Article by Professor Phil Graham QUT
This article has been peer reviewed prior to publication
People have been grasping about for a while now, searching for metaphors through which to comprehend the rise of streaming businesses. The three main rivals so far have been sales, radio, and licensing metaphors, allegiances to which have mainly depended on where one sits in the hierarchy of revenue sharing. In a recent and brief conversation with Michael Smellie, mainly centred around the issue of labels occupying seats on “both sides of the deal” when catalogue-level licensing agreements are done, the topic of the record club came up.
Thinking about my own engagement with that business model I had a similar memory to that which Snow (2013) relates:
One day I came across an advertisement in one of the magazines to which my parents subscribed. The advertisement was from Columbia House Records, and it promised ten 33-rpm record albums for a penny if you joined their record club. Such an offer I could not resist, so I clipped and filled out the ad, enclosed a copper penny, and sent it off. I was thrilled (and my mother was surprised) when two weeks later a package arrived containing 10 record albums. I explained to mother what I had done, reassured her, and settled back to listen to Gene Pitney, Neil Sedaka, Lesley Gore, and others. (Snow, 2013)
If we think about it, the record club is probably the closest metaphor for the streaming business model we have at least as far as past practices in the recorded music business go. Leaving aside the fact that it has its history firmly grounded in the commodity-based era of material “things”, it shares many aspects in common with the streaming model. It generated billions in revenue and created thousands of new jobs, it functioned to a degree as a discovery mechanism for consumers, it was based on a “freemium” model, and it was used by the majors to construct deals that were unfavourable to the artist, the retailer, and the customer. And while it is still an insufficient metaphor to understand what is currently going on with streaming, it does provide us with a model of what avid fans of music were a) willing to pay for music on a monthly basis (on average just under $50), b) what that might mean if money were fairly distributed to artists in current circumstances, and c) what “consumers” (subscribers) and retailers might expect from streaming services in future. 
the record club is probably the closest metaphor for the streaming business model we have,
A streaming paradox
According to Audiam Data (Price, 2015), which has just conducted a detailed and comprehensive analysis of the 6 largest streaming platforms operating in the US, there is a striking paradox built into the system: as platforms, including Spotify, collect more revenue from subscriptions, ‘artists, songwriters, publishers and labels’ are being paid less and less money per stream.
as the platforms, including Spotify, collect more revenue from subscriptions, ‘artists, songwriters, publishers and labels’ are being paid less and less money per stream.
The Spotify Premium monthly per-stream rates are calculated by dividing the money in the royalty pot (the Spotify Reported Gross Revenue) by the number of streams in that month. The decrease in the per-stream rate is occurring due to the number of streams per month growing at a more rapid rate than the revenue.
In other words, it appears anyone that pays $10 a month for unlimited music streams a hell of a lot of music. In addition, as the rates drop, the money is being spread over a larger number of artists causing the money to spread more “thinly”.
The end result in 2014 for Spotify’s Premium service is an artist needs increased streams each month, at what appears to be an untenable rate, to just stay even with were they were financially the previous month.
It also means that in many cases, they can have more streams than the previous month and make less money. But this isn’t just limited to Spotify, it can be seen more dramatically across all music streaming services we track. (Price 2015)
There are some intriguing mathematics, behaviours, and assumptions implied here. If subscribers are so enthusiastic for free music, as we told over and over they are, why would they stream many times more music simply because they paid ten dollars per month rather than nothing? Do we assume guilt on the part of the freemium user? That there is genuine discomfort built into the freemium interface? How much more music would be accessed by subscribers if the price were to rise to the average level reached at the peak of the record club average monthly subscription?
Fig 1. Overall streaming revenue vs overall per-stream royalties (from Price, 2015)
As far as artists are concerned, all this is to some degree just another red herring in the fair remuneration argument. That is evident in the language when we see the term “rights holders” used as if it were meant to imply that artists and composers were actually being paid, or when it is put forward as a general term simply to ameliorate the outrage about criminally low royalty payments. In such contexts it means, unequivocally, major labels and publishers. They are the ‘rights holders’ Daniel Ek, Spotify CEO, means when he talks about his business paying out billions to them. I for one believe him.
