$SPOT: The $GOOGL of Audio in the Making
SPOT 0.00%↑ at its simplest, is a music retailer. 75% of the money it makes today is kept by labels. For this reason, $SPOT has a relatively low gross margin of 25%.
$SPOT has a strong cultural dominance (way ahead of its competitors), is very well managed and is getting smarter every day (AI).
The above 3 components are brewing an intelligent global audio network, akin to the Google search engine, but for audio. I believe that higher margins will naturally emerge as $SPOT continues to move in this direction. Q2 2021 is showing some early signs of this happening.
As margins expand and MAU growth continues (not necessarily at breakneck speed), $SPOT´s valuation will naturally increase through time.
Low Margins, Low Multiples
Spotify at its simplest is a music retailer. It mostly buys music from labels and then resells the music to listeners, through a subscription service. Roughly 75% of the money that listeners pay to Spotify ends up going to labels as royalties, whilst 25% of the money is kept by Spotify. As you may be thinking, this amounts to quite a high cost of revenues for the company, with a gross margin of around 25%. This, together with a tough competition from $AMZN, $AAPL and $GOOGLE and some MAU (monthly active users) estimates misses, is what all the discussions about Spotify´s valuation are about today. The market is reluctant to assign generous multiples to Spotify primarily for these three reasons.
Labels, Independent Artists and Royalties
Before we go any further, allow me to dissect the raw material Spotify merchandises with, since it is foundational knowledge. Rights to a particular song are made of two parts: royalties for the print and royalties for the recording. The first go to whoever composed the song, whilst the second go to whoever recorded it. So for example, if an artist performs a song at a concert, royalties will be paid to the licensor of the print, but not to the licensor of the recording, because the song is being performed live. If the song is played on the radio or on Spotify, then recording royalties are due, as well as print royalties.
Labels usually are beneficiaries of the second type of royalty. Labels help artists produce their music and get it out into the world. Labels invest capital to make this happen and therefore carry the risk/reward of the artist in question turning out to be popular or not. Today Warner Music Group, Sony and Universal Music Group account for around 70% of the music on Spotify, whilst the remaining 30% is accounted for by Merlin (another label) and independent artists. Independent artists skip working with labels and publish their songs directly on Spotify - making 6x the money vs going with labels and presenting $SPOT with lower cost of revenues.
Spotify therefore greatly depends on labels. However, in the case of Warner Music Group, $AAPL and $SPOT alone account for 27% of their sales, so it looks like the dependence is mutual. Further, music streaming is the only part of the label´s business that is actually growing.
Before going public in 2018, Spotify struck a deal with labels that resulted in higher margins, hitting 20% in 2017. Since then, the gross margin has inched its way up to 26.39% in TTM. A similar and sudden margin hike is unlikely to happen again or at the least, a bull thesis should not count on the labels giving $SPOT a free meal down the line. Beyond the deal, it seems likely that gross margins have grown from 20% to 26% over time due to an increased conversion of ad supported users (with a gross margin of 11.3%) to premium users (with a gross margin of 30.8%).
The question going forward is, is $SPOT stuck with labels and hence with relatively low gross margins?
Spotify is Increasingly an Audio Retailer vs Purely a Music Retailer
$AMZN started off by selling books and for a while (from 1994 to 1999, sources cite) it must have looked like that was all it was going to sell. In hindsight, what Jeff Bezos was actually doing is kick starting a platform to sell anything online. Essentially, the mechanics that are required to sell books are the same than the mechanics to sell hats, TVs or pretty much anything else. I believe something similar will happen with audio as a category. Once the infrastructure for you to listen to music on Spotify is in place, the marginal cost of feeding you any other format of audio content, such as podcasts, audio books, guided meditations, e-learning courses and so forth, is relatively low.
