The Economics of the Music Industry and the Impact of Digital Technology
A little over 100 years ago, Emile Berliner invented the gramophone, the first consumer music playback system. Music composers had already already making considerable income by printing music scores so that people could play popular songs on their piano at home, which led to the concept of “publishing royalties”; they were followed by “mechanical royalties” which served the same purpose for recorded media. Soon an entire industry grew up around the sale of recorded music — and a very valuable industry it was. At its height in 1981, over 1 billion vinyl albums were sold in one year worldwide, along with another half billion music cassettes. At somewhere around $5-$10 per unit, that’s a lot of money.
An important point to remember, however, is that this is not money paid for music, it is money paid for a delivery method that contains a recording of music. That might seem like playing with semantics, but that subtle difference will become important later.
Digital Music Arrives
In 1982 Philips and Sony introduced the CD technology. This was a digital storage system designed specifically for music — although it would quickly be extended to cover any computer data via the CD-ROM format. The CD was introduced as a replacement for the vinyl record — and record companies loved it for two main reasons:
- CDs were a lot smaller and lighter than vinyl records, and thus could be stored, shipped, and merchandised much more cheaply and efficiently
- Record companies were facing a saturation problem — by the early 1980s many people had already bought all the “classic” albums they wanted, and they weren’t buying enough records by new artists. CDs enabled the record companies to sell those same “classics” again in a new format instead of having to finance the creation of new music.
This strategy worked incredibly well, and by the mid-90s record companies were selling over 2 billion CDs per year. Interestingly, although CDs had a lower per unit cost than vinyl albums, record companies quickly realised that the market was willing to pay more at retail for them, which made them even more profitable. However the royalties that record labels paid musicians for sales of their recordings on CD remained the same per unit as they were during the vinyl era. There are two vital points to note here:
- That record companies assumed that the CD was nothing more than a cheaper, better, more profitable album
- That musicians and their representatives did not work particularly hard to get paid better royalties, even though record companies were now making considerably more money selling their recordings
Both of these things were to have significant repercussions in succeeding years.
CD-R — The Kraken Awakes
The first CD-R (recordable CD) system was the Yamaha PDS recorder introduced in 1990. It was the size of a washing machine and cost $35,000. Within 5 years Philips has reduced its size to something that could fit inside a personal computer and cost less than $1000. It was sold as a data storage system, but computer geeks quickly realised something that record companies had not: that digital audio is just 1s and 0s like all other digital data and thus could be perfectly copied and reproduced infinitely without any loss of quality. Digital piracy was born.
It’s important to understand that piracy of recording is no new thing. As far back as 1906, the American composer John Philip Sousa had been complaining about pirated copies of his work; and in the 1980s the phonographic industry was insisting that “Home Taping is Killing Music” despite them selling billions of albums and pre-recorded cassette tapes during the period.
Yet, for all this, nobody in the business of selling music seemed to understand that once you digitised music you had created an infinitely replicable resource. They would get a rude awakening within the next years:
By 2001 40% of all CD music was pirated discs — each one sounding exactly like the original, but costing a fraction of the price. The record companies’ response was to increase the enforcement of copyright laws, mainly by lobbying governments and pressurising them to arrest pirates. This had the unintended effect of alienating the music fans who had previously been the record industry’s best customers. It may have been good for short-term business, but it wasn’t good long-term customer relations.
In 1995, the Fraunhofer Institute released WinPlay3, the first openly available software encoder/decoder for MP3. This was the first software system able to decompress and play back compressed audio files in real-time. Suddenly audio files were a lot smaller, could fit on hard drives, and be transferred successfully via home Internet connections.
Fraunhofer also published reference source code as a freely available ISO standard, enabling anyone to download the code and use it to create their own MP3 application — and many did just that.
By November 1997 mp3.com was offering thousands of files in MP3 format, and by 1999 Napster had launched. Engadget points out that:
In January, 1999, “MP3” became the pre-Google internet’s top search term, overtaking “sex.”
Yet three years later in 2002, the music industry’s own report on piracy still devoted 8 pages to CD piracy and only 2 pages to Internet piracy. The record companies’ standard response of using law enforcement also didn’t change — by 2002 they had sued, or attempted to sue, Napster, Morpheus, Kazaa, and several other “file-sharing sites”. Not a single company in the music industry made any attempt at embracing online technology. That would all change on April 28, 2003.
An Apple a day keeps the pirates away
Surprisingly, the man which would revolutionise the selling of music had no previous connection to the music business: Steve Jobs. Jobs told Steven Levy, author of The Perfect Thing:
When we first approached the labels, the online music business was a disaster. Nobody had ever sold a song for 99 cents. Nobody really ever sold a song. And we walked in, and we said, ‘We want to sell songs a la carte. We want to sell albums, too, but we want to sell songs individually.’ They thought that would be the death of the album.
