Spotify Form F-1
sec.gov
sec.gov
I also think they cracked music discovery. Their discovery weekly playlist is stellar.
My only wish is they'd get off Electron (EDIT Chromium - thanks jjgod) and go native. A music player is a frequently enough used piece of software to optimize for performance.
If it were any other application, I wouldn't care. But because this is the second most frequently piece of software I use (browser being first), I'd love for it to be as performant as possible.
Not to mention the SSD-destroying & CPU-hogging habits of the application. That being said, I still use Spotify because I really value the music discovery and algorithms... and nobody else offers a streaming service with a dark theme :)
$ ls -lh /Applications/Spotify.app/Contents/Frameworks/
total 0
drwxr-xr-x 5 jjgod staff 160B Nov 7 2016 Chromium Embedded Framework.framework
drwxr-xr-x 3 jjgod staff 96B Nov 7 2016 Spotify Helper EH.app
drwxr-xr-x 3 jjgod staff 96B Nov 7 2016 Spotify Helper.appI love Spotify, but I use Pandora sometimes _exclusively_ for music discovery, and then add to my Spotify.
There's no way to know if I listened to that track in full because I liked it or because I had it on in the background while doing other work and it wasn't distracting enough (usually the latter is the case for me). Furthermore if I skip a track they don't know if I skipped it because I didn't like it or I wasn't in the mood for that style of music, or I was just sick of that song cause I've heard it enough in the past X listens.
There's no way to easily signal to spotify why I'm taking an action, and yet their algorithms are trying to make assumptions over every little action.
For posterity, I have a little under 1000 tracks saved from 200 or so artists (of which I am following around half), around 50 favorite album playlists and about as many "I should give this album a listen" playlists.
So not really a huge collection in absolute terms. It is very much focused on hard rock and metal, though.
Are you worried about any Spotify terms of service issues?
Upon saying that, even with my Sennheiser HD 202s and Macbook built-in audio I can tell the difference between Spotify and lossless FLAC. My Australian internet is so bad that I probably couldn't even stream FLAC anyway.
For the mobile app, it's 96, 160 and 320 for normal, high and extreme, respectively.
Chromecast and the browser-based player both use AAC, at a maximum of 256kbps.
There may also be mastering differences between different versions of an album.
So the best thing is actually to do your own Ogg Vorbis encodes of a lossless album you have, and do the a/b testing on those files instead, so you're certain of the provenance.
I think the last time I tried to price a single one of my playlists, it'd cost me thousands to buy the songs and that's without trying to track down FLACs.
It wouldn't surprise me if they end up killing their golden goose out of sheer incompetence, though.
IIRC they moved to Chromium to ensure decoupled releases and faster iteration (starting at ~9:35: https://vimeo.com/85490944).
Is IKEA effect at play? Without Spotify, I would only start another artist when I finished the current one. So, with Spotify it becomes an information management problem due to the large amount of albums available at a click. Or it could just be that my life has changed and I don't have a time slot dedicated to music anymore.
I was able to recognise any song in my collection in less than 5 seconds, give you the artist, how I discovered it and potentially an anecdote about them.
Now I have playlists full of stuff I know nothing about. It definitively solved the music distribution issue, but I have trouble to be emotionally attached to the music I discover on Spotify.
I think your theory about the IKEA effect is probably spot(ify) on
My discover weekly playlist ALWAYS contains at least one song in Portoguese and at least two in German. I speak neither of those languages, and I've never willingly searched and played songs in those languages (except for "99 Luftballons" once).
I also get some terrible variation of "Here comes the sun" at least once a month despite not listening to The Beatles at all. [0] is the one that I hate the most.
Spotify, please sthap. I can't take any more Portoguese, German, and "Here comes the sun" cover songs.
I think it's a newer feature so maybe it just hasn't made its way to mobile yet.
As much as I love the notion, I have to conclude that Discover Weekly's mediocrity is just another piece of evidence that the world actually did end in 2012, and ever since, we've been living in a simulation. A simulation where the level of petty annoyance is carefully adjusted to a maximum, remaining ever so slightly within the bounds of believability. (In November 2016, someone spilled coffee on the controls.)
totally?
The discover weekly algorithm seems to latch on to outliers, I guess.
I've been wondering whether paying for a Bandcamp album every so often to match the cost of Spotify, and pirating music otherwise would actually benefit artists more.
I've read that Spotify pays fractions of a cent per play, where as Bandcamp pays out quite well, better than iTunes and Google Play.
If you want a streaming service more like Spotify you might also check out Resonate Cooperative https://resonate.is/ which has a stream-to-own model.
Discovering new music is low cost, repeated listens double until the ninth, upon which you own the track outright. You can then download or stream that track for free.
Still ramping up but I'm excited about their prospects.
Already signed up.
Personally I like a distinction between looking for music and having it DRM-free offline on my hard drive.
No. Your monthly $10, after Spotify and the labels get their majority cut, goes by percentage to the most popular artists on the service, with absolutely no regard to which artists you have personally listened to, so if you have even remotely obscure taste you are effectively not supporting the artists at all.
Bandcamp, on the other hand, takes a 15% cut of digital downloads and 10% for physical releases/merch. If you buy an album for $10 on Bandcamp, the artist gets $8.50. If you play an album by an indie artist on Spotify 100 times, they might get $0.0002 or thereabouts¤.
¤ Not actual numbers, but the point is it is effectively zero.
