15 February 2019 Global

EQUITIES Global Audio Revolution

40,000 10.0% The has changed 30,000 5.0% 0.0% 20,000 -5.0% 10,000 -10.0% Key points

0 -15.0%

2007 2015 2006 2008 2009 2010 2011 2012 2013 2014 2016 2017 2018E 2019E 2020E 2005  The convergence of advertising and consumer direct spend on audio entertainment disrupts broadcast radio in favour of streaming platforms.

Recorded music Radio Audiobook But the golden age of recorded music is behind us. Audiobooks and Streamcasts Audio yoy (RHS)  are gaining share, undermining labels’ competitive position. Source: IFPI, Magna, APA, IAB, Macquarie Inside  Our top picks are companies building a competitive advantage in the wider audio entertainment space: TME, SPOT and SIRI. Underperform on VIV.

Table of contents 3

Exec summary 5 New models in audio entertainment Redefining audio entertainment 16 Recorded music is to audio what films are to video entertainment. It’s premium Freemium convergence 20 content, but it only represents 36% of the total 2018E audio revenues in the US. Streaming revolution 29 OTT is the first technology that brings together all audio content (music, Content-distribution equilibrium disrupted 36 audiobooks and radio/podcasts) on a single distribution platform. And it reunites advertising revenues with consumers’ direct spend (downloads/ subscriptions). Vertical integration appeal/necessity 40 Further disruption comes from a new type of integration achieved by TME (social/ The buoyant Chinese market 42 entertainment ecosystem) and by (e-commerce/entertainment plus voice Smart speakers and voice platforms 46 platform). Our new global audio entertainment model 50 Streamcasts and audiobooks rise, music slows Regulation 53 We argue our holistic approach allows new investment opportunities:

1) Ad revenues shift from broadcast radio to on-demand streaming - We expect

commercial radio advertising to decline as streaming platforms grow into podcasts, the equivalent of unscripted TV shows in parallel with video Analysts content. recently announced a $400-500m investment envelope for Macquarie Capital () Limited podcasting content, including the acquisition of Gimlet/Anchor. Giasone Salati +44 20 3037 2670 [email protected] 2) Audiobooks the most interesting growth opportunity - Audiobooks are the

Jenny Chen +44 20 3037 5242 equivalent of premium scripted TV shows that have been so successful on [email protected] Netflix, but are mostly overlooked by consensus. Amazon is a leader with its

Audible platform. Macquarie Capital Limited

Wendy Huang, CFA +852 3922 3378 3) Music slowing, labels’ dominance challenged - We recently lowered our [email protected] revenue growth forecasts for Vivendi’s UMG to 8%, capping EBITA margin

Macquarie Capital (USA) Inc. improvement to 15%. Global music majors seem to have missed the chance

Amy Yong +1 212 231 2624 to grow into the wider audio space and are losing their bargaining power with [email protected] distributors, in our view.

Benjamin Schachter +1 212 231 0644 [email protected] Winners and losers

Tim Nollen +1 212 231 0635 Streaming is not just another music technology cycle like CD or iTunes [email protected] downloads. We expect OTT distribution will completely disrupt the current

equilibrium and result into a winner-take-all scenario. The business model Macquarie Capital Securities (Japan) Limited

Damian Thong, CFA +81 3 3512 7877 innovation we describe is the greatest disruption in the last 100 years of audio [email protected] entertainment and Asian markets are leading with a new diagonal integration

model which is little understood by most of the incumbents. Macquarie Securities Korea Limited

Kwang Cho +82 2 3705 4953 Our top picks are companies that are consolidating a dominant position across [email protected] the entire audio entertainment spectrum, either globally or at the local level. We

Macquarie Capital Limited, Taiwan Securities Branch have high conviction Outperform ratings on TME, SPOT and SIRI; we have an

Steven Yang +886 2 2734 7528 underperform rating on VIV (see page 4 and 12-15 for details). [email protected]

Please refer to page 74 for important disclosures and analyst certification, or on our website www.macquarie.com/research/disclosures.

15 February 2019 15 Macquarie Research Fig 1 Global Music landscape – Content owners and major digital retailers/distributors (2018)

Universal Music (Vivendi) Warner Music $5,150m (including ~US$1.4bn at Sony Music Music revenues: $6,873m $3,576m Entertainment Japan)

% of group revenues 44% 9% 100% Music EBIT margin: 14.1% 16.0% 6.2%

Recorded Music: 69%, Music Publishing: 12%, Rec/pub revs split: 85%/ 15% 84%/ 16% Visual Media & Platform: 19% Global share of Recorded: 29.7% 21.9% 16.2% Global share of Publishing: 19.5% 27.3% 12.0% Vivendi (VIV FP) owns 100% (plan to sell Sony (6758 JP) owns 100% of SME and Private (Len Blavatnik's Access Ind. owns Ownership: up to 50% by Dec 2019) SMEJ, ~90% of EMI 100%) Macquarie rating: Underperform (VIV FP) Outperform (6758 JP) n/a

Record labels Record Macquarie analyst: Giasone Salati Damian Thong n/a

Spotify / iTunes Amazon Prime Music/ Fire YouTube Music (Alphabet) Music revenues: $4,669m $4,618m $1,132m $2,687m % of group revenues 100% 2% 0.5% 2% Geography: Global Global Global Global Free streaming users: 116m none none 1,300m Pay streaming subs: 96m 40m 40m users out of 100m+ Prime subs n/a Premium tier price: $9.99 ($14.99 Family pack) $9.99 ($14.99 Family pack) $99/ year (Amazon Prime) $9.99 ($11.99 for premium access) Other products: Podcasts, TV programme clips Downloads, Online radio, Music videos Free delivery, Prime Video, etc. User generated content, Music videos Ownership: Spotify (SPOT) owns 100% Apple (AAPL US) owns 100% Amazon (AMZN US) owns 100% Alphabet (GOOGL US) owns 100% Macquarie rating: Outperform (SPOT US) Outperform (AAPL US) Outperform (AMZN US) Outperform (GOOGL US)

Globaldist. Macquarie analyst: Giasone Salati, Amy Yong Ben Schachter Ben Schachter Ben Schachter

Sirius XM Music Entertainment Netease Cloud Music

Music revenues: $5,425m $2,751m ~$180m $400m % of group revenues 47% 100% ~1.8% 30.0% Geography: US, Canada Asia Mainland China Free streaming users none 800m+ ~120m (20m+ IDs only 1min free sample songs) Pay streaming subs 33m (satellite radio) 35m (music streaming and social combined) ~4m 4m $15.9/m streaming 0R $10.9-20.9/m sat $2.2/ month for music streaming; $5-75/month Premium tier price: $1.7/ month $10.99/ month radio for social entertainment Downloads, online karaoke, livestreaming, Other products: News, Sports channels Downloads, ads Paid downloads copyright sublicensing

Ownership: Public (J. Malone's Liberty Media ~70%) Tencent (700 HK) owns 58% Netease (NTES) owns majority KaKao Corp owns 100% Revolution Audio Global Macquarie rating: Outperform (SIRI US) Outperform (TME US) Outperform (NTES US) n.a. Localdist. Macquarie analyst: Amy Yong Wendy Huang Wendy Huang n.a. *NTES Cloud music revenue, pay streaming subs are on our estimates. TME revenue is our 2018 forecast, free users and streaming subs are based on company released data for Sep 2018. Source: Macquarie Research, Factset, Company data, February 2019

2

Macquarie Research Global Audio Revolution

Table of contents

Exec summary ...... 5 Redefining the audio entertainment market ...... 5 1) Audio entertainment by distribution technology ...... 6 2) Audio entertainment by content ...... 7 3) Audio entertainment by business model: freemium convergence ...... 8 The vertical integration dilemma ...... 9 Terms of trade ...... 10 China the new focus...... 10 Smart speakers and voice platforms ...... 12 Top picks from audio disruption ...... 12 Apple ...... 13 Amazon ...... 14 Sirius XM-Pandora ...... 14 Sony ...... 14 Spotify ...... 14 Tencent Music Entertainment ...... 15 Vivendi ...... 15 Redefining audio entertainment ...... 16 Technology convergence ...... 16 Audio like video content: music is only the tip of the iceberg ...... 17 Freemium convergence ...... 20 Where is the growth and who benefits from it ...... 20 Audiobooks ...... 20 Streamcasts, the new podcasts people actually want ...... 23 Recorded music ...... 26 Commercial radio ...... 27 Concerts and festivals ...... 27 Streaming revolution ...... 29 Streaming/podcasting as a replacement to radio/physical ...... 29 Higher streaming ARPU means greater customer satisfaction ...... 30 Free users also growing, as a result of audio entertainment convergence ...... 33 Free-to-pay conversion keeps increasing ...... 33 Younger demographics equally likely to pay ...... 34 Content-distribution equilibrium disrupted ...... 36 Record labels concentration has been an advantage until recently ...... 36 Music distribution still relatively fragmented ...... 37 Homogeneous pricing suggests labels control music distribution ...... 38 Spotify’s pricing strategy different from Netflix’s ...... 38 Vertical integration appeal/necessity ...... 40 Spotify direct distribution and licencing deals ...... 40 Tencent Musicians Plan ...... 41 Korea case study ...... 41 The buoyant Chinese market ...... 42 Major consolidation in China’s music distribution ...... 43 East to west disruption ...... 43 The booming audio streaming market ...... 44 Distribution: upstream expansion is the norm in China ...... 45 Western labels collaboration ...... 45 Smart speakers and voice platforms ...... 46 Sizing the smart home market ...... 46 Smart speakers ...... 46 Apple an expensive and closed system compared to / Amazon ...... 47 Music important in smart homes, nice to have in connected cars ...... 48 Our new global audio entertainment model ...... 50 Accelerating audio growth, but not because of music ...... 50 15 February 2019 3 Macquarie Research Global Audio Revolution

Audiobooks and podcasts growth contribution underestimated ...... 51 Slowdown in recorded music growth well documented ...... 51 Regulation ...... 53 Online platforms likely to face tighter scrutiny from safe-harbour reviews ...... 53 European Commission and US Copyright Office aims to foster a “fairer and more balanced” framework for content owners ...... 53 Global Audio – review of main listed companies ...... 54 ...... 55 Strategy ...... 55 Latest trends ...... 57 Catalysts ...... 57 Apple Music/ iTunes (Apple) ...... 58 Strategy ...... 58 Latest trends ...... 59 AAPL and GOOGL mobile OS growing their regulatory risk ...... 60 Catalysts ...... 60 Sirius XM ...... 61 Strategy ...... 61 Latest trends ...... 61 Catalysts ...... 62 Sony Music ...... 63 Strategy ...... 63 Spotify ...... 64 Strategy ...... 64 Latest trends ...... 64 Catalysts ...... 65 (Vivendi) ...... 66 Strategy ...... 66 Latest trends ...... 66 Catalysts ...... 67 ...... 68 Strategy ...... 68 Latest trends ...... 68 Catalysts ...... 69 YouTube Music (Alphabet) ...... 70 Strategy ...... 70 Latest trends ...... 71 Catalysts ...... 71 HIM International Music ...... 72 Financial and valuations ...... 72 Latest news and development ...... 72

15 February 2019 4 Macquarie Research Global Audio Revolution

Exec summary Let’s start by addressing the elephant in the room. We believe most of the analysis of recorded music trends is flawed. It simply fails to consider recorded music in the wider space of audio entertainment. Moreover, the received wisdom on recorded music mostly neglects advertising and live streaming as a form of monetisation: Spotify’s long-term guidance implies faster growth for advertising revenue compared to subscriptions and TME generated over 70% of its revenue from social entertainment. In this report we look at audio content in its entirety and focus on the convergence at play in the distribution channel (physical, radio broadcast, downloads, etc.) towards streaming platforms. The inevitable shift to a freemium model (ad-supported + pay subscriptions), we conclude, will upset the existing equilibrium between content owners and distributors. We expect good (and bad) vertical integration will follow, as the industry moves towards a Netflix-like model. Our top picks are global, integrated distributors: Spotify (emerging leader in podcasts) and Amazon (leader in audiobooks), though audio is a relatively small part group revenues for the latter. We also like diagonally integrated local champions like Tencent Music Entertainment in China and innovative incumbents with a broad content reach like Sirius XM-Pandora in the US. We are cautious on record labels (e.g. Vivendi and Sony) that broadcast radio.

Redefining the audio entertainment market Audio streaming is generally considered a replacement only for physical music sales and iTunes downloads. But a much deeper disruption is ongoing and it involves the whole audio entertainment space: the new technology is imposing a new paradigm, by reuniting music with non-music audio entertainment on the same platform. Terrestrial radio and CD purchases were historically separated only due to technical reasons: when broadcast radio became widespread in the 1940s, it was not commercially viable to charge directly the ultimate radio listener. At the time, advertising was the only way to monetise Hertzian waves and premium content (e.g. music “on demand”) was sold separately in the physical retail channel (e.g. CDs, and more recently downloads). This separation was artificial and favoured broadcast radio in terms of reach (Fig 2). With the right technology enabler (i.e. set-top-box), video content evolved in a much more unified way and pay TV has achieved as large a reach as radio, reuniting all video entertainment much earlier.

Fig 2 Weekly reach of audio among adults 18+ (2018)

Radio

Streaming audio on smartphone

Podcasts

Satellite radio

Streaming audio on tablet

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Source: Nielsen, Macquarie Research, February 2019

The consumption of audio entertainment is more personal/portable, and it requires a more advanced technology which only became widely available with the spreading of smartphones and 3G networks.

15 February 2019 5 Macquarie Research Global Audio Revolution

Only streaming technology combines the accessibility of free service with the sophistication of the pay subscription, which is great for the user experience. Behind the scenes, it also allows content owners the full flexibility to distribute different content with different models: ad-supported for cheaper content and pay subscription for premium content. A unified distribution platforms begs a unified content offer and possibly vertical integration. Sirius XM’s acquisition of Pandora (2019) and Amazon’s acquisition of Audible (2008) are two examples of this trend, from very different angles. Spotify’s ambition in podcasting ($400-500m envelop for acquisitions, including Gimlet and Anchor) points in the same direction. And it is not just strategic buyers, private equity firm KKR announced last June the acquisition of RBmedia, a large producer of audiobooks (35k titles) and spoken content. Let’s now look at the wider audio entertainment space from three different angles1: first by technology (broadcast radio, physical sales, static downloads and live streaming). Second we will look at the breakdown by content (music, audiobook, podcasts, radio shows, etc.). Finally, we analyse the different business models across technology and content (ad-supported/free vs consumer direct spend in the form of download/purchase and streaming subscription). 1) Audio entertainment by distribution technology To these days, broadcast radio (advertising revenues) still accounts for over 40% of the total audio entertainment revenues in the US (Fig 3). Its importance is decreasing as audience fragments and shifts to more convenient on-demand services, a very similar trend to what we observed in video entertainment over the past few decades.

Fig 3 US audio entertainment market by technology – fading Fig 4 Broadcast radio advertising in structural decline radio’s leadership as audience moves to streaming (global revenues, USD m)

Physical (Music) 40,000 20% 17% 30,000 15%

20,000 10% Radio Downloads 43% (Music, 10,000 5% Audiobooks) 13% 0 0%

-10,000 -5%

-20,000 -10%

2008 2015 2000 2001 2002 2003 2004 2005 2006 2007 2009 2010 2011 2012 2013 2014 2016 2017 2018E 2019E 2020E 2021E 2022E Streaming 1999 (Music, Audiobooks, Streamcasts) 27% Radio Radio as % of global Radio growth yoy Source: American Publishing Association, Audio Publishers Association, Interactive Advertising Bureau, IFPI, Magna, Macquarie Research, February 2019 Source: Magna, Macquarie Research, February 2019

Terrestrial radio’s reliance on local advertising makes it somewhat more exposed than TV to the rise of internet advertising. While TV offers primarily mass audience which is difficult to replicate online, radio offers more targeted advertising, which is in more direct competition with online advertising. The result is broadcast radio revenues are declining in absolute terms and losing ground compared to other media (Fig 4). More broadly in the entertainment industry, we observe a shift in the way content is monetised, from advertising-supported, to direct consumer spend. So radio is losing share of advertising and in turns advertising is losing share of entertainment money. The combination of these two trends make for a difficult and long transition. Terrestrial radio is losing not just revenues to streaming platforms, but also talent: radio stars are increasingly enrolled as podcasters. In fact, the problem is not with radio content per se. We have already seen the likes of Apple hiring music radio editors to create radio-like channels on Apple Music. Following the acquisition of Pandora, Sirius XM also recently confirmed that it owns the right to distribute its “radio content” via streaming platforms (sport being the main exception for now).

1 At this point we primarily focus on data for the US market that are often not representative of trends in other geographies. We will however deep dive in the very interesting Chinese market further down in this report and highlight specific trends then. 15 February 2019 6 Macquarie Research Global Audio Revolution

Alongside the decline of both radio and audio physical sales, we observe the obvious rise in streaming (music plus podcasts). Downloads for us is less clear cut as we include not only music (iTunes downloads are plummeting), but also fast-growing audiobooks. The distinction between downloading and streaming content is becoming increasingly blurred, as most streamed content can be downloaded and vice-versa, depending on the connection speed.

2) Audio entertainment by content So many are the analogies between music and video that it is useful to talk about content by comparison. Starting from the premium end of content: recorded music is to audio, what films are to video entertainment. Looking at the mid-tier content: audiobooks should be seen as the equivalent to TV series (e.g. scripted), which incidentally were the major focus for new OTT entrants such as Netflix and Amazon. Finally, at the “cheaper” end of the range, commercial radio and podcasts are the equivalent to ad-supported free-to-air programming (e.g. non-scripted TV).

Fig 5 Audio/video content comparison Video Audio

Premium Featured movies Recorded music Mid-range TV series Audiobooks Basic Broadcast TV shows Broadcast Radio and Podcasts Source: Macquarie Research, February 2019

Taking our new definition of the audio entertainment industry further, we aggregate data from different sources to present the actual breakdown by content (Fig 6). We find that premium music revenues, including both recorded music and broadcast radio music, account for 56% of the total. Audiobooks, the equivalent of “quality” scripted content, account for only 11% of the total. Finally, “basic” or “non/scripted” audio entertainment actually represents the largest slice of the pie, as it was the case for video entertainment at the early stage of the transition to cable/satellite TV.

Fig 6 US audio entertainment market by content – Music (premium audio content) represents the largest slice, but faster growth in audiobooks (mid-range)

Basic (radio shows, podcasts) 34%

Premium (Music) 56%

Mid-rang (e.g. Audiobook) 10%

Source: American Publishing Association, Audio Publishers Association, Interactive Advertising Bureau, IFPI, Magna, Macquarie Research, February 2019

When we compare content segments using our holistic definition of audio entertainment, we find that premium/music growth has peaked and will remain stable over the next few years. This is not because we expect music streaming to grind to a halt, rather, we see streaming of other forms of audio entertainment as gaining share. OTT distributors are expanding into the wider audio space where it is easier to develop proprietary original content without entering in direct conflict with the music majors. Spotify now has over 185k podcasts titles, nearly on par with industry leader Apple and it recently acquired Gimlet and Anchor as part of a total $400-500m investment envelop in proprietary content. Last October, Spotify also launched Spotify for Podcasters, a platform for content creators to upload their content directly (10,000+ podcasters are using the portal every month). In February 2019, Spotify launched Spotlight, a new format adding visual elements to podcasts about news, politics and entertainment and we estimate the company is investing tens of millions a year to develop proprietary radio-like content.

15 February 2019 7 Macquarie Research Global Audio Revolution

Amazon was first-mover in audiobooks and it is now the unrivalled leader in this segment via Audible that offers the largest catalogue on a subscription basis. Since February 2019, Google also started offering audiobooks with an a-la-carte model. However, Google doesn’t have as clear brand/proposition as Amazon does, and it seems more tailored to the occasional audiobook fan, while Audible’s subscription model is more cost-effective for regular users. Music majors demonstrated little interest for alternative sources of audio entertainment, at risk of being marginalised, we argue. Vivendi recently acquired French consumer book publisher Editis, but it seems more inclined to consider synergies with its pay TV arm Canal+ (film adaptation) than with UMG (audio content consolidation).

3) Audio entertainment by business model: freemium convergence A loose definition of the freemium model involves free distribution of the basic version of a product with an implicit/explicit advertising component, and a more compelling pay version of the same product (typically a subscription) whose revenues should eventually cover the free version. Freemium has been dubbed as the marketing cost of start-ups, but it has grown to be a practice also used by more established companies. Until a decade or so ago, free audio entertainment (i.e. “basic” content distributed via broadcast radio) was rigidly segmented from pay audio (i.e. “premium” content such as CDs), and the usability of each was dramatically different (live vs on-demand, etc). On streaming platforms, users can access free and pay content which means the audio entertainment experience is now much richer: subscribers can switch from new release hits, to podcasts of radio content or even access audiobooks. This is a sort of Copernican revolution that consumers have only just started to take advantage of. Re-aggregating the data by monetisation type we can see subscription revenues are currently only less than 20% of the total, compared to advertising at nearly 50% and on-demand at about a third (including one-off purchases). As we have seen on video entertainment, subscription is the main area of growth across the whole industry – notice Spotify expects to grow faster on advertising that that is more than offset by commercial radio decline at industry level.

Fig 7 2017 US entertainment sales by business model – subscription to drive overall industry growth as a more efficient way to price-discriminate

On demand (Audiobook, downloads and physical) 34%

Advertising (Music and radio) 49%

Subscription (Music and streamcasts) 17%

Source: Bloomberg, PwC, Macquarie Research, February 2019

Over time, the overall industry will grow in size as subscription allows for much greater price segmentation than any other business model. Advertising and on-demand models typically charge the same “price” to all potential viewers/ buyers, resulting in under/over-charging. As we discussed in our previous reports on video entertainment, the subscription model allows to discover and match more precisely each user’s reserve price for specific combinations of content. Still, the shift from radio and single purchases to subscription bundles will completely disrupt the industry equilibrium. Radio stations are particularly exposed, unless they fold into streaming services. Music majors need to grow expand into non-music audio content to maintain the bargaining power with more powerful distributors. Across the spectrum, OTT distributors appear as winners, especially is they combine free and pay models.

15 February 2019 8 Macquarie Research Global Audio Revolution

Among the global streaming players, Spotify is the only player whose platform is set up to attract and monetise both free and pay users. Amazon and Apple use music mostly as a loss leader to retain traffic on their platform and to sell hardware (e.g. iPhones). Google’s pay service doesn’t seem to have ever been marketed with conviction and it remains a very small player in the industry. Among local players, Wendy Huang (Macquarie media anlyst in China) recently launched coverage of TME describing it as the equivalent of Spotify + Facebook + YouTube for China. The platform incorporates much more than a music service and can capitalise on multiple revenue streams, including both consumers direct spend and advertising. Users’ engagement on TME is the highest in the industry (70min a day per user vs 50min on Spotify): users use the platform for karaoke, to share short videos or to stream live music performances. We also argue TME will find it easy to add other types of audio entertainment given its quasi-monopolistic position. Moving to the US, Sirius XM also seems well positioned. Before the acquisition of Pandora the company collected only a meagre 3% of total revenues from advertising. But once the deal closes, advertising revenues should increase to a significant 29% of group revs and management presented a clear vision of combining the pay and free model to maximise reach and improve monetisation.

The vertical integration dilemma There is no proof that a vertically integrated business actually functions any better or worse. But, as soon player in the industry moves in that direction and creates a bottleneck in the value chain, everybody else has to follow to avoid unfair treatment either upstream (access to content) or downstream (access to distribution). The problem of the integrated model is that if one of the links fails the risk of contagion across the whole company is significant. So if production fails to deliver content hits, the distribution channel will typically be unfairly penalised. Similarly, if the distribution platform offers a poor user experience relative to peers, the content unit will reach a smaller audience as a result. At least this is what economic theory says. In practice though, for a pure player, the risk to face a higher concentrated supplier/customer segment compared to an integrated player is too high. So, for example, one could say that Disney had such a strong position in content production that it would not need to be a distributor. Indeed it resisted that temptation when Comcast bought NBC/Universal (2009-10). But things changed more recently and even Disney now has a fully flexed its direct-to-consumer strategy. Netflix is a fine example of how powerful vertical integration can be in terms of differentiation and in terms of bargaining power. The company has been extremely successful at both ends of the spectrum (content/distribution), and it is now forcing most of its competitors to follow the same model to compete effectively. In music there are some signs of a trend towards vertical integration with Spotify Direct servicing unsigned artists. Development of more original content away from music is also high on the agenda for the same Spotify (investments in owned and produced podcasts, acquisition of Gimlet and Anchor) plus Apple Music (mostly via curated channels) and Amazon (e.g. Audible, Prime Reading).