Nevertheless, soon all of the record labels signed with iTunes, in the expectation that it would be a tiny blip compared to their normal retail business. A mantra frequently heard at music business conferences in the months preceding the iTunes launch was “Who’s going to pay to download a song when they can just download it for free?” A lot of people, as it turned out. Apple sold 1 million downloads in the first week of operation of the iTunes store. They had sold 25 million by the end of 2003, and one billion by February of 2006, taking a 30% cut on every sale.
Earlier I pointed out that an important thing to remember was that the music industry didn’t actually sell music; what they actually had been selling were artefacts that contained a recording of music. After decades of record sales however, this distinction had been effectively forgotten — in the eyes and ears of almost everyone the record was an embodiment of the music. iTunes changed that. Although the terminology, and indeed the retail sales model, remained the same as before, with iTunes selling “singles” and “albums”, in reality there was no record. There was nothing a consumer could take home and hold in their hands. All they were buying was a piece of digital information. A piece of infinitely replicable digital information with no solid reality.
But the music business didn’t care. People were buying, and financially the record companies had never had it so good. CDs had been cheaper to produce than vinyl, but still sold for the same price, but MP3s went one better — they cost nothing! They were pure 100% profit (minus Apple’s 30% cut — but record labels had been paying that to physical distribution companies before anyway). And of course, just like they had done when CDs had been introduced, the labels didn’t pay the artists a higher percentage of the income — they continued to pay the same royalty as they had always done. So from the labels’ point of view, this was just another step down the same road as the CD revolution: keep selling records, keep lowering costs, keep increasing profits.
Tea. Earl Grey. Hot.
Pop quiz: what’s the currency that the Federation use in Star Trek?
Answer: there isn’t one.
How can that be?
Because the Federation invented the Replicator. When Captain Picard wants a cup of Earl Grey tea on the Starship Enterprise he doesn’t need to pop down to the Starbucks on Deck 32A. He just goes to the Replicator, tells it what he wants, and out it comes. There’s no need for currency, because everything is instantly available to anyone at any time.
A great deal of a good’s Value is based on scarcity. The less Supply there is of something, the more Value. An oil painting by an artist who is dead is worth more than when when he’s alive — because once the artist is dead he can’t paint any more. Gold is more valuable than sand, even though sand is probably more useful; because there’s a lot more sand than there is gold. Since Value is inversely proportional to Supply, then as Supply increases, Value decreases. Therefore once there is an infinite supply of something (like Earl Grey tea on the Starship Enterprise) it ceases to have any Value.
And this is where the record labels made their biggest mistake: in assuming that buying a download in an online “retail store” was the same as buying a CD in a bricks & mortar store. Yes, the sales model was the same, but the product sold was not. There was no physical good. And not only that, the product that they were selling was infinitely replicable for free. Apple was the Starbucks on the Starship Enterprise.
Convenience beats everything
So did this mean the end of download sales? Not quite yet. And once again, because of Apple:
- Apple had made buying music from iTunes sexy with highly effective and widespread high quality marketing
- Apple made clicking to buy a song file in iTunes convenient. They realised that consumers will pay for convenience more than anything, and even though it was free to copy an MP3 file, it still took time and energy, and many people would rather just click a button and pay instead.
So everyone involved in the industry was still happy, or at least relatively happy. The recording companies, while publicly applauding Apple, were secretly unhappy that they had effectively given Apple a worldwide monopoly on download distribution. If any of the labels had had the foresight and wherewithal to do it themselves, they could have kept an extra 30% of all of those billions of dollars in revenue. This issue would come to rankle the labels more and more as time went on.
A Torrent of problems
The thing that would kill Apple’s “convenience factor” had already been invented, but was not yet widespread: the BitTorrent protocol, first announced in 2001 by Bram Cohen in a post to Yahoo! Groups:
The quote at the end shows that Mr. Cohen was very aware of the likely economic impact of what he was doing, and the kind of results he was expecting.
Unlike systems like iTunes, which were based on a central server “store” and many client “customers”, BitTorrent was a peer-to-peer distributed protocol which allowed anyone to be both client and server at the same time. And because each new user became yet another node that could supply files to the existing users, each new user made the whole system yet more viable and useful. In succeeding years use of the BitTorrent protocol would snowball.
Pirates become Privateers
The strength of the iTunes retail model was obvious: it effectively replicated the model of a bricks & mortar store selling physical records, except with the major advantage of there being no physical store or physical product, thus massively cutting costs for the manufacturers (the record labels) and the retailer (Apple), while preserving retail prices to the customer, which added up to much higher profit margins for the industry. However the success of BitTorrent showed that (at least technically) there was a totally different method of consumption available, one that didn’t rely on buying an individually downloaded file. All that was needed was a way to monetise it, and some file-sharing services tried via subscriptions. Speaking about these early attempts at providing music on demand, Steve Jobs famously stated:
‘’…these services treat you like a criminal. We think subscriptions are the wrong path. We think people want to own their music.”