I know a couple of relatively obscure bands personally, and they do get something from Spotify. It's not much, but they do get a little bit. They consider streaming more as advertising that gets their name out there and their music easily accessible, but they make most of the actual money from shows and merch.
E: Artists can see their playcount stats and payouts here: https://artists.spotify.com and as far as I know, even niche artists are paid the $0.006 per song, not just the biggest players.
E2: I am a huge fan of the Bandcamp model, though. Most of the time I find music there that isn't on Spotify, and I've also bought albums on there even though they're on Spotify as well, because I want to support the artists directly. I have completely stopped buying physical albums.
There used to be a FAQ entry on spotifyartists.com about royalties and the complex formulae by which they are divided amongst the various parties entitled to slices of those precious fractions of a cent, but I can't find it now.
(That's right: on a promotional site directed towards artists, aimed to encourage people to make their work available on Spotify, there is nothing anywhere about how they intend to pay you. Because honestly, they don't.)
You also need to sell a hell of a lot of albums every month, to make minimum wage.
I do agree that they should pay artists more, but I consider it an issue with the record labels, rather than with Spotify.
Made up example: $100000000 * 0.0000001 of all listening activity = $10
Or $10 * 1 of all my listening activity = $10
Or are you saying that Spotify cuts off at a certain point and doesn't pay anything out, or that they take a fixed cut after allocation?
With the current model, if I do that, Artist X gets paid near zero. So you can see how the models are indeed different.
Which is exactly the current situation, by the way. Anyway, what I was doing was just demonstrating that the system is not the same.
Although I am admittedly bad at maths and statistics, it seems to me that the system I am proposing would be more likely to pay out at least coffee change to small-time artists. I might be wrong, though.
Basically I think it's unfair and unintuitive how everything I pay goes to artists I don't listen to. There has to be a better system.
Getting 100% of your subscription is the same thing as getting 1% from everyone out of 100 customers, where only one person (you) listen to that artist (and exclusively).
That 99% of "your" money went to an artist you dislike doesn't matter if 1% of everyone else's money went to the artist you do like. The money doesn't know where it came from, and the artist doesn't care. Only the aggregate distribution matters.
Now I don't know how they do their reward distribution, I'm merely pointing out the flaw in your reasoning.
Love Spotify, disagree strongly with this. My experience is terrible music that doesn't seem to have anything to do with any of the hundreds of songs I have on my playlist. I get this curious Millennial mix of soft rock music with long, boring instrumental guitar strumming, a disco beat and a sleepy, sarcastic-voiced gal (sometimes guy).
It might be because I live in Brooklyn.
And there's no Like or Dislike button like (beloved) Pandora.
edit: Don't think anyone compromised my account.
No idea why they hid it there, and there only.
When I first started listening to Spotify (about a year and a half ago), I found the dicovery really high-quality, but it just seems to have gotten worse and worse. Discover Weekly is particularly bad. I find that the first 5 or so songs are okay, but from there is rapidly decends into madness.
I have the weird habit (though I'm weening off) of keeping all the music I listen to out of curiosity on a big playlist together, whether I liked it or not, as a way of keeping track. I thought that might be throwing Spotify's algorithm off (and it probably still is), but it seems it's definitely not the only reason.
I've decided to just trust that Last.fm is doing it's juob and not add any more to that playlist.
I love the daily mixes, though thats mostly just sorting the music I normally listen to into genres. I still find it kinda cool. Especially that I always get a mix with my kids' tastes, since I often play music for them.
I disagree about the native app, though. PWDs are where it's at now. Native apps are doomed.* I didn't even download it on my new MBP, it was slow and buggy. The web version is good enough for me, combined with web-scrobbler for Chrome so I can still scrobble to Last.fm.
[*] https://medium.com/javascript-scene/native-apps-are-doomed-a...
There may be some odd connection made internally, between fans of some specific genres and Dutch rap, and you were unlucky enough to be hit by it.
I get why Spotify doesn't want to expose the connections their algorithms make (because that's their biggest competitive advantage), but it would be nice to dig through once in a while, to see the connections they've made.
It appears to me that they include Discover weekly consumption in determining preferences. Any personalised Spotify playlist should be excluded to avoid this kind of reinforcement.
Hah, imagine that, AI Cognitive Biases!
On the other hand what I do like is the playlist radio that comes on after you finish listening to a playlist. I find that a great way to find new music and new artists.
I do love it though, generally speaking although being a heavy daily user I'm not sure the discovery part is actually as good as it initially appears.
I listen almost exlusively to hard rock and metal, and all of my auto-generated playlists are absolutely stellar in their selection.
It seems that once Spotify lumps you in with a specific segment, it's hard to get out of it again.
My only issue with discover weekly is how heavily it weights the music you've listened to most recently. I guess thats part of the whole weekly aspect, but I wish they would give you recommendations based on your whole listening history.
I miss the forums, IRC rooms, and insane collection of obscure music though. Not sure that will ever be totally recreated.
> My only wish is they'd get off Electron (EDIT Chromium - thanks jjgod) and go native. A music player is a frequently enough used piece of software to optimize for performance.
I agree, but I also think it's probably the main reason we have a semi-official Spotify client on linux.
I also like the artist and song radio options which work well for office background music.
(They've even implemented a solid weekly "new music mix" discovery service... and the apps are native.)
I have no big issues with the desktop Spotify app. I find it better than iTunes.
It's the Android app that Spotify need to fix.