A potential tie-up between Vivendi’s UMG and Tencent Music Entertainment (TME buying up to a 50% stake in UMG), as widely discussed in the press, would be another example of vertical integration, but the geographical mismatch would limit the operational synergies and increase potential dis-synergies. But the most interesting type of integration is of a totally new breed: we call it diagonal integration. TME is breaking new ground with its wider strategy incorporating a number of adjacencies (live- streaming, karaoke, file sharing, etc.) alongside pure music listening. Amazon has seamlessly integrated its basic entertainment offer within the most successful e-commerce platform. And it’s built the leading innovation in smart speaker with its proprietary voice ecosystem. This type of integration is relatively common in Asian markets (e.g. Alibaba), but it is completely new to most players in the US and Europe. For this reason we believe the disruption will be even greater than in previous audio technology cycles (radio, audiocassette, CD, downloads).

15 February 2019 9 Macquarie Research Global Audio Revolution

Terms of trade A change in business model (the combination of ad-supported and subscription payment on the same platform), the reunification of audio content (premium music, audiobooks and radio-like streamcasts) and a general trend towards increased vertical integration, all together challenge the status quo for audio incumbents and force new entrants to adopt a fluid strategy. Streaming revolutionise distribution, which is where most of the disruption has already been observed. High street distribution (e.g. CDs) has already been decimated by the shift to digital and by competition from online platforms. Radio advertising revenues are already under pressure and risk seeing talent cost inflation alongside to accelerated audience losses. Music majors re-gained control of retail pricing: from Spotify to Apple, to Amazon Music Unlimited, all platforms are quite comparable in terms of content on offer and price ($9.99/m before promotions). But this market structure reminds us of the antitrust case brought against book publishers in 2013 (e-book price fixing). In any case, we argue negotiations with distributors are getting tougher. Indeed an investigation into music streaming price-fixing was launched in both the US and Europe in 2015. Music majors, radio broadcasting producers, and books publishers were also used to operate in relatively well-defined and highly concentrated (read not competitive) segment. Now they all need to adopt a more holistic content strategy, if they want to defend their competitive advantage. We argue demand for a wider variety of content will allow dominant distribution platforms to play content providers against each other and squeeze better terms over time. One clear example is Spotify where we expect the next round of negotiations with the majors to allow for margin improvement. It has already been the case with the last agreement which lead to a decline in the minimum guaranteed payment set for free users. OTT distributors, global or local, are the only clear winners. Spotify and Amazon have already proven to be the most innovative, embracing a holistic audio content strategy (e.g. podcasts and audiobooks, respectively). The combination of Sirius XM-Pandora has great potential and TME already achieved 20%+ operating margins, while still growing top line nearly 40% (CAGR 2018- 22E).

In spite of their significant size in audio entertainment, we consider Apple and Google marginal players in audio entertainment in the sense that both companies seem to treat audio as a loss leader to improve their ecosystem. For example, Apple was first-mover in podcasting, but never set up a system to monetise them directly.

China the new focus Much of this report focuses on US data because is generally available and because it offers valuable insight in the global market. China is however way more important in terms of growth potential and it is also the region where business models appear to evolve more rapidly. In this report we lean heavily on the expertise of Wendy Huang, Macquarie media analyst based in Hong Kong. China has historically been afflicted by a very high level of piracy, but things changed in 2001 with a historical licencing deal between the global majors (via One-Stop-China subsidiary) and , the largest search engine that until then had made free music available in a very convenient way. Piracy though still remains an issue accounting for 16% of total music listening time, over the global average of 7% (Fig 8-9).

15 February 2019 10 Macquarie Research Global Audio Revolution

Fig 8 China breakdown of music listening time still shows Fig 9 Global breakdown of music listening time shows 16% from piracy (2018) “only” 7% from piracy, but interactive not significant (2018)

Broadcast Other radio Other 14% 5% 13% Interactive 11% Broadcast Pirated radio 7% 29%

Pirated 16% Purchase 16% Audio streaming 14%

Free audio streaming Youku Purchase 16% 9% Video 20% Paid audio Other video streaming streaming 6% 17% 7%

Source: IFPI, Macquarie Research, February 2019 Source: IFPI, Macquarie Research, February 2019

The comparison of music listening time breakdown in China with the global average also highlights another interesting factor: the importance of interactive listening in China. That comes in the form of karaoke, sharing, etc. and contributes to much stronger user engagement with the platform, which bodes well for future migration of free users to pay subscriptions (TME currently has paying ratio of 4% vs Spotify’s 45%) China currently represents only ~2% of the global recorded music market (Fig 10), but we expect it will more than double in size over the next few years, moving from #10 to the top-five in global rankings. Beyond recorded music, China is as vibrant a market as the US when it comes to streamcasts and audiobooks.

Fig 10 China ranked among top 10 music markets by Fig 11 TME controls 70%+ of the streaming distribution revenue ($bn, 2017) market, following four big acquisitions

Xiami Migu Music, Qingting FM, MIUI Music (Alibaba), Lizhi, 1.3% 1.0% 7.0 (Xiaomi), 2.3% 2.1% 1.5% iMusic, 0.6% Changba, 6.0 2.5% Baidu Music Ximalaya, (Baidu), 0.4% 5.0 6.2% NetEase Kugou 4.0 Cloud Music, (Tencent), 10.3% 32.4% 3.0 Kuwo Tencent Radio (Tencent), 2.0 (Tencent), 11.8% QQ Music 0.8% WeSing (Tencent), 1.0 (Tencent), 25.6% 14.0% 0.0

Source: IFPI, Macquarie Research, February 2019 Source: Questmobile, Macquarie Research, February 2019

The Chinese music market has different competitive dynamics compared to Western countries. Streaming distribution is more highly concentrated: Tencent Music Entertainment (TME) controls 70%+ of the market (Fig 11), following a number of recent acquisitions in the wider music space, including karaoke platform WeSing, #3 by MAUs. Over the past few years, the three global majors all signed exclusive deals with TME, which effectively controls sub-licencing to other player in China. The government intervened to assure that Tencent “exclusive” content deals are made available to rival distributors (e.g. NetEase), but the devil is in the details.

15 February 2019 11 Macquarie Research Global Audio Revolution

The IPO of TME in the US provided greater disclosure and attracted more interest to this very interesting market. The Chinese market has only recently started the conversion of free users to pay subscribers, but it already has a more sophisticated pricing structure (e.g. charges for downloads) than elsewhere.

Smart speakers and voice platforms Alongside with convergence of audio entertainment on mobile devices, we can’t underestimate the role of smart speakers in conjunction with voice platform. At home, like on the road, consumers don’t necessarily need to proactively plan for audio entertainment. They can just shout, literally, “Alexa play read me a book”, or “Hey Google, play some music from Spotify”. Amazon was first in this segment and it currently has the largest share in the global market (32%), but Google is catching up relatively quickly, having already achieved a 23% share. Three local Chinese players occupy the next three spots reminding us of the importance of the Chinese market. Apple is only #6 globally (Fig 12), having pursued a more closed and premium strategy.

Fig 12 Global smart speaker shipment market share in 3Q18 Fig 13 US smart homes as a percent of total US households

30.0% 28.0% 25.5% 25.0% Others 22.5% 14% Apple 19.6% 5% Amazon 20.0% 32% 16.3% Xiaomi 15.0% 8% 12.5%

Baidu 8% 10.0% Alibaba Google 10% 23% 5.0%

0.0% 2016 2017 2018 2019 2020 2021

Source: Strategy Analytics, Macquarie Research, February 2019 Source: S&P Market Intelligence, Macquarie Research, February 2019

Again, using the US as a reference point, smart home penetration reached ~20% in 2018 and is expected to increase to nearly 30% by 2021. Smart speakers dominate the market for connected homes: in the UK almost 40% of homes first get connected via a smart entertainment device.

Top picks from audio disruption OTT distribution has had a deep impact on video entertainment, and we argue it will be even more disruptive in the audio entertainment industry. Bringing together different forms of audio content – from “premium” music to audiobooks, to “basic” podcasts - on the same platform and merging different business models - subscriptions and advertising revenues - is a much deeper structural change.

The audio industry is disrupted by new entrants and threatened by technology companies that want to use audio entertainment as a loss-leader to sell more hardware (Apple offers all podcasts for free) or to generate more traffic (e.g. Google’s YouTube). To be successful, it is paramount to reach scale, either in content or in distribution. For content owners it should be relatively easy to complete their offer acquiring and developing new exclusive copyright. Music majors for example should be proactively acquiring and creating audio IP like audiobooks (online literature is booming in China) and other forms of premium content. We are concerned though that this might happen only at a later stage via expensive acquisitions as labels appear focused primarily on growing recorded music revenues. For existing distributors, the challenge is much greater as witnessed by the bankruptcy of a number of high street music retailers (e.g. HMV bankruptcy in 2013 and 2018). Some of the new entrants appear cornered into just one of the revenue model, like Apple (premium subscription) or Google (chiefly ad-supported). But audio entertainment is not the main source of profit for them.

15 February 2019 12 Macquarie Research Global Audio Revolution

The best place to benefit from audio disruption is in the middle. Spotify rightly embraced both free and pay and expect its share of revenue coming from advertising to double over the next few years to 20% of total. Sirius XM recently completed the acquisition of Pandora, bringing together different technologies and revenue models. We also like Amazon and Tencent for their more holistic business models. Amazon’s leadership in voice platforms and in audiobooks means its role in audio entertainment can only grow. After a number of important acquisitions, TME boasts strong growth and high margins, testimony of its strong competitive position in China.

Fig 14 Global music players valuation metrics Current Target Upside/ % of music Music Market Cap EV/ EBITDA Company Analyst Ticker Rating Currency price Price downside revenues Top Pick ($m) 18E P/E 18E

Alphabet Benjamin Schachter GOOGL US Outperform USD 1,129 1,150 1.9% 2% 781,763 14.1x 24.4x Amazon.com Benjamin Schachter AMZN US Outperform USD 1,640 1,850 12.8% 1% 805,573 18.7x 63.0x Apple Benjamin Schachter AAPL US Neutral USD 170 149 -12.4% 1% 802,446 9.8x 13.9x n.a. n.a. n.a. n.a. n.a. n.a. n.a. 100% 1,160^ n.a. n.a. KaKao Corp n.a. 035720 KR n.a. KRW n.a. n.a. n.a. ~30% x n.a. n.a. NetEase Inc Wendy Huang NTES US Outperform USD 240 305 27.1% 47% y 31,435 14.5x 18.2x Sirius XM Radio Amy Yong SIRI US Outperform USD 6 7 16.7% 47% y 27,952 13.2x 25.2x Sony Damian Thong 6758 JP Neutral JPY 5,038 5,800 15.1% 9% 57,334 5.7x 14.5x Spotify Technology Any Yong, SPOT US Outperform USD 127 170 33.9% 100% Y 25,960 nmf nmf Giasone Salati Tencent Music Ent. Wendy Huang TME US Outperform USD 16.0 20.0 25.0% 5% Y 25,955 30.6x 34.6x Vivendi Giasone Salati VIV FP Underperform EUR 22.0 18.0 -18.2% 43% 32,606 10.0x 18.3x Warner Music Group n.a. n.a. n.a. n.a. n.a. n.a. n.a. 100% 3,300 n.a. n.a. Himalaya n.a. n.a. n.a. n.a. n.a. n.a. n.a. x 3,400* n.a. n.a. Notes: ^ Deezer EUR1bn valuation is based on latest news in 2018 (Link) ~Warner $3.3bn valuation is based on Access Industries acquisition in 2011 *Himalaya $3bn valuation is based on latest news on fund raising (Link) Source: Company Data, Macquarie Research, February 2019

We are more cautious on the music majors (Vivendi’s UMG and Sony) as the new focus on non-music audio entertainment means their bargaining power is decreasing quickly. Apple seems to have lost its innovation roots and it is playing catch up on connected speakers and voice platforms. Google will remain an important partner for music promotion, but its focus on harvesting consumer data makes it a less relevant player in audio entertainment.

Apple Apple continues to play a leading role in music and it is a close competitor to Spotify, and it is leader in podcasts (i.e. radio-like audio content) but its offer is somewhat less sophisticated than peers. In comparison to Spotify, Apple Music doesn’t have a free tier, nor an advertising strategy at this point even though we notice press reports of a potential acquisition of iHeart Radio, which may address that. In comparison to Amazon, Apple lacks integration of its audio offer with a wider ecosystem and it seems to have fallen behind in terms of voice platform Vs both Google and Amazon.

Last Apple Music reported 40m paying subscribers, and 8m more on their three-month free trial. The total increased by 2m to 50m users (paying and free trails combined) in May. There hasn’t been any official figures since, but AAPL did give a decent amount of details in prior years, exiting FY16 at ~17mm paying subs, and FY17 at ~30mm. For 2019 we expect a solid c40% growth in Apple Music revenues, though significantly slower than 60% in 2018E growth and we notice that music streaming margins are lower than the majority of the Services line, limiting its profit contribution. We recently downgraded AAPL to a Neutral rating and $149 TP. Our focus remains on Services, and while AAPL noted that Services revs of $10.8 billion in Dec (up 27.5% y/y) showed strength in every geo, we are quite concerned that Services growth is going to slow meaningfully beginning in the March qtr. There are simply too many concerns (beyond the pre-announcement) to ignore. Tough Services comps are coming and we have significant concerns about the sustainability of growth for the key Services drivers. Specifically, we think the top three Services drivers (Licensing, App Store, and Apple Care) are all likely to slow in FY’19 and the faster growing Services business (Music, iCloud, and Apple Pay) are not big enough to offset the slowdown (and Music margins are relatively low).

15 February 2019 13 Macquarie Research Global Audio Revolution

Amazon Music is a relatively small part of Amazon’s revenues, but the company is a big player in music: Amazon Music + Amazon Music Unlimited is approaching the range of Apple Music and to a lesser extent Spotify in terms of audience reach. We also argue Amazon has the most sophisticated ecosystem and pricing model. Subscriptions to AMU growth can be attributed to two key factors: Prime and Echo, both of which receive discounts from the regular $9.99/mo price. Prime members pay $7.99/mo under the Individual plan (and $14.99/mo for up to six members under the Family plan). Echo-only plans are $3.99/mo. Notably, Amazon taps new demographic of streamers through smart speakers – More than half of Amazon Music users have utilized Echo’s voice-first functionality. Amazon believes it is expanding the premium streaming market segment by leveraging the popularity of smart- and voice-activated speakers. Most of the new paid subscribers are coming from users of home devices and include a new, untapped demographic of streamers that are not necessarily mobile-first users. This new group of streamers (AMZN referenced the older generation) includes those slower to adopt listening to music on a smartphone, but appear to prefer the convenience of listening on a voice-activated smart speaker at home.

Sirius XM-Pandora In the US, SIRI-P also represents the richest offer in terms of audio entertainment, though it still misses audio books (Sirius discontinued audio books in July 2013: Sirius XM Book Radio, formerly known as Sonic Theatre, was a channel on the Sirius XM Radio network that specialized in playing Audio Books and Radio Dramas). We also see tailwinds from ongoing investments in video, telematics, and 360-L rollouts. Sirius XM is striking partnerships with Amazon to embrace the growth in smart speakers, among others (58m smart speakers in the US reaching 32% of households). This is all part of its Bring Us Home campaign. As management proceeds with the integration of Pandora, we expect a better articulation of strategy and top-/bottom-line synergies. Already, we see cross-promotional efforts, with Sirius XM advertising on Pandora.

Sony Sony views its various Music businesses as a key part of its portfolio. Excluding the Fate/Grand Order mobile game business (which sits in the Aniplex subsidiary of Sony Music Entertainment Japan), the Music businesses are set to account for 10% of Sony’s operating profit in FY3/20. Sony is a world market leader in recorded music, and with the acquisition of the remaining ownership stake in EMI for US$2.3bn, the company is also a leader in music publishing. We expect a 20% revenue CAGR in music streaming between FY3/18 and FY3/23 (to ¥500bn in FY3/23), offset by continuing declines in digital downloads and physical sales. Sony’s high exposure to the Japanese market, where physical sales remain overwhelmingly dominant, is a factor that moderates growth in recorded music. We estimate overall Music segment revenue and OP growth at 5% p.a. and 0% p.a. in FY3/18-23, largely due to our assumption of an eventual decline in the >¥40bn OP contribution from Fate/Grand Order. For Sony overall, we see stronger growth momentum in image sensors, and a relatively stable long-term profit trend in the Game & Network Services (PlayStation business).

Spotify SPOT shares currently trade at 2x P/S, nearly 2.5x turn lower than its peak in July ’18. Growth remains anything but slow. Holiday promos and market share continue, giving us confidence in its 2019 outlook. We model total/paid MAUs of ~204m/~95m paid, driven by holiday campaigns (Google, , and Hulu/Showtime) coupled with improving churn of 4.7% and healthy conversion rate. We expect ’19 will end at ~250m and ~119m, respectively, or growth of 22%/25%. We model ARPU of €4.72, relatively stable vs 3Q, as the impact of family/student plans to mix moderate.

15 February 2019 14 Macquarie Research Global Audio Revolution

Though competition is high in the US with Apple/Amazon, global markets remain an area of opportunity. We believe Spotify is gaining solid traction in markets where Android vs iPhones dominate (LatAm, Southeast Asia). In fact, Brazil/Mexico are among its top-ten markets. Secondly, Spotify recently expanded into the Middle East and North Africa in mid-November. Next is Russia and India, fragmented markets but big enough to support multiple players. Though premium subs remain the sweet spot, we believe advertising revenue is poised to see a ~24% three-year CAGR. Much of the GDPR-related issues are behind it and its self-serve platform is now launched in the US, UK, Canada, Australia, among others. Lastly, margins are set for structural improvement as label deals continue, artist relationships strengthen and the company expands into adjacent business. In Vivendi CEO Grainge’s holiday letter, he highlighted data and insights as tools to enhance performance. Spotify is key to supplying these, which could prove helpful during negotiations. However, NT accounting rules could veil progress around margins.

Tencent Music Entertainment TME was the first online music platform to turn profitable, back in 2016. Its sticky and massive user base and wealth of music IP have helped it to build a monopoly in the fast-growing Chinese music market. Commanding the top-four music apps in China, TME serves 800m unique monthly active users. Compared to global peers, TME relies less on any label companies and enjoys better economics via its diversified content source. Backed by Asia’s largest social networking company, Tencent, TME has been embracing live- streaming and social entertainment. In 3Q18, social entertainment contributed 71% of its total revenue and generated Rmb119 monthly ARPU, much higher than the Rmb8.5 ARPU for competing online music subscription services. We also see synergies between TME and Tencent’s video, literature and social businesses. This not only helps enrich its content library and increase user stickiness but also helps margins. Notably, TME enjoys superior margins compared to global online music peer Spotify or traditional music labels. In 3Q18, TME achieved 22% adjusted operating margin and we expect it to be maintained at or above 20% in the long term despite the company’s heavy investment in content.

In the long run, we also anticipate its paying ratio go up from its current 4% to 15% in five years’ time, albeit it is a low level compared to Spotify’s 46%. We also view TME as vertically integrated as it continues to invest in content production and expand its reach from online karaoke to podcasts. We recently initiated on TME with an Outperform rating and view it as the best content play in China.

Vivendi We recently downgraded Vivendi to Underperform due to increased capital allocation risk and operational disappointments at Canal+. The slowdown in recorded music growth also contributed to cap our valuation for UMG at €16bn, well below consensus. Universal Music Group (UMG) accounts for 43% of group revenue and 66% of profit (51% of our Sum of the Parts valuation). Vivendi intends to sell up to 50% of UMG, for cash, before the end of 2019 with proceeds to be used for a buy back. At €20bn consensus valuation, UMG would be valued at 3x sales FY19E, a hefty premium compared to Spotify’s 2.2x; or 19.1x EBITDA FY19E, more than twice as much as Disney the most successful content company. Our €16bn valuation for UMG still implies 15.7x EBITDA FY19E, which seems generous given Spotify recent multiple compression and TME’s mixed debut. Alibaba, Tencent, Google, Amazon and Apple have all been mentioned in the press as potential buyers of a stake in UMG, but it will be difficult for Vivendi to meet consensus and crystallise a high valuation without selling control of UMG.

We don’t see much downside to Vivendi (17% per 2/13 price on our TP, including dividend), but equally we don’t see a reason to hold the stock now. On the contrary, we highlight the risk of further cuts to earnings in February when Canal+ will announce a new wave of investments for new set-top-box and international expansion. We would turn more positive on Vivendi and UMG if the strategy was more clearly refocused on non-music audio content to secure its dominant position in the new converged audio entertainment industry.

15 February 2019 15 Macquarie Research Global Audio Revolution

Redefining audio entertainment Considering the developments of recorded music in isolation is a mistake. Streaming is collapsing together business models and practices that made the audio entertainment ecosystem unique for nearly a century. The segmentation of radio as a tool for promotion, from physical sales as a tool for monetisation is collapsing, in our view. In the wider context, recorded music (CDs, downloads and streaming) is a relatively small part of the whole market, accounting for less than a quarter (Fig 15). Advertising revenues from radio are actually twice as large and concerts ticketing has grown to over a quarter over the past two decades.

Fig 15 2017 US entertainment sales by category

Audiobook Streamcasts 8% 1%

Recorded music 20% Concert ticket sales 27%

Radio advertising sales 44%

Source: Bloomberg, PwC, Macquarie Research, February 2019

The recovery in recorded music has been widely publicised, but the fastest growing segments in audio entertainment are actually audiobooks and podcasts. Across the value chain, both content owners and distributors need to expand their reach across these relatively new areas, or risk being marginalised.

Technology convergence Decades ago, the introduction of the set-top box reunited most video entertainment under the same umbrella, adding great ease of use. The result was decades of sustained above-GDP growth for the whole pay TV ecosystem (content and distribution), as it captured a greater share of consumer wallets via subscription (direct spend) and advertising revenues (free TV). Audio entertainment is more personal and typically consumed on the go, which means it took more time for technology to evolve. Now streaming, think of it as over-the-top (OTT) distribution, meaning pay/free revenue models are converging here too. The gap between ad-supported distribution (e.g. radio) and direct consumer spend (e.g. CDs) disappears.

We re-aggregate the data presented above by technology in Fig 16, as a better way to understand the secular shifts. Our thesis is that audio entertainment is naturally shifting towards the delivery method that consumers find the easiest to use, streaming. The latter is only 22% of the whole, including podcasts, but we expect it will continue to gain market share from radio, physical and downloads over time.

15 February 2019 16 Macquarie Research Global Audio Revolution

Fig 16 Streaming will grow to replace not just CD/downloads, but radio too!

Physical (Music) 17%

Radio Downloads (Music, 44% Audiobooks) 13%

Streaming (Music, Audiobooks, Streamcasts) 26%

Source: American Publishing Association, Audio Publishers Association, Interactive Advertising Bureau, IFPI, Magna, Macquarie Research, February 2019

If the history of video entertainment is anything to go by, we can expect streaming to continue gain share. Indeed, pay-TV and streaming currently account for the largest share of US video entertainment industry, excluding cinema tickets. In this representation we have excluded concerts and live performances as a segment where disruption will be minimal in the next decade or so. However, there is evidence that live stadium attendance at sport events has been heavily impacted by the video entertainment convergence. First, wider reach contributed to attract more fans to a given sport and more viewers to the live event. In this phase ticket prices increased quite substantially as a result of higher demand and relatively inelastic supply. At a later stage though, video broadcasting/streaming can enter in competition with live concert attendance.