It currently looks like $SPOT just plays music, but I think that pretty soon we are going to see it playing anything that our ears want to listen to. Podcasts have recently become quite popular and in my opinion, are something akin to books 2.0. Providers of other types of audio are not in the entrenched position that labels find themselves in and by definition, this would tend to translate into better unitary costs and hence margins for Spotify. Getting to a scenario where other types of audio are widely available and consumed within the platform with better business economics for Spotify, however, is likely to be a long journey. I believe $SPOT has 3 main properties that give it a good shot at getting there:
A constantly improving AI engine
The above 3 components are the synthetic summary of quite a number of complexities within the business and they all retro-feed each other. $SPOT is top of mind in the audio space, is very well managed and increasingly knows better than anyone who wants to listen to what and when. This is yielding an audio network which will eventually have the intelligence to run rings around incumbents in the audio industry in general, planet-wide. At this stage, I believe higher gross margins will come naturally. Let´s dissect this.
$SPOT is the go to music streaming service for most people. I have purposefully asked folks that I know about why they do not use Apple or Amazon Music vs Spotify and the answer is “I never thought of it”. You may have had a similar experience, but luckily we have data to shed clarity on this:
Once something becomes top of mind in the internet for a given vertical, if well managed, it becomes quite hard to displace. This seems to be Spotify´s case, which is quite far ahead of the competition in terms of paying subscribers. I find this more indicative than any other metric in terms of cultural dominance, since it shows us where the money is going. Zooming out, I find the market’s disappointment with the MAU miss short sighted.
In my opinion, due to the social nature of content in general, this dominance (network effect) is actually Spotify´s moat. This moat is likely to continue compounding through time, as network effects continue bringing in more users and creators, spilling over into new audio categories. Spotify is laser focused on music/audio, whilst its competitors are not. I believe this focus explains much of the advantage Spotify has over its competition, since it is able to evangelize stakeholders in the industry.
I mention this a lot when I discuss brands that have caught my eye, but Spotify is another brand which has alligned its what, how and why, yielding a gravitas which $AMZN, $AAPL and $GOOGL will have a tough time catching up with in this space. Whilst they can slash prices in their music services and then funnel users into their core businesses, I just don´t see them winning here. People know Spotify loves music, whilst maybe $AMZN, $AAPL and $GOOGL not so much. This love for music is palpable in the product and in the entire user experience and quite intangibly, I believe, is the company´s fuel for growth. This recorded live event speaks for itself.
Firstly, the incentive of the management seems to be alligned with that of investors. Daniel Ek, CEO, owns 8% of the equity and had a salary of 480k$ in 2020. Martin Lorentzon, Director, owns 11% of the equity and had a salary of 380k$ in the same year. On a qualitative level, the management picture looks to me like the opposite that I have observed studying $GE´s 20 year long downfall. In $GE, management was incentivized to produce earnings that satisfied the market’s short term expectations, which eventually led to the destruction of value. In Spotify, the management is all about building value and it shows in a whole range of ways, beginning with the alignment of their incentives with the creation of value and them not throwing themselves around to nail earnings every time. Over the past year, $SPOT´s MAU miss led to some negative price action, but I did not see the management losing its way to “fix” that - they just carried on executing.
Furthermore Spotify, to end users, is ultimately an app. Apps are unbelievably hard to get to work well at scale and a churn rate of 4% screams good management. A low churn rate is actually the result of a good overall customer experience, which is itself the result of every little component of the company working well and therefore a good management. On the internet, cultural dominance eventually has legs of its own, but until you get there, it requires excellent product iteration skills, in order to produce something that people love and want to share with other people. In this case, a low churn rate points to an organization that works well.
Moving into other audio categories, in my opinion, is not something that cultural dominance or the AI engine will do on their own (yet). It requires intelligent investments to get there. In my opinion $SPOT´s acquisition track record is looking good and to me, the Joe Rogan podcast acquisition stands out particularly. $SPOT got the rights to the podcast for dirt cheap, at 100m$. Over the last year, we have seen $SPOT making acquisitions to move further into the podcast and live audio spaces. Time will tell how they turn out, but qualitatively I´m not perceiving bad decision making in this area at all.
A Constantly Improving AI Engine
Much of my perception regarding $SPOT´s recommendation system comes from studying neural networks extensively. Music files are simply 1s and 0s put together. Neural networks are able to learn what particular patterns of 1s and 0s you like to listen to, to a degree of precision above human performance. Having learnt what patterns you like, a computer can find other similar patterns of 1s and 0s that you may like aswell.