Here is a blog post that describes one of their machine learning models using neural networks. http://benanne.github.io/2014/08/05/spotify-cnns.html
I've reached out to Spotify's support over Twitter and they were like keep listening to more music and it will correct itself, but so far it hasn't. :(
I think it would be a good suggestion on their Community Ideas board, I would certainly vote for it.
Spotify provided an extremely convenient and inexpensive alternative, with the added bonus of curated playlists and automatic recommendations.
Short-sighted publishers be damned.
Spotify allowed me to listen what I want, where I want. I don't like their chromium based application, but if it's the only way that they continue developing their Linux client, I'll not whine about it.
However, nothing beats owning the albums. So if the album passes a certain value in my mind, I buy it either on CD or from iTunes for archival and listening on my proper audio system, because all the snake oil and hype aside, nothing beats warmth of a good old amplifier, a decent DAC, and some big, good speakers.
I would love to hear some recommendations for streaming services that can accommodate my tastes. I'm a fairly fluent in Japanese so Japanese-native services are fine too, but I'm not aware if any exists.
There are plenty of songs which I've added to playlists, only to find a year or two later are no longer available due to expired licenses. Some of my favorite artists and labels - Drag City is a big one - aren't available on Spotify at all (usually due to objections to the royalty structure). Many live albums, and more obscure and vintage releases are often nigh-impossible to find, or only have one or two songs from the album: likely those that have been licensed previously for use on TV etc.
Spotify is great, but it will forever be fighting the friction of music licensing and fair monetization.
1) Electron needs to be good about sharing memory between multiple electron processes. If I have Slack, VSCode and spotify open, then memory shouldn't be tripled.
2) Better app development. It is quite possible to make performant electron apps. Apart from the memory hit of spinning up blink, if you manage your views and JS well, both memory and CPU usage can be quite reasonable.
3) I really do like Electron apps because it's a very consistent experience, no matter the platform. Developing for Mac or Linux is not an after-thought. It's always in Parity. I can see how that is both a great user experience and makes great business sense. There is a reason why React Native is gaining so much popularity: Consistent experience, write once, deploy everywhere, much smaller and maintainable codebase.
Update: This nugget from the F-1 [1] alludes to the interesting to watch pricing
> Moreover, prior to the opening trade, there will not be a price at which underwriters initially sold ordinary shares to the public as there would be in an underwritten initial public offering. This lack of an initial public offering price could impact the range of buy and sell orders collected by the NYSE from various broker-dealers. Consequently, the public price of our ordinary shares may be more volatile than in an underwritten initial public offering and could, upon listing on the NYSE, decline significantly and rapidly.
[1] https://www.sec.gov/Archives/edgar/data/1639920/000119312518...
If you're not working for a huge bank, stay away from this until it stabilizes.
Spotify isn't pricing it, they aren't even selling shares. The initial pricing will come directly out of the order book whenever it lists. It's really unusual. Read the plan of distribution section:
https://www.sec.gov/Archives/edgar/data/1639920/000119312518...
> The Registered Shareholders may sell their ordinary shares covered hereby pursuant to brokerage transactions on the NYSE at prevailing market prices at any time after the ordinary shares are listed for trading thereon.
Plus:
> prior to the opening trade, there will not be a price at which underwriters initially sold ordinary shares to the public as there would be in an underwritten initial public offering
As I said, this should be fun to watch!
"We'll give you a share for $20"
"I'll take it!"
"No, we're not selling"
> Unlike an initial public offering, the resale by the Registered Shareholders is not being underwritten by any investment bank. The Registered Shareholders may, or may not, elect to sell their ordinary shares covered by this prospectus, as and to the extent they may determine. Such sales, if any, will be made through brokerage transactions on the New York Stock Exchange (the “NYSE”) at prevailing market prices.
So essentially early investors will be the ones [potentially] selling. The pricing should be highly volatile.
Let say it is valued at $20B with shares at $20. What is stopping an early investors selling 1 shares at $1000, market valued at 1 Trillion, and get someone to buy that 1 shares?
My point is since there is no new shares, unless the early investor are really rushing to offload. Selling would be limited.
> because Morgan Stanley is not acting as an underwriter, it will not have engaged in a book building process, and as a result, it will not be able to provide input to the DMM that is based on or informed by that process
$SNAP is probably as nearest neighbor as any for price-action historical context.
Market timing (H1 2019) may also be ill-favored generally for anything tech related if we see significant correction.
But I certainly hope I'm wrong. And will be looking to add positions in Spotify, Dropbox, Uber, Lyft, Docker, Snowflake, AirBnB, SpaceX, WeWork, Stripe etc if and when they become available ;)
Also, I would be hard pressed to imagine that they would not nominate a stabilization agent to help deal with the issues that you noted so that they manage price volatility in initial trading.
> As this listing is taking place via a novel process that is not an underwritten initial public offering, there will be no book building process and no price at which underwriters initially sold shares to the public to help inform efficient price discovery with respect to the opening trades on the NYSE. Pursuant to NYSE Rules, we have engaged Morgan Stanley & Co. LLC (“Morgan Stanley”) as a financial advisor to be available to consult with the designated market maker (the “DMM”) in setting the opening public price of our ordinary shares on the NYSE.
Morgan Stanley will be involved to help set the price with the market maker (sounds normal), but:
> However, because Morgan Stanley will not have engaged in a book building process, they will not be able to provide input to the DMM that is based on or informed by that process. Moreover, prior to the opening trade, there will not be a price at which underwriters initially sold ordinary shares to the public as there would be in an underwritten initial public offering.