Another point to make is that we have not included physical book sales alongside CD and vinyl, a somewhat arbitrary choice dictated by the desire to keep limit the reach of this report. Later in this chapter we discuss briefly the dynamics of physical books, e-books and audiobooks sales for the publishing industry.

Audio like video content: music is only the tip of the iceberg Audio entertainment is more varied than one would think and extends well beyond music. Older generations tend to listen to radio shows, including news, drama, comedy, documentaries etc. Younger generations are increasingly discovering the power of podcasts and audiobooks alongside music, during their travel time, exercise, etc. As we did above in the breakdown by distribution platform, we re-aggregate data from different sources to present a breakdown of audio revenues by type of content. We used radio share of listening time to attribute revenues and kept podcasts together to highlight it a sort of new genre, even if they could be broken down in the same categories as radio. Again, we exclude live concerts and print books. It turns out that music is approximately half of total audio revenues (Fig 17).

15 February 2019 17 Macquarie Research Global Audio Revolution

Fig 17 Including radio, Music is the largest type of content

Audiobook 11% 2%

Sports 9%

Recorded News, music (incl. weather etc. radio) 21% 57%

Source: American Publishing Association, Audio Publishers Association, Interactive Advertising Bureau, IFPI, Magna, Macquarie Research, February 2019

At this point it is useful to present audio content in comparison to video content. We see recorded music as the equivalent of films. It is the most valuable audio content and the one consumers are happy to listen to over and over again. Audiobooks come shortly after and are comparable to high-quality TV series so successful on Netflix. This segment is possibly the least considered and the one with the greatest potential. Our grandparents used to listen proper theatrical interpretations of books divided in many episodes and broadcast live over terrestrial radio. Audiobooks are the same and they are surging in popularity (see further down in this report). Sport would fit between music and audiobooks for its appeal to advertisers and its potential to attract subscribers. Streamcasts are the audio equivalent of TV shows, or non-scripted content. This is a cheaper and more ephemeral type of content, typically produced on a local level. Still, terrestrial TV and radio stations can attract large, if declining audiences that are best monetised via advertising. Lastly, there are some “basic” public service functions like news that may be quite costly to produce, but are difficult to monetise directly. Advertisers don’t like the idea of having their message associated with controversial political comments or, even worse, next to war scenes. There is however a brand benefit deriving from news programming and potential for upselling as consumers often carry on listening/watching the same channel after the news. Broadly speaking, we believe it is useful to visualise content as an inverted pyramid (Fig 18) with music/films at the top (premium) and audiobooks/TV-series just below followed by streamcasts/TV- shows and news at the bottom (commoditised end of the spectrum). The pyramid is upside down to represent that premium content overall dominates both in terms of listening/watching time and in terms of potential for monetisation.

Fig 18 Video/Audio content comparison Video Audio Comment

Premium Film Music ↕ TV series Audiobooks “scripted” ↕ TV shows Streamcasts “non-scripted” Basic News News Source: Macquarie Research, February 2019

To close with the analogy between video and music, we notice that the most successful video operators play across all segments. So, for example, Disney produces both films and TV series and via its networks it also controls sport and new content. In a similar way, Sky and Comcast have been collecting both subscription and advertising revenues across the whole spectrum of content. The audio industry will follow the same model, we argue. So for the rest of this report we will work on the assumption that a truly dominant content/distribution operator has to reach to adjacent segment to make its offer the most enticing for users.

15 February 2019 18 Macquarie Research Global Audio Revolution

Before achieving a fully vertical integration a-la-Netflix (see Terms of Trade chapter), we expect distributors Spotify will expand and succeed in adjacent genres of audio entertainment, alongside music. By the same token, we would expect labels to start dealing with comedians, sport commentators and radio-like comment too. Strong growth in podcasts (currently only 1% of total), supports our view.

Spotify in audio Last June, Spotify announced Dawn Ostroff (President of Condé Nast Entertainment) would join as Chief Content Officer. Ostroff will lead "all aspects" of content partnerships across music, audio and video. We see this as a confirmation that Spotify is now aggressively focused on original audio content other than recorded music. This way it can avoid entering in competition with its main suppliers, the music majors, while gaining control on at least part of content costs. When it comes to podcasts and curation, Spotify is playing catch up and we understand its current contracts with music majors somehow limit its ability to add and develop its non-music offer. Spotify recommendations also appear somewhat driven the aim of maximising margins (some content has more advantageous economics for retailers), rather than focusing exclusively on users engagement.

15 February 2019 19 Macquarie Research Global Audio Revolution

Freemium convergence OTT distribution will prove a much greater disruption for audio entertainment than it has been for video. The reason is that video has already been regrouped by DVRs (Digital Video Recorders). Even in their earliest incarnations set-top-boxes aggregated all available premium channels together with networks, effectively blending free (ad-supported) and pay (subscription) revenue models. The flexibility of cable/satellite delivery combined with some storage available on the boxes, also meant consumers could experience a very early phase of on-demand content. For audio content, streaming means removing the barrier between advertising supported distribution and consumers direct spend. The concept of being able to listen to free and pay content on the same distribution channel is completely new to audio entertainment and it will reverse the status quo.

Where is the growth and who benefits from it The most important trend over the whole of audio entertainment is that monetisation is shifting from advertising and ad-hoc purchases (e.g. CD), to subscriptions. A closer relationship with the consumer means we can forecast stronger growth for overall audio revenues for decades. But who will benefit from that growth? Streaming services such as Spotify, Tencent Music, Apple Music, Amazon, etc. are the most obvious beneficiaries. They own the direct relationship with the consumer and are uniquely positioned to maximise revenues via price segmentation. Usage data analytics can be used to engineer sophisticated bundles, targeted promotions and packages to best service a given consumer and charge the optimal amount. Audio content owners in general will also experience a long period of expansion. But here we need to be more careful identifying the winners. If it’s true that a raising tide will lift all boats, it is also true that we expect a much greater growth opportunity for non-music content. The shift to OTT means that consumers will have all types of content available in the same app and can easily switch from one to the other. The slowdown in recorded music growth and the surge of audiobooks and streamcasts supports our view.

Fig 19 Growth for different segments by technology

70% 60% 50% 40% 30% 20% 10% 0% -10% -20% 2013 2014 2015 2016 2017

Radio Streaming (Music and Podcast) Downloads (Music and Audiobooks) Physical

Source: American Publishing Association, Audio Publishers Association, Interactive Advertising Bureau, IFPI, Magna, Macquarie Research, February 2019

Audiobooks We argue the most interesting segment in terms of growth potential is audiobooks. Two reasons why: 1) audiobooks are the most comparable to TV series which are the greatest beneficiaries from OTT video, 2) audiobooks are still poorly monetised via a subscription system which resembles more download than proper streaming.

15 February 2019 20 Macquarie Research Global Audio Revolution

In the US audiobooks already account for 11% of the overall audio entertainment market (see Fig 17 above in this report) and we expect this segment to reach $3bn in revenues in 2018. That is over a third the size of US recorded music, but the segment receives hardly any coverage.

Fig 20 US audiobook market growing at c30% Fig 21 Growth in Audiobooks lifts overall book sales

$ bn 3.5 50% 10 9 3.0 40% 8 2.5 3.0 1.5 1.8 2.1 2.5 7 1.2 1.3 30% 6 1.1 2.0 1.5 1.5 1.6 1.4 1.1 1.0 5 1.5 20% 4

1.0 3 5.6 5.1 4.9 5.0 5.2 5.3 5.4 10% 2 0.5 1 0.0 0% 0 2012 2013 2014 2015 2016 2017 2018E 2012 2013 2014 2015 2016 2017 2018E

US audiobook sales YoY Physical eBooks Audiobook Source: NDP, American Association of Publishers, Macquarie Research, Source: NDP, American Association of Publishers, Macquarie Research, February 2019 February 2019

Audiobooks revenue growth trends are also better than recorded music with full double-digit growth over the past few years, accelerating to over 20% last year. Audiobook expansion rests on stable physical books sales, but it is interesting to notice that e-book sales have been heavily declining: -60% since the peak in 2014 (Fig 21). The overall US books market has had a 4% CAGR since 2014, which appears like a remarkable achievement for such a mature industry competing with so many new forms of entertainment from social networks to videogames. Audiobooks have already achieved double the size compared to peak e-books. The rise of audiobooks is not just a trend in the US. In China, the audiobook market is growing even faster at 40% per annum and has already achieved RMB4.5 bn ($740m) in size (Fig 22).

Fig 22 China audio book market growing at c40% Fig 23 UK audiobook sales also growing 25%+

RMB bn £ mn 5 50% 35 50.0%

30 4 40% 40.0% 25

30.0% 3 30% 20

15 2 20% 20.0% 10 1 10% 10.0% 5

0 0% 0 0.0% 2015 2016 2017 2018E 2013 2014 2015 2016 2017

UK - Total audiobook download sales YoY growth Sales, RMB bn YoY Source: NDP, American Association of Publishers, Macquarie Research, Source: Publishers Association, Macquarie Research, February 2019 February 2019

Finally, in the UK, audiobooks remain the fastest-growing area in digital publishing with a 25% CAGR 2013-17. And the number of people who bought/listened to an audiobook increased to 5.7m in 2017, an increase of 7% on 2014 according to Nielsen’s 2017 UK Books and Consumer Survey. The report also shows that audiobooks continue to be most popular with men aged 25–44, a relatively young demographic when it comes to book consumption. Audiobook purchases now account for 5% of consumer book spending in the UK. 15 February 2019 21 Macquarie Research Global Audio Revolution

Amazon Audible The positive trends in audiobooks consumption have not gone unnoticed with the largest global book retailer. In 2008 Amazon bought Audible ($300m), a leading producer and distributor of audio entertainment. Audible was funded in 1995 as a proprietary platform to sells audiobooks, radio and audio versions of magazines and newspapers. In 2003 Audible signed an exclusive deal to make its library available on iTunes Music Store and in 2008 it began producing exclusive science fiction and fantasy audiobooks under its imprint "Audible Frontiers" over time. In the table below we summarise some of the most interesting facts about Audible.

Fig 24 Overview of Amazon’s audiobook company Audible Company Audible

Users Undisclosed but growing at double digits Revenue generation Subscription or a la carte purchase, no free users Subscription pricing $14.95/mth (or $149.50/yr) for one audiobook per month, or $22.95/mth (or $229.50/yr) for two audiobooks per month; both packages come with a free one month trial as well as selected free content Ownership 100% owned by Amazon since 2008 Valuation $300m based on Amazon’s acquisition Market share 41% of all audiobooks sold Content 400,000+ titles Compatibility Fire, Kindle, Kindle Fire, iOS, Android, Audible Cloud Player, PC & Mac computer Other features Audible Channels, which is Amazon’s podcast service is available for separate subscription of $4.95/mth Audible books can be shared with another adult user within Amazon Household account, however Audible Membership Benefits cannot be shared Availability Available in 38 languages Dedicated online marketplace stores in US, Canada, UK, Ireland, Australia, New Zealand, France, Belgium, Switzerland, Germany, Austria, Japan, Italy Source: Company data, Macquarie Research, February 2019

As of October 2018, Audible has a number of subscription options in the US, in many cases more expensive than Netflix.

1) $4.95/m plus tax (Audible Channels Plan) for unlimited listening to 70+ channels 2) $7.00/m plus tax (Silver plan) to get one audio credit every other month. Roll over up to five credits at a time for as long as your membership is active. 3) $14.95/m (Gold Monthly) for one audio credit each month. Rollover up to five credits at a time as for Silver Plan. 4) $149.50/y upfront, equivalent to $12.42m for 12 months (Gold Annual) for 12 audio credits each year, all at once. Rollover up to six credits at a time for as long as your membership is active. 5) $22.95/m (Platinum Monthly) for two audio credits each month. Rollover and up to 10 credits at a time for as long as your membership is active. 6) $229.50/y upfront, equivalent to $12.42m for 12 months (Platinum Annual) for 24 audio credits each year, all at once. Rollover up to 12 credits at a time for as long as your membership is active. Members also have access to two Audible Originals every month (to keep without limitation of time), a 30% discount off on any additional audiobook and unlimited streaming access to original premium podcasts in Audible Channels. Audible Studios is the largest producers of audiobooks and it lists nearly 30,000 original titles/programs in the UK version of the site. Revenues at this point are mostly skewed to audiobooks, but the breakdown of original content by type shows the tilt towards radio content (Fig 25). This is consistent with our thesis of converging audio content, as a result of delivery via streaming.

15 February 2019 22 Macquarie Research Global Audio Revolution

Fig 25 Audible content breakdown (by volume of titles/programs) highlight the trend to replace terrestrial radio: Radio and TV programs account for 32%

Speech 32% Performance 14%

Newspaper & Magazine 2% Periodical 0% Lecture 5% Article 2% Misc 3%

Radio & TV Program 32% Audiobook 10% Source: Bloomberg, Macquarie Research, February 2019

Now that all audio content is available on a platform supporting both free (i.e. advertising) and pay (i.e. subscription) we anticipate an increase in monetisation especially for audiobooks, the equivalent to TV series that mostly benefited from the shift to OTT distribution in video entertainment.

Streamcasts, the new podcasts people actually want The word podcasts is a combination of Pod, referring to portable devices like the Apple’s iPod, and cast, from broad/narrowcast. The term actually pre-dates Apple’s podcasting service and it implies specifically that new content is automatically downloaded on a given device. Before the wider roll- out of robust mobile data networks, that was a precondition to be able to listen to audio content on the road.

With the proliferation of high quality mobile data connectivity, e.g., 3G and 4G network, and the more recent introduction of smart home devices (e.g., connected speaker), the download of new content on one single device has become a somewhat redundant feature. Consumers are now moving towards a more flexible streaming on-the-go model. Podcast is still the term commonly used to refer to this segment of audio entertainment, typically non-music. But we prefer streamcasts to highlight the convergence of all content on OTT platforms. The top-four most common content genres in the US generate more than half the revenues: Arts/Entertainment (17%), Technology (15%), News/Politics/Current Events (13%) and Business (11%). According to American Publishing Association, 40% of the US population (85m) has listened to a podcast in 2017 – up from 36% in 2016 (Fig 26). Over the same period the penetration of streamcast listeners more than doubled from 12% in 2013 to 26% in 2018E, showing how quickly this format is being embraced by people.

15 February 2019 23 Macquarie Research Global Audio Revolution

Fig 26 85m Americans listen to streamcasts (monthly basis Fig 27 US podcast revenues, USD m

90 30% 26% 700 100.0% 80 24% 25% 90.0% 70 21% 600 80.0% 60 18% 20% 500 70.0% 16% 50 400 60.0% 12% 15% 40 50.0% 300 40.0% 30 10% 200 30.0% 20 20.0% 5% 100 10 10.0%

0 0.0%

2011 2015 2019E 2013 2014 2016 2017 2018E 2020E 0 0% 2012 2013 2014 2015 2016 2017 2018E

US population listen to podcast monthly (million) Podcast (US) Growth yoy % of US population listen to podcast monthly

Source: Nielsen, Macquarie Research, February 2019 Source: PwC, Macquarie Research, February 2019

In terms of revenues, streamcasts passed the $300m mark in the US in 2017 and PWC forecast they will grow by nearly 30% per annum to reach nearly $700m in 2020 (Fig 27), still a relatively small part of the audio entertainment market, but an important one in terms of the shift from terrestrial radio to streaming. Almost all of the companies involved on the retail side of audiobook distribution see podcasts as a gateway to audiobooks. This has prompted Audible, TuneIn, Overdrive and Scribd to all launch their own in-house podcast unit. Podcasts are also booming in the UK. Nearly six million people now tuning in each week and the number of weekly podcast listeners has almost doubled in five years – from 3.2 million in 2013 to 5.9m in 2018, according to Ofcom. The increase is across all age groups, with the steepest growth among young adults aged 15 to 24. Around one in five people from this age group now listens to podcasts every week. We believe this should be a strong support of faster growth trend in podcast in the mid to longer term.

Fig 28 Recent survey suggests increasing podcast Fig 29 UK consumers’ choice of audio entertainment by popularity in UK format

Have you listened to a podcast in the past month? Change in frequency with Which UK Consumers Listen to Audio, by Format, 2017 % of respondents ● 21% of those who listened only Yes started in the last 6 months 23% Personal Music Collection… Speech-based Radio Stations ● On average, listeners spend 3.6 hrs/wk Music Radio Stations Music Videos 2/3 of new listeners are aged 16- 34 Personal Music Stored on Digital Devices Audiobooks 70% of listeners have heard No Digital Music Services podcast ad and 3/4 of those have 77% taken action as a result Podcasts 0% 20% 40% 60% 80% 100% Less Often About The Same More Often Source: Acast Audio Intelligence Report, Macquarie Research, February 2019 Source: eMarketer, Ofcom UK, Macquarie Research, February 2019

It is also worth noting that radio and TV broadcasters are embracing the medium. Podcast versions of BBC radio programmes such as Desert Island Discs and Kermode & Mayo’s Film Review feature regularly in Apple’s iTunes podcast chart. TV broadcasters are also increasingly using podcasts as a source of material for TV shows, or as an extension of established series, such as ITV’s podcast Love Island: The Morning After.

15 February 2019 24 Macquarie Research Global Audio Revolution

Apple Podcast Apple recently rebranded iTunes podcasts to Apple podcasts to align it with Apple Music and exploit its first mover advantage, and expanded since the beginning put more emphasis on the curation of its playlist hiring already back in 2015 BBC Radio 1 star DJ Zane Lowe. More recently (February 2019), Apple also reportedly hired BBC hip-hop & grime editor Ryan Newman. In general Apple is establishing itself as the leader in curating music content that brings it the closest to the radio experience. In general, music majors have demonstrated limited appetite to expand in audio entertainment beyond music. Investing in popular radio speakers, journalists and commentators to build a new stream of original content may appear risky and costly, but it represents an opportunity to consolidate in the wider audio entertainment space and gain bargaining power. If the majors don’t do it, the distributors should not miss this chance to integrate vertically, the same model which is behind the success of Netflix.

Fig 30 Streamcasts only 2% of total listening, but … Fig 31 … but replacing AM/FM radio for in their market

TV Music Other Channels (e.g. TV Music Podcasts 1% Channels (e.g. SiriusXM Music Choice) 2% 5% 5% Music Choice) SiriusXM 9% 7%

Internet Streaming Radio/Music Audio Podcasts (Pandora, (Pandora, 30% Spotify etc.) Spotify etc.) 12% 12% AM/FM Radio 56% AM/FM Radio 21%

Owned Music Owned Music (CDs, Digital (CDs, Digital music , music files, etc.) etc) 17% 23%

Source: Edison, Macquarie Research, February 2019 Source: Edison, Macquarie Research, February 2019

Google Podcast for Android We take the launch of Google Podcasts last July, as a confirmation that a number of players are now looking at audio entertainment as the next frontier to disrupt. Until recently, Apple Podcasts dominated this segment with over 80% of listening (2015 report from Clammr) on its app. But Google’s Android has 80% share of the global smartphones market, so it can drive much wider adoption.

Fig 32 Comparison of top downloaded podcast apps on Launch Date Review Number of Name (estimated) Platform Downloads Rating Reviews Pricing

Audible n.a. Google Play 100,000,000+ 4.5 599,427 Subscription and a la carte 2016 Google Play 5,000,000+ 4.7 87,358 Free and in-app purchases Podcast Addict n.a. Google Play 5,000,000+ 4.6 425,051 Freemium Podcast Republic Aug-16 Google Play 1,000,000+ 4.5 68,307 Free and in-app purchases Player FM n.a. Google Play 1,000,000+ 4.6 67,665 Freemium and in-app purchases Podbean Oct-16 Google Play 1,000,000+ 4.6 51,340 Freemium and in-app purchases Google Podcasts Jun-18 Google Play 1,000,000+ 3.4 15,098 Free Pocket Casts Jan-11 Google Play 500,000+ 4.6 51,950 £3.99 app download Source: Company data, Macquarie Research, February 2019

The ambition of this initiative for Google should not be underestimated: “Our team’s mission is to help double the amount of podcast listening in the world over the next couple years,” says Google Podcasts Product Manager Zack Reneau-Wedeen. Google has a significant advantage as it can easily drive traffic by integrating podcasts in search results, which can greatly boost visibility and accessibility of audio streams. The full integration with the Assistant also means listening can easily be shared across connected devices such as Google Home, to continue listening on speakers once at home. 15 February 2019 25 Macquarie Research Global Audio Revolution

Google effectively wants to address the huge imbalance between iOS and Android in terms of audio entertainment, with the advantage of being able to monetise traffic three ways: advertising revenues as Youtube, subscription revenues as Apple/Spotify and lastly, but not less important for Google, by harvesting consumer data for better profiling.

Recorded music Not much has changed on recorded music trends since our last two global pieces. The replacement of legacy formats with streaming and the growing number of users is well documented (Fig 33). Our forecasts show that we are still in the steepening part of the adoption curve, represented in the chart by the S-curve for streaming as % of total recorded music.

Fig 33 Global recorded music revenues breakdown

Recorded revenues roller-coster: from $40bn in 1999, 45,000 to $15bn in 2015 and back to $30bn by 2025E 90.0% 40,000 80.0% 35,000 70.0% 30,000 60.0% 25,000 50.0% 20,000 40.0% 15,000 30.0% 10,000 20.0% 5,000 10.0%

0 0.0%

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018E 2019E 2020E 2021E 2022E 2023E 2024E 2025E

Streaming Download and other Physical streaming as % of total

Source: IFPI, Macquarie Research, February 2019

Excluding performance rights, streaming now accounts for 54% of the total (Fig 34) and it is growing at a much faster rate so we expect it will account for 73% by 2025. Going back 30-40 years, recorded music revenue growth was typically associated with a change in technology and a replacement cycle (vinyl, audiocassettes, CD, etc.). But streaming is different.

Fig 34 Global recorded music revenue by segment 2018

Physical 31%

Streaming 54%

Download and other 15%

Source: IFPI, Macquarie Research, February 2019

Streaming is a radical change of the user experience and the revenue model. Streaming transitions music from ownership to access, similar to what Uber does for auto, Airbnb does to properties. Even education publisher Pearson recently introduced a similar rental-only model for print books.

15 February 2019 26 Macquarie Research Global Audio Revolution

The lack of ownership guarantees a perpetual monetisation of content even if users keep consuming the same songs in the same format: effectively users tend to become hooked to the streaming service as witnessed by the low churn rate at Spotify. Following the same model of video entertainment, the most sophisticated distributors will continue to add new features (upselling) and segment the product more finely to convert the greatest share of population at the highest possible price for each individual. In other words we are saying that streaming growth will not only boost revenues for the next ten years, but also support strong growth for decades after that. Please see the chapter specifically dedicated to streaming further down in this report for more details.

Commercial radio The next big change in audio entertainment will deeply affect radio distribution, we argue. Radio advertising revenues have been declining as a % of total advertising revenues since 2001 and Magna forecasts negative growth in the coming years, before the impact of any cyclical downturn (Fig 35). Fragmentation of audience and the shift to online targeted advertising explains most of this trend. But there is another factor at play, which will gather momentum in the coming years, in our view.

Fig 36 US radio revenues consistently declines while other Fig 35 Global radio advertising revenues (USD m) audio entertainment options take market share (USD m)

40,000 20% 30,000

30,000 15% 25,000

20,000 10% 20,000

10,000 5% 15,000

0 0% 10,000

-10,000 -5% 5,000

-20,000 -10%

2008 2015 2000 2001 2002 2003 2004 2005 2006 2007 2009 2010 2011 2012 2013 2014 2016 2017 2018E 2019E 2020E 2021E 2022E 1999 0 2011 2012 2013 2014 2015 2016 2017 Recorded music Radio advertising sales Radio Radio as % of global Radio growth yoy Concert ticket sales Podcast

Source: Magna, Macquarie Research, February 2019 Source: IFPI, PwC, Macquarie Research, February 2019

Advertising is losing share to direct consumer spend. This is a phenomenon we can also observe in video entertainment where consumers increasingly shift to higher quality content distributed on more sophisticated pay platform (satellite/ cable first and more recently OTT).

Radio is undergoing the same transformation as TV: a shift from linear broadcasting to OTT narrowcasting. But in a way, audio technology retarded this transformation because audio has always been more portable than video entertainment.