The problem is what happens to the money next. A recent flurry of leaks from the major record businesses, starting with the leak of Sony’s 2011 contract with Spotify, have led to a similarly energetic flurry of defence tactics from the other two majors, all of which have highlighted the depth and thoroughness of the streaming deals’ cynical approach to artist remuneration. Here is Sony’s initial response:
Sony Music historically has shared digital breakage with its artists, and voluntarily credits breakage from all digital services to artist accounts.
Under the Sony Music ‘Breakage Policy,’ SME shares with its recording artists all unallocated income from advances, non-recoupable payments and minimum revenue guarantees that Sony Music receives under its digital distribution deals.
This applies to all revenue under digital catalog distribution agreements, whether or not the guarantees, advances or ‘flat’ payments can be associated with individual master transactions. (Sony, 2015, in Resnikoff, 2015).
Warner responded with an equally vague statement about what happens to advances paid by streaming services, also emphasising ‘digital breakage’ (however that might be philosophised into existence as commercial value) and other ‘unallocated’ incomes (Ingham, 2015).
The advance of the advance
The language in the Sony contract is remarkable for its brazenness. Here is an example:
The Contract Period 1 Advance, Contract Period 2 Advance and Contract Period 3 Advance shall be recoupable solely against Label Fees payable in respect of each corresponding contract period. (Sony Spotify Contract (SSC), p. 18)
Given that the total advances being paid to Sony by Spotify total $42.5 million over 3 years in the contract (2011-2014), along with an extra $9 million in advertising value, the above paragraph, and the fact that Sony agreed to a $0.00255 royalty rate, makes quite obvious that the label aims to keep the ‘label fees payable’ (essentially the aggregate of price calculated per stream) in any period as low as possible so as to avoid having to distribute any advances to artists (Ulloa, 2015). The language of “the advance” in the contract is to some extent a new meaning. It is framed as a guarantee by the platform of a specific revenue level from the platform regardless of numbers of plays, subscribers, advertisers, or any other income.
In such a deal advances can never be exceeded by royalties because the royalty rate is so low. Therefore the advances need never be distributed to artists since actual income (generated by streams) has no chance of equalling the advance paid by the platform (the income level guaranteed by the streaming platform). Add to this that many, if not most, artists signed to majors remain “unrecouped” on their own deals with their label, the paucity of streaming income further guarantees the unlikelihood of ever recouping against their personal advances in the streaming economy.
Therefore the issue of advance and recoupment is seen to be at the centre of the emerging streaming payment scandal, the overall result of which is as follows, at least for major label artists in on two of the largest streaming services:
Fig 2 ( from Holmes, 2015, see also Masnick, 2015, and Techdirt)
These charts, based on label-funded studies of Spotify and Deezer, show the labels taking ‘almost 75% of the post tax revenue’ (Holmes, 2015). Masnick points out that because the study that generated these charts was
paid for by SNEP, which represents the major labels, it then tries to spin this as being not only perfectly fair, but a good thing for the artists themselves. … The report claims that 95% of that money that goes to the labels goes to cover all of the “expenses” … [of making and selling a record] (Masnick, 2015)
The other issue of sustainability in the model is the fact that none of the largest platforms have ever made a profit. In fact as Spotify increases its revenues to $EU1Billion, its losses continue to grow at faster rates than ever .(Dredge, 2015).
none of the largest platforms have ever made a profit. In fact as Spotify increases its revenues to $EU1Billion, its losses continue to grow at faster rates than ever.