However what neither Jobs nor anyone else in the industry seemed to fully understand was that once you take away the physical product, you are effectively selling nothing but an experience — the experience of being able to hear the music you want, when you want it. That equated, in their minds, to owning the music. However this is where the point made earlier comes fully into play: the record companies and Apple were never selling music. What they were selling was a delivery method that enabled the user to hear the music.
While he was CEO of popular BitTorrent client µTorrent in 2006, Daniel Ek realised that what people really wanted was access to music. Up until the digital revolution, that access had been synonymous with owning a record. But now there was no record. For most “torrent freaks” it was a simple button push and the music played, effectively instantly. So consumers didn’t really want to “own a record”; what they really wanted was the ability to hear music on demand. The only problem was that streaming music directly was still illegal. All Ek had to do was change that part of the equation — and thus in 2008, Spotify was born; a system that offered on-demand playback of compressed digital audio streams, entirely legally and licensed by the record labels. This was a remarkable breakthrough and something of a business triumph.
Ek realised that record company executives were looking at companies like Apple and other online retailers (e.g. Amazon) and seeing them make not only tremendous amounts of money, but also seeing them with huge valuations in the stock market. The labels had been reduced to the status of being little more than basic suppliers to these retail giants. The record companies wanted some of that action. Ek’s genius move was to not only offer to pay a streaming license for music played on Spotify, but to offer the labels significant equity in his company, and guaranteed cash “advances” each year, thus promising them a full share in its success. The labels, still smarting from getting “conned” by Apple, accepted Ek’s offer enthusiastically.
Enthusiastically, but quietly, because the equity deals were kept private from the public, and most especially from the artists who created the music. The record labels were happy to provide music at a massively reduced streaming royalty rate, in exchange for equity and cash advance deals. And again, the artists creating the music simply accepted the tiny per-stream rate that was imposed upon them. What the artists completely failed to recognise was that Spotify, and the streaming services that grew up in its wake like Deezer and Tidal, guaranteed to pay labels a much higher annual cash advance than the labels were obliged to pay to the artists — and the labels got to keep the excess cash as pure profit instead of having to pay the artists their share. This, in essence, motivated the labels to negotiate lower royalty rates per stream instead of higher — because the lower the royalty rate, the more advance cash was left over for the labels to keep for themselves.
The Black Box
But what exactly does a “streaming royalty” mean? To most people, including most musicians, it means that if someone clicks on your song and listens to the stream, you get paid X amount of money for that stream. Simple, right?
No. Nothing in the music industry is ever that simple.
David Byrne of the band Talking Heads explains:
“‘…about 70% of the money a listener pays to Spotify… goes to the rights holders, usually the labels, which play the largest role in determining how much artists are paid. The labels then pay artists a percentage (often 15% or so) of their share. This might make sense if streaming music included manufacturing, breakage and other physical costs for the label to recoup, but it does not. When compared with vinyl and CD production, streaming gives the labels incredibly high margins, but the labels act as though nothing has changed.’
…However the singer discovered that artist royalties from streaming disappear into a pot called “Black Box revenue” — a name given to income collected by record companies from streaming which is not directly linked to a particular artist or song. Byrne said the money was distributed arbitrarily, citing Sam Smith’s massive hit Stay With Me as an example. “They might give him 3% (of gross revenue) — or 10%. What’s to stop them?”
Not only that, but streaming royalty payments do not stay constant but continuously vary according to the whim of the streaming service. And this variance has seen a pronounced downward trend, even as revenue earned by streaming services goes up:
Making money. Or… not
So, Spotify now had a fully legal system for supplying an effectively infinite amount of music, by subscription. But as Steve Jobs had pointed out years before, nobody was used to paying subscription fees. The solution: make it free for people to use, and hope they will pay a subscription fee later for more advanced premium features: the freemium model. But wait — surely if the streaming company has to pay a per-stream royalty rate, plus pay massive cash advances well into millions of dollars every year, and provide their service to most people for free, won’t they lose money?
Yes. Lots of money.
Spotify’s revenue has massively increased in every year of its operation, but its net losses haven’t decreased, they have increased. Yet it’s seen as the most successful streaming music company, with an annual revenue approaching €2 billion. Other streaming services show a similar trend, often significantly worse:
Profit vs. product
This immediately begs the question: when there is no product, can there ever really be a profit? Currently, for some people the answer is still yes — record labels are still making money. How long they will continue doing that, and how much profit they can retain is debatable however. The IFPI’s annual report for 2016 states:
“Total industry revenues grew 3.2 per cent to US$ 15.0 billion, leading to the industry’s first significant year-on-year growth in nearly two decades. Digital revenues now account for more than half the recorded music market in 19 markets.