Now there is a solid solution to stream audio on smartphones and there are a number of distributors looking at expanding from recorded music into adjacent audio content. Podcasts have been around for a few years, but they have only started to become widespread more recently.

Concerts and festivals In our previous global music report we pointed at the strong growth in the live music (concerts and festivals) as a sign that consumer demand for music increased over the past 20 years, despite the collapse in recorded music sales. This trend continues with concert ticket sales growth approaching again 10% in 2017 (Fig 37).

15 February 2019 27 Macquarie Research Global Audio Revolution

Fig 37 North America concert ticket sales, USD m

10,000 25.0%

8,000 20.0%

6,000 15.0%

4,000 10.0%

2,000 5.0%

0 0.0%

-2,000 -5.0%

2006 2000 2001 2002 2003 2004 2005 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Concert ticket sales % growth yoy

Source: Pollstar, Macquarie Research, February 2019

Here the comparison with video entertainment brakes down, as cinema tickets sales declined steadily through the past 40-50 years with viewers preferring at-home entertainment (pay TV, DVD, etc). We argue live music, including festivals, should really be considered as events/experiences closer to dining out/holidays than media entertainment. Consumers’ engagement in live music events is much superior to listening to music on the mobile phone and it is even possible that more sophisticated discovery engines will generate new fans for new bands, in turn increasing the volume of tickets sold. Software and hardware integration leads to multi-setting listening and more diverse listening options, more convenience, real on-demand consumption. Leaders are automobile companies, smart speaker, smart watches etc.

15 February 2019 28 Macquarie Research Global Audio Revolution

Streaming revolution A survey by Nielsen shows that 90% of people in the US had listened to music in 2017, up from 86% a year before. Listening time also increased by 20% from 26.6hrs to 32.1hrs a week. Streaming has grown from 32% share of listening time to 41% (Fig 38) and it is also slowly gaining share from terrestrial and satellite radio.

Fig 38 Share Of Consumer Listening Time 2015 Fig 39 Share Of Consumer Listening Time 2017

Other Other 7% 10% Physical Formats Physical Streaming Streaming 12% 32% Formats 11% 41%

Digital Digital Downloads Downloads 14% 23% Terrestrial And Satellite Radio 26% Terrestrial And Satellite Radio 24%

Source: Nielsen, Macquarie Research, February 2019 Source: Nielsen, Macquarie Research, February 2019

Streaming/Podcasting are as important innovation for the music/audio industry as the set-top box has been for the video industry. Consumers’ response is so positive that we forecast the entire industry will grow to new highs. The new apps are so rich in content and easy to use that many consumers spontaneously moved from pirated services to ad-supported platforms, and from these to premium pay subscriptions: the. free-to-pay conversion rate for Spotify is really high, coupled with low churn. The migration is picking up pace, and it will eventually lead to a fuller and more sophisticated monetisation (e.g. price segmentation) of music consumption.

Streaming/podcasting as a replacement to radio/physical For nearly a century, since the advent of commercial radio, audio entertainment has been monetised via physical sales and advertising revenues. Radio broadcasting had much larger audience and it was the main source of revenues for audio entertainment other than music (talk shows, comedy, news, sport, etc.). On the other side, recorded music revenues mostly derived from physical sales (e.g. vinyl, audiocassette, CD, etc.) to a relatively small number of buyers, notwithstanding the powerful role of radio in promoting new acts and even genres. The two models had different economics and where highly segmented: Virgin is the only example we can think of a retailer also having radio stations. This market structure is changing. Podcasting and streaming are converging towards an integrated distribution platform that will be able to collect both subscription and advertising revenues from recorded music and audio entertainment more in general. It is a similar process that brought pay TV platform to provide both ad-supported channels and on-demand videos. The reunification of audio entertainment in one single channel has implication on the industry economics. The pricing model for CDs was too rigid, forcing all consumers to pay the same price for the same , regardless of how much they were ready to pay. A uniform price level means that genuine fans pay less than they would and that other potential buyers are discouraged by a price that is too expensive. The same reasoning goes for radio listeners, who have to listen to music they don’t necessarily like alongside the music they wanted. Some potential listeners were put off by excessive levels of advertising. The result was a loss of revenue at both ends.

15 February 2019 29 Macquarie Research Global Audio Revolution

Streaming resolves this issue with a more continuous approach to pricing. At this early point, we already have free tier, standard tier and promotions that favour a remarkably high conversion to pay subscription. Over time, better segmentation of music listeners means that everybody will end up paying the price they want for the level of access they desire. In economic terms that equals charging each user close to his/her reserve price for music. Pay TV packages (including promotions) are the greatest example of this continuous price differentiation.

Each platform is additive to pay streaming growth The number of pay subscribers has grown over the past few years (Fig 40) with new entrants typically adding to the whole market growth (e.g. Apple Music). Technology companies with consumer exposure see music as a relatively easy way to reduce churn and improve customer experience. Tencent recently acquired the top four players consolidating the Chinese market jumping with a 70%+ share.

Fig 40 Growth in the number of subs continues to accelerate

Apple Music and Tencent Music are incremental 250.0 to overall streaming subs 180.0% 160.0% 200.0 140.0% 120.0% 150.0 100.0% 80.0% 100.0 60.0% 50.0 40.0% 20.0% 0.0 0.0% 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018E Spotify pay Apple Pandora Tencent music Amazon Deezer Other growth %

Source: Company data, Macquarie Research, February 2019

Amazon Music Unlimited (the standalone offer closely comparable to Spotify and Apple Music) didn’t attract a vast number of subscribers. Still, Amazon is a large player in music streaming via Amazon Prime subscriptions that include a relatively small, but still interesting library of songs. The bundled nature of the subscription and limited disclosure by the company on music specifically, leaves us with approximate estimates, but we still included them in the chart above for completeness. Apple Music launched in 2015 with a focus on curation, which is appealing to the broader masses of “lean-back” listeners. According to press reports, the company caught up with Spotify and it is now the leader in paid music streaming, at least in the US. A remarkable result given there was a 4m gap in subscribers only a year ago and that Apple doesn’t take advantage of the free tier as a marketing tool.

Apple’s ascent is remarkable when considering that the company also has a limited trial period and fewer discounts, bar the $14.99 family plan. Some data also show Apple’s user base is more engaged; for example, it produced a higher number of streams on the first day of Drake’s new album Scorpion, despite having a smaller number of subs at the time.

Amazon music subscribers in the chart above represents estimates of subscribers for both Amazon Prime music and Amazon Music Unlimited. The estimate is based on Midia research and company statements. Higher streaming ARPU means greater customer satisfaction Spotify’s reported ARPU shows a downward trend (Fig 41), which was once attributed to the strong competition for land grab with cash-rich tech competitors, that may even run their music business as a loss leader if it helps maximising hardware sales (e.g. Apple) or web traffic (e.g. Amazon). The geographical mix also plays a part now as international expansion comes at lower ARPU.

15 February 2019 30 Macquarie Research Global Audio Revolution

ARPU is based on the purchasing power of each market. We outline our Premium ARPU assumptions in Fig. 41. We expect the rate of decline will slow as a result of FX and sub mix. Though certainly not in our estimates or near-term expectations, if Spotify demonstrated pricing power, we believe it could lead to a significant re-rating of the shares.

Fig 41 Spotify Premium ARPU estimates through ’20

(€) 8.00 16% 6.20 5.32 6.00 4.75 4.51 4.38 12% 4.00 8% 2.00 4% - 0% (2.00) -4% (4.00) -8% (6.00) -12% (8.00) -16% 2016 2017 2018E 2019E 2020E

ARPU yoy growth

Source: Company Data, Macquarie Research, February 2019

The majority of the sub growth for the Premium segment comes from the conversion of ad- supported users into Premium subs; as of 4Q17, ~50% of MAUs became Premium subscribers within 36 months of being in the ad-supported tier. Spotify tries to engage its ad-supported users by highlighting features they could unlock if they opt for the Premium service. These include product links, targeted campaigns, and performance marketing on social media.

Past Spotify ARPU decline due to promotions

While our estimates show a declining ARPU for Spotify, we believe the decline mostly comes from the changing market mix, meaning higher subscribers growth in emerging markets than established markets. This should not be a surprise as market prices are generally lower in emerging markets than European countries, where Spotify has a stable user base. According to FT, Spotify also confirmed that growth in Latin America and the “rest of the world” had outpaced its “more established markets” in 2017. We argue that longer term though there is strong evidence ARPU will grow higher. Excluding the distortion of early day promotions, pay streaming subscriptions already exhibit a higher ARPU than previous technologies achieved at their respective peaks: 1999 for physical, 2012 for downloads (Fig 42). We find the higher spending per user easily justifiable, given the remarkably higher ease of use and vast catalogue available on most streaming services. The all-you-can-listen bundle allows consumers to discover new music at zero incremental cost, so there is a higher perceived value for money, which also reduces churn.

15 February 2019 31 Macquarie Research Global Audio Revolution

Fig 42 ARPU is greater for Streaming services than it was for downloads and physical (global)

Average spending per music buyer ($, yearly) 7.0

5.8 6.0 5.3 5.0 4.0 4.0

3.0

2.0

1.0

0.0 Pay Streaming (2016) Physical (1999) Downloads (2012)

Source: Company data, IFPI, Macquarie Research, February 2019

Our calculated industry ARPU of $5.8 is nearly half the advertised price of $9.99 for Spotify and Apple Music partly because of the Family discounts but also because of the common three-month- free promotions. In any given year, we calculate that acquisition costs may reduce revenues by over 10%. Another factor which artificially depresses the calculated ARPU is the high number of so-called streaming zombies. Midia Consulting estimates that there are in excess of ten million inactive pay subscribers globally, typically from telecom bundles. In 2016, Deezer, the highest offender in this field, last reported 3.3m of “inactive bundle subs”, out of a total of 6.3m. Some of these inactive subscriptions may have zero associated revenues, depending on the agreement with the telecom operator.

Fig 43 Spotify Total MAUs By Region

Rest of the World 11%

Europe 36%

Latin America 21%

North America 32% Source: Company data, Macquarie Research, February 2019

With strong growth in pay subscribers, retailers are in no rush to raise the cost of subscriptions, but in Korea MelOn has already proven that consumers are remarkably inelastic to price increase. The standard price for MelOn, the retail name for LOEN’s successful streaming service (#1 market share with 2.6m subs) increased from 3,000 won ($3) in 2004 to 8,000 won ($7) in 2015, excluding the additional cost for downloads. In other words, there is room to see ARPU improving across the industry.

15 February 2019 32 Macquarie Research Global Audio Revolution

Free users also growing, as a result of audio entertainment convergence The survey from Nielsen already mentioned above also finds that 45% of people prefer to stream music for free, compared to 29% already subscribing to a pay service. Free streaming is a very important source of growth. The concept of a free subscription (e.g. Spotify) may appear an oxymoron, but it is particularly important for the which has been struggling with piracy for over a decade. Being able to transition most, if not all, users to a legitimate platform is a milestone per se. For the same reason, music majors have also come to accept that most of their content is available for free on YouTube, the largest player in this segment (Fig 44).

Fig 44 Global free users by platform 2017: YouTube dominates the segment Fig 45 Spotify free subs growth improves vs pay subs

(m) 350 250% Pandora Netease Xiami KKBOX 3% 3% 1% 0% 294 300 Spotify 246 200% 4% 250 SoundCloud 203 8% 200 150% 157

150 123 100% 100 100 60 37 50% YouTube 50 27 Tencent 55% 12 18 Music

0 0%

2011 2012 2013 2014 2015 2016 2017 2018E 2019E 2020E 26% 2010

Total subs Free subs growth Pay subs growth

Source: Tencent Music free subs taken from Oct 2018 IPO prospectus, Company data, Macquarie Research, February 2019 Source: Company data, Macquarie Research, February 2019

Bringing usage on a legitimate platform has also the advantage of being able to collect advertising. Ad revs only represent c.10% of Spotify revenues currently, but the company believes that will increase to 20% in the medium term. Growth here is function both of strong volume growth (free subs on Spotify are still growing nearly 30%) and pricing: Spotify recently signed an agreement with Nielsen to provide an independent audience measurement and advanced analytics. Ahead of its IPO, Spotify negotiated a lower minimum royalty payment for free subs in exchange for more stringent targets in terms of conversion to premium pay tier. But monetisation of free listeners via advertising is still a point of contention between the music majors and YouTube (see comments on the so-called Value Gap later in the report). Over time, free streaming has become the equivalent of radio broadcasting in terms of its importance for promotion. So music majors fully realise that even when they can withdraw material from, say, YouTube, it may be against their own interest as it would hinder the wider distribution of a certain album. In fact, free options may become even more widely spread as the transition from broadcast radio to streaming progresses. Free-to-pay conversion keeps increasing Free-to-pay conversion typically ranges between 2%-4% (number of paying subscribers/total number of users). Some very successful apps like Mailchimp and Evernote passed the 10% mark, but this remains the exception more than the rule. In this context, Spotify’s current 45% conversion rate (Fig 46) is striking and serves as additional proof of how much consumers enjoy music streaming once they try the free tier.

15 February 2019 33 Macquarie Research Global Audio Revolution

Fig 46 Spotify’s free-to-pay conversion rate stands out (2017) Fig 47 Spotify’s conversion rate keeps growing (2017)

45.2% Free-to-pay conversion rate 50.0% 45.2% 50.0% 45.0% 45.0% 40.0% 36.9% 40.0% 35.0% 35.0% 28.0%

30.0% 30.0% 25.0% 13.3% 25.0% 13.0% 25.0% 18.9%

20.0% 8.3% 20.0% 7.5%

7.0% 14.3%

4.1%

4.0% 4.0% 15.0% 3.2% 15.0% 10.0% 0.5% 10.0% 5.0% 3.4% 5.0%

0.0%

Logmein Evernote Music Tencent Spotify Hulu Mailchimp Pandora Skype Linkedin Dropbox Drive Google 0.0% 2009 2010 2011 2012 2012 2013 2014 2015 2016 2017

free/pay conversion

Source: Company Data, Factset, Macquarie Research, February 2019 Source: Company Data, Factset, Macquarie Research, February 2019

Spotify’s conversion rate also looks very strong from a dynamic standpoint with continued improvements over time, from as low as 3% in 2010 to the record level in 2017 (Fig 47). Perversely, one could argue that conversion is so high because consumers are very happy to pay for music and there should be no free tier, or at least nothing as rich as the plan Spotify offers. On one hand, music majors are very keen to make the free tier plan less appealing to encourage an even higher conversion to the most profitable premium service. On the other hand, Spotify’s free tier certainly served the company well in the early land-grabbing phase of expansion and it also served the majors well to migrate users out of illegitimate services onto legal ones. The transition to a more differentiated free/pay offer started very cautiously so that some new releases are now available exclusively on the premium pay tier for up to two weeks. We expect more segmentation and windowing as the streaming model matures. Younger demographics equally likely to pay The demographics of streaming free users are highly skewed towards the youngest with over 70% of total subscribers below the age of 34. The breakdown by age of paying subscribers is also similar with a higher share of users in the 25-36y range, but overall again >70% are less than 35y (Fig 48). In Sweden, which has one of the longest histories of streaming, cohort analysis reassuringly shows that young subscribers remain on Spotify as they grow older.

Fig 48 The majority of Spotify free/pay subs are from young demographics (2015)

50% 46% pay/free demographics are 45% very similar: >70% below 34y 40% 35% 36% 35% 30% 25% 25% 20% 18% 14% 15% 9% 10% 7% 4% 4% 5% 2% 0% 0% 13-24 25-34 35-44 45-54 55-64 65+

Pay subs Total subs

Source: Company data, Macquarie Research, February 2019

15 February 2019 34 Macquarie Research Global Audio Revolution

The concentration of pay subscribers in the younger demographics seems in contrast with the common perception that young people will prefer piracy to legal sites and free options to pay subscriptions. On the contrary we argue that streaming facilitates discovery and as such it is more appreciated by younger listeners. Added functionalities, like the ease to share playlists and accessing premium content, will likely continue to push users across all ages to upgrade to pay subscriptions. Spotify recently introduced a new function to skip ads, Active Media, in Australia. This functionality allows users to skip as many ads as they like, regardless of the ad, according to Ad Age. At the same time, it also provides cost benefits for advertisers as they would only have to pay for completed listens and views through the platform. The move could bring another upgrade to Spotify’s free tier, which potentially leads to growing paying subscribers. The current change is only implemented in Australia and there has been no announcements on expanding the new advertising policy to other regions yet.

Fig 49 Spotify introduced a new advertising policy in Australia Current New

Ads skip functionality   Applicable users Global excl. Australia Australia only Cost for advertisers Pay for all ads Only pay for completed listens/ views Personalised experience for users Less More Source: Company data, Macquarie Research, February 2019

15 February 2019 35 Macquarie Research Global Audio Revolution

Content-distribution equilibrium disrupted Before the streaming revolution, the audio entertainment industry evolved into a complex web of highly segmented silos with “localised” dominant positions controlled by either content owners or distributors. The shift to a single platform (streaming reunites radio, physical and digital downloads) and to a converged freemium business model (streaming allows for both advertising and subscription revenues to be collected on the same app), has the potential to overhaul terms of trade for the whole industry. The tension between music majors and streaming platform on royalty payment is probably the most relevant for investors, so we will focus on that. But new tensions are emerging for radio talent (e.g. Apple hiring BBC radio DJs/presenters) and book rights (e.g. Amazon bulk buying for highly anticipated new books).

Record labels concentration has been an advantage until recently Over the past two decades, during the most difficult years for the music industry, the number of majors reduced from six to only three that now control over two-thirds of the global market (Fig 50 and 51). Following the merger with Polygram, Universal played the most active part in consolidating the market with the acquisition of EMI (2012). Sony bought BMG (2008) and more recently, EMI’s publishing arm (2018).

Fig 50 In 1998, six majors controlled just over 75% of the Fig 51 In 2017, the top-three labels control 68% of the global recorded music market market, and many independents regrouped under Merlin

Universal Universal Bertelsmann 6% (UMG) (BMG) Polygram 30% 14% 17% Independents 32%

Warner Sony SMG (WMG) 17% 15%

Sony (SME) Indies Warner 22% EMI 16% (WMG) 15% 16%

Source: Music&Copyright, Macquarie Research, February 2019 Source: Music&Copyright, Macquarie Research, February 2019

As of 2007, Merlin Network emerged as the leading global digital rights agency, representing over 20,000 independent labels with the aim of negotiating collective deals for them. Including Merlin, which effectively acts as one major when it comes to licensing agreements with the likes of Apple and Spotify, the four largest record entities accounted for c.85% of the recorded music market. Music majors have also been growing their presence in the publishing segment, with the top three players now accounting for 59% of the global market (Fig 52): Sony/ ATV 27%, UMPG 20% and Warner Chappell 12%.

15 February 2019 36 Macquarie Research Global Audio Revolution

Fig 52 Music publishing segment as consolidated as recorded music, with the top-three players controlling 59% (2017)

Sony/ ATV 27%

Independents 41%

UMPG 20% Warner Chappell 12%

Source: Music&Copyright, Macquarie Research, February 2019

The combined strong position in both recorded and publishing gives Universal, Sony and Warner the most complete view of the overall market, ideal diversification and a strong bargaining position with retailers.

Music distribution still relatively fragmented In 2003, digital music distribution was highly concentrated following the success of the iTunes Store, with Apple reaching 75% share of global downloads. Apple’s close ecosystem was at the time the only viable legitimate platform, leading to radical changes like album unbundling and a $0.79 fixed-price policy for all songs, a restriction which lasted until 2009. But looking at the whole picture, including streaming and the physical sales (Fig 53), the landscape for distribution now appears much more fragmented.

Fig 53 Global recorded music distribution fragmented in Fig 54 Proliferation of platform in streaming distribution three similar-size retail models (2018E) means higher fragmentation (2018E)

Rhapsody/ Kakao SoundCloud KKBOX Netease Napster 2% 1% 1% 1% 0% Xiami 2% Pandora 0% (free&pay) 3% Physical Deezer 31% 4%

Amazon Music 7% Spotify (free&pay) Streaming YouTube 36% 54% 8%

Download and other Tencent Music 15% (free&pay) 15% Apple Music 19%

Source: IFPI, Macquarie Research, February 2019 Source: Company Data, Factset, Macquarie Research, February 2019

Apple’s leadership in paid downloads has been significantly weakened by Amazon MP3, Google Play and local online distributors (e.g. FNAC in France). Amazon Prime also emerged as a strong physical retailer with a 15-20% market share in some countries (e.g. the UK). And within streaming, there is the juxtaposition of Spotify and Apple Music, before we account for the rise of Amazon.

15 February 2019 37 Macquarie Research Global Audio Revolution

Revenue figures are not available for each of the streaming distributors, but we can build a complete picture by combining share of users in the paid/free models with the proportions just highlighted (Fig 54). The breakdown shows Spotify as the largest player, followed by Google/YouTube and Apple with about half the market share in 2015 (c.15% each). Including downloads, Apple would still be the dominant player across all digital, but Amazon is also a strong and growing physical retailer. Both of these adjustments would actually be backwards looking though because of the fast rate of decline in both physical and downloads. Overall, the proliferation of distribution platforms and the direct competition between Spotify and Apple on one side plus the emergence of Amazon as an alternative to YouTube on the other side, still leaves the majors in a strong bargaining position, at least when it comes to music alone.

Homogeneous pricing suggests labels control music distribution One sign of the majors’ strength is the lack of price differentiation with most streaming services offered at a $9.99/ month equivalent (Fig 55). The main reason is the labels’ desire to control retail pricing and the shift to a revenue sharing model with distributors rather.

Fig 55 Rigid pricing strategy around $9.99/ month in most countries Free tier Mid-tier Premium tier Top tier

Apple No No $9.99/m Spotify Yes No $9.99/m YouTube Music No No $9.99/m Tidal No No $9.99/m $19.99 for HD Amazon Prime Music No $3.99 Echo, $9.99/m $7.99 w/ Prime Deezer Yes No $9.99/m $19.99 for HD (LOEN) No No $10.99/m Tencent Music Yes $1.3/m $2.2/m $1.3/m for 300 premium songs $2/mth for 500 premium songs Pandora Yes $4.99/m (Plus) $9.99/m SiriusXM No No $15.9/m Source: Company Data, Macquarie Research, February 2019

Music has been used in the past as a loss leader to sell hardware (e.g. Apple), attract web traffic (e.g. Google) and/or support ecommerce platforms (e.g. Amazon). For this reason content owners were nearly forced to take control of retail pricing to avoid excessive discounting. Promotions for family and for students have been very popular, but in general, there is still resistance to adding a lower entry price and offer more segmented pricing options. Looking forward we also expect more experimentation with higher price points (above $9.99/month), beyond the higher quality offer by Tidal for $19.99/month. To imagine what could come next, we can take inspiration from the video industry. At the higher end we could have live concert streaming, special editions and exclusive preview of new ahead of the general public. At the lower end we can imagine single music genre being sold separately (rap, rock, electronic, etc.) or pure library services (no new releases).

Spotify’s pricing strategy different from Netflix’s To better understand the music streaming retail pricing, it is interesting to compare it to video streaming pricing by Netflix. Across different regions, Spotify appears to have a much more differentiated pricing strategy (Fig 56), which seems to contradict our argument of homogeneous pricing set by the majors.

15 February 2019 38 Macquarie Research Global Audio Revolution

Fig 56 Spotify's subscription price varies significantly across countries, USD Fig 57 Netflix price by country, USD

Sweden Switzerland

United Kingdom Sweden

Germany Taiwan

United States Germany

Japan Indonesia

Canada United States

Hong Kong United Kingdom

Poland Malaysia

Mexico Canada

Malaysia Japan

Indonesia Mexico

0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16 Source: Company Data (FX adjusted using 2017 FX rate), Macquarie Source: Company Data (FX adjusted using 2017 FX rate), Macquarie Research, February 2019 Research, February 2019

But looking more in depth, different pricing is typically the result of a different catalogue with the largest majors having a much smaller share of total in smaller countries. In comparison, Netflix’s library has grown much more homogenous over time, so it is somehow natural that pricing also has to converge to avoid VPN “piracy” (users subscribing in a lower-cost country). The fact that Netflix is has been available virtually everywhere in the world since January 2016 supports our interpretation. Netflix has an increasingly higher proportion of original content for which is owns worldwide rights, as opposed to Spotify relying predominantly on local licencing deals with the majors and local content owners. For example, the Western labels control only c20% of recorded music in China.