At the same time Spotify’s market value (its value based on share price) has grown to be larger than the entire North American recorded music business (Kaye, 2015). Kaye notes the paradox in the existence of a streaming service that is worth ‘more than all music sales combined’ by an order of about 150 per cent, an extraordinary mathematical prospect:
Streaming income, which includes paid subscriptions and ad earnings, accounted for 27% of the recorded music industry’s $6.97 billion earnings. Yet Spotify, just one company that makes up that percentage, is actually worth more than every single US retail music revenue source combined. Given, Spotify is an international company, but either way the math is fascinating. (Kaye, 2015)
That puts Spotify’s market capitalisation at about $8.4 Billion (Kaye, 2015). This is where the double dealing aspect of the whole problem comes into more pointed focus for artists because, according to most estimations, the major companies own around 20 percent of the streaming companies (Resnikoff, 2014a). By rough calculation that puts the major labels’ proportion of the latest Spotify valuation at around $1.76 Billion, or $588 million per major assuming an even share of the Spotify pie for each (which is not quite the case but close enough for our purposes here). So, for example, in addition to Sony’s $42.5 million guarantee (in 2011 dollars), the amount of value that they have extracted from being part-owners of the platform itself is more than 10 times that amount over roughly the same time span. Despite the entire value of the service relying on music, artists are unlikely to ever get a share of that “other side” (market capitalisation) of the deal.
Streaming economy revenues in the US
Yet, again, all of this obscures our question:what is a fair payment to the artist?
all of this obscures our question: what is a fair payment to the artist?
How do we work it out? Further complications come into play. So far we have been concerned with the machinations between the major labels and the fast-emerging array of new streaming services. Price (2015) does musicians the excellent service of analysing the annual value of the streaming economy revenue in the US in 2014 (Price, 2015). It is $660,098,066.52, just over a third of the combined record club revenue at the turn of the 21st century ($1.5 billion). That represents a 50.29% growth rate in revenue from 2013. That revenue was generated by 74,122,004,238 individual streams, which is an 80. 71% increase from the previous year. $445 million was paid to ‘rights holders’ and $36 million was paid to ‘public performance’ fees (Price, 2015). Outgoing revenue for mechanical royalties was 6.29% of Gross and for public performance royalties it was 5.57% of Gross. But again, this is all to do with the relationships between the platform and the major recording and publishing businesses. There is a gap in calculations about what percentage of streaming revenue is generated by DIY, or “non-major” label music, which, in sheerly quantitative terms far outweighs that of the major labels. Recent rumblings around Apple’s move into the streaming space, beginning with a 0% royalty rate, have proven interesting. The issues for indie and major artists alike are summarised succinctly by Max Willens (2015).
Earlier this week, Spotify released a graph suggesting that indie artists on its service weren’t just surviving – they were making pretty decent money, and they were about to become millionaires. As part of a posting announcing that its service now has 20 million subscribers, Spotify said that a “indie/niche artist” typically earns $700,000 a year from its service, a total that would rise to $1.2 million with so many new subscribers now paying into the company’s royalty pool.
“More people listening on Spotify means more payouts to the creators of the music you love,” the post read. “As we grow, the amount of royalties we pay out to artists, songwriters and rights holders continues to climb faster than ever.”
For many artists, as well as for advocates who have long complained that Spotify does not pay a fair amount for streams, that number came as a surprise. “That was pretty shocking,” said Kurt Feldman, a sound designer and engineer who spent many years living as a working musician with the bands The Pains of Being Pure At Heart and the Depreciation Guild. “I don’t know what they consider indie.” (Willens, 2015)
The Spotify announcement said that the platform pays out between $1.4 and $2.6 million per successful heritage artist per year and between $0.7 to $1.2 million per year to ‘niche industry artists’ (Willens, 2015). Those figures shocked artists from both categories.
Fig 3. Spotify payments per artist according to Spotify (from Willens, 2015)
More for less and less for more when less means more
We have already seen that more paying subscribers actually means less money for artists. But we can let that pass here. The main thing is that Spotify defines as ‘indie’ by a usefully indicative example: ‘a U.S. artist’ selling ‘between 100,000 and 300,000 albums a year’ (Willens, 2015). This to greatly underplay the number of indie records that are available in the streaming economy, since there were only about 200 artists in the US who managed to sell that many albums last year (Willens, 2015). For example (and my figures are old here, I realise, although I cannot conceive of the proliferating DIY trend having reversed in the meantime), I have previously pointed out in this forum (Graham, 2012) the massive proportion of songs in Gracenote that are not accounted for in the licensing numbers that IFPI cites. In 2011, according to IFPI there were ‘400+ licensed music services involving 13 million tracks’ but, ‘according to MusicHype CEO Kevin King,’ there were ‘97,000,000 songs in Gracenote’ (Graham, 2012). More recent numbers of licensed tracks in IFPI’s are between 30 and 37 million, depending on how titles are being counted (IFPI, 2014). That still leaves unaccounted for tens of millions of tracks that are part of the total streaming environment but which are not counted in the IFPI statistics. The Gracenote numbers undoubtedly include many items with zero streaming potential, such as soundtrack cues, production music, incidental music, and so on, but the vast number of tracks generated globally by the DIY sector, as evidenced by YouTube, SoundCloud, Bandcamp and others, is clearly in the tens of millions and by now far outweighs by sheer number the amount of tracks owned by the majors.
the vast number of tracks generated globally by the DIY sector, as evidenced by YouTube, Soundcloud, Bandcamp and others, is clearly in the tens of millions and by now far outweighs by sheer number the amount of tracks owned by the majors.
The easy argument against those DIY numbers being of any relevance is that they make up such a small percentage of the tracks streamed that they are not worth consideration. That may be true in terms of revenue percentages generated per track, although low transaction costs should not preclude the smallest artist receiving unit-based remuneration. The net result is that, in Resnick’s (2014b) words, ‘streaming has failed artists and independent labels’. That is somewhat to overlook the elephant in the room. The sheer fact of the size of the independent sector contributes enormously to both the perception and reality of massive choice and diversity available on the streaming services. This is another essential element in the calculation of fair return. Sebastian Chase puts it this way:
Nobody would subscribe to a service that only had one label’s acts on it. The majors tried it and it failed outright. The punters know that there is an endless ocean of music out there and they can get it for free. More than they can ever listen to. The thing that makes the streaming platforms viable is the sense the music fan gets from them that everthing is on there and they can listen to it whenever they want. Even if they only listen to a few thousand tracks in their whole life – and that’s a lot – it’s the perception of choice that makes the platforms valuable. That’s why they launch a PR campaign every time an artist makes noise about withdrawing their rights. (Sebastian Chase, personal communication, 2013)
What this points to is an approach that recognises the relationship between sheer mass and diversity and the ongoing viability of streaming platforms. Such cannot be had without independents and DIY artists who are the majority holder of copyrights in numerical terms on the net (if the difference between IFPI and Gracenote figures are to be taken seriously). In 2013, 80 percent of tracks on Spotify had ‘been streamed at least once’ (Rochelle, 2013). That means at least 4 million tracks had never been played at all (Rochelle, 2013). Yet they did their part in terms of perceived value. They contributed to the impressiveness of the 20 million tracks in Spotify at the time (16 million being a much less impressive a figure), even in their failure to attract individual attention. Whatever value the platform has, each track listed there contributes to it passively by contributing to its perceived value, mass, diversity, and related viability to subscribers. That implicates ether a “share of ownership” model that extends to each artist on a per track basis, or a basic annual fee per track simply for being part of the platform, with extra payments for actual streams paid on a cost-based, per-stream model.
Spotify might argue that is currently the case, and in some ways the Sony-Spotify contract would support such an assertion. But for Spotify it would be true only after advances are paid to majors. The advance and recoup model currently in place needs to be reversed for it to be fair to anybody, including the streaming platforms. That is, the revenue raised by a platform should be first split into payments per stream, with advances to any party being based on a figure that can realistically be recouped within the term of the contract with that party and relating specifically to tracks belonging to that entity having been streamed in previous years or quarters. That in turn would necessarily mean gross income being seen as identical to total streams per artist minus costs of operating the platform plus an agreed, and probably capped, profit percentage. Otherwise everybody loses.
Making the magic pudding
At their height, the combined revenues of the large record clubs were around $1.5 Billion per year.
At their height, the combined revenues of the large record clubs were around $1.5 Billion per year. Discourses of complaint, accusation, and defence around their business practices two decades ago have recognisable resonances today:
Record clubs have been under attack from retailers, who charge that club offers of “12 CDs for the price of one” devalue music product. In March, at the National Assn. of Recording Merchandisers’ annual convention, a group of merchants met to consider pursuing legal action to remedy what they consider to be onerous record club practices. But that effort appears to have reached an impasse, because four of the largest accounts are said to be opposed to taking legal action.
In addition to music merchants, some artist managers have been critical of record clubs because they pay reduced royalty rates and use a significant amount of free goods to market the club. (Christman, 1996)
The punchline of this particular article about the fight between retailers, record companies, and artists is Columbia House CEO Richard Wolter’s closing comments that could have come straight out of the “Age of Streaming PR Strategy Book For Interested Parties”: ‘We continue to be a unique distribution channel, which has helped the industry grow … We are an important income stream for artists and labels … and we are good for the consumer’ (Christman, 1996).
The double dealing aspect of the new streaming models, which sees the major labels on both sides of the deal – that is, as major shareholders in the services that are being licensed by the labels and the main licensing entities – were also a feature of the record club business model. Here is the outline of a suit that was launched by retailers at the turn of the century and which was eventually successful:
Last year, Sony and Time Warner Inc.’s Columbia House record club bought online retailer CDNow. Sony and Time Warner – which is linking with London-based EMI Group PLC in a joint venture that creates the world’s biggest recorded-music company – each own 37% of CDNow.
The lawsuit alleges that the creation of CDNow and the links to its online store from the enhanced CDs provide Sony with “substantial pricing and promotional advantages” over competitors and illegally “misappropriates” consumer information from retailers.
Sony illegally provides identical sound recordings to CDNow at a lower price than it charges retailers, the suit says. This creates artificial barriers to entry for competitors who want to compete for online recording sales, the suit alleges. (Hendrie, 2000)
Nobody is in the record club business any more. The record retailers have also largely disappeared in the meanwhile. The “musical middle class” of professional players, composers, and musicians has withered away and music has become all but free to produce and distribute thanks to digital technologies. Sales of music have fallen dramatically since the turn of the century and have only been partially replaced by a streaming model that for many musicians simply represents a form of legalised piracy, with the game being rigged at the table before anyone gets to the door of the casino.
Back to the future – again
Let’s go back to our opening tale about the excitement and magic of joining Columbia House record club and getting ‘ten new LPs for a penny’ and see what happened next to Steven Snow, excited teenage customer of Columbia House:
Things went along quite well until a few weeks later when I returned home from school to face my angry mother, who displayed to me a bill from Columbia House Records for $84. You have to understand that in those days $84 would buy several weeks of groceries for our entire family. To this day I don’t recall exactly what went wrong with my new record club arrangement. In hindsight I probably missed the mailing from Columbia House Records to buy the current month’s record, or perhaps I simply missed the fine print in the ad. But upon seeing my mother’s distress, I did something I had never done before or, for that matter, had ever seen my parents do before—I called a lawyer. (2013, pp. 10-11)
And this is the seemingly inevitable, expensive, and unsatisfactory end to the streaming tale currently facing major recording businesses and streaming services alike. With the Sony contract out in the open, legislatures in the US and Europe are beginning their enquiries. Simultaneously, Apple are being threatened with legal action from existing services, and threats from independents to pull their tracks from the service. This follows Apple’s moves first to end the freemium model as they attempt to enter and conquer the streaming market then doing an about face with a zero royalty policy that maxes out at 58 per cent. Whether “Free” simply has to end or not is a whole other question with its own parameters around revenues from advertising and other sources. What does have to end is the current advance-recoup arrangements that are structured to incentivise the lowest possible payouts.
What does have to end is the current advance-recoup arrangements that are structured to incentivise the lowest possible payouts to artists, composers, and independent labels; the destruction of the recorded music business; and the commercial ruin of the streaming platforms themselves.
The next move, based on past practices, will probably be some convoluted billing strategy that reintroduces the concept of a “sale” in an “opt out” billing model or some such. If that happens it will probably spell the end of streaming. It is only at the point at which revenue and costs are made transparent across all platforms, and the point at which all music included in all platforms is remunerated as an intrinsic part of the platform’s value, will we be at a point where artists can be fairly remunerated.
Christman, E. (1996). Record club ups hold-back period. Billboard, 108 (26): 8.
Dredge, S. (2015). Spotify financial results show struggle to make streaming music profitable. The Guardian. Available online at: http://www.theguardian.com/technology/2015/may/11/spotify-financial-results-streaming-music-profitable
Graham, P. (2012). Digital value chains for music promotion, licensing, and sales. Music Australia. Available online at: http://musicinaustralia.org.au/index.php?title=Digital_Value_Chains_for_Music_Promotion,_Licensing,_and_Sales
Hendrie, P. (2000, Feb 01). Music retailers sue Sony music over hyperlinks group claims buyers are being steered to Sony online stores. National Post. Accessed online at http://search.proquest.com/docview/329520484?accountid=13380
Holmes, D. (2015). Labels, not Spotify, are screwing over artists and breaking the music industry. Here’s how to fix it. Pandodaily. Available online at: http://pando.com/2015/02/06/labels-not-spotify-are-screwing-over-artists-and-breaking-the-music-industry-heres-how-to-fix-it/
Ingham, T. (2015). Warner pays artists share of Spotify advances… and has for 6 years. Music Business Worldwide. Available online at: http://www.musicbusinessworldwide.com/warner-pays-artists-a-share-of-spotify-advances-and-has-for-6-years/
IFPI. (2014). Digital Music Report 2014. IFPI. Available online at: http://www.ifpi.org/downloads/Digital-Music-Report-2014.pdf
Kaye, B. (2015). Spotify is now worth more than the entire US music industry? Consequence of Sound. Available online at: http://consequenceofsound.net/2015/04/spotify-is-now-worth-more-than-the-entire-music-industry/
Masnick, M. (2015). Yes, Major Record Labels Are Keeping Nearly All The Money They Get From Spotify, Rather Than Giving It To Artists. Techdirt. Available online at: https://www.techdirt.com/articles/20150204/07310329906/yes-major-record-labels-are-keeping-nearly-all-money-they-get-spotify-rather-than-giving-it-to-artists.shtml
Price, J. (2015, June 11). The more money Spotify makes, the less artists get paid. Digital Music News. Available online at: http://www.digitalmusicnews.com/permalink/2015/06/11/the-more-money-spotify-makes-the-less-artists-get-paid
Resnikoff, P. (2014a). The Major Labels Are Trying to Sell Spotify for $10 Billion, Sources Say. Digital Music News. Available online at http://www.digitalmusicnews.com/permalink/2014/06/11/major-labels-trying-sell-spotify-10-billion-sources-say
Resnikoff, P. (2014b). The Music Industry Has 99 Problems. And They Are… . Digital Music News. Available online at: http://www.digitalmusicnews.com/permalink/2014/09/02/music-industry-99-problems
Resnikoff, P. (2015). Sony Responds: “We Share All Advance Income With Artists”. Digital Music News. Available online at http://www.digitalmusicnews.com/permalink/2015/05/26/sony-responds-we-share-all-advance-income-with-artists
Rochelle (2013). There Are 4 Million Songs on Spotify That Have Never Been Played Once. Digital Music News. Available online at: http://www.digitalmusicnews.com/permalink/2013/10/11/songsonspotify
Snow, S. E. (2013). Musings of a small-town lawyer. Religious Conviction 3, (14). Available online at: http://digitalcommons.law.byu.edu/life_law_vol3/14
Ulloa, N. (2015). $51.5 Million Reasons Why Spotify Isn’t Paying Artists. Digital Music News. Available online at: http://www.digitalmusicnews.com/permalink/2015/05/19/spotify-gave-sony-51-5m-in-advances-and-ad-space-over-three-years
Willens, M. (2015). Is Spotify Making “Indie” Artists Rich? Depends On How You Define “Indie”. International Business Times. http://www.ibtimes.com/spotify-making-indie-artists-rich-depends-how-you-define-indie-1964507
 Note that I am not including the new commission-based majors here, such as Kobalt and BMG, whose business models are organised around commissions rather than advances. There are many arguments for getting more specific about sectors and sub-sectors involved in streaming. This is a broad-brush treatment of a major problem and here is not the place to go into the many kinds of alternative arrangements and deals around streaming about which little is known and which the problems canvassed here literally dwarf.