So they didn't get hired to do a road show and won't have interest from institutional investors. As for having a stabilization agent? It sure doesn't sound like it:
> This lack of an initial public offering price could impact the range of buy and sell orders collected by the NYSE from various broker-dealers. Consequently, the public price of our ordinary shares may be more volatile than in an underwritten initial public offering and could, upon listing on the NYSE, decline significantly and rapidly.
https://www.bloomberg.com/view/articles/2018-02-21/your-ipo-...
I'm a happy Spotify customer, I hope they continue their growth and become a pillar of the music industry.
If you contrast this with 51 -> 92 for ad supported MAUs in 3 years it seems like they have been successful in converting a lot of ad supported users to premium ones.
This may be one of my first "buy and hold" stocks simply for the personal reason of loving the service.
The only alternative service I even debated about was Apple Music when they first announced their streaming service, but there is no clear advantage to their service and I hate the Apple Music UI.
I hope they find a way to help a bigger % of the revenue get into the hands of artists and start cutting out the big record companies who just middleman everything. But other than that, no complaints.
I have to wonder how many users Spotify picked up when Apple blew the christmas HomePod release and Amazon flushed out Alexa-enabled devices at heavy discounts.
It might not be a lock-in, but given my track time on a non-Apple ecosystem and the value I'm getting out of Spotify compared to Apple Music...I'm not going back any time soon.
And, yeah: all it took was a $30 device.
If Apple Music was supported in as many places as Spotify I would drop Spotify.
I am subbed to GPM for a few years now. I did it first because Spotify was and still isn't available in my country, but I don't think I would switch if they become available tomorrow.
Second is the interface in general. Spotify realizes that especially on mobile, you have limited screen real estate. I really dislike Google's choice to overly emphasize giant thumbnails of album art I'll never care about or recognize, at the expense of not providing a list view for all levels of the hierarchy. This also leads to situations where artist names are truncated beyond recognition because there's only a few characters space for them.
The GPM app on my Pixel XL also takes seemingly forever to load and play, podcasts often lose their place, etc. Spotify just works.
Menu ("..." on left in web) > Browse stations > Activities
It's right there!
I like the service but there is many things in the UI that I don't like:
- No desktop app
- Album listing is a mess: . It doesn't differentiate between singles and albums so the album list is sometimes very long for no reasons . You can't mentally filter them yourselves as the number of tracks and the album's length are not displayed for each "album" . It is somehow sorted by release year, but not entirely (you can't chose anyway). . On mobile the year is not even displayed so if you want to quickly see if an artist has a new album or simply listen to the more recent one, you can't really be sure of what is displayed.
- On mobile the search button is not always accessible, you have to navigate / "go back"
- Features disappearing: . You used to be able to filter new releases by genre. Now it's a general category of just popular artists. . Same for popular tracks and albums. For my tastes this entire feature is basically useless.
- Radios: the UI used to suggest auto-generated radios based on artists you listen to, it was my favorite feature that helped me discover many great artists and songs. Somehow it got moved and it took me a long time to find it again (it's now only a link on the artist page I guess ? I never think to do that)
- Bonus nitpick : the upload feature is great, sadly your music is not available to other people on the family plan
I don't really like the Play Music UI (but I don't hate it either, the new parts are very nice, but the album pages could use a bit more love).
I dislike the spotify UI though (why is it THAT dark and depressive ?) and play music removes the ads from youtube (us user) and pays youtubers, so that's an easy choice.
Oh, the winning feature of Play Music for me is that I can add my own album, even if it does not belong to their catalog. It still shows with all it's metadata and cover as an album.
Also the Youtube Music app.
I've found their catalog to be really complete, I've rarely not been able to find something specific I'm looking for.
Only advantage I'm hearing from Spotify music users is that the algorithmic selections sound really good. Meh, I put in about an hour a month of surfing different Play Music radio stations and recommendations. Works for me.
Idk how that’s even possible but it happens once a month for me
If you have songs in your playlist that someone else has in their playlist it'll find a song in that other playlist that you don't have. Offload the complexity of matching to humans already doing it.
For instance, the Fresh Finds playlists - music that's good but very new or undiscovered - are made by curators who sift through songs and artists listened to each week by "tastemaker users". Spotify combs through their listener data and identifies users who have a habit of listening to music that later blows up and becomes much more popular. They tag those users anonymously and use their taste to inform the Fresh Finds playlists, which also feed the rest of the editorial department.
Really cool stuff.
https://news.spotify.com/us/2016/03/02/introducing-fresh-fin...
I am also not a fan of Spotify's UI. I've tried in order, Spotify, GPM, and am now on AM. I may try Spotify again, since my first try was a few years ago and I didn't care for the UI and became annoyed after I built a playlist and half the songs disappeared one day (rights issue I guess).
GPM is a great deal because it also includes YouTube Red and YouTube music if that's your thing.
You can also get a deal on AM if you buy for the year for $99 and use an iTunes gift card that you can usually find for ~$80.
The AM UI is not the best, but since I'm on iOS, the integration is nice.
I tried Spotify a couple years ago and the experience was terrible and I haven't been back. I don't recall exactly: there was tons of hype, I had to download and run an installer, then create an account and validate email, and then I could play a song, so I picked a song and got some audio ads and subsequent more ads and suggested songs that were not at all related to the song I picked. It wasn't just not good enough to get me to switch, the onboarding was repellent.