Another important difference with Netflix is on the potential for price increase: we expect Netflix growth to accelerate again in 2018 also thanks to the c10% price hikes from 4Q17 flowing through the P&L. In 2018 Spotify increased prices by 10% in Norway, one of its strongholds, for both new and existing subscribers and across all plans, including student and family. This is the first price increase for Spotify since it launched 10 years ago, and is an important test for a wider move. But in other countries price increase for music streaming subscriptions may be much more difficult to put through. Homogeneous retail pricing avoids unfair competition, but creates a barrier for any of the players to be the first one to increase prices, as that could create a wave of churn towards competing services. Most streaming platforms offer a very similar catalogue of music. Equally, a concerted price increase by Spotify, Apple, Amazon and Google, would probably attract antitrust attention. Apple paid a $400m fine in 2016 after losing an antitrust case in the US over price-fixing of e-books.

15 February 2019 39 Macquarie Research Global Audio Revolution

Vertical integration appeal/necessity Netflix’s success in creating its own original content has forced most video content players to adopt a direct-to-consumer strategy. The model isn’t new and historically it has not been the most successful. But if one large competitor achieves a significantly higher level of vertical integration, all competitors have to follow in a defensive way to avoid being cut off from access to content and distribution. In the 1930-40s, most Hollywood studios expanded downstream to control cinema theatres in order to secure access to distribution for their content. The strategy backfired with a very high number of bankruptcies due to high leverage and excessively stiff operations: failure to produce strong content, for example, has a double hit on the distribution arm too. Still, the new wave of vertical integration in video entertainment is likely to spread to audio entertainment like a contagion. One example is Vivendi’s outspoken interest in selling part of UMG to a strategic partner (it has to be a distributor given UMG’s concentration in content). Indeed talks between Tencent and Vivendi have been ongoing as widely reported by the press.

Apple has also recently been reported to be looking at expansion in content, and it is not the first time. In 2003, when Vivendi was on the brink of insolvency, the LA Times reported Apple was in talks to buy Universal Music for $5-6bn. Things are different now with UMG expected to grow revenue by high-single digits (we have 8% organic growth in 2018). Alibaba, Tencent, Google, Amazon and Apple have all been mentioned in the press as potential buyers of a stake in UMG, but it would be difficult for Vivendi to meet consensus and crystallise a high valuation without selling control of UMG. Even so, we argue that recorded music growth has peaked and multiple compression may continue, leading to a disappointing end for the UMG sale saga. Away from music content, Apple is a leader in podcasts and Amazon in audiobooks, showing that vertical integration is already much more advanced in the wider audio entertainment space.

Spotify direct distribution and licencing deals Content owners have been relatively shy in terms of organic expansion downstream, but streaming platforms appear to be much more aggressive in terms of upstream expansion. For example, both Spotify and Tencent Music Entertainment have started signing direct deals with new artists. Spotify’s CEO claims the company “has no interest in becoming a label”. Still, it offers very attractive payouts to unsigned artists, compared to the majors. Spotify’s new direct upload feature launched last summer and puts more pressure on distributors like Tunecore than on the labels themselves, but it reaches closer to direct licencing deals. According to MBW, Spotify pays 50% of net revenue to artists for direct distribution deals, which compares to 25-45% when going through a label/distributor. The latter would also help maximise the artist’s visibility, but for artists that can organically build a good fan base via social networking, the offer is compelling. A little bit less is known about Spotify direct licencing deals when Spotify tries to disintermediate the majors directly. But judging from the labels’ hostility and push-back, Spotify is definitely trying to expand there too. An example is the delay in Spotify’s launch in India, where the company is being held back by reluctant majors. For reference, Apple and Amazon already offer a music service in India.

Fig 58 Spotify and Tencent stepping up collaboration initiatives with artists Spotify Direct Tencent Musicians Plan

Service Spotify signs contract directly with artists that retain copyright Tencent signs contract directly with artists and provides benefits and is able to license their work to other platforms under including online live concerts and copyright protection separate agreements Availability Worldwide China only Artist base Global Majority Chinese Platform integration None Full integration including a Karaoke app Payout to artists c50% of streaming revenue n/a Artists registered n/a 60,000 Albums available n/a 100,000 Source: Company Data, Macquarie Research, February 2019

15 February 2019 40 Macquarie Research Global Audio Revolution

Tencent Musicians Plan TME, which operates in a geography where western majors are less dominant (combined less than 20% of Chinese recorded music market), had announced back in early 2017 its plan to partner with new artists. The programme, Tencent Musicians Plan, aims to help over 60,000 unsigned artists (i.e. without a label contract), to build a fan base and expand their audience, exactly what a major would normally do. Tencent is also able to leverage its social network platform in China. So artists will be able to use the streaming platform as a social media channel to connect with fans in new ways including online live opportunities, integrated social media. On this front, traditional music labels simply cannot compete, explaining why it seems so important for companies like UMG to strike a deal with Tencent in China. Here are some examples of Tencent’s licencing agreements, showing how innovative the deals can be and how big Tencent’s reach is:

 Lu Han ( ) – a Chinese pop star, released his debut digital album “Reloaded I” exclusively on QQ Music, which broke China’s digital music sales record, selling more than 3.4m copies. Since then, Lu Han has released seven digital albums via QQ Music with total sales of over 16.5m copies and accumulated over 8.2m followers. Lu Han’s popularity was boosted by various online social events in connection with his album releases, organised by Tencent.

 Ada Zhuang ( ), pop star – Ada started out as a talented singer on Tencent’s live- streaming platform. A few months later, she released her debut album on Kugou Music. Since then, Ada has released over 200 songs that have won numerous music awards. A live- streaming session hosted by Ada on Tencent platform in 2018, where she performed her debut album, recorded a peak viewership of over 100,000, and more than one million copies of the album were sold within just one month of release. Ada has collected over 4.3m followers on the Tencent platform.

 Ai Chen ( ), a young singer from Tencent’s WeSing online karaoke – To date, Ai Chen has released three digital albums. More than 100,000 copies of his debut album were sold within one hour of release on WeSing, and sales reached 300,000 in the following 24 hours. Total streams of his songs on Tencent’s platform has exceeded one billion. Ai Chen currently has approximately 9.5m followers on WeSing.

Korea case study Korea is an interesting example of how a more vertically integrated music market works. There, Kakao and control the streaming market, venturing first into distribution and then further into content. Kakao Corp owns the leading streaming platform (MelOn) in Korea, by users, a music distribution arm (including marketing, production of CDs and physical distribution) and music production (traditional label model). MelOn’s streaming market share in Korea is over 60% and Kakao’s market share in distribution is ~33%. We estimate Kakao’s market share in content is only ~5%, but it tends to include must-have acts. By regulation, Kakao has to make all of its in-house music library available to competing streaming platforms, if the same terms are offered. But those terms are often difficult to accept for global players, who tend to have set global revenue share guidelines. Indeed, Spotify and Amazon are not in Korea yet, and Apple Music (2016 launch) is struggling. We estimate Apple only achieved tens of thousands of subscribers in nearly two years, compared to 4.7m at MelOn and 3.1m at Genie Music. Difficulties in accepting the specific terms of Kakao’s content deals greatly reduce its appeal to the wider public in Korea. Korean labels are not necessarily happy with the situation, but plans to launch an independent platform have stalled.

Beyond content and distribution, Korean entertainment companies have largely relied on revenue from concerts, product endorsements, star merchandises and special events, in a model more similar to Disney than the traditional music label.

15 February 2019 41 Macquarie Research Global Audio Revolution

The buoyant Chinese market The fast popularization of mobile phones and mobile app usage in China has led to strong growth in the Chinese online music streaming market. According to IFPI, China’s recorded music revenue grew by 35.3% in 2017, driven by a 26.5% rise in streaming revenues and it is now ranked among the top-10 countries by music revenue (Fig 59).

Fig 59 China ranks among top-10 music markets by revenue in 2017

RoW 19% China US 2% 34% Brazil Australia 2% 2% Canada 2% South Korea 3% France 5% UK 8% Japan Germany 16% 7%

Source: IFPI, Macquarie Research, February 2019

With revenue of $0.3bn, China’s music market still has a lot of catching-up to do with more mature markets (US $5.9bn, Japan $2.7bn and South Korea $0.5bn). However, streaming activity is at an all-time high and industry experts see future growth in China’s music market far exceeding the global average (Fig 60-61).

Fig 60 China outperforms rest of the world in audio Fig 61 Projected Music and Radio Revenues growth in entertainment ex-audiobooks (sales CAGR 2017-2022E) Selected Countries (2015-2020E)

8% 14%

7% 12% 10% 6% 8% 5% 6%

4% 4% 2% 3%

0%

Indonesia India Nigeria China Argentina Brazil Sweden Africa South USA Vietnam 2%

1%

0% Global Mainland China Global

Source: PwC, Macquarie Research, February 2019 Source: PwC, Macquarie Research, February 2019

Sony Music COO Kevin Kelleher feels there is more optimism on China, which became a top-10 market in the global recorded music business last year. (Link) “We think in five years China could be a top-5 market”. Tencent Music Entertainment (TME), one of Tencent’s subsidiaries recently listed in the US, is a major distributor in that marketplace accounting for over 70% of music streaming traffic (700m+ monthly users, including 25m paying subscribers). Warner and Sony recently bought $200m of shares in TME and Vivendi’s UMG also has a minor stake (we estimate less than 1%).

15 February 2019 42 Macquarie Research Global Audio Revolution

Major consolidation in China’s music distribution In conjunction with the introduction of paid service and the drive to start monetising fast-growing online traffic, Tencent, NetEase and Alibaba made several acquisitions in music streaming, resulting in a rather consolidated market. Tencent integrated the #1 and #3 players (Kugou and Kuwo) into its digital music business QQ Music to become the undisputed market leader (acquisition of China Music Corporation, 2016). Netease is now a relatively distant number-two followed by Alibaba, which operates the XiaMi and Ttpod internet portals (Fig 62).

Fig 62 Top-five music streaming service MAU in April 2017 Fig 63 Top 15 China mobile music apps by MAU (February (China Internet Watch) 2019)

250 Migu Music, 229 Xiami Qingting FM, MIUI Music (Alibaba), Lizhi, 1.3% 1.0% 211 (Xiaomi), 2.3% 2.1% 1.5% iMusic, 0.6% 200 Changba, 2.5% Baidu Music Ximalaya, (Baidu), 0.4% 6.2% 150 NetEase Kugou Cloud Music, (Tencent), 108 10.3% 32.4% 100 Kuwo 63 Tencent Radio (Tencent), (Tencent), 11.8% QQ Music 50 0.8% WeSing (Tencent), (Tencent), 25.6% 14 14.0%

0 Kugou Music QQ Music Kuwo Music Netease Xiami Source: China Internet Watch, Macquarie Research, February 2019 Source: Questmobile, Macquarie Research, February 2019

Tencent, with over 1 billion active users on its mobile social platform Wechat, is at the forefront of the digital revolution in China. Across its three brands, Tencent is now becoming the most important channel for Western pop stars such as Katy Perry and Taylor Swift to sell music in China. This explains Vivendi’s UMG interest in finding some kind of deal with Tencent, hoping to get more direct access to what may eventually become the second largest music market. The company also signed a deal with NetEase to share 99% of all content to which they respectively hold the rights. Tencent Music has also signed “exclusive” licensing contracts with over 20 music majors including Sony, Warner Music, Universal Music, YG Entertainment and partnership agreements with over 200 record companies. It owns the largest licensed content library in China with over 17M songs. In contrast, we notice that the three largest global labels combined control less than 30% of local Chinese music content, compared to circa two thirds of the total elsewhere.

East to west disruption Tencent is more than just a music streaming platform, it is actually a much more deeply integrated audio platform, a model that we believe could become more common outside China. Tencent Music has WeSing (karaoke app with 228m MAUs as of February 2019) that lets users livestream their own versions of top hits.

Performers can receive digital gifts and tips from their fans, and Tencent earns a percentage. Ng, vice president for Tencent Music outlined Tencent Music’s future potential in online interactions through music: “Tencent’s future lies with fans not only listening to music, but also watching video, singing along and even creating their own music to share.” Leveraging Tencent’s QQ app and WeChat, two of China’s most popular instant messaging and social networking platforms (over 1b monthly unique users), Tencent Music’s paid subscription package (green diamond membership) has strong social features built into the products. Besides ad-free music streaming and more generous download quotas, Tencent Music’s green diamond members also have access to exclusive rights on QQ IM. Compared to the simpler pricing model forced by the majors in most other regions, we also notice a higher sophistication in price segmentation, more akin to mature pay TV package tiers. The price for green diamond membership is RMB15 per month (~US$2.5). For general downloads, the price is RMB8 per month (~US$1.3) for 300 song downloads and RMB12 per month (~US$2) for 500 song downloads. Digital download prices ranges from RMB5-20 per album (US$0.80-3.00). 15 February 2019 43 Macquarie Research Global Audio Revolution

TME appears to still be in the initial phase of converting free users to pay subscribers. Notably, its revenue was only US$1.4bn in 2017, well below that of US$4.1bn for Spotify, though revenue growth is nearly 3x faster for TME (Fig 64). The September MAU for the company’s online music combined reached 800m, compared to Spotify’s 207m MAU.

Fig 64 Tencent Music financials 2016-2018E Rmb bn 2016 2017 2018E

Revenue 4.4 11 17~18 yoy% 152% 73% Adjusted net income 0.43 1.90 4.0 yoy% 347% 108% Adjusted net margin 10% 17% 21% Source: Company data, Macquarie Research, February 2019

While we have limited information about TME’s ad revenue, we believe TME is also exposed to internet ad dollars and radio ad dollar spending in the long term. According to GroupM, total ad revenue in China was Rmb569bn, of which internet ads were Rmb329bn and radio ads were Rmb11bn. For reference, Spotify is currently generating 10% of its revenue from ads with guidance of that increasing to 20% in the long term.

The booming audio streaming market Founded in 2012, Himalaya boasts more than 400 million total listeners and a 73% market share in China. In February 2019, Himalaya FM closed a RMB4 billion ($580 million) financing round from investors including Tencent, Goldman Sachs, and General Atlantic. The latest funding round sees Himalaya’s valuation go up to RMB24 billion ($3.48 billion) ahead of a possible IPO in the second half of 2019, note the company has denied media reports about its listing plans. The company generated ~$300m sales in 2017, of which half comes from subscription payments and the other half from ad revenues as well as hardware sales. As the company gains fast popularity among Chinese listeners, it has expanded into the hardware business to develop an eco-system for its users, allowing users to easily switch between different locations or devices. Users can easily stream Himalaya content on their mobile or in the car while commuting to work, and continue when getting back home, using the AI smart speaker “Xiaoya”.

Fig 65 Overview of Himalaya : China’s most popular audio book/podcast platform Company Himalaya FM

Users Over 400m (~8.5m paying users according to revenue/ARPU proportion) Mobile phone users Over 330m MAU c40m Daily active users c6m ARPU 90 RMB c$18 per year Valuation $3.4b Investors Tencent, Goldman and other PE investors Rumored IPO date 2H 2019 Revenue c$300m Revenue split 50% from user payments; 50% from ads and hardware Content 16 categories including business commentary, talk shows, history and culture, audio book, children sections and health education etc. provided by known Chinese publishers and 3000+ KOLs, 10000+ paid courses Partnerships Formed strategic partnership with 60 big automotive brands including BMW, Ford, Cadillac etc. to provide in-car entertainment systems Source: United Media Solution, China Today, Macquarie Research, February 2019

Yu Jianjun, CEO and co-founder of Himalaya, expects the market volume for audio books and audio streaming in China to increase to RMB300 billion (US$47.7 billion), as he told China Daily in mid-2017.

15 February 2019 44 Macquarie Research Global Audio Revolution

Distribution: upstream expansion is the norm in China As China's independent music heads towards mass market, major digital streaming platforms are trying to build stronger relationships with artists. Many platforms have expanded activities into offline events, including exclusive concerts, sponsoring talent shows and music award ceremonies. These events provide independent artists a wider exposure opportunity to fans and also have great symbolic significance for music streaming platforms as they explore new ways of vertical integration in the music market. Tencent Music Entertainment, has publicly announced its partnership plan with grassroots artists in early 2017. The program, Tencent Musicians Plan, aims to help over 60,000 “original artists”, meaning with no label contracts, to strengthen their relationship with their fans and potential audience. Artists will be able to use the streaming platform as a social media channel to connect and promote works directly to fans. This enables new experiences and live opportunities that include online live concerts and offline fan events. The social media integration and value-added music experiences should give Tencent Music more leverage compared with labels in attracting independent artists.

NetEase also introduced initiatives to work directly with artists. NetEase Cloud Music held a special NetEase Cloud Music Original Artists Ceremony in Beijing in 2017. As China's first ceremony for independent artists, this ceremony has a rather symbolic importance towards the transformation of the roles that China's music streaming platforms play in the music market. As independent music is booming globally today, NetEase Cloud Music is not only opening a bigger market for independent artists locally, but also exporting Chinese independent music to the international market. The highly concentrated distribution channel compared to the highly fragmented content channel means that in China will likely see the vertical integrated model pioneered by Netflix, becoming the status quo with the role of pure content owners weakening further.

Western labels collaboration There is no reliable data on the exact split, but industry sources suggest 80% of the market is local Chinese language, another 10% is K-Pop and J- with only 10% made up by international artists. The major internet companies in China have started to move to secure licensing deals with major recording companies. This is seen as a benefit to both parties as Chinese music streaming platforms tries to expand into the western world, while major global labels are eager to build stronger interest from fans in the Chinese music market. Sony, Warner and a number of leading Taiwanese independent labels signed an agreement with Tencent in 2014. In 2017, UMG has also partnered with TME by signing a landmark licensing agreement: both parties will work together to find new ways to develop artists, to innovate business models and to reinforce a robust copyright protection environment. In addition, TME will support UMG artists to promote their music, leveraging Tencent online properties and other media channels.

15 February 2019 45 Macquarie Research Global Audio Revolution

Smart speakers and voice platforms A smart home is equipped with network-connected products (i.e., "smart products," connected via Wi-Fi, Bluetooth or similar protocols) for controlling, automating and optimizing functions such as entertainment, temperature, lighting, security or safety. A connected car uses mobile internet technology to control key functions remotely. It contains OEM hardware, which allows networking of vehicles in different versions. Safety and security applications, such as e-Call systems that alert emergency services in the event of an accident, will be the most common services supported by connected cars, plus entertainment. In this report we are interested in the potential for audio entertainment in smart homes/ autos, and in particular when it is associated with voice platforms.

Sizing the smart home market There are globally c.200m of smart homes with an annual growth rate ~15% (Fig 66 North America benefits from favourable regulation and robust telecom infrastructure (4G LTE/ 5G) and exhibits a smart home penetration of 20% (Fig 67). Asia Pacific is the fastest growing market owing to growing connected car devices in passenger cars and rise in digital services such as cyber security.

Fig 66 Global number of households with smart systems, mn Fig 67 US smart homes as a percent of total US households

350 35% 30.0% 28.0% 300 30% 25.5% 25.0% 250 25% 22.5%

200 20% 19.6% 20.0% 150 15% 16.3%

15.0% 100 10% 12.5%

50 5% 10.0% 0 0%

5.0%

2016 2014 2015 2017

2020E 2018E 2019E 2021E 2022E 2023E

Number of households with smart systems yoy 0.0% 2016 2017 2018 2019 2020 2021

Source: Strategy Analytics, Macquarie Research, February 2019 Source: S&P Market Intelligence, Macquarie Research, February 2019

If, on one hand, it is true that growth appears to be slowing overall, on the other hand, it is also true that the definition of smart homes and cars is blurring as it is sufficient to connect a mobile phone to an existing car/home device (say a sound system) to achieve similar effects.

Smart speakers Smart speakers are a special category that emerged only in the past few years. The popularity of devices such as Alexa and Google Home derives from the fact that the speaker comes with an integrated voice platform which can help search for information (news, weather, etc), command other connected devices (lights, curtains, alarm, etc.) and, most importantly, play audio content. Amazon Alexa is leader in the category with nearly a third of the global market, but Google appears to be catching up with 23% share of sales, followed by Alibaba, Baidu and Xiaomi (Fig 68). The last three combined account for 26% of the total, serving exclusively the Chinese market, which goes a long way showing how fast this geography is progressing. Incidentally we notice how Apple accounts for only 5% of global shipments.

15 February 2019 46 Macquarie Research Global Audio Revolution

Fig 68 Global smart speaker shipment market share in 3Q18 Fig 69 Global smart speaker shipment volume comparison

mn 8 7.2 7 Others Apple 14% 6 5.2 5% Amazon 5 32% 5 Xiaomi 4 8% 3.3 3 Baidu 2.2 8% 1.9 1.9 1.9 2 1.1 Alibaba Google 0.6 10% 23% 1 0.1 0.1 0 Amazon Google Alibaba Baidu Xiaomi Apple Others

Q3 18 shipments (m) Q3 17 shipments (m)

Source: Strategy Analytics, Macquarie Research, February 2019 Source: Strategy Analytics, Macquarie Research, February 2019

China emerged this year as the second largest market for smart speakers after the US and 2018 shipments are in line to grow in excess of 10x, thanks to more affordable devices recently launched. In the US, Amazon Alexa has much higher penetration, dominating nearly two-thirds of the market followed by Google controlling nearly a third, and gaining share (Fig 70). The pattern is very similar in the UK (Fig 71).

Fig 70 US smart speaker market penetration by brand Fig 71 UK smart speaker market penetration by brand (Others include Apple HomePod and Sonos) (Others include Apple HomePod and Sonos)

66.6% 68.0% Amazon Amazon 63.3% 63.8%

29.5% 26.2% Google Google 31.0% 29.5%

8.3% 7.0% Others Others 12.0% 10.3%

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0%

2018 2019E 2018 2019E

Source: eMarketer, Macquarie Research, February 2019 Source: eMarketer, Macquarie Research, February 2019

Overall, it appears clear Amazon is a global leader, with the exception of China, which is dominated by even more vertically integrated global players. Music is often used as a Trojan horse to attract or retain consumers on a given ecosystem and the integration with voice platform appears to be key for further expansion by technology players.

Apple an expensive and closed system compared to Google/ Amazon Most voice platforms are comparable in terms of understanding natural language questions, but there are subtle differences. For example Google is context-aware in the sense that it can use information from the previous question to answer the following one. Google is also fully integrated with more sophisticated functions giving it an advantage when it comes to getting relevant answers from web searches. Amazon’s Alexa appears more widely integrated with a number of third parties. Via its so-called “skills” it is possible, for example, to call a taxi, and add items to a shopping list (an interesting area for Amazon Prime). There are also is difference in terms of language availability, though we anticipate major languages to be covered by all voice platforms in the near future.

15 February 2019 47 Macquarie Research Global Audio Revolution

Apple’s smart speaker is by far the most expensive, having focused on sound quality rather than smart home functionalities. The lack of a more affordable version (cheaper versions exist for both Google and Amazon) also poses a major obstacle to wider adoption. Overall, Apple also appears to be a much more closed system compared to its competitors (Fig 72).