However, there is a fundamental weakness underlying this recovery. Music is being consumed at record levels, but this explosion in consumption is not returning a fair remuneration to artists and record labels. This is because of a market distortion resulting in a “value gap” which is depriving artists and labels of a fair return for their work.”
If trends continue that “value gap” is going to keep increasing rather than decreasing. The streaming services continue to report significant and growing losses, and one way they can mitigate that is by continuing to adjust royalty payments downward — and there’s effectively nothing that the rest of the industry can do to combat that. And because of cash advances from the streaming services, coupled with the fact that the record labels hold equity in those services, it’s also arguable that the labels have little motivation (at least in the short term) to want to. Over the longer term of course, that strategy may lead to them cutting their own throats, but in the modern world of “making quarterly earnings” there may be little incentive for executives to look closely at those long-term implications.
The wider issue of course, is the one raised right back in that first BitTorrent post by Bram Cohen: “The market can remain irrational longer than you can remain solvent”. How long can the streaming services like Spotify, Deezer, Tidal etc. keep on losing money and still survive?
Note that I haven’t mentioned Apple Music in this discussion yet. Apple finally bit the bullet and launched their own streaming service in 2015. Like the others, it’s also losing considerable amounts of money. However unlike most of the others, it can afford to. As of February 2017 Apple has $246.1 billion in cash, so it doesn’t need to make money on its service to stay in business. For Apple, its streaming music service is not a business in itself — it’s a value-add that keeps people within the Apple ecosystem. As long as it helps ensure that people keep buying iPhones instead of phones from Apple’s competitors, Apple Music’s losses don’t have to matter.
Similarly, Amazon are strongly positioning their Prime Music streaming as a value-add to its Amazon Prime service. This is made very clear from the outset on the Amazon Prime Music web page:
Amazon Prime membership also includes free one-day delivery on Amazon shopping orders, access to a huge library of free digital books, and a powerful streaming video service, all for the inclusive price of $10.99 a month. With Deezer, Spotify, and Tidal subscriptions costing $9.99 per month just for streaming music, the consumer value proposition seems clearly weighted in favour of Amazon’s offering. Admittedly dedicated music streaming services do offer significantly bigger music libraries than Amazon’s relatively paltry 2 million songs, but with data suggesting that many songs on streaming services are never actually played by users anyway, we have to ask ourselves how important those numbers really are to potential customers.
This is hardly a new process either. If we look back at the early days of broadcast in the UK, listeners originally had to buy a license to receive BBC radio, but when BBC television licensing was introduced in 1946, radio was included in this price as a value-add. Separate radio-only licenses continued to be issued until 1971, but from that point on BBC radio broadcasts were looked upon as so unimportant financially in comparison to TV that they became completely free. It’s quite possible that the same thing could happen with some Internet streaming music services.
Looking to the future
There are huge revenues being made in the digital music industry, but the foundations those revenues are built on are shaky. For example Soundcloud, once considered a hot property, was recently forced to admit to investors that it may run out of cash this year.
One option for reducing costs is for music streaming services to adopt the Netflix / Amazon Prime Video model of circumventing traditional producers in favour of commissioning their own content, but this runs into the danger of upsetting the record companies that they still depend on for the huge bulk of their content. However there is some evidence that the streaming services may be quietly exploring this option. With pressure from the other direction from artists and their management sick of getting tiny royalty payments, it can only be a matter of time before both sides begin to question what place record labels have in the equation going forward.
A strong argument could be made that music streaming services may not ever be able to turn a profit and that the only real future lies in being a value-add to another service or product. Personally I think this is practically inevitable, and I assume that the current standalone streaming services are actively pursuing acquisition or merger deals with other services and companies. Soundcloud has been looking for a buyer since 2014, but their asking price of $1 billion was widely seen by analysts as being far too high — an estimation that has subsequently been proven correct. There is now the possibility that potential buyers may simply wait until Soundcloud runs out of money and try to pick the company up for pennies on the dollar — a fate that has already befallen services such as Rdio and Songza.
The one constant in the history of digital music is that it’s a race to the bottom, and that race is far from over. Basing your income on a product that has effectively no intrinsic market value is not a good position to be in. Of course, music does have value to consumers, but that value is based on emotional connection, not on market forces. Whether the “music industry” as an entity can survive at all is not certain — it may well be subsumed into being a minor supplier to the larger media industry as a whole. Where that leaves the musicians is extremely unclear. The only thing certain is that people will still be listening to a lot of music somehow in future, but they won’t be paying a lot for the privilege — if at all.