I probably won't be their customer but I'm happy to see them succeed for a variety of reasons: customers love it, they aren't FAMGA, and that slide deck on their organizational squads and tribes [1] was significant in my work life.
[1]: https://blog.crisp.se/2012/11/14/henrikkniberg/scaling-agile...
Different tastes. I love their green and black theme. it's focussed and clean.
Now, I'm using Google Play Music. There's no doubt in my mind that everything else outside of this limit is better on Spotify.
Please, Spotify, fix this! Let me help inflate your valuation!
1. https://community.spotify.com/t5/Accounts/Library-Song-Limit...
I guess I'm in a niche? For me, there are tons of hiphop mixtapes that will never be on Spotify that I definitely need in my music player of choice.
Hell, Spotify won't even let me import local MP3 files from my phone!
I fought it for a while with local playlist syncing and such, but just wasn't worth it over time. There's enough other music out there that I gave up trying.
https://support.spotify.com/is/using_spotify/features/listen...
Or are you referring to something different?
I pay them $10/month for service, so why can't they reserve a couple gigs of space to store my files that they don't have in their library?
I scrobble to last.fm so I can still get decent recommendations over there.
This happened at spotify too and I assume it's due to legal issues. When an artist released a certain album on one label and only that album disappears.
It's very annoying and is pushing me back to good old piracy.
At least Spotify has an open API, so you (maybe) can solve your own problems given some time.
(10k is really not that much at all if you're into a few genres.)
I have a playlist folder called "favorite albums" specifically for this.
At a (low!) estimate of 4 minutes per song, if you listen to Spotify for 3hrs/day, it would still take you 222 days to listen to all of them once.
Is this a form of data hoarding? If not, how do you use that many songs?
I have 15k songs in my iTunes library, and that doesn't include the music I listen to via Apple Music and SoundCloud. Given that I've been building up this music collection for about two decades now, there's quite a bit of stuff in the long tail, though I do listen to near all of it over the span of a few years.
If Spotify want to be around for another decade, I think it's reasonable that they let people build up a respectable library.
Use playlists as well. Each playlist has its own limit of 10K tracks AFAIK. So I save only absolute favorite individual tracks (and take some stuff off the list once in a while), and I save favorite full albums as playlists.
It's still streamed (unless you save it for Offline, which it supports) but it's waaay better for organization. Have you ever forgetten the name of an artist or forgot an artist exists until you scroll past them in your music library?
I've actually changed my model primarily around playlists and don't actually use the library feature of spotify hardly at all.
I'm sure there's a good reason why people want to use their Library instead and why the "playlist" way is inefficient since I've seen a lot of users annoyed with the limit, but anecdotally I've never had a problem with the playlist-first way of using the app.
Actually at this point if they put the limits on playlists and had an unlimited Library I'd probably be jumping ship to something else myself.
I imagine if I hit the 10K limit this inability to use shuffle on ALL of my tracks at once would annoy me.
That way I can still shuffle my collection of favorite awesome tracks, but I also have easy access to my favorite albums, where I still consider the whole album to be good, but only a few of the tracks stand out as absolute favorites.
I can see some use cases where it would occur though.
https://www.rogers.com/customer/support/article/share-everyt...
The challenge for Spotify will be the next few years, seeing how they withstand the substantial onslaught coming from Apple's music subscription service (which is booming as well). I'm skeptical Spotify can stay in the fight financially over time. There's nothing they can do in music that will ever produce the kind of profit required to support a $25 billion market cap (critical given they're about to open themselves up to public shareholder scrutiny). If they can't, the public shareholders will eventually force a sale of the company.
They'd need 250-300 million paying subscribers, most likely, to get to ~$800m in net income (~5% net income margins), assuming they can ever actually make money to begin with. That'd be a generous ~31 PE. It's essentially impossible.
Apple, Google and Amazon on the other hand, never need to earn a penny of profit in music. It'll be perpetually in their interests to hold music service margins on the floor. The music industry won't be so stupid as to harm Spotify, given they'll want the leverage vs Apple & Co? Well they successfully killed off Pandora on margin squeezing, so sure they will. Their view is there's always another company to replace the last one, and that their music rights are the value that's core and eternal.
Your last paragraph is one of the best takes I've read on this.
I think about half of their 2017 net loss was due to something to do with Convertible Notes, that they're explaining on pages F-43 and F-44.
I'm trying to calculate patio11's "fundamental equation of SAAS" ( https://twitter.com/patio11/status/965985835790184448 ) for the numbers for churn, subscribers, and ARPU on pages 67 to 70ish. Will update if I actually figure it out.
Ok, here is my attempt for their premium service:
pp68, 2017 ARPU = 5.32 EUR/mo.
pp69, 2017 Churn = 5.5%/mo.
pp79, 2017 Customer Acquisition, 48 Million at end of 2016, 71 Million at end of 2017. (71 - 48) / 48 * 100% = 48%. To convert to monthly, if I just divide by 12 it is 4%/month. If I do (1.48)^(1/12) it is 3.3%. I guess this is net of churn, So their fresh acquisition I would estimate at 5.5% + 3.3% = 8.8%.
I'll leave the math as an exercise to the reader, partly because I don't know what is being summed over in the formula.
I'd be curious to know at what number of users do they break-even if their cost of revenue, sales, and marketing are linear to users and their other non-financial costs remain fixed.
