Everyone hurts – the problem with ‘fixing’ streaming

Apple’s Q4 2022 revenue fall was further illustration that the global economic environment is affecting everyone. During such times, companies look for ways to avoid the worst of the impacts, partially through ‘efficiencies’ but also through growth, by exploring new income streams and improving deal terms. The music industry is no exception. With global streaming revenues slowing – despite a strong performance from Spotify– there is growing pressure on music rightsholders to identify new growth drivers. This is especially the case for major labels, who have new institutional investors who have become acclimatised to rapid growth. All of which leads to streaming royalties taking centre stage. But the problem is that everyone in the streaming ecosystem has problems with the model. So, can any fix make everyone happy? [TL;DR, no]

To heavily oversimplify, streaming has three main constituents:

• Creators (songwriters, artists, etc.)
• Rightsholders (labels, publishers, distributors, CMOs, etc.)
• Streaming services 

At the start of 2023, all three have issues with streaming:

1. Songwriters continue to push for higher royalties while long and mid-tail artists cannot make streaming economics add up
2. Publishers continue to lobby for higher rates while UMG is now advocating for a new royalty system
3. Spotify just reported a net loss of nearly half a billion dollars for 2022

Then add in all the perennials: too much music being released; no artist longevity; the commodification of music; listening fragmentation; the decline of superstars etc.

We have a streaming market in which none of the stakeholder groups feel entirely content with the current market and all would like a larger share of the revenues to flow to them. Because they all extract value from the same revenue pot, the arithmetic is simple: one stakeholder’s gain is another’s loss.

None of this is an argument for, or against, the relative merits of the case of any of the three main interest groups. But it does mean that any change to the system will leave someone unhappy. This is the impossible equation that must be balanced.

What further complicates matters is that market benefits to different stakeholders can be perceived as negatives to others. For example:

• Streaming helped democratise the means of production and distribution. Long-tail and mid-tail artists benefit, and superstars lose their share
• Streaming helped make music the soundtrack of daily routines. Suppliers of mood music benefit, traditional artists, and labels lose listening share
• Streaming helped level the playing field, making it easier for smaller labels to compete. Larger labels faced stronger competition

The debate around new royalty regimes has been around for some time, but momentum is picking up. When the CEO of the world’s biggest record label weighs in, then you know that change is going to come. But as the above illustrates, what might make a major label happy, has the potential be detrimental to other stakeholders. There is no ‘make everyone happy’ fix.

Here are two pragmatic alternatives:

Lean forward premium 

One of the cleanest fixes would be to create a two-tier royalty system based on the nature of the plays:

1. Lean forward plays (higher royalty): when a consumer plays from their own collection or seeks out a song to play it
2. Lean back plays (lower royalty): when a consumer listens to music in an algorithmic ‘radio’ channel or listens to curated playlists

As with all streaming ‘fixes’, the approach would not be without problems. Mood-based music would certainly find itself generally collecting a smaller share of royalties, but also, many of streaming’s hits (including those from majors) rely on driving larger numbers of streams in curated playlists and ‘stations’ – which in turn help fire up the algorithms and power songs to further success.

Penny per stream

Another approach would be a fixed stream rate, which would effectively mean metered streaming. For example, if every stream generated $0.01, a subscriber would be able to listen until their subscription fee was used up, with the ability to top up to listen further or upgrade to a higher capacity tier. This would certainly help drive increased ARPU (something all parties want) but could deter some subscribers as it would mean an end to the all-you-can-eat (AYCE) proposition. But maybe it is time for that. Music is not a scalable resource in the way that, say, mobile data is. Everyone’s song is someone’s creation. Also, there would need to be a solution for free streams.

Don’t forget the listener, ever

Of course, there is a massive missing detail in all of this, the missing stakeholder in the streaming economy: the listener. Crucially though, for all the problems creators and rightsholders face, consumers are not complaining en masse. They are content with a proposition that not only represents exceptional value for money but that also evolves to meet their tastes and behaviours. 

Streaming’s problems are supply side issues, not demand-side. All industry stakeholders should be careful about pushing solutions that could favour the supply side without proper consideration of the demand side. The history of business is littered with the corpses of companies that did not properly consider the needs of their customers.

