Spotify's Big Pivot

In its most recent Investor Day, Spotify revealed for the first time its podcasting revenue, detailing that the company has generated over $200mn in 2021, but it’s not revenue they’re after. Spotify is growing beyond its music roots and transforming into a streaming platform for all audio content. Why? restructuring costs.

“Over the next 3-5 years we believe podcast margin should top 30% and our long term view is that this business could reach 40-50%” -Daniel EK, Spotify CEO

Spotify currently generates a little less than 30% and expects its music gross margin to top out at 35% over the long-term, that’s a 15% margin differential with podcasts. Currently though, the company’s far from profitable, distributing music at an operational loss and continuously weighed down by the high content costs.

"We're paying too much in royalties, but we can probably offset that with the user growth" - Overheard at Spotify HQ (circa 2012)

Spotify bears a striking resemblance to SiriusXM and its tumultuous path to profitability. Going back to 2009 SiriusXM, had emerged as a newly formed entity following the merger of Sirius Radio and XM the newly formed entity had not been generating any profits since its inception in 2000. And for investors hoping for a late-stage dotcom play, the stock had no momentum. It had been trailing the S&P by 20% that year and looked like not even a merger could save it.

A few months after the merger, the new combined entity went bankrupt, thanks to a massive drop in car sales during the 2008 financial crisis. But after a chapter 11-repellent convertible loan from Liberty Media, the satellite radio company would re-emerge as a profitable business. All thanks to the company’s stranglehold on the satellite radio niche that would give them the pricing power to pass on royalty fees to users. Following the merger, SiriusXM was able to slash content costs all while maintaining its subscriber base. Between 2008 and 2011, the company brought down its content costs as percentage of revenue from a high of 39% all the way down to 25%. On March 9th, 2009, SiriusXM's 4th quarter earnings results came out, with one thing made clear, SiriusXM recorded its first positive adjusted operating income and the battle with royalty fees is all but over.

Unlike SiriusXM, Spotify operates in a highly competitive environment and is forced to pay much higher royalty fees. Spotify doesn't break down its cost of revenue, but it primarily includes royalty fees paid for streamed music as well as amortization of podcast content and, constituting 74% of revenue as of 2020.

But it's not just competition and royalty fees hurting Spotify’s cost structure, they also have new and unexpected challenges to deal with. Here’s an unusual one: music itself has changed the past couple of decades in a way that makes it more expensive for Spotify to operate. Specifically, songs are getting shorter. A typical Billboard Hot100 track would on average be 4 minutes in length in 2000 is now likely to be only 3.1 minutes. While Spotify’s payments to rights holders depend on how much revenue Spotify generates, the number of subscribers and the amount paid per subscriber, artists are still paid per stream and the cost per stream matters. As users are increasingly able to stream more songs in less time, this creates pressure on Spotify to keep per stream rates stable.

Song durations have been trending
downwards for the past two decades

Trend is expected to continue in the coming years

Faced with these dynamics tugging at Spotify's cost structure, the company began making moves in the podcasting space. As positioned by Spotify's CEO:

“...we found so many times before is that the more people engaged, the more likely they are to pay. And the same is true with music as it is with podcasts, too. It’s really all about getting them onto the platform and starting to expose them to this entire ecosystem of creators and amazing content that we have on the platform. And once that happens, we know people eventually will convert into paying customers." -Daniel Ek

I couldn’t buy into the argument that podcasts are a way to convert free users to paid ones. This is not about premium revenue or a honeypot for user acquisition. It's rather a honeypot for spiralling costs. Spotify is in a marginal cost trap; an issue that plenty of analysts have gone over, so I'm not going to harp on it. I am however interested in how Spotify is trying to get out of that trap.

A marginal cost trap is essentially a squeezed operating margin, which leaves a company with one of two choices; raise prices or cut costs. SiriusXM did both, but Spotify can't do either as they don’t control their costs and they can’t afford to lose out to competition offering lower pricing. When a company can't expand margins on its existing product offering, there's only one thing they can do and that is introduce a new product.

I grew fond of the podcast format back when the word did not register with most people until you tell them "Oh, it's that purple icon on your iPod Touch", since then they've grown into their own media category and garnered the attention of the tech incumbents, with Spotify taking center stage. The company's $250mn acquisition of Gimlet, along with other acquisitions and licensing deals would at first glance look like a company investing in an audio medium that's caught renewed attention. But on closer inspection, it was made clear to me how Spotify is leveraging podcasts to anchor their new cost strategy; one that rests on capping marginal costs from royalties through altering the revenue mix.

I wasn’t surprised at the number or frequency of shows and episodes released by Spotify in 2020, but I did notice that Spotify's episodes are getting shorter. At first, I thought that doesn't make sense; wouldn't Spotify want longer episodes with more ad breaks? But then I realised that the goal is not just the podcast ads. Spotify wants to onboard casual listeners and a 1-hour episode is kind of intimidating, but a 20-minute episode is more accessible. It's also in line with changing tastes; in an age of ever-increasing distractions and ever-shortening attention spans, users want to get to a song's chorus faster and they want to hop from one podcast to the next.

Spotify drastically shortened
its original podcasts

Spotify originals averaged less than 30 minutes in 2021

This isn’t the whole picture though. Data could be swayed by short, yet more frequent, low budget podcasts, listenership might be concentrated in a few popular podcasts with a long tail of less relevant shows. I also don’t have insight into Spotify’s budget allocation in podcast production, so Spotify might still be favoring longer-form podcasts and they might prove to be more popular with listeners, but the overall picture shows that they’re catching up to the ever-shortening attention spans in an age of content abundance.

Gauging the impact of these insights on Spotify's financials is no easy feat given the lack of any granular metrics beyond the headline figures. To assess Spotify's strategic shift to podcasts, I needed to get an idea of what the content split is right now, so here's what we know:

Based on Spotify’s latest disclosure that 7% of listening hours come from podcasts, I am assuming that the same is true on average on a per user basis. But Spotify’s podcasts business is far from profitable. According to Paul Vogel, Spotify’s CFO, the company is losing 57% on a gross level in the podcast business. While on the surface this seems steep, I am more interested in what Spotify will look like once it realises the full potential of podcasts and their positive impact on margins.

In order to assess the impact of podcasts on Spotify’s financial situation, I opted for a bottom-up approach to the problem. Financially, Spotify was able to hit operating margin profitability in 2021, albeit just barely (1% EBIT margin), but continued to incur a net loss. Management’s end-goal is clear; get more users to consume podcasts, monetize more podcasts and generate higher gross margins. But how much podcast content do users need to consume for that strategy to pay off? Here’s what I could gather and piece together

With these figures at hand, I ventured into trying to picture what the content split looks like for every user, so I synthesised a dataset to reflect the inverse relationship between the content split and spotify’s direct costs/hr. The dataset -reflected in the below scatter plot- shows a sample of users’ content split in a 1 hour streaming session, with each marker representing a 1 hour user session. Stretching podcast consumption for the average user to 35% of all content, brings down costs by more than 30% and lifts Spotify’s operating margin by more than 20%, that would put Spotify in line with Netflix’s operating margin (21% in 2021). This is however a hypothetical analysis with a lot of caveats. One being that Spotify is only monetizing 14% of its podcast inventory, while my analysis here assumes all of the margin gains are coming from cost optimizations. In reality, Spotify could hit the 20-something operating margin levels just by doubling podcast consumption to 14% with a lot of upside.

Podcast streaming lowers
Spotify's cost of revenue

The average Spotify user spends 3-4 minutes listening to a podcast in any given hour

Spotify has also been eyeing verticals beyond music and podcasts, recently closing its acquisition of Find a way, an audiobook distributor with more than 325k titles up its sleeve. This is another margin play, with Daniel Ek expecting the business to generate margins above 40%. Spotify’s been at this for a long time to know that streaming music doesn’t pay and diversifying away from royalties is the right way to go, but it’s far from easy. Just last a few months ago, Spotify announced it’s shutting down Spotify Studios, its namesake production studio. Creating good, resonating podcasts is hard, but convincing users to listen is even harder.