For the years ended December 31, 2015, 2016, and 2017, we generated €1,940 million, €2,952 million, and €4,090 million in revenue, respectively, representing a compound annual growth rate (“CAGR”) of 45%.
For the years ended December 31, 2015, 2016, and 2017, we incurred net losses of €230 million, €539 million, and €1,235 million, respectively.
For the years ended December 31, 2015, 2016, and 2017, our EBITDA was €(205) million, €(311) million, and €(324) million, respectively.
For the years ended December 31, 2015, 2016, and 2017, our net cash flow (used in)/from operating activities was €(38) million, €101 million, and €179 million, respectively.
For the years ended December 31, 2015, 2016, and 2017, our Free Cash Flow was €(92) million, €73 million, and €109 million, respectively.
Apple Music on the hand hasn't even verified my artist profile— which I need in order to view analytics. So I don't even know if people listen to me there.
As a listener, I'm quite happy about it too. Sure, a lof of there is a bit meh to me, but I've easily discovered 50 new artists that I like a reasonable lot through that channel.
I do wish there were better ways to shape what songs end up in there as sometimes it can be a bit of a miss.
I built Amplitude [1], which uses Spotify's recommendations API to create artist based discovery playlists, if you need more new music than Discover Weekly provides and also Rediscover [2], to save your discover weekly playlists so you don't lose songs.
Spotify used to have apps which had something like this (... I forget the name of the app), it's good to have a replacement.
I would happily pay $1 a month for this. Where is the $12 a year subscription?
Even if it's based on Electron, their Linux desktop client at least exists and this is reason enough for me to keep being a subscriber. The web player is working but is missing features and has some major bugs (e.g., I cannot "dislike" a song on a playlist). My one wish would be for it to get on the same level of playlist management that Amarok, Foobar2000 and similar music players were on like 15 years ago...
Amazon Music's desktop client, despite also being Electron-based, is only available on Windows and Mac and its usability is so bad, I cannot even describe it. There is a reason why Amazon Music is a lot cheaper: it's worse in every metric beside costs that I were evaluating.
Electron does not make cross-compilation easy. It's about as cumbersome as building any other C++ project for multiple platforms.
That said, I'd very skeptical of investing in Spotify. Apple Music has a real chance of becoming the default service everyone subscribes to on their iPhone without thinking twice. Apple Music just have to be good enough.
Only in US.
I wouldn't worry about Apple too much, the market will become saturated in some countries after some years where they are popular. But in other countries, where they only use Android, who do you think will win the market?
Although I've graduated, Apple seems to think I'm still in school and is still only charging me $5/month. The moment that ends, I'm switching to Spotify. Also, the Apple Music app on iOS is horrible. Spotify's app is so much better.
I know Spotify has a similar student deal but for whatever reason I went with AM and now that I've graduated I've just kind of stuck with it since they still charge me the student rate.
Apple Music is incredibly awful at music discovery, and even fairly bad when it comes to just showing me music I might be interested based on what I've listened to. For example I might listen to one song somewhere without necessarily liking the genre of it, and then AM repeatedly shows me music for that genre. Not very intelligent.
With Apple Music and human curators, I can actually find new music better. And I can get my rock fix when I need it.
Edit: Cuz steaming? lol
I subscribe to both because one is not a subset of the other, which I dislike.
But hey, it’s been great, and I’ve actually found new music I love on Spotify/Tidal, unlike every other platform which has attempted recommendations.
I can definitely hear the difference between 96/160/256. Between 256/320/lossless I'd probably fail an ABX test [1] and I'm very doubtful of anyone claiming to hear the differences between 320/lossless regardless of how much of an audiophile they claim to be if they refuse to take an ABX test. I'd maybe believe telling 256 and lossless apart but still be doubtful without any ABX results.
[0] https://support.spotify.com/us/using_spotify/system_settings...
A confounding variable is that Spotify may use different bit rate heuristics than Tidal when streaming, so I’d need to preload the albums to be compared on both platforms, but I’m willing to conduct it blind.
A bigger complaint for me, honestly, is that Spotify’s navigation does not do what I want it to. (e.g., when trying to go to an artist’s page from a song, it just won’t do what I want.)
Or at least they claim they can. I've yet to know anyone with any expensive systems who claims to be an audiophile who can actually pass an ABX test with any more accuracy than random guessing would have given them. Largely because they'd need superhuman hearing - able to hear things that are literally outside the frequency of normal human hearing. Not saying these people can't exist... just that it's extremely unlikely.
In the end, Spotify's content discovery was worth more to me, but if they finally release some lossless version for twice the price, I'll be happy to subscribe. Even at that price, one month of Spotify is stil cheaper than the first R.E.M CD I bought in 1991.
Spotify has been streaming 1411kbps for most tracks through their browser player in Chrome and only then I notice little difference between Spotify and Tidal HiFi. The Masters tracks still are noticeably better than the Spotify on Chrome app playing at 1411kbps.
There is an upside to streaming from the Spotify Chrome app for 1411kbps which doesn't incur extra cost but this is all functionality Spotify is not advertising. The downsides are this only currently works in the US (IIRC) and navigation in the Chrome Browser app is very limited compared to the desktop app.
I'm very surprised you were not able to notice a difference with such a great stack as the HD800s + Mojo and suspect it was something else in the chain limiting the potential. I suggest giving even just Spotify direct in the Chrome webapp a try again if not Tidal and ensure you have the chain configured properly.
You cannot compare bitrates directly like that.