Streaming was built for yesterday’s music business

The saying goes that in a good compromise, no one is truly happy. So, there is an argument that streaming is already the balance of compromise. Against this though, streaming was built for an industry that is very different than today, so it is only logical that the model needs honing to catch up, and many of streaming’s second-order consequences cannot be undone. On the demand side, music consumption has become commodified, transformed from a largely artist-centric fan experience (radio excepted) into an audio soundtrack to everyday life. On the supply side, there are simply more people than seats at the table.

Any significant ‘fix’ is going to come at one, or more, stakeholder’s expense. And even then, increased royalties will only go so far. For example, an independent label artist might expect to earn around $2,000 from a million streams (after distribution and label deductions). Members of a four-piece band would thus take home $250 each. Even doubling the standard royalty rate (which could not happen without breaking the entire model) would still only mean $500 each, which is not going to turn streaming into a living wage for most mid-tail artists, let alone the long-tail. So, ‘fixes’ will only go so far. Perhaps it is time to double down on building new things on top of and around streaming, and nurture those that already exist (Bandcamp, etc.). 

Absolutely continue to focus on improving streaming economics but do so alongside building a new industry infrastructure that is built to meet the needs of today’s creators and business rather than those of the noughties. In short, grow the pie rather than simply look at how to re-slice it.

What Spotify Can Learn From The Roman Slave Trade

OK, you’re going to have to bear with me on this one, but let me take you back to 2^nd century Rome….

Roman Slaves

The Roman Empire was at the peak of its powers. Its borders stretched from Scotland down to Syria and across to Armenia, and across its dominions Rome spread its culture, language, administration and of course, military prowess. It brought innovations such as under floor heating, running water, astronomy and brain surgery but the consensus among many modern day historians is that the Roman Empire could have been much more. Rome was fundamentally a military, expansionist state. Its endless conquests produced a steady flow of captured people that fuelled Rome’s most important economic interest: the slave trade. By the mid 2^nd century around 1 in 4 Romans were slaves. It was common for wealthy citizens to have 40 or more household slaves while the super-rich had hundreds.

The Importance Of Economic Surplus

The problem was that the over-supply of labour meant that wages were horrifically low for the masses while the rich over spent on slaves to keep up with the neighbours. The net result is that the Roman Empire was not able to create an economic surplus across its population, which meant that there was insufficient investment in learning, science and culture. If that surplus had been created, Rome would have spawned a generation of innovators, inventors and entrepreneurs that should have created an industrial revolution. This raises the tantalizing possibility of steam power and steel emerging before the middle ages, which in turn could have meant that today’s technology revolution might have happened hundreds of years ago by now.

Instead, the Roman Empire eventually crumbled with Europe forgetting most of Rome’s innovations, paved roads weeding over, aqueducts running dry and heated floors crumbling. We had to wait until the second half of the 18^th century for the Industrial Revolution for the change, which crucially followed and overlapped with the Age of Enlightenment, a period of learning unprecedented since the Renaissance (when everyone busied themselves relearning Rome’s lost secrets) which was fuelled by Europe’s economies have developed sufficiently to create enough surplus for more than just the aristocracy to learn, invent and create. 

So, Rome inadvertently held back human progress by half a millennium because of its obsession with slaves. But what does that mean for Spotify? The key lesson from the Roman experience is that being saddled with too large a cost base may not prevent you from becoming big but it will hold you back from fulfilling your potential and from building something truly lasting. You can probably tell now where I am heading with this. Spotify’s 70% rights cost base is Rome’s 1 in 4 are slaves.

Product Innovation Where Are You?

Spotify has made immense progress but it and the overall market have done too little to innovate product and user experience.  There’s been business and commercial innovation for sure but looking back at the streaming market as a whole over the last 5 years, other than making playlists better through smart use of data and curation teams, where is the dial-moving innovation? Where are the new products and features that can change the entire focus of the market. Compare and contrast how much the likes of Google, Facebook and Amazon have changed their businesses and product offerings over that period. Streaming just got better playlists. Musical.ly shouldn’t have been a standalone company, it should have been a feature coming out of Spotify’s Stockholm engineering team. But instead of being able to think about streaming simply as an engine, Spotify has had to marshal its modest operating margins around ‘sustaining’ product development and marketing / customer acquisition.

Post-Listing Scrutiny

Spotify will likely go public sometime next year as a consequence. But once public it will need to be delivering demonstrable progress towards profit with each and every quarterly SEC filing. Growth alone won’t cut it. Just ask Snap Inc. Spotify does not have a silver bullet but it does have a number of different switches it can flick that will each contribute percentages to net margin and that collectively can help Spotify become commercially viable and in turn enable it to invest in the product and experience innovation that the streaming sector so crucially lacks.  Spotify hasn’t done these yet because most will antagonize rights partners but it will be left with little option.

Spotify The Music Company

To say that Spotify will become a label is too narrow a definition of what Spotify would become. Instead it would be a next generation music company, encompassing master rights, publishing, A+R, discovery, promotion, fan engagement and data, lots of data. If Spotify can get a couple of good quarters under its belt post-listing, and maintain a high stock price then it could go on an acquisition spree, acquiring assets for a combination of cash and stock. And the bigger and bolder the acquisition the more the stock price will rise, giving Spotify yet more ability to acquire. This is the model Yahoo used in the 2000s, with apparently over-priced acquisitions being so big as to impress Wall Street enough to ensure that the increase in market cap (ie the value of its shares) was greater than the purchase price. Spotify could use this tactic to acquire, for example, Kobalt, Believe Digital and Soundcloud to create an end-to-end, data-driven discovery, consumption and rights exploitation music power house.

What other ‘label’ could offer artists the end-to-end ability to be discovered, have your audience brought to you, promoted on the best playlists, given control of your rights and be provided with the most comprehensive data toolkit available in music? And of course, by acquiring a portion of the rights of its creators though not all (that’s where Kobalt / AWAL comes in) Spotify will be able to amortize some of its content costs like Netflix does, thus adding crucial percentages to its net margin. It will also be able to do Netflix’s other trick, namely using its algorithms to over index its own content, again adding crucial percentages to its margin.

Streaming Is The Engine Not The Vehicle

The way to think about Spotify right now, and indeed streaming as a whole, is that we have built a great engine. But that’s it. We do not have the car. Streaming is not a product, it is a technology for getting music onto our devices and it is a proto-business model. While rights holders can point to areas where Spotify is arguably over spending, fixing those will not be enough on their own, they need to accompany bolder change. Once that change comes Spotify can start to fulfil its potential, to become the butterfly that is currently locked in its cocoon. While rights holders we be understandably anxious and may even cry foul, they have to shoulder much of the blame. Spotify simply doesn’t have anywhere else to go. Unless of course it wants to end up like Rome did….overrun by barbarians, or whatever the music industry equivalent is…

Spotify Earnings: Growth Comes At A Cost

Spotify has published its much anticipated 2016 revenues. Because the company is under so much analytical scrutiny, there is little that is particularly surprising but there is still plenty we can learn from the results:

• Growth maintains momentum: Spotify recorded revenues of €2.9 billion in 2016, up 51% from €1.9 billion in 2015. Although that was a lower growth rate in % terms (80% for 14/15), it was a bigger net add in revenue terms (€989 million net new revenue in 2016 compared to €863 million in 2015). Spotify still has some way to go before it challenges Netflix’s $8.2 billion streaming revenue, but it is making clear progress.
• Spotify is getting ready for public reporting: The 2016 accounts featured heavy restating of previous year figures and many line items from last year’s accounts were no longer reported. All of which points to an organization getting its reporting structures in place for a public listing of some kind.
• ARPU is a mixed story: Spotify’s total monthly user ARPU increased from €1.82 in 2015 to €1.94, driven by a small increase in ad supported user APRU and, more importantly, a higher share of paid users (38% in 2016 compared to 31% in 2015). However, that increased paid conversion has come at the price of lower paid ARPU, with $1 for 3-month trials etc., pushing down paid ARPU from €5.16 in 2015 to €4.58 which in turn is more than an entire dollar a month less than the €5.85 paid ARPU figure Spotify enjoyed in 2014.
• Losses are widening again: Spotify reported losses before tax of €539 million against revenues of €2.9 billion (i.e. 18% of revenue). This was up from 12% in 2015 although it had been as high as 17% in 2014. In order to keep up with the market, Spotify is having to spend heavily, and this is all without any major product or territory launch in 2016. You need deep pockets to play at streaming’s top table.
• Rights costs may be on a positive trajectory: Spotify’s Cost of Sales (previously reported as Royalty Distribution and Other Costs) were €2.5 billion, or 84.6% of revenue, down slightly from 85.5% in 2015. The shrinking share of the loss-making ad supported user base is most likely the key contributor here. Though the new UMG and Merlin deals will help sustain this path.

In Search Of A Margin

So, what do Spotify’s results say about the economics that we didn’t already know. In truth, not much. The market has lots of growth in it yet; competing is expensive, growth has to be incentivized and rights are the main cost component.

As Spotify nears a public listing or an acquisition by Alibaba or Tencent, it remains the benchmark for the health of the streaming economy. With the underlying fundamentals remaining largely unchanged in 2016 despite stellar growth, here are a few thoughts on how the economics of streaming might change:

An often repeated argument from record labels is that streaming services will hit profitability when they reach scale. So, when does that happen? 48 million subscribers can lay a good claim to being ‘scale’, but it isn’t driving profit. While the market establishes itself, streaming services have to overspend on product innovation and marketing (and then, later, on user retention). So, these costs will likely rise in relative terms. Meanwhile, rights are always going to remain largely in line with revenue (though the UMG and Merlin deals reward growth with some discounting, which is a welcome innovation). But even these deals will not change the fact that rights will be large enough to challenge margins and will largely scale with growth. Which means no truly meaningful scale benefits. So here are a few alternative ways in which streaming margins might be improved:

• Doing a Netflix: Because Netflix owns much of its own content, it is able to use its recommendation algorithms to ensure that content over-indexes, improving margin. It also amortizes costs against those content assets to help it register a profit. Spotify could do the same but is unlikely to do so anytime soon. It cannot afford to antagonize its major label partners, each of whom has a UN Security Council type power of veto (Spotify would falter if any one of them pulled out). Someday, Spotify probably will become a label, though not in the way most people would understand the term. However, it will wait for more scale and confidence before flicking the switch on that strategy.
• Ecosystems: Apple has long demonstrated the value of competing right across value chains. Now Amazon is following suit (e.g. Amazon Video covers rights, infrastructure and distribution). Exercising control across the value chain gives a company more places to extract margin. Perhaps Alibaba or Tencent (or some other Chinese giant) could buy a major label and a streaming service? Access Industries is already on this path, wholly owning WMG and more than half of Deezer (though there doesn’t seem to be much in the way of dots being joined yet). And then the wildcard is a streaming service becoming so big that it can buy a major or a collection of big indies. Or of course Apple deciding to any of the above. Should this feel like wild conjecture, do not forget that it was not so long ago when an ISP (AOL) bought WMG, and a water and sewage conglomerate (Vivendi) went on a media company acquisition spree and bought UMG.
• Ancillary revenue streams: The most pragmatic solution though is not a silver bullet, but instead a blended strategy of new revenue streams. These can range from B2B (e.g. Spotify selling its data to live companies like Live Nation and AEG to help them get more cost effective with better targeting), through premium user add-ons to new formats such as limited capacity, pay-per-view artist live streams.

Spotify played the starring role in streaming’s biggest year yet and looks well on track to do the same in 2017. But at some stage, the losses need to narrow. The industry needs to help ensure this happens, unless it wants the market to end up being dominated by companies that simply do not have to have streaming turn a profit because they are making money elsewhere.

 

Why Netflix Can Turn A Profit But Spotify Cannot (Yet)

Having just celebrated its 10^th (streaming) birthday, Netflix followed up with a strong earnings release, announcing 5.8 million net new paid subscribers in Q4, sending its share price up by 9%. This wraps up a stellar year for Netflix, one in which it doubled down on original programming and delivered acclaimed hits such as Stranger Things and The OA, shows that don’t fit the traditional TV mould. In fact, Stranger Things was turned down by 15 TV networks before finding a home at Netflix and The OA’s oscillating episode lengths (from 1 hour 11 mins to 31 mins) would have played havoc with a linear TV schedule (not even considering its mind bending plot).

Netflix closed 2016 with 89.1 million subscribers and the temptation to benchmark against Spotify’s equally strong year is too strong to resist. Spotify (which celebrated its decade in June 2016) closed the year with around 43 million subscribers, 48% the size of Netflix. But a closer look at the numbers tells another growth story.

Read the full post on the MIDiA blog by clicking here.

The 2 Spotify Charts You Need To See

Tuesday’s media scrum around Spotify’s financials illustrate that whatever ground Apple and Tidal may have made in recent months, Spotify clearly remains the poster child / bellwether for streaming. The stories oscillated between the broken nature of the underlying economics to how streaming is the future of the music business. Both are true. But a closer look at the numbers reveal some even more important findings.

Rights costs are Spotify’s Achilles Heel. Rights and associated costs accounted for 83% of Spotify’s 2015 revenue, up from 81% in 2014 and this resulted in a dramatic fall in Spotify’s gross margin per user: down from $4.20 in 2013 to $3.45 in 2015. This is particularly challenging for a model with already wafer thin margins. A number of factors underpin this decline:

• Discounted promotions: Promos such as the £0.99 for 3 months have supercharged Spotify’s growth for the last 18 months. But as labels only contribute part of the cost this means that Spotify loses more margin with every new promo user
• Advanced label payments: When Spotify strikes its licensing deals with labels it makes advanced payments and guarantees based on its expected growth. This means that for a growth stage company like Spotify, booked rights costs will always be higher than current booked revenue. This has obvious cash flow implications. Also, should Spotify’s growth slow and it miss those targets, it will still have to pay the monies guaranteed to labels, at which point the rights costs share will rise even further
• Publisher rates: Over the last couple of years, music publishers have been asserting their role in the digital music value chain, pushing for more equitable rates. The net result is that publishing rights costs can now range up to 15%, depending on the deal, up from a low of 10% in some cases. This upward momentum will continue, and as labels aren’t decreasing their rates, it means less margin for Spotify and other streaming services

As Spotify edges towards an IPO it is doing everything within its power to get its house in order. It is investing in video to show Wall Street it is attempting to lessen its dependence on the labels and it is improving is cost ratios virtually everywhere else in its business, other than rights. Between 2013 and 2015, the Average Cost Per User (ACPU) for Research and Development fell from $2.12 to $1.61 and for Marketing it fell from $3.23 to $2.77. But Rights ACPU grew from $17.59 to $18.35. In fact, even in terms of costs as a % of revenues, every single expense Spotify reported fell except Rights (and Depreciation and Amortization which increased slightly). It is rising rights costs that are keeping Spotify from commercial sustainability.

There is another really important part of Spotify’s growth story: subscriber ARPU has fallen from $79.09 in 2013 to $62.30 in 2015. This is a result of multiple efforts to drive growth, including the price promos, telco bundles and student discounts. All of which are viable tactics but the fact they are necessary to drive Spotify’s growth underscore a point I have been making for years: 9.99 is not a mass market price point, and Spotify’s subscribers agree. By transforming the ARPU into an effective monthly retail price, Spotify’s average price point is now just $6.49, down from $8.24. It is about time that the music industry stopped pretending that this isn’t the reality of the market and instead starts pursuing proper pricing innovation rather than by stealth via discounting, which only serves to confuse consumers about long term value.

The music industry is in a transition phase. In such periods, the old and new worlds co-exist and collide. There are statistics that both sides of any argument can hold up in their defence, in fact they can often hold up the very same numbers to support opposite perspectives. Similarly, the comparisons you chose to benchmark with, can paint entirely different pictures. Such is the nature of transitions of human and business behaviour. For example, 83% of Spotify’s gross revenue going to rights is clearly too high and unsustainable, yet $0.00098 per song going to artists is also clearly too low and unsustainable. Something needs to give, for both ends of the value chain.

Maybe if/when Spotify gets to 50 million subscribers it will feel it has enough clout to compel rights holders to rethink licensing economics. Perhaps it will take Spotify getting to a 100 million to make that happen. Perhaps it will never happen. But if it doesn’t, the economics of streaming will remain so broken that only companies with ulterior business objectives will remain viable players, enter stage left streaming’s Triple A: Apple, Amazon and Alphabet (Google). The labels need to ask themselves whether that is the streaming future they want…