Fig 72 Smart speaker feature comparison

Apple HomePod Google Home Amazon Echo Price $349 $130 $180 Release date Feb-18 Nov-16 Nov-14 Voice assistant Siri Alexa Size 6.8x5.6x5.6 5.6x3.8x3.8 9.3x3.3x3.3 Weight 5.5 pounds 1.1 pounds 2.3 pounds Connections Wifi and Airplay Wifi and Bluetooth Wifi and Bluetooth Always listening Yes Yes Yes Multiple users support Not specified Yes, can recognise up to 6 Yes different voices at once OS support iOS Android and iOS Android and iOS Customisable appearance No Yes No Music services officially supported Apple Music Amazon Music YouTube Music Spotify Spotify iHeartRadio Pandora (US only) Audible TuneIn Pandora iHeartRadio TuneIn Deezer Smart home control Works with Apple's 50+ Works with 10000+ smart home Works with 10000+ smart HomeKit partners mainly devices and 1000+ brands, home devices and 2000+ related to security, heating including Toshiba, Philips, Sony brands, including Honeywell, and cooling, lights and and Bang & Olufsen etc Belkin, Philips, Lifx, Logitech entertainment etc Selling points Best sound quality Integrates with Google services Widest support for third-party Stronger privacy protections and devices apps and smart-home devices Google Assistant is smartest Integrates with Amazon with general knowledge services More affordable Better sounds quality than Customizable design Google Home Source: Company data, Macquarie Research, February 2019

While the smart speaker market in North America and Europe is primarily a two-horse race, a three-way battle is now emerging in China as Baidu significantly ramped up shipments in 3Q following the launch of its first entry-level smart speaker in June. Low prices and heavy discounting have been key drivers of growth for Alibaba and Xiaomi in China and so Baidu’s move makes clear its ambition to rapidly scale is DuerOS voice platform in the home.

Music important in smart homes, nice to have in connected cars Though smart devices can be used for very similar tasks at home and in connected cars, consumer’s perception is quite different. Smart speakers dominate the market for connected homes – in the UK, almost 40% of homes get first connected via a smart entertainment device. But when it comes to connected cars, safety and efficiency are most important. According to a recent survey in the US, smart speakers are primarily used to listen to music, weather, news, etc (Fig 73) – which is quite similar to overall radio usage and supports our view of an integrated audio entertainment market. Streamcasts and audiobooks don’t appear explicitly in the stats, but we have all reasons to believe they will become more common and rise to the same ranking as music, if usage on mobile phones is anything to go by.

15 February 2019 48 Macquarie Research Global Audio Revolution

Fig 73 Music is the most common use for smart speakers at Fig 74 Music is a less important feature in connected cars: home (US survey) safety and efficiency key

Listen to music 90% Driving analytics 44% 35% 16% 2%3%

Emergency assistance 51% 31% 12%4%3% Search for real-time information 81% (e.g. weather, traffic) Entertainment connectivity 42% 38% 13%4%3% Search for factual information (e.g. 75% trivia, history) Safety alerts 46% 30% 14% 6% 4%

Listen to news 68% Internet-enabled navigation 49% 30% 13%4%4%

Vehicle diagnostics 39% 35% 16% 5% 5% Chat with voice assistant for fun 68% 0% 20% 40% 60% 80% 100%

Use alarms, timers 68% Very important Somewhat important Neutral Not very important Not at all important 0% 20% 40% 60% 80% 100%

Source: Nielsen MediaTech Trender, Macquarie Research, February 2019 Source: Nielsen, Macquarie Research, February 2019

Another Nielsen survey (Fig 74) actually shows that entertainment ranks relatively low as a functionality for connected cars, though over 40% of respondents still say it is “very important”. We argue the gradual shift to autonomous cars will probably increase demand for entertainment, while other more practical functions will become more commoditised. The same happened with mobile telephony where the actual mobile phone is increasingly undifferentiated and what consumers care about is what entertainment they can access on it.

15 February 2019 49 Macquarie Research Global Audio Revolution

Our new global audio entertainment model We expand our industry model to reflect the shift from music to the wider audio entertainment space. In the previous part of this report we have broken down the overall audio revenues by technology, by content and by business model. The best way to model industry revenues though is to stay close to the reporting sources for recorded music, radio audio, audiobooks and podcasts (Fig 75).

Fig 75 Audio industry breakdown (US sales, 2018E)

Streamcasts 2% Audiobook 13%

Recorded music 36%

Radio 49%

Source: IFPI, Magna, APA, IAB, Macquarie Research, February 2019

Reliable global data are available for the two largest categories (recorded music and radio advertising), but for the fastest growing segments of audiobooks and podcasts we could only find good data for the US. Still, the companies we follow are mostly global so we use the US market as a proxy for the global market, with some adjustments.

Accelerating audio growth, but not because of music With most of market commentary focused on recorded music growth, we actually highlight much of the further disruption will come from the shift of advertising revenues from radio to podcasts and the resurrection of audiobooks, an old format that is quickly gaining popularity. Looking now at the US data in particular, we can plot a 20-year history of audio revenues (Fig 76). The structural decline of recorded music revenues since 1999 was accentuated by the cyclical drop in radio revenues during the great financial crisis. From 2010, the industry remained broadly flat and only returned to growth in 2016.

Fig 76 Audio revenues growth boosted by new formats, while recorded music and radio slow

40,000 20.0%

30,000 15.0%

20,000 10.0%

10,000 5.0%

0 0.0%

-10,000 -5.0%

-20,000 -10.0%

-30,000 -15.0%

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018E 2019E 2020E

Recorded music Radio Audiobook Streamcasts Audio yoy (RHS) ex. Podcasts&Audiobooks

Source: IFPI, Magna, APA, IAB, Macquarie Research, February 2019

15 February 2019 50 Macquarie Research Global Audio Revolution

Over the past two decades, audiobook and podcasts added nearly 200bps of growth to the overall industry (illustrated by the solid/dotted line gap in the chart). And with the slowdown in music streaming adding to the decline in radio advertising, we forecast the gap to increase to 600bps over the next five years.

Audiobooks and podcasts growth contribution underestimated Breaking down our forecasts for each format shows they are far from extreme. The shock effect comes simply from the way we combine them and we analyse music and radio for the first time in the wider context. In fact, it is generally expected that radio advertising revenues will continue to decline slowly as audience shifts online with podcasts (that also includes radio shows, but they will be monetised by different distributors). It is also generally expected that recorded music revenue growth peaked in 2017 and we still expect reasonably strong 13% growth, on average, for the next few years. The acceleration in growth for audiobooks and podcasts is more of a surprise (Fig 77), but very much in line with recent trends and structural audience shifts on streaming platforms.

Fig 77 Growing growth differential in various audio formats

120.0% 100.0% 80.0% 60.0% 40.0% 20.0% 0.0% -20.0%

-40.0%

2009 2018E 2010 2011 2012 2013 2014 2015 2016 2017 2019E 2020E

Recorded music Radio Audiobook Streamcasts

Source: Macquarie Research, February 2019

We forecast audiobooks to grow c30%pa building on strong consumer interest for smart speakers both at home and in connected cars. Amazon is vertically integrated with a dominant position in both content (Audible) and smart speakers (Echo), which provides a smooth customer experience and maximum flexibility on bundles. Spotify built a podcast library as large as industry leader Apple and it guides advertising revenues to grow to 20% of total over time (no specific date), from the current 10%. Podcasts are also starting from a smaller base than audiobooks (10% of the audio revs), so we forecast some modest acceleration of revenue growth to 60%-70% (was 40%+ over the past five years).

Slowdown in recorded music growth well documented The latest IFPI data show global recorded music revenues grew by 8.1% in 2017 (constant currency), just short of our forecasts. The mix of revenues is becoming increasingly favourable with streaming accounting for nearly 40% of total (Fig 78). Though streaming growth in 2017 was “only” 41.1%, slower than the five-year average and much slower than the 60% reported in 2016. That was somewhat expected with streaming now growing from a much higher base ($6.5bn in 2017 vs just over $1bn in 2012).

15 February 2019 51 Macquarie Research Global Audio Revolution

Fig 78 Global recorded music peaked in 2018 Fig 79 In 2018 streaming increased to 54% of total

45,000 10% 40,000 5% 35,000 30,000 0% Physical 31% 25,000 -5% 20,000 15,000 -10% Streaming 10,000 54% -15% 5,000

0 -20%

2008 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018E 2019E 2020E 2021E 2022E 2023E 2024E 2025E 1997 Download and other 15%

Streaming Download and other Physical growth %

Source: IFPI, Factset, Macquarie Research, February 2019 Source: IFPI, Macquarie Research, February 2019

Excluding performance rights and synchronisation revenues, streaming accounted for 54% of the total, compared to physical at only 31% and downloads at 15%, both declining (Fig 79). Last March Sony guided recorded music revenues ex-Japan to grow only 4% in 2018, compared to +10.5% in 2017 (Fig 80). The slowdown is function of higher decline in physical (projected at -35% compared to -20% in the past two years and slower growth in streaming (+29% compared to +51% average over the past two years).

Fig 80 Sony guides for a significant slowdown in global recorded music

18.0 12.0% 15.0 9% 10% 10.0% 2.1 2.2 12.0 2.1 8.0% 1.8 2.0 9.0 1.9 2.7 4.5 6.5 8.2 6.0% 6.0 3.9 3.6 4% 4% 4.0% 3.2 2.7 3.0 2.1 2.0% 4.0 3.8 3.4 3.2 2.6 0.0 0.0% -3.0 -2.0% -2.6% -6.0 -4.0% 2014 2015 2016 2017 2018E

Physical Digital (excl. streaming) Streaming Performance rights/sync Total growth

Source: Sony IR Day 2018 presentation, Macquarie Research, February 2019

Vivendi’s UMG also reported a slowdown in organic music growth from 10% in 2017 to 9% in the first nine months of 2018, with us/consensus expecting a weaker 4Q due to seasonally heavier physical sales in the Christmas season. And for Spotify we forecast premium subs growth of 23%/21% in 2019-20E, a modest deceleration from 28%/32% in ‘17/’18E (at this pace, we expect it could hit 500m MAUs by ’23). The slowdown is partly due to a more established distribution segment lacking the incremental boost associated to new big launches like Apple Music or Amazon Echo. An acceleration in the decline of legacy format also have a negative impact in a given year. In our last global report on music (June 2016) we argued for an acceleration of growth driven by streaming. Today we report growth probably peaked in 2017, and it will follow a more erratic depending on promotions/price increase together.

15 February 2019 52 Macquarie Research Global Audio Revolution

Regulation Regulation can affect the audio industry in two main ways: by setting rules to regulate royalty fees, or by protecting IP at different levels. Here below we provide some details on the two main initiative that we expect could have an impact on how g revenues are divided between distribution platform and content owners. But we expect no material change any time soon. Online platforms likely to face tighter scrutiny from safe-harbour reviews Much of the legal framework regulating the relationship between music content owners and streaming distributors revolves around the safe-harbour regime established in the early days of the internet to protect online platforms against users uploading illegal content. So if an un-licensed track is uploaded by a user on, say, YouTube, the latter has no liabilities for the subsequent illegal streams/downloads. Online platforms effectively only have the obligation to take down the track in question upon request, leaving the music labels with the difficult task to scan the web for illegal tracks and request each one of these to be taken down. The problem is that as soon as a track is removed, the same one is uploaded by another user, making it an infinite loop. To help music majors and artists, YouTube has set up a system to proactively identify un-licensed tracks and offer majors the opportunity to keep them up and monetise them according to the specific licensing agreement in place or to take them down. Still, content owners feel that the terms of these agreements poorly reflect the value Google/YouTube extracts from the traffic generated by music consumption and that they would be able to negotiate better terms if the platform was legally responsible for illegal content posted on its website. European Commission and US Copyright Office aims to foster a “fairer and more balanced” framework for content owners In 2015, the European Commission acknowledged the issue and the need for a more effective regulation. The US Copyright Office has also started a review of the safe-harbour regime last year. The final draft of the EC’s ongoing Digital Single Market proposal discussion (February 2019), that will need to be approved by EU parliament, has been agreed on 14th February. Its controversial article 13 would require internet platforms like YouTube and Facebook to proactively block illegitimate uploads of copyrighted material. The provision has been supported by European producers and the music industry. Article 13 would require internet services with at least 5 million average monthly users to “demonstrate that they have made best efforts” to prevent the upload of content flagged as copyright-protected by rights holders, per the final text of the rules. The intent of Article 13 is to let creators and actors in entertainment space “have more control over the use of their content uploaded by users on these platforms and be remunerated for it,” according to the EU. This proposed statement essentially strengthens the current safe-harbour policy. However, it does also introduce a certain extent of responsibility for content sharing providers to demonstrate best efforts to prevent acts of copyright infringements. The proposal also discusses the evolution of licensing agreements and the need for a fair and balanced approach: “Over the last years, the functioning of the online content marketplace has gained in complexity. Online content sharing services providing access to a large amount of copyright protected content uploaded by their users have developed and have become main sources of access to content online. Legal uncertainty exists as to whether such services engage in copyright relevant acts and need to obtain authorisations from rightholders for the content uploaded by their users who do not hold the relevant rights in the uploaded content, without prejudice to the application of exceptions and limitations provided for in Union Law. This situation affects rightholders' possibilities to determine whether, and under which conditions, their content is used as well as their possibilities to get appropriate remuneration for it. It is therefore important to foster the development of the licensing market between rightholders and online content sharing service providers. These licensing agreements should be fair and keep a reasonable balance for both parties. Rightholders should receive an appropriate reward for the use of their works or other subject matter.” While Google and other companies in the same position may still not be held liable for copyright infringements, the increased scrutiny should raise significant awareness around copyright and licence agreements for the industry. 15 February 2019 53 Macquarie Research Global Audio Revolution

Global Audio – review of main listed companies Page intentionally left blank

15 February 2019 54 Macquarie Research Global Audio Revolution

Amazon Music Music has been a very small part of Amazon revenues, but Amazon is a big player in music (largest physical retailer, second largest in downloads and growing in streaming) and it is growing at an extraordinary rate. Amazon offers two music streaming services, Amazon Prime Music (2m songs offered with Prime subscription package) and Music Unlimited (40m songs by purchasing a separate subscription plan). We believe Amazon Music has the most sophisticated pricing strategy, including the best value offer as part of Prime subscriptions. The Amazon Echo wireless speaker has been well received, opening the door for more potential subscribers to Amazon Music through the smart home segment at accessible price points.

Fig 81 Amazon only offers 2m songs as part of Prime subscription, but could grow 2018

Music revenues: $1,132m % of group revenues 0.5% Geography: Global Free streaming users: none Pay streaming subs: c40m users out of 100m+ Prime subs and Music Unlimited subs Premium tier price: $99/ year (Amazon Prime) Other products: Free delivery, Prime Video, etc. Ownership: Amazon (AMZN US) owns 100% Macquarie rating: Outperform (AMZN US) Macquarie analyst: Ben Schachter Source: Company data, Macquarie Research, February 2019

Strategy Selling music was the second thing Amazon ever did after selling books. The company is currently the largest physical music retailer (CD, vinyl) and it is still growing in this segment. Via its Kindle Fire, Amazon is also the second-largest retailer after Apple for music downloads with an estimated 20%+ market share in the segment (up from 8% in 2008). Excluding streaming, Amazon was the largest music retailer in the UK in 2016 (Fig 82).

Fig 82 Amazon is the largest music retailer in physical/ downloads in the UK (2016)

Other 16%

Amazon Sainsburys 28% 5%

Tesco 6%

Asda 6% Apple HMV 19% 20%

Source: BPI/ Kantar, Macquarie Research, February 2019

When it comes to content, Amazon’s strategy seems to be evolving from a value proposition (lowest prices on books and DVDs) to a more balanced tiered offer made up by a relatively generous basic bundle (Amazon Prime) fully integrated with a more expensive on-demand service for the wider catalogue.

15 February 2019 55 Macquarie Research Global Audio Revolution

Specifically in music, Amazon Prime subscribers can only access ~2m tracks, a remarkably smaller catalogue compared to the c.40m on available on dedicated services like Spotify or Apple Music. However, the Music Unlimited plan, which offers 40m songs for an additional cost, has gained fast growth over the past year, surpassing Prime growth rates. The growth seems to largely driven by the increased popularity of Echo device as the cost of Music Unlimited drops from $7.99 to $3.99 per month with an Echo device. Amazon is also increasing its efforts on negotiations with labels. Most recently, Andre Stapleton was hired from Sony Music as Head of US Label Relations. Right now, Amazon has the best strategy in terms of customer segmentation and price differentiation while streaming retailers are mostly concentrated around the $9.99 mark. Completing a well organised offer across different types of content and a cleverly segmented pricing strategy, Amazon also extends into hardware. The Amazon Fire Phone was a fiasco, but Amazon (Kindle) Fire continues to fare well in the competitive tablet segment and it works well with music downloads. Late in 2014 Amazon launched Echo, a WiFi-connected always-on voice- controlled speaker, which seems set to become a success. Currently Echo is offered at three accessible price points, $50 for the Echo Dot, $99 for Echo and $149 for Echo Plus. Apple also offers a similar product but at much higher price point of $349, called HomePod, which is integrated with Siri.

Amazon Music Unlimited Competes Through Prime and Echo Amazon Music announced in February 2019 that it has “tens of millions” paying subscribers. The updated figure suggests that Amazon Music + Amazon Music Unlimited is approaching the range of Apple Music and to a lesser extent Spotify in terms of audience reach. Subscriptions to AMU had more than doubled in the past six months with growth attributed to two key factors: Prime and Echo, both of which receive discounts from the regular $9.99/mo price. Prime members pay $7.99/mo, and Echo-only plans are $3.99/mo. Notably, when including physical music as well, Amazon is the largest global distributor/retailer of music.

Amazon taps new demographic of streamers through smart speakers More than half of Amazon Music users have utilized Echo’s voice-first functionality. Amazon believes it is expanding the premium streaming market segment by leveraging the popularity of smart- and voice-activated speakers. Most of the new paid subscribers are coming from users of home devices and include a new, untapped demographic of streamers that are not necessarily mobile-first users. This new group of streamers (AMZN referenced the older generation) includes those slower to adopt listening to music on a smartphone, but appear to prefer the convenience of listening on a voice-activated smart speaker at home.

Alexa AMZN will continue to significantly invest in this opportunity and use Alexa to drive more benefits for Prime members. Prime members remain the core competitive advantage for AMZN, and we simply don’t see a competitive challenge to continued Prime growth. In addition to device sales, we expect Alexa to broadly drive the Prime ecosystem while increasing music subscriptions, retail sales, home automation and security services, as well as other emerging businesses. AMZN’s priority is on expanding Alexa’s reach to places and experiences where it can be useful, particularly with smart home devices.

Echo’s promotion of Amazon Music counter to Apple’s strategy for HomePod An interesting distinction between AMZN and AAPL’s strategy with Echo/Music Unlimited and HomePod/Apple Music is that AMZN is leveraging its position of strength in smart speakers to gain new music subscribers, while AAPL is attempting to leverage its position of strength in music streaming to sell its new smart speaker. The result is that AMZN is less directly competing with the likes of Apple, Spotify, and Pandora, and may more easily reach users outside of the traditional mobile-first approach of the competition.

15 February 2019 56 Macquarie Research Global Audio Revolution

Latest trends Amazon Prime’s main target are young families, where much of the consumption is around the pop, rock and country genres, making its negotiations with labels also narrower compared to Spotify and Apple Music targeting wider demographics. Children’s music consumption on the platform is relatively small at less than 5%, but much larger than the industry average. Amazon CEO Jeff Bezos announced that the company has “tens of millions” of music subscribers as of 2017. Amazon Music VP Steve Boom also confirmed in an interview that the number of paying subs to Amazon Music Unlimited had doubled in six months (Oct 17-Mar 18). According to Steve Bloom, the growth in the adoption of smart speakers has opened up the home and unlocked a new, untapped demographic of streamers, which is where Amazon has seen most of its new paid subscribers come from. That includes both older listeners who are traditionally slower to adopt new technologies, and country-music listeners, another genre that has been slower to make the switch to streaming. According to the company, streams of country songs on Amazon Music are 2.5 times greater than the industry average. Catalysts Amazon is still in its early stages of growth phase in the music segment we believe. With one of the most attractive price points and bundled service offerings with Prime and Echo, Amazon Music is well positioned for an extended period of fast growth in its customer base.

15 February 2019 57 Macquarie Research Global Audio Revolution

Apple Music/ iTunes (Apple) With the launch of test integration of Apple Music and Android platform, Apple Music may play a more important role to Apple’s consumer strategy, as previously it has been regarded as a way to maximise hardware sales. Music revenues per-se are still insignificant at the group level, making Apple less interesting as a play on music growth, though its moves in software and hardware integration with Apple Music could have direct repercussions on all industry players (e.g. Spotify and Amazon).

Fig 83 Apple is the second-largest retailer in paid-for streaming plus largest player in downloads 2018E

Music revenues: $4,618m % of group revenues 2% Geography: Global Free streaming users: None Pay streaming subs: 40m+ Premium tier price: $9.99 ($14.99 Family pack) Other products: Downloads, Online radio, Music videos Ownership: Apple (AAPL US) owns 100% Macquarie rating: Neutral (AAPL US) Macquarie analyst: Ben Schachter Source: Company data, Macquarie Research, February 2019

Strategy AAPL promoted new head of Apple Music in February 2019

Apple promoted Oliver Schusser to head Apple Music Worldwide in February 2019. Schusser’s new official title is VP of Apple Music & International Content. Hired in 2004 by Eddy Cue, SVP of Internet Software and Services, Schusser has led international content initiatives for the past 14 years, including the App Store and iTunes Movies and TV. Schusser, who also has a background in tech acquisitions, recently played a key role in the $400 million planned acquisition of Shazam, Apple’s largest purchase since its $3 billion Beats Electronics acquisition in 2014. Schusser relocated from the UK to California and will continue to lead international teams, reporting directly to Cue. Continues its push for original video content

Jamie Erlicht and Zach Van Amburg, hired in June 2017 from Sony Pictures TV, continue to lead Apple’s worldwide video programming. Apple recently won a competitive project to develop a TV series adaptation of Isaac Asimov’s influential “Foundation” science fiction novels, adding a high- profile show to its list of original video content for its streaming platform. Apple Music grows to over 50 million paying subscribers in May. Up to 50 million paid subscriber milestone in February 2019 The memo announcing Oliver Schusser’s promotion to head Apple Music in February 2019 also included the latest subscription count of over 40 million paid subscribers, up from 30 million in September 2017. In addition to the paid subscribers, Apple Music had another 8 million users currently on free trial periods. That number was increased to a total of 50mm users (paying and free trails combined) in May. There haven’t been any official figures since, but AAPL did give a decent amount of details in prior years, exiting FY16 at ~17mm paying subs, and FY17 at ~30mm. However, while revenue continues to grow nicely, music streaming is a structurally lower margin business than the majority of the Services line, limiting Apple Music’s profit contribution.

15 February 2019 58 Macquarie Research Global Audio Revolution

Fig 84 Apple Music subscribers (paying and trial) reached 50m in three years

60 50 50 40 40 35 30 30 27 23 20 17 20 15 13 11 10 7 0

0

Sep-16 Jun-17 Jun-15 Sep-15 Dec-15 Mar-16 Jun-16 Dec-16 Mar-17 Sep-17 Dec-17 Mar-18 Jun-18

Source: Company data, Macquarie Research, February 2019

Latest trends Apple Music launched in July 2015 and has already achieved 50m subscribers (Fig 84) across 113 countries. We recently downgraded AAPL to Neutral Rating and $149 PT. Our focus remains on Services, and while AAPL noted that Services revs of $10.8 billion in Dec (up 27.5% y/y) showed strength in every geo, we are quite concerned that Services growth is going to slow meaningfully beginning in the March qtr. There are simply too many concerns (beyond the pre-announcement) to ignore. Tough Services comps are coming and we have significant concerns about the sustainability of growth for the key Services drivers. Specifically, we think the top three Services drivers (Licensing, App Store, and Apple Care) are all likely to slow in FY’19 and the faster growing Services business (Music, iCloud, and Apple Pay) are not big enough to offset the slowdown (and Music margins are relatively low). We expect Apple Music revenue to grow at a 43% CAGR from an estimated $2.4B in FY 2017 to $7.0B in FY 2020.