1411.2kbps is uncompressed CD audio, while the 320kbps is a highly developed and advanced psychoacoustic format that aims to be completely equivalent to CD quality to the human ear.
So it's not "getting less than 1/4th of the music", like some audiophiles claim, based on faulty or non-existent understanding of audio codecs.
Also, you are absolutely not getting 1411.2kbps from the Spotify Chrome app. That is simply a display issue, or the actual decoding happening at another step in the chain, and it's either giving you Ogg Vorbis at 320kbps or AAC at 256kbps.
I am a Soundcloud Go+ (or whatever their premium tier is called by now) subscriber, and it is the worst. They don't have the same catalogue available like Spotify, the alogrithm constantly shows me outright garbage (young aspiring musicians? More like 15 year olds with a mic in the basement) and their app... I don't know how a company that old can have an Android app that is so shitty. It even misses the subscription status and won't let me play my collection, even when 2 minutes earlier (and later) it all worked (works) fine.
Can't wait to end my subscription there and move to Spotify/Apple Music.
Sure, there are a lot of unknown/aspiring/indie artists on Youtube and nowhere else, some of them are even quite talented, but if I can find audio literally anywhere else, I will use that instead.
But yes, absolutely buy merch and go to shows, and support artists through Patreon. For the sheer amount of enjoyment these people give people, I think most can afford to support their favorite artists a little extra.
Today, Spotify is thriving as a company and the major record labels have their hands in it. Sure, they have some work to do to figure out how to become profitable and sustainable, but they've come a long way.
Yet I never moved on to Spotify, rather started browsing and buying on bandcamp (who have a great weekly podcast, just enough for me).
I actually think the Apple Watch is a pretty interesting product and I'm a bit interested in buying one, but the lack of Spotify support has turned me off the idea for the time being.
From Idea to Execution: Spotify's Discover Weekly
And remember that not every project is a rock band. Think of an ambient, experimental artist creating calm music for home listening and an adult audience - those artists won't fill any venues or impress their listeners with colorful t-shirts. They rely on album sales.
And a lot of musicians don't make anything, or are actually losing money. But they keep doing it for the love of the art.
Would I like to see these talented individuals properly compensated? Hell yes! But between predatory record labels and the sheer vast number of people out there also producing music, it's not easy to break through.
2015: € 6.84
2016: € 6.20 (9% decline)
2017: € 5.32 (14% decline)
Also, the number of hours consumed per MAU as of December of each year is also increasing. I should have taken the average over the year for both but I am interested in the trend.
2015: 191 hours per year (17.4 billion/91 million)
2016: 217 hours per year (26.7 billion/123 million)
2017: 253 hours per year (40.3 billion/159 million)
So could be possible if you like music commonly listened to in Latin America and don't use Spotify's playlists. Otherwise I'd pay more to stay with the service of your home country.
> No public market for our ordinary shares currently exists. However, our ordinary shares have a history of trading in private transactions. Based on information available to us, the low and high sales price per ordinary share for such private transactions during the year ended December 31, 2017 was $37.50 and $125.00, respectively, and during the period from January 1, 2018 through February 22, 2018 was $90.00 and $132.50, respectively, in each case excluding the Tencent Transactions (as defined herein).
> As of February 22, 2018, we had 176,976,280 ordinary shares outstanding. Except as otherwise indicated, the number of ordinary shares outstanding excludes (i) 14,095,254 ordinary shares issuable upon exercise of stock options outstanding as of February 22, 2018 at a weighted-average exercise price of $49.02 per ordinary share, (ii) 191,985 ordinary shares issuable upon the settlement of restricted stock units (“RSUs”) outstanding as of February 22, 2018, and (iii) 6,720,000 ordinary shares issuable upon the exercise of warrants outstanding as of February 22, 2018, at a weighted average exercise price of $59.92 per ordinary share.
So their private valuation is somewhere between $15B and $23B?
But nowadays I prefer to use smaller sites like http://www.bandcamp.com and the sites of small labels.
For discovery I use http://www.gnoosic.com an AI recommender created by a fellow HN member.
For me, music consumption feels more adventurous this way.
But 90% of my listening time is spotify, because I no longer really keep a personal library outside of spotify.
I download files off bandcamp, and upload them to spotify
and Music-map.com (gnoosic) is amazing. I didn't think it was AI though? I thought it was just simple: if you like this, other people who liked this, liked that.
I couldn't imagine renting my music and I don't think it's a working business model for anyone but Apple, Google & Co., who can handle a loss.
I discovered them recently. The model appears to be (paraphrasing) 'give us non-exclusive world-wide rights to distribute your music freely, and we'll work on ways to monetise later'.
Of note: high-profile artists Wyclef Jean and Bonobo appear to have signed up.
There's no point if I'm still going to have to pirate half the stuff I watch or listen to, or pay twice for a subscription. (If you can subscribe in the first place, that is; last I heard one couldn't even watch game of thrones legally in the Netherlands if I wanted to.)
Well, and it also doesn't help that I've heard you need to unroot your android device (I will not be an unprivileged user on my own hardware) and enable DRM in firefox to watch Netflix. But the main reason is that every time I check, they don't have at around a third of what I would want to watch.
While Netflix's originals also offer a distinct advantage over their competitors and increases the cost of switching, Spotify will likely have a very profitable future as a company, and I'm curious to see if their market cap will ever approach that of Netflix.
Additionally I don't think anybody's cracked the code on how to make reliably well-received music yet. You can't distract the audience with good visuals and special effects if the plot is no good like in a movie.
