) got phased out, and the new stuff didn't incorporate it.
Employees changed, and moved on to other companies, and the new that came didn't think it will improve things. etc... and the feature got forgoten.
I think a lot of people in here are asking a 'reset my recommendations' feature. And that did exist as well at some point. Not sure why it got removed/it is not in production.
The idea of resetting recommendations is a minority feature ask from customers as well and I wish companies (including my own) handled that better. Sadly its a pretty expensive project for such a minority stake.
It was just constantly frustrating to battle them, so I switched to Tidal, and it's been amazing. The UI is just simple and lets me play music, and nothing has broken in the last year. The only changes have been minor improvements that make it even easier to do the one thing that I want: to play music.
Instead of having a discussion about how it could be improved, maybe we can have one about which alternatives exist and how good they are (tidal, is it?)
EDIT: I tried Tidal, that app has a confusing mess of an interface just like Spotify. Plus they have a "convenient" bug during sign up that may sign up students with more expensive non student accounts.