At the individual level, this does not seem like something too powerful, but when we are dealing with 300M+ monthly active listeners, this sort of intelligence begins to capture the intelligence of the overall market. With this kind of intelligence, $SPOT may be able to predict things like what patterns of 1s and 0s make a super hit in South Korea in a particularly sunny month of May, or in Spain in a colder winter than usual, for instance. The more listeners, the more intelligent it gets, specially relative to its competitors.
On a similar note, it may be able to learn what parts of the podcasts you listened to this week you liked the most. In this way, it may be able to recommend new podcasts that you may like or perhaps, let an emerging podcaster know what topics are likely to drive engagement this week. The value that Spotify can drive with this sort of intelligence is pretty much endless within the market and as a result, I believe margins will naturally emerge. Spotify´s value proposal here is radically different to “we play music” and begins to be “we know audio better than anyone”. Margins are likely to emerge not just from better costs of revenues, but from being smart enough to add incremental value to participants in the market.
Additionally, AI is likely to play out exponentially in the audio industry. $NFLX, for instance, deals with video files, which by definition have way more dimensions than do audio files. At any given point in a video there’s the audio component plus an entire evolving three dimensional space, with an infinite number of possible arrangements. For this reason, it is much harder to understand what a consumer liked about a particular video file than a particular audio file. $NFLX recommendations are rarely “wow”, but $SPOT ‘s are amazing quite often. This is just early innings. In 5 years time we are going to see $SPOT get very smart and in a further 5 years time, incredibly smart.
Unknown Specific Path
The best way I can explain it is that the above 3 components are kind of quantum-entangled with top line growth and higher margins. How $SPOT will get there specifically is unknown, but my feeling is that as Spotify continues to consolidate its cultural dominance, execute well overall and continue to get smarter, at the other side a highly intelligent global audio network with higher margins is likely to emerge.
Cultural dominance translates into user (MAU) growth, which directly enhances the network´s intelligence, because the more active users, the more Spotify learns about music / audio in general. Good management defends cultural dominance by ultimately optimizing churn and fosters the continuous improvement of the AI engine, which in turn reinforces cultural dominance, by making the platform increasingly relevant to all participants.
Eventually, monetizing the resulting value is about careful trial and error, evaluating different approaches, which a good management would be capable of doing so cost-effectively. This would involve things like further developing the two sided marketplace they have been experimenting with recently. This marketplace enables labels and artists to pay $SPOT to promote their content to users that the algorithm thinks may like it. The path is largely unknown, but all the key components are there to make this happen so long as the company continues on the current path.
The moment any of the 3 examined properties goes sideways, the bull thesis is greatly impaired.
A Quantitative Clue
Firstly, I find that the uptick in gross margin reported in Q2 2021 is quite interesting. In the graph below, this uptick is made evident by the straight red line.
According to the company, this uptick in gross margin is due to “cost of revenues efficiencies” and accrual releases. Meanwhile, podcast revenue is up 627% year over year. Today, Spotify offers mostly music and podcasts meaning that the optimization of cost of revenues is most likely meaningfully attributable to the early traction that podcasts are showing. Unless the company has experienced an uptick in the overall engagement of content produced by independent artists or has favorably modified a contract with one of the big labels (and it has disclosed neither), this is the first quarter that we may be seeing results of the company turning into an audio company vs a music company. The company has also disclosed that users who listen to podcasts then turn out to be more engaged with the platform in general, which seems to mirror what we see in the graph above.
This may also be largely attributable to the increase in prices in the UK and the US, pointing to pricing power. At this stage this is quite qualitative, but I believe $SPOT has a lot of demand elasticity to capitalize on going forward. I know a lot of people that would be happy paying 2X what they pay now for the service, specially if it included premium quality sound and things of the sort.
The company has enough cash to cover its long term debt and capital leases. I see a very non-menacing balance sheet. Definitely not buying or selling $SPOT for the balance sheet.
