Lessons from Spotify
stratechery.com
stratechery.com
Not saying it isn't a risk, just that there are accepted strategies for protecting profits as a distribution channel.
How can possibly Spotify diversify its income though?
[1] https://www.quora.com/Why-do-people-buy-merchandise-of-YouTu...
I'd say his stated rationale for not caring about money (artists getting funded) is at best a low priority and at worst total fluff - they don't care because they make money on phones/hardware, and it also long-term allows them to price competitors (like spotify) out of the business (I don't think it's a coincidence that he gave this interview in the days before spotify's F-1 went public). to apple, streaming is an ongoing marketing campaign for their hardware and hey, if it breaks even great, but even at a loss they can make it up in hardware profits.
I would not be surprised if apple is planning to waiting to announce a significant price or other change to Apple Music right before spotify's first or second quarterly earnings report - not very nice, but it would be a pretty smart thing to do.
[1] https://www.fastcompany.com/40525409/why-apple-is-the-worlds...
Honestly, Netflix's selection is poor enough that I know quite a few people who have gone back to piracy. Spotify is in an even worse position.
Notice how little power Spotify and Apple Music
have; neither has a sufficient user base to
attract suppliers (artists)
Edit: Just remembered that Spotify may already be doing something like this[0] for certain areas where name recognition and back catalog don’t matter. A very clever approach, if they commissioned more tracks like these and filled their playlists with them, and concentrated on growing the areas of their user base that listen to those kinds of playlists, they could decrease their dependence on the record labels.
[0]: https://www.theverge.com/2017/7/12/15961416/spotify-fake-art...
But on researching this comment it would seem I may be way off base https://insights.spotify.com/no/2017/11/02/listening-diversi....
I really hope that is the case. I have more than a hundred artists in my saved list and a huge spread of genres and styles. I enjoy Spotify constantly for giving me some of these insights. I am just as likely to get a suggestion for an artist with 1000 followers versus someone with 500,000 thousand. The algorithm doesn't seem to give much weight to popularity and I love that. Popularity just isn't a strong indicator for quality music and people have been musing over that point for centuries.
I think for many, music is just background noise or social fodder. Used solely for those ends then popularity matters. Otherwise, it's a useless metric and you are just as likely to find a genius with 200 followers as you are a vapid repeater box with 500,000. The hurdles serve as enough of a barrier for people posting just garbage to the platform, but they don't seem to hinder those who are creative but unpopular.
That's me 90% of the time.
So names doesn't matter much AND I'd actually be happy if the artists got paid better, directly and without feeding the middle men in todays music industry.
Why would that need management folks? Wouldn't a simple signup form be enough?
I love the music industry.
/s
No reason it has to be the same entity providing everything that the labels used to.
maybe it means 'exclusively'...
i.e. it's not the absolute size of user base that matters but relative
as long as there is no clear winner, artists would not want to sign up exclusively to one of the services
Meanwhile I'd say that maybe 2/3rds of my playlists are tracks from the past decade or older. Having only access to a subset of the music being produced today is not really enough for me to justify paying 10 euros a month. When I hear some music that I like on the radio I fully expect to be able to find it on Spotify and most of the time I'm right. If suddenly I have to use 3 different streaming services to access all the music I like then it'll be back to pirating for me. How do you even compile a playlist in these conditions?
I really don't think that "pulling a Netflix" would make any sense for Spotify at all, and it would be a terrible thing for the users.
But it is a bad comparison. Start with the consumer and work back.
Consumers expect their music streaming service to have any song they are looking for. For video they don’t have the same expectation.
Music has a greater element of “I want to listen to this particular song/album because it is my favorite”. Video is more about “I want to find something new to watch”.
That fundamentally changes the value of having a complete catalog in music and puts Spotify in a weaker spot relative to Netflix.
I think this has more to do with the economics of pirating movies versus music.
When Netflix started, people were not used to torrenting (or Napster-ing, or Kazaa-ing, etc) movies. They were already at the mercy of Blockbuster, so being at the mercy of Netflix instead wasn't that big a deal. Pirating movies is more popular today, but people are still generally OK with having a limited movie selection.