They don't need it. With Apple Music, Google Play and Spotify majors can afford to destroy Spotify with royalties as they please, and users will just migrate to other platforms and continue to pay.
It's kinda funny once you think about it. Back in early 00's, when Napster appeared, everybody was "dude, record labels are dead, new business models, bla-bla-bla".
And now we are 18 years later, and majors are quite fine. And they did literally nothing for 18 years. No investments, no development, nothing. They just milked different internet services with their royalties.
Well, except enabling you to buy your music from anywhere, DRM-free, playable with the software of your choice.
Is it really their fault that an all you can eat service for 10 bucks doesn't actually feed anyone, including the artists and the platform itself? It's simply too cheap.
We pay for Spotify but we use Pandora on the TV because of the app quality.
Also, some of the songs that I've saved over the past few years have been removed from Spotify's library - just yesterday I noticed that The Safety Dance is greyed out in my client. (Others include I Walk the Line, Save Tonight and Ambarsariya)
That said, I'm a Spotify lifer. I can't see myself leaving and as for the selection, I live in eternal hope that they'll improve it with time.
I once complained to Sufjan Stevens' label about his albums missing in Spotify. They basically told me the deal with their distributor in Europe expired and they didn't look for a new one yet. "woops". It took them many months to get the tracks back.
But I REALLY hope they succeed, because I've been a customer since they first announced US support.
There is a few things that bothers me though. One is that zero effort is spent on solving developer issues in regards to their public API. Some issues has been open far too long on github and the community is unhappy. The other is that they bought up EchoNest, which provided everything Spotify was lacking in terms of a music API, only to close the service and implement none of those features into their public API.
That's it.
For anyone considering the same, beets (python) is your best friend
It's the only credible alternative to Foobar2000 on Linux.
But further, could IP and acquisitions be split out and components attributed to this?
They file plenty of patents, and acquisition of entities such as Soundtrap could possibly have elements allocated to R&D. Is this standard/recommended practice?
Also, RIP Rdio.
Spotify has some huge lawsuits underway for not clearing proper rights for music as well, I believe one is for 1.8 billion. They do have huge investment (ownership chunk) from at least one major label though I believe?
I don't get all the hoopla about Spotify's bad discovery stations. I notice this similar reality distortion field on reddit, feels like marketing.
And I will be keeping my hard drives full of media and cds.
I work away from home in another city weekdays and still in the family plan. If your brother never visits, I understand their rationale though.
We've gone from digital citizenship to digital serfdom...
And nobody has actually abandoned local files, you can buy and download more music than ever before, and own it forever (if you have backups).
Subscriptions have existed for hundreds of years, for anything from newspapers to soap. Priced correctly, subscriptions are perfectly fine. You pay a recurring fee for a service, and in return you don't have to worry about logistics or storage.
A subscription model could even work perfectly fine for cars (and does, in many cities), as long as it's correctly managed and fairly priced. Pay a set monthly fee, and you get access to a range of cars, parked in a central location near where you live. You don't have to worry about maintenance or depreciation, the car sharing company handles all of that. In return, the monthly rate is lower than what most people would pay to actually own and maintain a car of their own. It works because the costs are shared across a lot more people.
I know a lot of people prefer to own stuff, and I get it in a lot of cases, especially when you know you'll need the item in question for a long time, need to modify or otherwise do something that's outside of a normal use case. On the other hand, I think this race towards owning as much stuff as possible is driving western society to destruction.
I agree except it is only through ownership that you are free of others' (e.g. your rentors') control. When that control is used almost universally to expand the financial relationship between rentor and rentee, or to enrich the rentor at the further expense of the rentee, it is imperative for individuals to remove themselves from its influence.
Also, if you load 320kbps MP3 and 14.4khz WAV into a spectral analyzer you will find differences, which will be detectable to anyone with the right critical listening skills.
>Also, if you load 320kbps MP3 and 14.4khz WAV into a spectral analyzer you will find differences, which will be detectable to anyone with the right critical listening skills.
Do you listen with your eyes? A spectrogram can't tell you what your ears will hear.
The only way to know, is to do a proper double blind ABX test on files you encode yourself from known original files. You can use Foobar2000 with the ABX Comparator plugin for this purpose. I think you'll be surprised at just how low you can take the bitrate even on an ancient format like MP3, before you can tell a difference.
"For the years ended December 31, 2015, 2016, and 2017, we incurred net losses of €230 million, €539 million, and €1,235 million, respectively"
How is it better than Twitter?
EBIDTA is also negative.
In other words, there's a clear function being offered and a clear path to giving them money in exchange for that function.
While both Twitter and Snap are functional, their only real offering is a slight twist on you general social network, and their only way to make money is by advertising to their vast consumer base.
As an investor I'd be far more interested in the revenue streams provided by Dropbox and Spotify.
Spotify is also insanely cheap in terms of getting basically all the music you would want in the world for around 10 euro's. It's a solid product with a good reason to exist, which fills a niche for many people.
Twitter is selling user generated content, which costs nothing and not going anywhere. Twitter is as popular as ever. Trump, et al.
Spotify is selling something that belongs to major recording labels; Spotify just license it. Majors can change the licensing rules and torpedo the whole business in a one year, and there is nothing Spotify can do about it.
Of course they won't do it, it reckless, but I think they will be adjusting the licensing rules so Spotify will have almost zero profits for the years to come.
As investor, I would sleep much better as a Twitter shareholder.