Fig 85 Apple Music revenue to be more significant as % of total services revenue

Source: Macquarie Research, February 2019

15 February 2019 59 Macquarie Research Global Audio Revolution

AAPL and GOOGL mobile OS growing their regulatory risk For the past few years, Spotify has been a vocal critic of Apple’s 30% commission, arguing publicly and to various US and European regulators. Spotify has been moving away from using the app stores as a channel for distribution. We believe this speaks to its own scale and size as a music streaming company. In Spotify’s F-1 filed on February 28, 2018, the company states that “Apple and Google also own application store platforms and are charging in-application purchase fees, which are not being levied on their own applications, thus creating a competitive advantage for themselves against us.” We believe that this may become a critical issue as apps grow and the platform holders attempt to introduce more services themselves. In its F-1 filing to go public, Spotify states that AAPL and GOOGL are leaving others at a competitive disadvantage by charging in-app platform fees while exempting their own music streaming services. Statements like these make us more nervous about AAPL and GOOGL being open to liability/regulation as they add additional services that directly compete with app developers paying to be on the platforms. This is an area to watch. While in the past, the 30% fee from the App Store and Google Play could have been considered simply a fee that must be paid to reach customers on those platforms, with music plans from AAPL and GOOG, as well as video and other potential services, competitors such as Spotify may be able to draw more attention to the fact that competitor services will be structurally more expensive to consumers than those from AAPL and GOOGL (and even potentially AMZN) because of the fee. Some might try to argue that this is anti-competitive and harmful to consumers. Catalysts The testing of an integration with Apple Music would help Apple open up to a much bigger audience and likely to accelerate its growth in subscribers. (Android market share 85% in 2018) With the launch and recent updated HomePod and the more recent Apple Watch increased level of integration of music, but an even more interesting development would be a dedicated effort to tackle home integration with new IoT hardware in direct competition with Amazon Echo. Indeed Apple was among the first one to facilitate playing music in multiple rooms via wireless at home (i.e. Airport Express). Lastly, we have to consider the potential for Apple to complete its end-to-end integration including a predisposed Apple car, on which the company has been working on for years. For reference we notice that next-generation car manufacturer Tesla does not offer the option to install a satellite receiver on the majority of its models. In any case, new cars are generally “connected” and designed for smartphone compatibility making streaming apps easy to play.

15 February 2019 60 Macquarie Research Global Audio Revolution

Sirius XM Sirius XM is executing extremely well at the increasingly busy crossroads between AM-FM broadcasting and pure music streaming. But the borders are blurring with Spotify diversifying into non-music video content and Sirius XM offering a streaming-only subscription. Lack of exposure to growth outside the US is also a limitation compared to global competitors.

Fig 86 Sirius XM dominates the North American market for satellite radio subscription and targets nearly 5% growth in subscriber base in 2016 2018

Music revenues: $4,764m % of group revenues 83% Geography: US, Canada Free streaming users none Pay streaming subs 34m (satellite radio) Premium tier price: US$15.99/US$20.99 for Select/All Access packages Other products: Sports channels (NFL, MLP, NBA, NHL, PGA) Ownership: Public (LSXMA owns 71%+) Macquarie rating: Outperform (SIRI US) Macquarie analyst: Amy Yong Source: Company data, Macquarie Research, February 2019

Strategy

LSXMA, ultimately controlled by John Malone, owns 71%+ of Sirius XM, which resulted from the acquisition of XM Satellite Radio by Sirius Satellite Radio in ‘08. Sirius XM is narrowly focused on satellite radio and differentiates itself from pure music streaming services thanks to a wide portfolio of programming including music, news, talk and sports. Historically, one of the biggest draws has been Howard Stern who signed a five-year deal in Dec. ’15.

As management prepares to welcome Pandora, we expect a better articulation of strategy and top/bottom-line synergies. Already, we see cross-promotional efforts with Sirius XM advertising on Pandora. Additionally, we applaud management for preserving its balance sheet through this deal. Latest trends

Sirius XM continues to strengthen its OEM relationships including renewals with Audi, MB, and Volvo; progress on 360-L rollouts should pick up in ’19 into the ’20 model year. Its connected vehicle initiatives are also taking off and now contribute to double-digit revenue growth and positive cash flow. Finally, the Bring Us Home campaign and partnership with Amazon will help SIRI thrive outside of the car (58m smart speakers in the US reach 32% of households). We model ’18 self- pay net adds of 1.3m vs guidance of 1.275m driven by new car sales of 17m and stable churn of 1.8%.

Fig 87 We model steady adj. EBITDA growth through ‘20

US$000s 3,000

2,500

2,000

1,500

1,000

500

- 2017 2018E 2019E 2020E Adj. EBITDA

Source: Company data, Macquarie Research, February 2019

15 February 2019 61 Macquarie Research Global Audio Revolution

Catalysts

What drew investors to Sirius XM’s standalone playbook remains intact while the addition of Pandora provides a healthy boost to its terminal value. Though early, the potential to cross- promote SIRI-P is clear as well as the ability to monetize/address non-converters.

15 February 2019 62 Macquarie Research Global Audio Revolution

Sony Music Excluding revenues and profits from the hit smartphone game, Fate/Grand Order, Sony’s music businesses will account for ~10% of revenues and profit in FY3/20. The music businesses remain a key part of Sony’s overall portfolio, with the company intent on building up its content and channel assets.

Fig 88 Sony Music is the second-largest major FY3/20

Music revenues: ~US$5.5-6.0bn for music recordings and publishing businesses alone; US$6.5- 7.0bn in total including ancillary businesses including live music venues, artist management and licensing % of group revenues ~10% of revenues Music EBIT margin: ~15% including Fate/Grand Order / ~12% excluding Fate/Grand Order Rec/pub revs split: Recorded Music: 54%, Music Publishing: 16%, Visual Media & Platform (including Fate/Grand Order): 30% Global share of Recorded: 22.6% Global share of Publishing: 28.3% (Sony/ATV with 2.16m songs catalogue + 2.05m in the EMI catalogue) Ownership: Sony (6758 JP) owns 100% of SME and SMEJ Macquarie rating: Neutral (6758 JP) Macquarie analyst: Damian Thong Source: Company data, Macquarie Research, February 2019;

Strategy

Sony’s Music segment comprises two distinct groups – the US-based Sony Music Entertainment (SME) and the Japan-based Sony Music Entertainment Japan (SMEJ). Sony/ATV, the music publishing arm, is a subsidiary of SME. In addition, Sony jointly owns (with Universal Music Group) – a leading music video streaming service. Sony continues to place priority on recurring revenue streams, and hence has continued to re- invest in its music business, most recently with Sony/ATV lifting its ownership in EMI to 100%, and with taking a stake in Tencent Music. Sony has no interest in building its own streaming service, and instead has invested in service providers like , Spotify and Tencent Music.

Fig 89 Sony Music segment revenue and operating profit margin

1,200 30%

1,000 25%

800 20%

600 15% ¥ ¥ bn

400 10%

200 5%

0 0%

FY3/15 FY3/09 FY3/10 FY3/11 FY3/12 FY3/13 FY3/14 FY3/16 FY3/17 FY3/18

FY3/19E FY3/20E FY3/21E FY3/22E FY3/23E

Music segment revenues (LHS) OPM (RHS)

Source: Company data, Macquarie Research, February 2019

SMEJ, which accounts for ~45% of Music segment revenues in FY3/19, is Japan’s second largest music recordings company after Avex. SMEJ has developed into a broad-based entertainment company, with businesses in recorded music, live entertainment, anime and mobile games. One of SMEJ’s recent successes has been the hit smartphone game Fate/Grand Order, which at the time of writing remains a Top-10 smartphone game globally, generating for Sony almost US$0.4bn in annual profit.

15 February 2019 63 Macquarie Research Global Audio Revolution

Spotify Spotify is exiting the year with some clear successes in share gains, global market expansion, and building an ad platform. The backdrop though, has remained tough with multiple contraction and a mixed debut of Tencent Music Entertainment. We remain bullish on Spotify’s dominant yet growing position in audio entertainment as well as its ability to alter industry dynamics.

Fig 90 Spotify is the largest retailer for pay subscriptions 2018E

Music revenues: $5,281m % of group revenues 100% Geography: Global Free streaming users: 109m Pay streaming subs: 95.4m Premium tier price: US$9.99 Other products: Podcasts, TV programme clips Ownership: Public Macquarie rating: Outperform (SPOT US) Macquarie analyst: Amy Yong & Giasone Salati Source: Company data, Macquarie Research, February 2019

Strategy Spotify launched in ’08 and in less than a decade has become the largest global music streaming service by active users, with a presence in 65+ countries and territories. The business model is two-sided: 1) Premium subscription-based; and 2) ad-supported. Spotify currently counts 191m Monthly Active Users (MAUs) and 87m Premium subscribers. We believe that Spotify is one of the few players that can capitalize on the global mobile/digital evolution. Connectivity is on the rise and there is a universal enjoyment of music. Its advantage, in our view, is not only its ability to penetrate 65+ countries and reach an elusive/desirable demo, but eventually to structurally change the industry from ownership to access, in keeping with the broader trend in entertainment. Spotify is going even further adding attractive video clips from sources conspicuously not about music (in the UK short clips from Vice, Ted, Sky Sports, Comedy Central, BBC and FT). Spotify is now looking at investing $400-500m in acquisitions to strengthen its position specifically in podcasting content where we see the most interesting market opportunity given that Apple doesn’t have an ad-supported model which is the most appropriate for this radio-like content. Latest trends Holiday promos and market share gains continue, giving us confidence in its ’19 outlook. We model ’18 total/paid MAUs of ~204m/~95m paid, driven by holiday campaigns (Google, Samsung, and Hulu/Showtime) coupled with improving churn of 4.8% and healthy conversion rate. We expect ’19 will end at ~250m/~119m, respectively, or growth of 22%/25%. We model 4Q ARPU of €4.72, relatively stable vs 3Q, as the impact of family/student plans to mix moderate.

Though competition is high in the US with Apple/Amazon, global markets remain an area of opportunity. We believe Spotify is gaining solid traction in markets where Android vs iPhone dominate (LatAm, Southeast Asia). In fact, Brazil/Mexico are among its top ten markets. Secondly, Spotify recently expanded into the Middle East and North Africa in mid-Nov. Next are Russia and India, fragmented markets, but big enough to support multiple players.

15 February 2019 64 Macquarie Research Global Audio Revolution

Fig 91 We model steady improvement in gross profit through ‘20

(€m) 2,500 30% 2,073 25% 2,000 1,712 20% 1,500 1,336 15% 1,000 849 10%

500 401 226 5%

- 0% 2015 2016 2017 2018E 2019E 2020E

Gross Profit Gross Profit Margin

Source: Company Data, Macquarie Capital (USA), February 2019

Catalysts Margins are set for structural improvement as label deals continue/artist relationships strengthen and Spotify expands into adjacent business. In Vivendi CEO Grainge’s holiday letter, he highlighted data and insights as tools to enhance performance. Spotify is key to supplying these, which could prove helpful during negotiations. However, NT accounting rules could veil progress around margins.

15 February 2019 65 Macquarie Research Global Audio Revolution

Universal Music Group (Vivendi) We recently downgraded Vivendi to Underperform. We believe recorded music growth peaked in 2017 and Vivendi has taken no significant steps to strengthen its position in the wider audio entertainment space. In spring 2018 management announced its intention to list UMG, but it has subsequently narrowed down its options to a partial disposal (up to 50%) to one or more strategic partners. We believe consensus expectations of €20bn + valuation for UMG are optimistic, in the current market environment. Beyond music, Vivendi is also struggling to turn around Canal+ France and its opportunistic capital allocation has led to mixed results in Italy (minority stake in both Telecom Italia and Mediaset) plus unnecessary exposure to structurally challenged advertising agencies (Havas bought from Bollore’ Group in 2017).

Fig 92 UMG’s market share has deteriorated within an already slowing music market 2018

Music revenues: $6,873m % of group revenues 43% Music EBIT margin: 14.1% Rec/pub revs split: 85%/ 15% Global share of Recorded: 29.7% Global share of Publishing: 19.5% Ownership: Vivendi (VIV FP) owns 100% Macquarie rating: Underperform (VIV FP) Macquarie analyst: Giasone Salati Source: Company data, Macquarie Research, February 2019

Strategy UMG is one of only two assets retained by Vivendi following the major reorganisation of its portfolio by Vincent Bolloré, controlling shareholder since 2012. As such UMG is core to Vivendi’s content strategy. Universal represents 59% of group revenues and 44% of its SoTP, based on our estimates. The management team has remained broadly unchanged, since the arrival of Mr Bolloré, a strong sign of support for what is in fact one of the best run majors, globally. Since 2011 Lucian Grange is the Chairman and CEO of UMG. He championed Universal’s progress in digital music and is widely regarded as one of the most powerful people in the music business. Following the acquisition of EMI in 2012, Universal has consolidated its position as the largest of the majors, but its market share declined from a peak of 34% of global recorded music to 30% more recently (from 23% to 20% in music publishing). Over the last few years, UMG has expanded in many adjacencies of recorded music from ticketing websites (Digitick and See Tickets) to performance venues (L’Olympia in France plus tens of CanalOlympia concert halls in Africa). Following the acquisition of advertising agency Havas (2017), more emphasis has been put on the potential to monetise data collected via streaming and to use them as effective tools to fine-tune marketing and A&R spending. Stricter cooperation with Havas (60% controlled by Mr Bolloré) is also intended to combine these proprietary data with consumer profiling used to target digital ads. Latest trends One of the main reasons investors hold Vivendi, we understand, is the growth story in recorded music and the potential crystallisation of a high valuation for UMG. Vivendi recently confirmed it is seeking to sell up to 50%, adding it would only accept cash, roughly half of which could be distributed back to shareholders via a buyback. When the plan was first announced, Vivendi’s own valuation for UMG was “at least as much as Spotify”, which had a market cap of $30bn+ at the time, but consensus has moved towards €20bn more recently (we have €16bn, based on 16x EBITDA 19E). Finally, we argue the recent listing of Tencent Music Entertainment at the bottom of the IPO range ($13-15) doesn’t bode well for UMG disposal.

15 February 2019 66 Macquarie Research Global Audio Revolution

Fig 93 Universal music benefited the acquisition of EMI in 2013, but the company has been losing market share recently

7,000 18% 6,000 16% 14% 5,000 12% 4,000 10% 3,000 8% 6% 2,000 4% 1,000 2%

0 0%

2014A 2001A 2002A 2003A 2004A 2006A 2007A 2009A 2010A 2011A 2015A 2016A 2017A

2005PF 2008PF 2012PF 2013PF

Recorded music Music publishing EBITA margin

Source: Company data, Macquarie Research, February 2019

With >10% EBITA margin it is the most profitable of the majors, a level achieved via continuous restructuring charges over the past 16 years. We forecast EBITA% to expand to 15% by 2020 (currently 14%), but further improvement seems unlikely due to the tension highlighted in the wider audio entertainment industry and Vivendi’s excessive focus on the purely recorded music. Catalysts We believe the main catalyst for Universal in 2019 is the potential disposal of up to 50% stake by Vivendi. At €20bn consensus valuation, UMG would be valued on 3x sales 19E, a hefty premium compared to Spotify’s 2.2x; or 19.1x EBITDA 19E, more than twice as much as Disney the most successful content company. Our €16bn valuation for UMG still implies 15.7x EBITDA 19E, which seems generous given Spotify recent multiple compression and TME’s mixed debut. Alibaba, Tencent, Google, Amazon and Apple have all been mentioned in the press as potential buyers of a stake in UMG, but it would be difficult for Vivendi to meet consensus and crystallise a high valuation without selling control of UMG. We don’t see much downside on Vivendi (13% on our TP, including dividend), but equally we don’t see a reason to hold the stock now. On the contrary, we highlight the risk of further cuts to earnings in February when Canal+ will announce a new wave of investments for new set-top-box and international expansion. We would turn more positive on Vivendi and UMG if the strategy was more clearly refocused on non-music audio content to secure its dominant position in the new converged audio entertainment industry.

15 February 2019 67 Macquarie Research Global Audio Revolution

Warner Music Group Debt-laden Warner Music Group plays a somewhat marginal role, being the smallest of the majors. WMG’s market share performance is generally strong, but that comes at the cost of the lowest margins compared to peers.

Fig 94 Warner is the smallest of the majors, which may weaken its bargaining power with large retailers 2017

Music revenues: $3,576m % of group revenues 100% Music EBIT margin: 6.2% Rec/pub revs split: 84%/ 16% Global share of Recorded: 16.2% Global share of Publishing: 12.4% Ownership: Private (Len Blavatnik's Access Ind. owns 100%) Macquarie rating: n/a Macquarie analyst: n/a Source: Company data, Macquarie Research, February 2019

Strategy In 2011, Warner Music Group was taken private by Access Industries, a conglomerate holding ultimately controlled by Len Blavatnik, for $3.3bn (1.14x trailing sales). Mr Blavatnik had already served on the board of WMG in 2004-08. Ukraine-born, U.S.-educated Len Blavatnik was the richest man in Great Britain in 2014 with an estimated net worth of $20bn. WMG has had and still suffers from relatively high financial leverage with 4.6x net debt to EBITDA in 2017 down from over 6x in 2015. As such, WMG took minimal part in the consolidating process over the past two decades with the exception of the acquisition, after Universal was forced to divest it as part of the conditions for the EMI acquisition in 2013. Even with that, WMG remains the smallest of the labels, both in recorded and publishing music. Like Universal and Sony, Warner Music is also shifting its focus onto expanded-rights deals to capitalize on ancillary revenues, from merchandising, fan clubs, sponsorship, concert promotion, and artist management. Warner Music announced in August that it has sold its entire stake of c2% in Spotify, realizing $504m, of which $126m in proceeds will be shared with its artists as part of the label's latest royalty payments. WMG in May said it had sold around 75% of its Spotify stock for $400m in proceeds, and in its latest earnings call confirmed the sale of the rest of its stake. Latest trends Reported revenues, included small acquisitions (e.g. Parlophone in 2013), improved since 2012 at just over $3.5bn in 2017. Net income has also reached the positive inflection point in 2016, mostly due to moderation in interest charges. The most distinctive point on Warner Music compared to its peers is its relatively low EBIT margin, which has been around the mid-single digits for nearly a decade. Profitability could be an additional surprise to our above-consensus estimates for the whole industry.

15 February 2019 68 Macquarie Research Global Audio Revolution

Fig 95 Warner Music revenue returned to positive growth trend post 2012 (incl Acquisition of Parlophone in 2013), but mid-single-digit margins are the lowest among peers

4,000 3,516 3,491 3,576 8.0% 3,383 3,246 3,500 3,176 2,988 3,027 7.0% 2,886 2,780 2,871 2,966 3,000 6.0% 2,500 5.0% 2,000 4.0% 1,500 3.0% 1,000 2.0% 500 60 30 149 1.0% 0 0.0% -500 -21 -56 -100 -112 -91 -1.0% -143 -205 -198 -308

-1,000 -2.0%

2006 2007 2008 2009 2010 2011PF 2012 2013 2014 2015 2016 2017

Revenues Net income EBITA margin

Source: Company data, February 2019

Catalysts Warner has been particularly active in its portfolio, first with the acquisition of Indonesian label ISS, then with the disposals of Chrysalis records, part of the post-Parlophone planned divestments. More recently it has acquired X5 Music group, a company specialised in licensing deep catalogue and specific music genre.

15 February 2019 69 Macquarie Research Global Audio Revolution

YouTube Music (Alphabet) With its estimated reach of 1,300m monthly free subscribers, YouTube is also the largest music promotion channel globally. Music is only a small part of group revenues, but the traffic and data generated by music fans is valuable for advertising across all of Alphabet’s assets. YouTube is under increasing pressure to pay out a bigger share of royalties to artists and labels, with a pending potential review of the safe harbour regime.

Fig 96 YouTube is the largest distributor for the ad-supported model 2018

Music revenues: $2,687m % of group revenues 2% Geography: Global Free streaming users: 1300m Pay streaming subs: n/a Premium tier price: $9.99 ($11.99 for premium access to all content) Other products: User generated content, Music videos Ownership: Alphabet (GOOGL US) owns 100% Macquarie rating: Outperform (GOOGL US) Macquarie analyst: Ben Schachter Source: Company data, Macquarie Research

Strategy YouTube was acquired by Alphabet in 2006 and it has grown to be the most widely used platform globally for content, including user-generated content. In keeping with its parent company’s core strategy, YouTube has chiefly focused on its ad-supported model. Music is one of the more popular categories on the platform, consisting of 15 of the top 50 most-subscribed channels with representation from English, Spanish and Hindi speaking artists. In February 2018, YouTube announced its revamped paid-for subscription plans, YouTube Premium and Music Premium. This is unsurprising given unimpressive results from previous paid subscription models (Google Play and YouTube Red) along with months of rumours on the coming new plans. The two new YouTube plans split out the previous benefits of YouTube Red of: 1) premium long and short-form video content, and 2) the removal of ads and the ability to listen to music with the screen off (due to no ads and therefore no need to verify ad impressions), into their own separate subscription offerings more focused on their separate use-cases. There is no official subscriber number data yet, however the company has been much more aggressive in pushing these packages to users, note that Alphabet is seeming more comfortable with these offerings for the long term. Over the past 90 days, YouTube Premium has consistently been a top-5 grossing app on iOS, and YouTube Music has consistently ranked in the top-25, according to data from App Annie. Compared to other music retailers/distributors, YouTube currently has one of the most supportive regulatory frameworks. The safe harbour regime protects online platforms against any liability for illegal content uploaded by their users. Majors are effectively forced to closely monitor millions of tracks to identify the illicit uploads and require they are taken down. YouTube claims that its Content ID technology is 99% accurate and it offers majors and artists the option to remove the content or monetise it. But record companies complain the system effectively protects YouTube forcing them to accept unfavourable terms. According to data from the IFPI, ad- supported users represent 93% of total music listeners, but they only contribute 20% of the recorded music industry.

15 February 2019 70 Macquarie Research Global Audio Revolution

Fig 97 Ad-supported “free” users are 93% of total (2018)... Fig 98 ... but contribute only 20% of revenues (2018)

Paid-for 7% Ad-supported 20%

Ad-supported Paid-for 93% 80% Source: IFPI, Company data, Macquarie Research, February 2018 Source: IFPI, Company data, Macquarie Research, February 2018

Latest trends GOOGL remains very challenging to model, in our view. Despite its impressive sustained run of 20%+ FX adj growth, we are increasingly struggling to estimate GOOGL’s advertising business with accuracy or confidence in the underlying rationale. While GOOGL remains bullish on the LT prospects of its emerging businesses, advertising is still the vast majority of revenue, and our concern is that these lower-margin businesses may not make up for the inevitable slowing of the higher-margin ad business. While it has always been difficult, we feel we are approaching a point where we (and we believe The Street collectively) are not able to accurately size YouTube vs search vs programmatic, which may lead to increasing volatility. Catalysts While Google and other websites in the same position may still not be held liable for copyright infringements, the increased scrutiny should raise significant awareness around copyright and licence agreements for the industry.

15 February 2019 71 Macquarie Research Global Audio Revolution

HIM International Music HIM International Music (8446-TW) is the only listed company in Taiwan’s music industry. The company has 40 artists, including S.H.E. and Power Station, 1,700 recorded songs, 520 MVs, 1,460 copyrights on music and lyrics and >110 patents. In 1H18, 36% of revenue came from music royalty, 60% from artist management, and 4% from physical products. HIM has seen a 33% earnings CAGR over the past five years (2013-2017), driven by expansion of both gross margin (from 42.5% to 41.7%) and operating margin (from 11.7% to 19.1%), mainly thanks to an increased profit contribution from the music royalty business. Strategy  New growth drivers for the music royalty business. After ending its three-year contract with Alibaba (Mar’15-Mar’18), HIM began its new partnership with the second-largest Chinese music streaming service provider, NetEase Cloud Music, in February 2019. According to management, NetEase Cloud Music now has over 400m users, compared with Alibaba’s c100m. With a larger user base driving music royalty growth, HIM’s revenue was up 24% YoY in 8M18, and management expects the growth momentum to continue in 2H18. In addition to NetEase, HIM is also planning to partner with Spotify to gain more market share in the streaming music market. Management indicated that larger exposure and higher popularity of HIM’s artists will help grow its artist management business from more commercial events.

 Margin expansion on better product mix. Risks derived from the booming digital music market. The global digital music market grew 19% YoY in 2017, with streaming music up 41%. While digital music helps stimulate the industry, it also forces music producers like HIM to shift marketing strategies and sell channels from offline to online. In addition, some investors are worried that digital streaming players who are setting up their own artist agency business may become potential threats, and management believes HIM possesses key advantages and a good reputation in producing music and cultivating artists, which are not easily duplicated.