There is not much nonsense going on in $SPOT´s IS. R&D expenses have rarely exceeded 11% of revenue, which is very much in line with the average R&D spend of the software industry. The only other component of its OPEX is Selling G&A. OPEX overall is lean and it does not make much sense to make a case for $SPOT to get leaner, since the competitive landscape demands considerable investments in R&D to stay ahead. A simple glance over the IS statement reveals that the way forward for $SPOT is margin expansion.
Despite revenue rising considerably over the last few years, operating income remains close to 0. Per se, there is no optimization in earnings from continuing operations that can make up for this, so the way forward is indeed margin expansion.
FCF generating exhibits a similar behaviour to operating income - the company is being prudently managed but evidently needs higher margins at the top of the IS. When margins expand, the company will generate healthy amounts of FCF.
A Simple Model
With the above snapshot, we can do a very simple back of the napkin kind of model to see where the stock might be heading.
In TTM, $SPOT has had a revenue of 10,224.1m$ with an average of (320+345+356+365)/4 = 346,5m MAUs. Yearly MAU growth used to be at 40% a few years ago for $SPOT and is at less than 25% today, whilst management is guiding for 20$ for FY 2021. Let´s assume that for the next 10 years it does an average of 10% YoY MAU growth. 1.1^10 = 2.59, so in 10 years time, $SPOT would have 945.35m MAUs, or 1bn MAUs for simplicity.
With an average yearly revenue of 10,224.1 / 364.5 = 28.04$/MAU today, 1bn MAUs would equate to 28bn$ for FY2030. It would not be far-fetched at all to hypothesize a gross margin of 35% by then, with podcasts well into the mix by then hovering at a gross margin of 40% and a range of software services such as promoted content (two sided marketplace) and so forth at 90%. This would yield a gross profit of 9.8bn$ in FY2030. Today $SPOT has 191.36M shares outstanding and if we take a YoY growth of shares outstanding of 3.5%, in 2030 we get 1.41 * 191.36M = 269.8m shares outstanding.
If we assume 10% of the 9.8bn$ goes off into continuing operations expenses, we get a net profit of 8,820.0m$ between 269.8m shares, yielding an EPS of 32.78$/share. Applying today´s SPX PE (35.05), we get a price target of 1150$/share in 10 years time, or about 4.5X times today´s price. It would not surprise me to see an accelerated appreciation of the company´s valuation as it begins to clearly show further quantitative results of the above thesis in the coming quarters, perhaps compressing the time for that 4.5X return.
$SPOT today is reasonably priced at around 4 times sales, so no crazy entry price here. It is worth noting that the above model assumes no pricing power, but rather the ability to maintain prices constant from here. Factoring in the pricing power that I suspect will be made obvious in the coming year or two, the upside is considerably larger.
There are plenty, but the ones I am looking out for most are the three below.
Web 3.0 is coming at us really fast. Web 3.0 is powered by the blockchain and it is going to effectively disintermediate the transaction of value in the internet. As it refers to content creators, it is going to give them all the tools to directly interact with their audience and then easily monetize it. It is unclear to me at this stage whether companies like $SPOT, $NFLX and so forth will be able to harness web 3.0 or will be totally disrupted by it. Keeping an eye on this, as I continue to get deep into developing web 3.0 apps myself and so forth to stay up to date with the tech.
$AMZN, $AAPL, $GOOGLE
These guys can slash prices anytime and justifiably so by then funneling music users into sales of their core business. It is my opinion that they are not going to win, because they cannot really compete with $SPOT´s 100% dedication to the space, but they can cause a lot of damage along the way. Keeping a close eye on this too.
Further, $SPOT depends a lot on $AAPL´s app store. Recently $AAPL announced it is going to allow $SPOT and $NFLX to bypass the App Store by linking users to an external website to make payments.
Labels going it alone
The 3 big labels account for 70% of the music in $SPOT, which is still its main raw material and will remain a key asset going forward, no matter how well podcasts do. Labels could get together and produce their own streaming platforms a la HBO, jointly or separately. It would not be without risks for them, but it could happen and this is also worth keeping an eye out for.
If you enjoyed this article, remember to subscribe to my newsletter for more!
You can also reach me at: