Written evidence submitted by Will Page

16th November 2020

Economics of Music Streaming Inquiry

Introduction

This submission was prepared by Will Page, a visiting fellow at the London School of Economics Marshall Institute,  regular contributor to leading music industry publications, consultant to music industry leaders including Spotify and author of the forthcoming book, ‘Tarzan Economics: Eight Principles in Pivoting Through Disruption’, which will be published by Simon & Schuster in the United Kingdom in April 2021.

This inquiry comes at a critical time for the British music industry, to which Covid-19 has delivered an asymmetric shock. While streaming has shown remarkable resilience, live music has been all but silenced. As artists typically earn the majority of their income from live music, reliance on streaming income has never been greater. In such an unprecedented economic environment, it is a timely call to examine the economic impact of streaming on artists.

Executive Summary

Publicly available economic data for the music industry both in the UK and internationally support the following key points:

●       Since 2015, the British recorded music industry has been growing, largely thanks to the success of streaming. At the same time, the number of artists claiming a share of monies generated by streaming has increased dramatically; thus there are far more artists seeking payment for their work.

●       A fact frequently overlooked in the discussion of the economic structure of the music industry is that streaming services do not pay artists directly but instead pay intermediaries (i.e., record labels, CMOs and publishers), all of whom the artists elect to represent their copyrights on privately negotiated terms. It is these agreements between artists and intermediaries that determine an individual artist’s income. Streaming services cannot directly affect artist pay-outs from intermediates, nor do they have visibility into how much artists ultimately earn.

●       Streaming services typically pay over 70% of their gross revenue to such intermediaries through “revenue sharing” agreements, yet for every $100 collected from the consumer, the songwriter can expect to see just 7.1% and the artist 6.6%.[1]

●       Everyone benefits from the streaming-led recovery, but some more than others – with artists and songwriters feeling short-changed.[2] For labels, the music industry is thriving. Between 2015 and 2019, the streaming-led recovery boosted UK major label turnover by 21% and operating profit margin increased from 8.7% to 11.8%.  The recorded music business not only got bigger, but also much more profitable for record labels. Artists, however, have not received proportional benefit.

●       Much of the public discussion about the cause of the perceived inequitable distribution of digital music revenues focuses on streaming services, but an analysis of the economic structure of the music industry points to the key role played by intermediaries and the distribution of rights holder income.

Background

When illegal file sharing first grew rampant at the turn of the millennium, the UK music industry saw revenues fall dramatically, shedding almost half of its value between 2003 and 2015. Streaming ushered in a robust recovery, as label revenues have returned to strong and stable growth since 2015. More money is flowing into the business, but there are also far more ‘mouths to feed’. Since 2009, the number of British artists and songwriters has more than doubled. Understanding the economic impact and long-term implications of streaming hinges on a clear understanding of this fact.

As to whether the current state of the music industry is equitable, it is important to dispel a common misconception: streaming services do not pay artists at all. Instead, they pay only aggregators. Artists and songwriters are therefore indirect beneficiaries of streaming income, wholly dependent on privately negotiated contracts with these intermediaries.

Furthermore, unlike traditional brick-and-mortar retailers (who made more profit on every single CD sold than a streaming company makes per subscriber), streaming services don't pay just the record labels, but must also directly pay publishers and collective management organisations (CMOs). In addition, they must cover their operating costs including investment in innovation and engineering, staff, advertising and other expenses.

Making music is inherently a collaborative effort, in which producers, artists, performers, managers, CMOs, labels and publishers all play essential roles. Following the money from the consumer to the creator reveals an array of costs, deductions and time-delays that reduce the amount artists and songwriters ultimately receive. From the way an intermediary recoups the advance paid to the creator (where they can make a profit long before the creator is out of debt), to the time it takes for the creator to get paid (for artists this is typically months; for songwriters it can be years), the supply chain appears antiquated compared to other technology platforms that match buyers and sellers.

It’s not just following the money that matters, but also understanding what this money paid through to artists represents. There is still widespread confusion within the industry about the value of a stream. One on-demand stream worth half a penny may seem paltry compared to a radio spin worth £100, until you divide that royalty cheque by the radio audience of, say, five million, which results in a comparative value of a radio stream of just £0.00002.

An analysis of BPI and Spotify data reveals that, since 2015, streaming is responsible for almost all growth in the British music industry, offsetting the collapse in downloads and CDs and overshadowing all other digital channels. Among the UK’s three major labels, turnover has increased 21% since the recovery began in 2015, whilst operating profit has grown by 64%. Operating profit margin has also increased, from 8.7% to 11.8%, meaning the labels are making more profit from more money. Streaming has dramatically reduced their costs (for example by reducing or eliminating the manufacture and transport of millions of CDs) while revenues have grown sharply.

Terms of Reference

What are the dominant business models of platforms that offer music streaming as a service? 

The prevailing music streaming business models provide licensed content (for which costs are frequently based on a percentage of the revenue such content yields) either free of charge to the user with advertising, such as Spotify Free and Deezer Free, or by charging a monthly subscription fee such as Spotify Premium, Amazon Music and Apple Music.

The intention of a service’s free tier is twofold: to offer music to fans who could not otherwise afford to enjoy music, and to demonstrate the value of the service. The reduced functionality of the free service – online-only, limited interactivity – is meant to encourage consumers to upgrade to premium, where they have offline listening, full interactivity and no ads.

As for paid subscription models, the UK has a reported 22.6m active music streaming subscription accounts (see the chart below). Among the UK’s free, ad-funded offerings, the user count of Spotify’s free package is dwarfed by YouTube, which reaches around two-thirds of the UK population. Two additional user-generated content (UGC) services are worth noting: Twitch, Amazon’s video streaming platform, and TikTok, the mobile social media app. Both of these services offer short form content and are reportedly developing their music content strategies.   

Have new features associated with streaming platforms, such as algorithmic curation of music or company playlists, influenced consumer habits, tastes, etc.?

Streaming services routinely offer personalised recommendations based on a wide variety of inputs related to user behaviour. These recommendations appear in playlists as well as suggestions for listening.

Playlists are typically derived from three sources: algorithms, human curation and users. A large majority of playlists are created by users. Algorithmic playlists are a small minority of all playlists. There are, however, synergies at play here as the algorithmic playlists are able to offer scalable curation on platforms which are now ingesting 40,000 new tracks every day[3]. With such vast amounts of new music, some degree of reliance on algorithmic control is a necessity. By way of example, this mix of factors means that artists with the greatest number of followers will have the highest chance of being featured on playlists (e.g. an artist’s new release is likely to feature in its fans’ playlists).

The impact of playlists has been to grow the number of artists who are in the ‘top tier’ (i.e. artists whose tracks in aggregate constitute 90% of all streams on the platform). For example, in 2020 Q2, the number of artists making up Spotify's top tier was 43,000 – up 43% on the previous year. This is part of a longer-term trend: in 2015, the equivalent figure was just 15,000. The Top 40 artists of yesteryear have become the Top 40,000[4].

What has been the economic impact and long-term implications of streaming on the music industry, including for artists, record labels, record shops, etc.?

The most important and exciting fact is the music industry is growing again, in absolute and relative terms. Between 2014 and 2019, spend on recorded music, driven by paid streaming, grew 40% – only to be outpaced by live music, which ballooned by 50% over the same period. The relative success of the British music industry is best illustrated by the increase in the total spend as a share of disposable income (the red line plotted below). Since both recorded and live music took off in 2014, gross disposable income has increased by 20%, whereas music's share of wallet has increased by 25%. In other words, consumers have more money to spend and are spending more of it on music.

The chart below points to another encouraging trend:  a causal link between growth in music streaming subscriptions and growth of spend on live music streaming. This suggests an important spill-over impact of music streaming in the UK, insofar as consumers who pay £120 per year to access music are more inclined to see the music they discover performed live.

The chart below illustrates the sources of revenue contributing to this incremental growth in the revenues to UK record labels between 2015 and 2019. Audio streaming payments from services like Spotify, Apple and Amazon to record labels alone was enough to offset the collapse in both physical revenues and contraction in all other digital revenues combined. Overall, the market grew by over two hundred million pounds.

A key question is whether this five-year recovery in UK recorded music revenues is being shared equitably across all of the industry’s stakeholders. For songwriters and publishers, PRSforMusic reported around 50% growth in total collections between 2015 and 2019, with online revenues up over 300% – both outpacing the equivalent growth rates in the recorded music sector many times over.  

For labels, the music industry is thriving. Aggregate information disclosed in the published accounts of the three UK major labels shows increases to turnover and operating profit (i.e. labels’ earnings less all costs of production including payments to artists, overhead and depreciation but before interest and tax), and to operating profit margin on turnover[5]. The chart below shows that, between 2015 and 2019, disclosed major label turnover increased by 21% whilst operating profit grew by an astounding 64%. Operating profit margin has increased from 8.7% to 11.8%. Clearly, the recorded music business has not only gotten bigger, but also much more profitable for record labels.

As for the implications of streaming for artists, there is more money available but also many more mouths to feed. For context, streaming has transformed the supply side of the British music industry beyond anyone’s expectations. In 1984 (the earliest available data), the UK music industry released just 5,000 singles and 6,000 albums in one calendar year. By 2015, this had increased to a total of 60,000 albums per year. Presently,, Spotify is on record in uploading onto its global digital shelf 40,000 tracks every day. Streaming services are also democratising demand: again using Spotify as an example, in 2015, only 16,000 artists made up the top 90% of streams; in 2020, 43,000 artists make up 90% of the streams. 

This explosion in the supply of new music has been accompanied by a staggering increase in the number of new British-based artists and songwriters. Since 2009, PRSforMusic has seen membership grow from 65,000 songwriters to 140,000 (up 145%) and PPL, which represents artists and recorded rights holders, has seen membership grow from 47,000 performers to 115,000 (up 115%). One of the perverse outcomes of this “long-tail” growth is that if there are more voices expressing concerns about the economics of streaming, it is because there are more artists and songwriters participating in the market than ever before.

Amidst the Covid-19 lockdown, streaming has been a rare success story. Whilst live music has been all but erased due to the virus, revenues from recorded music have remained strong thanks to subscription streaming. In the Billboard article ‘Examining COVID-19's Impact on Britain's Live and Recorded Music Industries’, a model of the potential impact of COVID-19 on an artist's income shows an 80% decline of live music revenues (allowing for some concerts at the start and end of the year), suggesting that the average UK artist’s 2020 income will have fallen by at least 70%.

Access to the £1.57bn Culture Recovery Fund announced by DCMS in August is crucial for individual artists contemplating a savage 70% cut in the income trickling down to them in 2020. However, this assistance will not directly affect artists, insofar as it reaches only the intermediaries that are eligible to apply (e.g. venues, promoters and independent labels). In any event, it will take considerable time for these funds to reach artists.  Highlighting the urgency of this problem, the latest Musicians Union Covid-19 impact poll reveals that 34% are considering abandoning their career in music.[6]

As this data shows, while much of the public discussion about the root cause of the perceived inequitable distribution of digital music revenues focuses on streaming services, an analysis of the economic structure of the music industry points to the key role played by intermediaries and the distribution of rights holder income.

As the market develops, how can the Government protect the industry from knock-on effects, such as increased piracy of music? Does the UK need an equivalent of the Copyright Directive?

Whilst forms of piracy do continue (e.g. stream-ripping and account-sharing), piracy does not ravage the music industry as it once did. Thanks to the success of the on-demand “freemium” models, which offer would-be illegal music pirates a frictionless and free entry point into legal streaming, the problem of piracy has been 'old news' for quite some time.

As far back as 2011, Sweden reported that once licensed streaming services became available, the number of people who pirated music fell by about a quarter between 2009 and 2011. Soon afterwards, similar conclusions came from studies in Denmark, the Netherlands and Norway.

In the UK, the most recent trends also paint an encouraging picture. Since the music industry returned to growth in 2015, the percentage of the adult population downloading unlicensed music has more than halved. Complacency by policy makers is the biggest threat: pirates know where they came from and can easily go back.

Do alternative business models exist? How can policy favour more equitable business models?

Alternative business models should be considered from the perspectives of consumers, creators and intermediaries. Viewed as a simple value chain, it is the consumer who puts ‘money in’, the creator who takes ‘money out’ and the intermediaries who facilitate this exchange. Options for alternative business models and potential implications for these three components of the value chain are explored below.

Alternative licensing models for bringing ‘Money In’

Are there commercially viable alternatives for getting ‘money in’ from the consumer for paid streaming? Probably not. Across the various streaming services, the consumer is presented with remarkably similar propositions, i.e., convenient access to the same 60 million songs for the same fee of £9.99 per month. There is no explicit discount for streaming fewer tracks or for listening to less-established artists, nor is an explicit premium charged for doing the opposite.

Features of the service, however, such as personalisation and ease of use, vary widely. Consumers make their choices based on such features, as well as on other variables including licensing (content acquisition) policies, marketing strategies and sound quality.

Alternative distribution models for paying ‘Money Out’

Are there viable alternatives for getting ‘money out’ to songwriters, publishers, CMOs, record labels and performing artists? In theory, yes; in practice, not yet.

To determine the distribution of monies (directly and indirectly) to both songwriters and artists, streaming services may do so using the “pro rata” or the “user-centric” model[7].

The ‘pro rata’ distribution model

‘Pro rata’ distribution is the predominant model industry-wide. This system aggregates all consumption and spend from a country’s subscriber population  and distributes royalties to rights holders ‘pro rata’ to their streams, i.e. if your tracks accounted for 1% of  all subscription streams in that country in that month, you would receive 1% of all the streaming company’s net distributable revenue (NDR) from that country.

This system is inherently fair insofar as it produces an efficient and transparent outcome where every stream is worth the same to the rights holder, and the model is relatively cost-efficient to manage. It does not, however, recognise that no two consumers necessarily value their streams equally – so those who stream more would effectively be subsidised by those who stream less.

The 'user-centric’ distribution model

'User-centric’ distribution (commonly referred to as User-Centric Payment System or UCPS), by contrast, isolates each consumer's subscription fee and allocates it exclusively to the particular tracks streamed by that user. This system may be perceived as fairer by some creators. However, UCPS arguably increases administrative and operational costs for the numerous intermediaries, not least due to the hugely-increased complexity introduced by the variance of the value of each individual stream. For example: one stream could be worth as much as £4 (if a subscriber were an exceptionally light user and streamed just one song per month), whereas another stream of the same song could be worth a fraction of the current half-a-penny rate (if the subscriber were an exceptionally heavy user and streamed all day every day).

A common misapprehension is that the choice of 'pro-rata' or 'user-centric' distribution systems alters the total sum paid by the streaming service to rights holders. This is not the case, as total ‘money out’ remains largely unchanged: it is not about 'more money', but rather about how to allocate and distribute the same amount of money. To creators as an aggregate, this is at best a zero-sum outcome. Also, it should be stressed that, contrary to popular perception, this allocation and distribution is done by country – so the UK’s ‘money in’ is not pooled with ‘money out’ from any other country.

Another common misapprehension is that the user-centric system would make the rich worse off and the poor better through progressive redistribution of wealth. On the contrary: industry analysis suggests that UCPS would have only a minor impact on the allocation of net distributable revenue among different tiers of artists.[8] Under UCPS, a subscriber who listens solely to a blockbuster artist’s new album would make that artist even richer. What would increase, however, is the cost (to streaming services, labels, publishers and CMOs) of developing and deploying new accounting and IT systems to implement a user-centric model. Moreover, the sunk costs and marginal costs of this would disadvantage smaller players in the value chain who lack economies of scale.

Alternative intermediary models

The question of equitable (and efficient) business models has been addressed by significant market developments since 2009. Independent labels responded to market inefficiencies by creating Merlin, a non-profit collective licensing body that allowed for a better division of labour;  the independent labels’ market share of streaming has increased every year since. With this resurging independent sector, a new model, referred to as ‘label services,’ has also emerged, under which the artist retains ownership of their copyright and receives a more equitable share of revenue in return for a greater financial risk. Additionally, an increasing number of artists are choosing the Do-It-Yourself (DIY) route of not signing to a label at all, but instead using a distribution platform such as CD Baby, EmuBands or Distrokid, whereby the artist pays a one-off fee, retains their rights and keeps 100% of their royalties. The COVID-19 crisis has accentuated the need for and popularity of such DIY tools, so much so that Distrokid has sent almost three times the volume of tracks to streaming services as all three majors combined since May 2020.

It is clear from the recent increase in copyright catalogue transactions that there is plenty (and an increasing amount) of money to be made from the 70% of revenue paid from streaming companies to rights holders. In August of this year, Variety magazine reported that music catalogues ‘attracted what some characterize as astronomical multiples, to the tune of 12 to 15 times what those rights earned annually.’[9] More recently, a transaction in November was worth over $300m and represented 18 times the catalogue’s historic annual income.[10] This market development appears to be driven by the fact that intermediaries’ profits are growing faster than turnover. Notably, neither the artist nor the songwriter directly benefit from such catalogue sales.

The various alternative models discussed above can be viewed as a response to the need for more equitable methods to allocate the 70% of revenues that streaming services pay rights holders. If the last decade saw ‘Disruption 1.0’ transform the consumer experience in the music industry, this DIY trend (‘Disruption 2.0’) indicates that the next decade will transform the creator experience and the role of intermediaries in the digital music value chain, and usher in a new golden era for artists and fans alike.

The submission was prepared by Will Page, a visiting fellow at the London School of Economics Marshall Institute, regular contributor to leading music industry publications, consultant to music industry leaders including Spotify and author of the forthcoming book, ‘Tarzan Economics: Eight Principles in Pivoting Through Disruption’, which will be published by Simon & Schuster in the United Kingdom in April 2021. Will served as Chief Economist of PRSforMusic between 2006 and 2012, where his annual report Adding Up the UK Music Industry  became an essential tool for all music industry stakeholders. He also showed that the UK was one of only three net-exporters of music in the world and that music was seven times more valuable to UK plc than first thought. As Spotify’s Chief Economist,, he published work showing how UK artists are, on average, seeing four Spotify streams abroad for every one at home. The author would like to thank Shannon Nitroy, Tom Connaughton and Bryan Johnson (Spotify); Tom Frederikse (Clintons); David Safir (Music of Economics); Cliff Dane (Red Snapper Music); Sam Blake (dot.LA); Liz Martins (HSBC Economics); MIDiA; PRSforMusic; Entertainment Retailers Association; Phonographic Performance Limited (PPL); British Phonographic Industry; Sound Diplomacy and the Office of National Statistics for their helpful comments and data.