Financial and valuations  Mgmt expects 2018 full-year gross margin (GM%) to expand on higher revenue contribution from music royalties. Generally, HIM’s music royalty business enjoys a much higher GM % (70- 90%) compared with the artist management business (20-35%). HIM is scheduled to hold less concerts this year (~15 in 2018 vs 28/42 in 2016/2017), while seeing stronger growth in the music royalty business driven by its partnership with NetEase, thus leading to better product mix. Sales contribution from music royalty has increased from 26/25% in 2016/2017 to ~40% in 1H18.

 HIM is trading at 2018E 13.4x PER, 29.4% ROE, and 2017A 4.5% div yield based on Bloomberg consensus. Corresponding figures for the TAIEX are 13.9x PER, 13.8% ROE and 3.8% yield. As of 2Q18, it had a net cash position of NT$891m and a net cash-to-equity ratio of 71%.

Latest news and development  After ending its three-year contract with Alibaba (contract period: Mar’15-Mar’18), HIM began its new partnership with NetEase Cloud Music in February 2019. According to mgmt, NetEase Cloud Music is one of the fastest growing streaming music platforms in China, with users growing from zero to >400m in the past five years.

 According to local news in July, there was some speculation that one of HIM’s famous, Ella, did not intend to extend her contract with HIM, which led to market concerns and the recent volatile stock price.

 According to mgmt, HIM is currently planning to partner with Spotify, the largest global streaming music platform, in order to gain more market share in the streaming music market.

15 February 2019 72 Macquarie Research Global Audio Revolution

Fig 99 HIM – revenue breakdown by business, 1H18 Fig 100 HIM – revenue breakdown by region, 2017

Physical product Malaysia others 4% 1% 5% Singapore 1%

Music royalty 36% Taiwan Artist 35% management China 60% 58%

Source: Company data, February 2019 Source: Company data, February 2019

Fig 101 HIM – product mix by business, 2011–2017 Fig 102 Global recorded music industry revenue, 2008–2017

NT$m US$bn 1,800 18.0

1,600 16.0 1.3 2.4 0.3 1.3 1,400 0.4 2.3 14.0 3.4 1.4 1.4 1.6 0.4 1.8 1.9 2.0 1,200 3.7 0.6 1.0 12.0 1.4 3.9 1.9 2.8 4.7 6.6 1,000 4.2 10.0 4.4 800 4.3 8.0 4.0 3.8 3.2 600 2.8 6.0 11.9 400 10.4 4.0 8.9 8.2 200 7.6 6.7 6.0 5.7 5.5 5.2 0 2.0 2011 2012 2013 2014 2015 2016 2017 0.0 Artist management Music royalty Physical product 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Physical Digital (excl. streaming) Streaming Performance Rights Sync

Source: Company data, February 2019 Source: Company data, Macquarie Research, February 2019

15 February 2019 73 Macquarie Research Global Audio Revolution Important disclosures: Recommendation definitions Volatility index definition* Financial definitions Macquarie – Asia, USA, Canada, Europe and Mazi This is calculated from the volatility of historical All "Adjusted" data items have had the following Macquarie (SA): price movements. adjustments made: Outperform – expected return >10% Added back: goodwill amortisation, provision for Neutral – expected return from -10% to +10% Very high–highest risk – Stock should be catastrophe reserves, IFRS derivatives & hedging, Underperform – expected return <-10% expected to move up or down 60–100% in a year IFRS impairments & IFRS interest expense – investors should be aware this stock is highly Excluded: non recurring items, asset revals, property Macquarie - Australia/New Zealand speculative. revals, appraisal value uplift, preference dividends & Outperform – expected return >10% minority interests Neutral – expected return from 0% to 10% High – stock should be expected to move up or Underperform – expected return <0% down at least 40–60% in a year – investors should EPS = adjusted net profit / efpowa* be aware this stock could be speculative. ROA = adjusted ebit / average total assets Note: expected return is reflective of a Medium Volatility ROA Banks/Insurance = adjusted net profit /average stock and should be assumed to adjust proportionately Medium – stock should be expected to move up total assets with volatility risk or down at least 30–40% in a year. ROE = adjusted net profit / average shareholders funds Gross cashflow = adjusted net profit + depreciation Low–medium – stock should be expected to *equivalent fully paid ordinary weighted average move up or down at least 25–30% in a year. number of shares

Low – stock should be expected to move up or All Reported numbers for Australian/NZ listed stocks down at least 15–25% in a year. are modelled under IFRS (International Financial * Applicable to select stocks in Reporting Standards). Asia/Australia/NZ/Canada

Recommendations – 12 months Note: Quant recommendations may differ from Fundamental Analyst recommendations

Recommendation proportions – For quarter ending 31 December 2018 AU/NZ Asia RSA USA CA EUR Outperform 53.56% 57.51% 47.06% 48.65% 69.08% 51.23% (for global coverage by Macquarie, 4.12% of stocks followed are investment banking clients) Neutral 31.09% 30.24% 34.12% 46.22% 26.32% 39.41% (for global coverage by Macquarie, 1.92% of stocks followed are investment banking clients) Underperform 15.36% 12.25% 18.82% 5.14% 4.61% 9.36% (for global coverage by Macquarie, 0.47% of stocks followed are investment banking clients)

Company-specific disclosures:

Important disclosure information regarding the subject companies covered in this report is available at www.macquarie.com/research/disclosures.

Analyst certification: We hereby certify that all of the views expressed in this report accurately reflect our personal views about the subject company or companies and its or their securities. We also certify that no part of our compensation was, is or will be, directly or indirectly, related to the specific recommendations or views expressed in this report. The Analysts responsible for preparing this report receive compensation from Macquarie that is based upon various factors including Macquarie Group Ltd total revenues, a portion of which are generated by Macquarie Group’s Investment Banking activities. General disclaimers: Macquarie Securities (Australia) Ltd; Macquarie Capital (Europe) Ltd; Macquarie Capital Markets Canada Ltd; Macquarie Capital Markets North America Ltd; Macquarie Capital (USA) Inc; Macquarie Capital Limited, Taiwan Securities Branch; Macquarie Capital Securities (Singapore) Pte Ltd; Macquarie Securities (NZ) Ltd; Mazi Macquarie Securities (RF) (Pty) Ltd; Macquarie Capital Securities (India) Pvt Ltd; Macquarie Capital Securities (Malaysia) Sdn Bhd; Macquarie Securities Korea Limited and Macquarie Securities (Thailand) Ltd are not authorized deposit-taking institutions for the purposes of the Banking Act 1959 (Commonwealth of Australia), and their obligations do not represent deposits or other liabilities of Macquarie Bank Limited ABN 46 008 583 542 (MBL) or MGL. MBL does not guarantee or otherwise provide assurance in respect of the obligations of any of the above mentioned entities. MGL provides a guarantee to the Monetary Authority of Singapore in respect of the obligations and liabilities of Macquarie Capital Securities (Singapore) Pte Ltd for up to SGD 35 million. This research has been prepared for the general use of the wholesale clients of the Macquarie Group and must not be copied, either in whole or in part, or distributed to any other person. If you are not the intended recipient you must not use or disclose the information in this research in any way. If you received it in error, please tell us immediately by return e-mail and delete the document. We do not guarantee the integrity of any e-mails or attached files and are not responsible for any changes made to them by any other person. MGL has established and implemented a conflicts policy at group level (which may be revised and updated from time to time) (the "Conflicts Policy") pursuant to regulatory requirements (including the FCA Rules) which sets out how we must seek to identify and manage all material conflicts of interest. Nothing in this research shall be construed as a solicitation to buy or sell any security or product, or to engage in or refrain from engaging in any transaction. In preparing this research, we did not take into account your investment objectives, financial situation or particular needs. Macquarie salespeople, traders and other professionals may provide oral or written market commentary or trading strategies to our clients that reflect opinions which are contrary to the opinions expressed in this research. Macquarie Research produces a variety of research products including, but not limited to, fundamental analysis, macro-economic analysis, quantitative analysis, and trade ideas. Recommendations contained in one type of research product may differ from recommendations contained in other types of research, whether as a result of differing time horizons, methodologies, or otherwise. Before making an investment decision on the basis of this research, you need to consider, with or without the assistance of an adviser, whether the advice is appropriate in light of your particular investment needs, objectives and financial circumstances. There are risks involved in securities trading. The price of securities can and does fluctuate, and an individual security may even become valueless. International investors are reminded of the additional risks inherent in international investments, such as currency fluctuations and international stock market or economic conditions, which may adversely affect the value of the investment. This research is based on information obtained from sources believed to be reliable but we do not make any representation or warranty that it is accurate, complete or up to date. We accept no obligation to correct or update the information or opinions in it. Opinions expressed are subject to change without notice. No member of the Macquarie Group accepts any liability whatsoever for any direct, indirect, consequential or other loss arising from any use of this research and/or further communication in relation to this research. Clients should contact analysts at, and execute transactions through, a Macquarie Group entity in their home jurisdiction unless governing law permits otherwise. The date and timestamp for above share price and market cap is the closed price of the price date. #CLOSE is the final price at which the security is traded in the relevant exchange on the date indicated. Members of the Macro Strategy team are Sales & Trading personnel who provide desk commentary that is not a product of the Macquarie Research department or subject to FINRA Rule 2241 or any other regulation regarding independence in the provision of equity research. Country-specific disclaimers: Australia: In Australia, research is issued and distributed by Macquarie Securities (Australia) Ltd (AFSL No. 238947), a participating organization of the Australian Securities Exchange. Macquarie Securities (Australia) Limited staff involved with the preparation of research have regular interaction with companies they cover. Additionally, Macquarie Group Limited does and seeks to do business with companies covered by Macquarie Research. There are robust information barriers in place to protect the independence of Macquarie Research’s product. However, recipients of Macquarie Research should be aware of this potential conflict of interest. New Zealand: In New Zealand, research is issued and distributed by Macquarie Securities (NZ) Ltd, a NZX Firm. Canada: In Canada, research is prepared, approved and distributed by Macquarie Capital Markets Canada Ltd., a (i) member of the Investment Industry Regulatory Organization of Canada and the Canadian Investor Protection Fund, and (ii) participating organisation of the Toronto 15 February 2019 74

Macquarie Research Global Audio Revolution Stock Exchange, TSX Venture Exchange & Montréal Exchange. This research is distributed in the United States, as third party research by Macquarie Capital Markets North America Ltd., which is a registered broker-dealer and member of Financial Industry Regulatory Authority and the Securities Investor Protection Corporation. Macquarie Capital Markets North America Ltd. accepts responsibility for the contents of reports issued by Macquarie Capital Markets Canada Ltd. in the United States and sent to US persons. Any US person wishing to effect transactions in the securities described in the reports issued by Macquarie Capital Markets Canada Ltd. should do so with Macquarie Capital Markets North America Ltd. This research is intended for distribution in the United States only to major Institutional Investors (as such term is defined in SEC 15a-6 and Section 15 of the Securities Exchange Act of 1934, as amended) and is not intended for the use of any person or entity that is not a major institutional investor. Research analysts of Macquarie Capital Markets Canada Ltd. are not registered/qualified as research analysts with FINRA. The Research Distribution Policy of Macquarie Capital Markets Canada Ltd. is to allow all clients that are entitled to have equal access to our research. United Kingdom: In the United Kingdom, research is issued and distributed by Macquarie Capital (Europe) Ltd, which is authorised and regulated by the Financial Conduct Authority (No. 193905). Germany: In Germany, this research is issued and/or distributed by Macquarie Capital (Europe) Limited, Niederlassung Deutschland, which is authorised and regulated by the UK Financial Conduct Authority (No. 193905). and in Germany by BaFin. France: In France, research is issued and distributed by Macquarie Capital (Europe) Ltd, which is authorised and regulated in the United Kingdom by the Financial Conduct Authority (No. 193905). Hong Kong & Mainland China: In Hong Kong, research is issued and distributed by Macquarie Capital Limited, which is licensed and regulated by the Securities and Futures Commission. In Mainland China, Macquarie Securities (Australia) Limited Shanghai Representative Office only engages in non-business operational activities excluding issuing and distributing research. Only non-A share research is distributed into Mainland China by Macquarie Capital Limited. Japan: In Japan, research is Issued and distributed by Macquarie Capital Securities (Japan) Limited, a member of the Tokyo Stock Exchange, Inc. and Osaka Exchange, Inc. (Financial Instruments Firm, Financial Bureau (kin-sho) No. 231, a member of Japan Securities Dealers Association). India: In India, research is issued and distributed by Macquarie Capital Securities (India) Pvt. Ltd. (CIN: U65920MH1995PTC090696), 92, Level 9, 2 North Avenue, Maker Maxity, Bandra Kurla Complex, Bandra (East), Mumbai – 400 051, India, which is a SEBI registered Research Analyst having registration no. INH000000545. Malaysia: In Malaysia, research is issued and distributed by Macquarie Capital Securities (Malaysia) Sdn. Bhd. (Company registration number: 463469-W) which is a Participating Organisation of Bursa Malaysia Berhad and a holder of Capital Markets Services License issued by the Securities Commission. Taiwan: In Taiwan, research is issued and distributed by Macquarie Capital Limited, Taiwan Securities Branch, which is licensed and regulated by the Financial Supervisory Commission. No portion of the report may be reproduced or quoted by the press or any other person without authorisation from Macquarie. Nothing in this research shall be construed as a solicitation to buy or sell any security or product. The recipient of this report shall not engage in any activities which may give rise to potential conflicts of interest to the report. Research Associate(s) in this report who are registered as Clerks only assist in the preparation of research and are not engaged in writing the research. Macquarie may be in past one year or now being an Issuer of Structured Warrants on securities mentioned in this report. Thailand: In Thailand, research is produced, issued and distributed by Macquarie Securities (Thailand) Ltd. Macquarie Securities (Thailand) Ltd. is a licensed securities company that is authorized by the Ministry of Finance, regulated by the Securities and Exchange Commission of Thailand and is an exchange member of the Stock Exchange of Thailand. The Thai Institute of Directors Association has disclosed the Corporate Governance Report of Thai Listed Companies made pursuant to the policy of the Securities and Exchange Commission of Thailand. Macquarie Securities (Thailand) Ltd does not endorse the result of the Corporate Governance Report of Thai Listed Companies but this Report can be accessed at: http://www.thai- iod.com/en/publications.asp?type=4. South Korea: In South Korea, unless otherwise stated, research is prepared, issued and distributed by Macquarie Securities Korea Limited, which is regulated by the Financial Supervisory Services. Information on analysts in MSKL is disclosed at http://dis.kofia.or.kr/websquare/index.jsp?w2xPath=/wq/fundMgr/DISFundMgrAnalystStut.xml&divisionId=MDIS03002001000000&serviceId=SDIS03002 001000. South Africa: In South Africa, research is issued and distributed by Mazi Macquarie Securities (RF) (Pty) Ltd, a member of the JSE Limited. Singapore: In Singapore, research is issued and distributed by Macquarie Capital Securities (Singapore) Pte Ltd (Company Registration Number: 198702912C), a Capital Markets Services license holder under the Securities and Futures Act to deal in securities and provide custodial services in Singapore. Pursuant to the Financial Advisers (Amendment) Regulations 2005, Macquarie Capital Securities (Singapore) Pte Ltd is exempt from complying with sections 25, 27 and 36 of the Financial Advisers Act. All Singapore-based recipients of research produced by Macquarie Capital (Europe) Limited, Macquarie Capital Markets Canada Ltd, Mazi Macquarie Securities (RF) (Pty) Ltd and Macquarie Capital (USA) Inc. represent and warrant that they are institutional investors as defined in the Securities and Futures Act. United States: In the United States, research is issued and distributed by Macquarie Capital (USA) Inc., which is a registered broker-dealer and member of FINRA. Macquarie Capital (USA) Inc, accepts responsibility for the content of each research report prepared by one of its non-US affiliates when the research report is distributed in the United States by Macquarie Capital (USA) Inc. Macquarie Capital (USA) Inc.’s affiliate’s analysts are not registered as research analysts with FINRA, may not be associated persons of Macquarie Capital (USA) Inc., and therefore may not be subject to FINRA rule restrictions on communications with a subject company, public appearances, and trading securities held by a research analyst account. Information regarding futures is provided for reference purposes only and is not a solicitation for purchases or sales of futures. Any persons receiving this report directly from Macquarie Capital (USA) Inc. and wishing to effect a transaction in any security described herein should do so with Macquarie Capital (USA) Inc. Important disclosure information regarding the subject companies covered in this report is available at www.macquarie.com/research/disclosures, or contact your registered representative at 1-888-MAC-STOCK, or write to the Supervisory Analysts, Research Department, Macquarie Securities, 125 W.55th Street, New York, NY 10019. © Macquarie Group

15 February 2019 75

Equities

EMEA Research

Heads of Equity Research Financials Real Estate

Christine Farkas (US, Europe) (1 212) 231 6668 Robert Sage (London) (44 20) 3037 5144 Property Trusts & Developers Rowan Goeller (South Africa) (27 11) 583 2131 Neil Welch (London) (44 20) 3037 4272 Nazeem Samsodien (Cape Town) (27 21) 813 2771 Alternative Energy & Utilities Fatima Laher (Johannesburg) (27 11) 583 2113 Mahir Hamdulay (Cape Town) (27 21) 813 2705 Qaqambile Dwayi (Johannesburg) (27 11) 583 2229 Jose Ruiz (London) (44 20) 3037 1912 Larissa van Deventer (Johannesburg) (27 11) 583 2519 Find our research at Keegan Kruger (London) (44 20) 3037 5215 Industrials Macquarie: www.macquarieresearch.com/ideas/ Disruptive Tech & Consumer Capital Goods Thomson: www.thomson.com/financial Reuters: www.knowledge.reuters.com Sreedhar Mahamkali (London) (44 20) 3037 4016 Rowan Goeller (Johannesburg) (27 11) 583 2131 Bloomberg: MAC GO Guy Peddy (London) (44 20) 3037 4509 Transportation & Infrastructure Factset: http://www.factset.com/home.aspx Andreas Inderst (London) (44 20) 3037 2629 Rowan Goeller (Johannesburg) (27 11) 583 2131 CapitalIQ www.capitaliq.com Bob Liao (London) (44 20) 3037 2868 Olivia Peters (London) (44 20) 3037 5004 Contact [email protected] for access Giasone Salati (London) (44 20) 3037 2670 Pete Vanns (London) (44 20) 3037 2938 requests. Tim Nollen (New York) (1 212) 231 0635 Yusuf Sabir (London) (44 20) 3037 4412 Ruisi Liu (London) (44 20) 3037 2331 Materials Jenny Chen (London) (44 20) 3037 5242 Email addresses Natasha Moolman (Johannesburg) (27 21) 813 2774 Chemicals/Containers, Packaging/Paper & [email protected] Fatima Laher (Johannesburg) (27 11) 583 2113 Forest Products, Construction Materials Energy Rowan Goeller (Johannesburg) (27 11) 583 2131

Cooley May (Houston, US) (1 212) 231 2586 David Hewitt (London) (44 20) 3037 5492 James Knight (London) (44 20) 3037 5396 Giacomo Romeo (London) (44 20) 3037 4445 Gerhard Engelbrecht (Johannesburg) (27 11) 583 2407 James Carmichael (London) (44 20) 3037 4282 Naisheng Cui (London) (44 20) 3037 4062 Metals & Mining Aditya Suresh (Hong Kong) (852) 3922 1265 Grant Sporre (London) (44 20) 3037 5019 Gerhard Engelbrecht (Johannesburg) (27 11) 583 2407 Ioannis Masvoulas (London) (44 20) 3037 2314 Danielle Chigumira (London) (44 20) 3037 4269 Gerhard Engelbrecht (Johannesburg) (27 11) 583 2407 Yatish Chowthee (Johannesburg) (27 11) 583 2208 Hayden Bairstow (Perth) (618) 9224 0838

Equities ETF Sales / Trading South Africa Equity Syndication

Andrew Downe (Global Head) (65) 6601 0591 Bachir Binebine (London) (44 20) 3037 4680 Franco Lorenzani (Johannesburg) (27 11) 583 2014 Daniel Kaye (London) (44 20) 3037 4924 Emanuela Salvade (London) (44 20) 3037 4750 South Africa Sales Sarah-Jane Wagg (Johannesburg) (27 11) 583 2000 US Sales Trading European Execution Services Ed Southey (Johannesburg) (27 11) 583 2026 JT Cacciabaudo (New York) (1 212) 231 6381 Atish Jogi (Johannesburg) (27 11) 583 2252 Jason Maniloff (44 20) 3037 4983 Mike Gray (New York) (1 212) 231 2555 Jesse Ushewokunze (Johannesburg) (27 11) 583 2024 Matthew Hanley (London) (44 20) 3037 4949 Chris Reale (New York) (1 212) 231 2555 Sven Thordsen (London) (44 20) 3037 4864 Richard McDonald (London) (44 20) 3037 4908 EU Cash Sales South Africa Sales Trading Samantha Matthews (London) (44 20) 3037 4995 Darren Swabel (London) (44 20) 3037 4836 Matt Randall (London) (44 20) 3037 4967 Harry Ioannou (Johannesburg) (27 11) 583 2015 Richard Feldman (London) (44 20) 3037 4784 Richard Alderman (London) (44 20) 3037 4875 Welcome Plessie (Johannesburg) (27 11) 583 2058 Matt Rubens (London) (44 20) 3037 4824 Dilip Shah (London) (44 20) 3037 4903 Martin Hughes (Johannesburg) (27 11) 583 2019 Danny Want (London) (44 20) 3037 4847 Andrew Archer (London) (44 20) 3037 4865 Marcello Damilano (Johannesburg) (27 11) 583 2018 Holger Klees (London) (44 20) 3037 4991 Jonathan Mathews (London) (44 20) 3037 4869 Commodity Hedge Fund Sales Aaron Sawyer (London) (44 20) 3037 4982 Iain Whiteley (London) (44 20) 3037 4771 Simon MacKenzie-Beevor (London) (44 20) 3037 4910 Chris Wilson (London) (44 20) 3037 4925 Chris Looney (New York) (1 212) 231 0836 Cameron Wilson (London) (44 20) 3037 4757 Toby Ingram (London) (44 20) 3037 4957 Iain Lindsay (London) (44 20) 3037 4825 Martin Turner (London) (44 20) 3037 4767 John Clemmow (London) (44 20) 3037 4603 Guy Keller (Singapore) (65) 6601 0303 Andrew Macintosh (London) (44 20) 3037 4960 Benoit Rabu (London) (44 20) 3037 4862 Commodity Corporate Sales Ayrton Jacobson (London) (44 20) 3037 4758 Holger Hoepfner (Geneva) (41 22) 818 7777 Nael Noueiri (London) (44 20) 3037 4913 Ken Kane (London) (44 20) 3037 4786 Chris Baildon (New York) (1 212) 231 0370 David Hemming (London) (44 20) 3037 4909 Doug Stone (New York) (1 212) 231 2606 Rohan Khurana (Singapore) (65) 6601 0308 Nick Bryan (London) (44 20) 3037 4768 Juliette Lafille (New York) (1 212) 231 0977 Commodity Investor Products Chris Charbonnier (London) (44 20) 3037 4760 Dominic Prowse (New York) (1 212) 231 0781 Arun Assumall (London) (44 20) 3037 4953 Marc Crome (London) (44 20) 3037 4778 Shawna Giust (New York) (1 212) 231 0904 Catherine Littlefield (New York) (1 212) 231 6348 George Sampson (London) (44 20) 3037 1732 Global Sales David Meachin (London) (44 20) 3037 4821 David Goodman (London) (44 20) 3037 5133 Cate Park () (822) 3705 8604 Graham Cook (London) (44 20) 3037 2229 Derek Kwan (Hong Kong) (852) 3922 2191 Paul McGee (London) (44 20) 3037 5655 Angus Bottrell (Sydney) (612) 8232 5959 Ben Mooney (London) (44 20) 3037 5643

This publication was disseminated on 15 February 2019 at 06:10 UTC.