How Daniel Ek runs Spotify

I’m obsessed with Spotify. It’s not just because I love the product and find it thoroughly indispensable.  It’s because I’m fascinated by the company, it’s vision and how it’s leadership makes decisions. In an extremely cut-throat music streaming market, Spotify seems to be thriving. Their Q3 2019 results announced last month sent the stock climbing. The business is growing at 28% year-over-year on top of a billion dollar run rate, posting revenues of $1.88 billion. They beat estimates on gross margins, which have consistently been cited as a major weakness in their business model, and continue to generate free cash flow. 

Spotify’s success, in addition to being a very disciplined business, is in large part down to their product vision. They understand how instant access to the world’s music from any device is changing our relationship with music better than anyone else. If iTunes unbundled the music album, Spotify is at the forefront of another powerful bundling and distribution mechanism – the playlist. By focusing on the playlist experience very early on, Spotify was able to leverage their user’s listening habits to create the richest data set for music on the planet. When I first heard their Chief R&D Officer Gustav Soderstrom talk about this earlier this year, my mind was blown at the foresight and vision that went into creating this product experience. They’ve used this data set to build differentiated products like Discover Weekly that keep users hooked to the service. Revenue from premium subscribers grew 29% year over year and the number of premium subscribers grew from 87 million to 113 million. 

That’s why when I saw Daniel Ek was a guest on Invest like the Best, I sat down and listened to the entire podcast in one go. I found myself taking notes as I was listening because of how knowledgeable and insightful the conversation was. 

Here are a few ideas from their discussion that really stood out to me:

Daniel thinks globalization, automation and digitization are the 3 macro trends shaping business today.

According to him, the implication of these trends is that there will only be two types of business models in the future — those that operate at a massive global scale and others that go deep within a niche. The middle has completely dropped out. This view is similar to the one articulated by Josh Wolfe when he talks about there only being the ‘megas and the minnows’ anymore. Ben Thompson also made a similar observation a few years back when he discussed the Internet Rainforest

The jobs-to-be-done framework is at the heart of a lot of his product decisions. 

Clay Christensen first developed this framework to complement his theory of disruptive innovation as a way to help companies understand how to avoid being disrupted. Daniel made a couple observations here that I hadn’t heard before. Even if the job to be done is the same or similar, it may not make sense for the same company or product to do that job if the user persona is fundamentally different.This is why Spotify decided to launch a separate app for kids.

From the press release:

Beyond the content, the entire Spotify Kids user experience looks and feels different from the Spotify app. And that’s intentional. It’s built for kids, with their specific cognitive skills in mind, and exudes a fun, familiar, playful, and bright atmosphere. This look and feel also varies by age group—for example, the artwork for younger kids is softer and character-based, while content for older kids is more realistic and detailed. 

The second observation he made was complementary to the first in a lot of ways. What might seem as distinct jobs to be done initially can turn out to be similar once you consolidate the supply side. Spotify acquired Gimlet, Anchor and Cutler this year because they don’t see music and podcasts as different jobs to be done. Their market is audio, more precisely the ~3 billion people that want to ‘hire’ audio experiences to do the job of entertaining and educating them. By integrating podcasts into the core Spotify product experience, they hope to become a one stop shop for their users audio needs. I wonder if they’ll add audiobooks next. 

Spotify’s commoditized complements.

Spotify’s complements, or complements to audio experiences, which tend to be products like bluetooth headphones, phones, smart speakers and data services are only getting cheaper and better. Gwern Branwen explains how this works in this excellent post. Essentially, there are two types of products in the market when viewed relative to your product – substitutes and complements. Substitutes compete with your product for the same job to be done. Complements help, or might even be necessary for people to buy your product. 

Why becoming a label doesn’t make sense for Spotify.

A lot of people think Spotify needs to become a music label to reduce or eliminate the wholesale transfer pricing power the labels have over them, but he doesn’t think that’s the right long term approach. It’s not just that he wants to avoid upsetting the labels (which he does), but that the music and audio market is fundamentally different from video. Artists today make most of their revenue through ticket sales, which means their music needs to find the widest distribution possible to drive people to their concerts. Signing artists directly and restricting their music distribution to the Spotify platform would be severely detrimental to their business model. While It is tempting for Spotify to emulate the Netflix approach, that would amount to reasoning by analogy and not by first principles.  

He does see Spotify investing more and more in original podcasts and making them exclusive to Spotify Premium subscribers, but they need to think very carefully about the genre and format because people expect podcasts to be free and most podcast creators monetize via advertising. He thinks exclusive content makes more sense for genres like true crime, documentaries and short stories for example. VICE is developing three original podcasts exclusively for Spotify. 

How Spotify got into the music streaming business. 

Similar to how Apple convinced the labels to partner with them on iTunes, Spotify could license music from the labels in Sweden because their business had been completely decimated by piracy. This lead to an interesting discussion on whether identifying ‘weakened or fragmented’ suppliers can be a viable strategy to get a wedge into an industry and build a business. Without piracy, Spotify may not exist.    

Working at Spotify lets you see culture emerge in real time.

Daniel tracked the rise of Reggaeton into a global sensation over the past five years. Today he’s seeing a resurgence in gospel music, especially in Brazil. Is that why Kanye’s new album is all gospel? 

Measuring product performance at massive scale.

The larger you become the more important it is to move away from looking at averages and towards looking at distinct segments within your business. Spotify probably has hundreds of artist and listener archetypes. Metrics that use averages across their user base tend to not be useful in helping you understand product performance and the impact of your decisions.