Whereas by the time iTunes Music rolled around, people were already comfortable with having the whole world of music at their fingertips. You can't undo something like that easily. So the only way forward has been to try to have everything.
Perhaps the most interesting part about all this is that there was never (to my knowledge) such a thing as a store where you could rent records, tapes, or CDs -- Blockbuster for music.
Can you provide a source for this that adjusts for the increase in worldwide internet use as well as bandwidth?
When Netflix started, it was an overnight DVD delivery company, not directly competing with online movie piracy. Pirating movies in 1997 meant making large sacrifices over quality and convenience, not because people "weren't used to it".
By the time Netflix entered the online movie streaming market to compete with piracy, the DMCA had been in effect for years. The world was very used to filesharing before the DMCA went into effect. And Blockbuster was already suffering poorly not just because of Netflix but because of a general shift in how people consume their movies.
> Whereas by the time iTunes Music rolled around, people were already comfortable with having the whole world of music at their fingertips
Through YouTube? Because otherwise I don't know of a truly comprehensive legal streaming service. Unless you're referring to filesharing?
> Perhaps the most interesting part about all this is that there was never (to my knowledge) such a thing as a store where you could rent records, tapes, or CDs -- Blockbuster for music.
Music-ripping tools were ubiquitous at the time (everyone had a CD drive), movie ripping tools were not nearly as popular. It didn't economically make sense.
Why? The fact that it was physically hard to pirate movies in 1997 is part of why people didn't do it.
Through YouTube
iTunes Music predates YouTube being a viable music platform. I'm talking about filesharing.
Music-ripping tools were ubiquitous at the time
I'm talking about pre-CD-drive.
I don't disagree but I also find these foundational to your argument and they appear to be perceived opinion.
Not a store, but you can rent records, tapes and CDs at most libraries.
That is because it is illegal. 17 U.S.C. § 109(b).
It seems one could start a private non-profit library whose purpose is teaching about music history that rents CDs and promotes musical performances. I could call it "Musicbuster" perhaps. I'm assuming no one has done this because there isn't really a market for it? Or is that not a clear and easily exploited loophole in law?
Even if there is a market for it, the nature of the legal limitation to a non-profit library is that the usual capitalist incentives for serving such a market are removed.
Because Redbox rents games
Netflix has been able to track exactly what types of movies consumers are into so that it can deliver more in the same vein. Originally, it would send consumers to other movies it licensed. Now, as a producer, it can direct consumers towards its own flow of productions.
Spotify's recommendation engine has long been a draw for many users. I believe Spotify is setting itself up to be able to increasingly direct users into its own Spotify-flow of artists and musicians. The recent change to playlist mechanics (so that playlists autoplay "recommended similar songs" once completed) will make this even easier.
Having a huge catalog is important, but controlling the flow of what's popular may trump the back catalog in the long run.
edit: Spotify is not as far along in the process as Netflix, obviously, and the entrenched nature of the music business doesn't help, especially with the market practice of locking artists into 7-year exclusive contracts. I believe it will go in the same direction - it's just going to take a little longer.
Netflix's "recommendations" have been pretty terrible from day 1 -- it's mostly random garbage, and suffers from serious issues like grouping sci-fi with horror (seriously? Star Trek TNG belongs in the same category as Hellraiser?) -- and since they started producing their own content their recommendations have quickly become thinly-veiled attempts to push their own mostly-poor content. Add in their horrible UI, replete with autoplaying previews as you browse tiles paired with obnoxious stock music, and Netflix has become something painful to use. I've been moving back to renting/pirating depending on availability and price this year because Netflix doesn't have anything worth watching that I haven't already seen.
Now I open Bandcamp first when I want music.
They are both media-consumption services but I think most people treat them like
It perhaps sounds borderline tautological, but I think the reasoning under Aggregation theory is simply: if they could attract suppliers directly, they would have, and therefore because they haven't, they can't. I think this is reasonable where Aggregation theory holds, since being an Aggregator is an incredibly strong position in the market.
To make a reductio ad absurdum example, as a musician, I believe you could currently just upload your recordings to Spotify, and skip the record label, in the same way that Youtube video producers upload their content directly. (Youtube is an example of a true Aggregator in Thompson's terminology.)
However it seems quite clear that the Spotify-only approach wouldn't be sensible, because Spotify doesn't get you onto radio, tv commercials, and all the other channels that a record label has available. And Spotify revenue alone is less than Spotify-plus-record label. (e.g. see Taylor Swift, who gets (by a very swift Google search) $160m in revenue [2], vs. $2m paid by Spotify [3] (and split between her and her record label). Obviously some of that revenue comes from non-music revenue, but the album sales numbers suggest that she's netting at least tens of millions per year.
[1]: https://stratechery.com/2015/aggregation-theory/ [2]: https://www.forbes.com/sites/zackomalleygreenburg/2016/07/11... [3]: http://www.businessinsider.com/how-much-taylor-swift-earned-...
- People very very infrequently watch the same movie/tv show twice.
- People very very frequently listen to the same song
therefore
- The past is important for songs since at any point in time there are more songs in the past that you want to listen than new songs that have been just released
- The future is more important for movies since you are looking fwd to watching something new
therefore
Netflix >>> Spotify
While probably true now, this will change over time. Most people I know (mid/low 20s) use netflix as a replacement for the radio - putting on The Office / parks and rec just as background noise while doing other things. Those shows have been on repeat in my house everyday for a few years.
In contrast, the ridiculous ads on Pandora is what sent me to Spotify. I can only assume Pandora has low inventory, because I'm getting ads for a tractor company 30 miles away. I'm a Sr dev, listening to Drake, and I don't have a lawn or a farm.
Spotify is trying hard to make free users upgrade to premium, while enlarging the user base. I don't think they have any interest in playing real ads (except some music-related ones that maybe keep the labels a bit more quiet); that's the reason you get tolerable ads. But in the longer term, Spotify might decide to kill the free service (or just make it a timed trial.)
I doubt this will happen. A lot of teenagers right now have only have a free spotify subscription because they don't have any money, and spotify probably wants to keep them as subscribers so in the future they do become customers.
Also, screw Apple Music and Tidal for "exclusives."
Also, I'm curious, what issue do you take with exclusives? Do you also find fault in netflix, hulu, and amazon creating media and making it exclusive to their service?
I can be wrong but wouldn't something like this affect Spotify's bottom line in short term already?
That said, the situation might be even worse. As pointed by the top post from an earlier Spotify IPO discussion, record companies might squeeze Spotify even more if it starts becoming profitable:
Secondly, I don’t think the cost is a problem. if it IPOs, the proceeds will be used to fund (my guesss) customer acquisition which would be in line with their strategy of incurring larger costs now in return for future gain.
Labels on the other hand have other places to go. Labels have options, Spotify doesn't. It's not hard to see who has more bargaining power.
One tangent is that I'm totally dissapointed with the dominant "content aggregators" like kindle, netflix, spotify, apple and such.
Amazon wrapped up the ebook market quick. But, even though it's a revolutionary change in medium, very little changed.
The book publishing industry was ripe for change. Physical publishing gradually got less expensive and less risky. From physical typesetting to short run modern printing. ebooks bring that capital cost down to $0. In response, publishers maintained their "venture capital" status with advances (money upfront to authors) and (supposedly) "marketing," the impression that a publisher can make a book sell.
The point on publishers' marketing importance is dubious, especially in the context of amazon & ebooks.
Advances, editing and the remaining jobs publishers do... The value/cost does not add up to 80-90% of the revenue. Typical royalty rates (5%-20%) are absurd. Amazon could have disrupted this, easily. It would have completeley changed the industry, and made writing a viable living for many more people, producing many more books.
Also, mediums. Magazines famously spawned whole new formats and genres. So did things like paperbacks, pamphlets and other medium changes. Serials and shorts stories. Comic books and gossip collumns. The internet gave us blogs, tweets, infographics and PG essays.
Ebooks gave us digital versions of paper books. The impact has been minimal.
It is as if amazon did their very best to revolutionize books, but change as little as possible.
Spotify did something similar, maintaining and reinforcing the norms of artist-label dynamics, even though they are completely unsuited to the modern industry.
My second tanget is a point on the cost that does scale linearly: customer acquisition costs. This generally means FB & Google. They started as free discovery (customer acquisition from the customers' perspective) and became paid customer acquisition tools.
Today, they are mostly paid customer acquisition tools. This means (a) that these guys charge a toll on certain sectors and (b) that ability to pay this toll is a major factor of success.
It sucks to be in a market with head-t-head competition. These markets are efficient (especially on adwords) and the price of acuiring a customer tends to stabilze near marginal revenue per customer. IE, whatever your margin is on customer, it goes mostly to google.
If you earn less per customer (per impression, more accurately) than the competition, these channels are closed to you. This results in a uniformity of business models. Examples are many, but think of all the subscription food delivery, maid services, laundry, etc.
If you are offering one-off housecleaning and your competition is selling subscriptions, they will probably outbid you on adwords. You will need to change to a subscription model to stay on adwords.
Free discovery channels (eg, HN) have (had?) a totally different dynamic then the paid ones. I think this is an important thing to realize, if you're starting a something these days.
This has actually been fairly clear (if you have exposure to ad buying) since early days, but I've heard relativley little talk of it.
They put a lot of effort into making ad "ROI"easier for advertisers to track at a granual level. Micro1001 will tell you that firms spend upt to that point where marginal spend yields $0 marginal profit. In most markets this is pretty academic. Spending isn't that granual and return isn't that transparent.
In the simple economic model, the assumption is that marginal profit on the 2nd to last dollar spent is more than $0, and the dollar before that is even better. On adwords, it's typically more "efficient." Initially adwords leaned into this, with their "bid-for-ranking" approach. Pay more per click, get more clicks. You could bid to optimize for profit, profitability or volume, as you wished. An 8th place ad might cost a fraction of a 2nd ranked one.
Then, they dropped the RHS ads and focused on getting more clicks to the highest paying advertisers. This has brought adwords close to an on/off decision. There is no more "bid-for-volume" dynamic. The highest bidder takes the whole pie.
Th upshot of all this is that average roi and marginal roi are the same, approaching $0 in competitive markets.
There are a lot of markets where this is completely played out. That is, Google get a cut (anywhere from 10% to 50%) equivalent to the average margin that advertisers earn on whatever they're selling. The advertiser is whoever has the highest margin. Insurance, for example.
Sounds like actual insurance has been commoditised, whereas Google is still making monopoly rents on advertising. There's some kind of analogue of Amdahl's law: the cost reduction from commoditising part of the work of producing a product is smaller than you would think, because the cost will increasingly come to be dominated by the parts that can't be commoditised.
Eh maybe. These days I read a fair number of what I assume are self-published-on-amazon ebooks, as recommended to me by amazon's system. I assume the publisher has been completely cut out, and the length expectations have certainly changed (serials of more shorter books). Maybe people are still figuring out the medium. If you look at the early days of television you'll see "television plays" which at first were literally theatrical actors performing a play to the camera, and these plays were limited to a maximum of one rebroadcast and destroyed after by agreement with the actors' union. Television only became its own distinct medium (with the corresponding economic changes) over time.
Self-published books with non-traditional page counts, at wildly different price points, discovered via recomendation engines is exactly the kind of thing I was hoping to see.
From what I can tell though, this is still a marginal phenomynon, less impactful to the median author than twitter or blogging or somesuch. I don't think this is a coincidence. Amazon has allowed for it, but not really pushed on it, or gambled on it.
When kindles came out, I read a Seth Godin (blogger/writer) comment that eachone should have shipped with 1,000 books, including his. He'd have agreed, in order to get to readers (with his back-catalogue).
Maybe (hopefully) you're right on the timing. Maybe we're just discovering the medium. But, (a) I think things move faster these days, and 10 years should have been sufficient. (b) amazon controls most of the levers, and I think it won't happen unless they decide that it happens.
I admit I've been paying less attention over the last couple of years, but historically his content has been well worth reading. I was fortunate to catch one of his talks in person in Chicago, was quite interesting.