May 21, 2023 · 1h 24m · lennys-podcast
The science of product, big bets, and how AI is impacting the future of music | Gustav Söderström
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth interview, Spotify Co-President, CPO, and CTO Gustav Söderström breaks down the transition to generative AI, the organizational evolution beyond the famous squads model, and the rigorous frameworks behind Spotify's product strategy and design decisions.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Lenny holds 17.1% of the talking time here. How this is scored →
Lenny as informed peer 5.1 Guest teaching 5.9 Guest disagreement 1.6 Lenny pushing back 1.3
speaking balance: gold is Lenny, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:06:35 Rejecting the idea of product as inexplicable art
Gustav forcefully challenges the common belief that product intuition is art or magic, arguing that leaders must explain their decisions scientifically.
Hardest push from Lenny ▶ 34:33 Lenny pressing on what VP autonomy looks like in practiceLenny pushes Gustav to move past theoretical organizational charts and explain exactly how authority is delegated to leaders like the VP of Podcasting.
Biggest teaching moment ▶ 51:00 Deconstructing user intent: recall versus discoveryGustav breaks down the exact quantitative and behavioural reasons why Spotify's home redesign triggered backlash by confusing a 90% recall surface with discovery.
Lenny holds their own ▶ 43:12 Lenny connecting organizational extremes to business valueLenny synthesizes Gustav's analysis of Apple and Amazon by pointing out that staying at organizational extremes may offer clear strategic advantages over being stuck in the middle.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Lenny as informed peer | Guest teaching | Guest disagreement | Lenny pushing back | Why |
|---|---|---|---|---|---|---|
| Gustav Söderström’s 14-Year Leadership Journey at Spotify | 4 | 3 | 1 | 1 | Lenny opens with an overview of Gustav's 14-year career progression. Gustav recounts transitioning from mobile entrepreneur to product/CTO and now co-president, reflecting on how scaling felt like joining a new company every year. | |
| Using Podcasts for Internal Culture and Creator Empathy | 5 | 4 | 1 | 1 | Lenny explores Gustav's internal and external podcast series. Gustav explains how internal storytelling broke down leadership opacity and built empathy for creator workflows. | |
| From Recommendation to Generation and Spotify's AI DJ | 5 | 8 | 2 | 1 | Gustav delivers a detailed framework on the three eras of digital media: curation, recommendation, and generation. He details the concept of fault-tolerant user interfaces and how AI DJ solves the zero-intent listening use case. | |
| The Impact of Generative AI on Music and Royalties | 5 | 7 | 2 | 2 | Lenny asks about AI-generated tracks clone catalogs. Gustav contextualizes AI as a musical instrument analogous to digital audio workstations (DAWs) like Avicii used, explaining why unique originality remains difficult. | |
| Why Great Products Must Pull Off a Magic Trick | 6 | 4 | 1 | 1 | Lenny and Gustav discuss the necessity of product magic tricks, connecting Spotify's AI DJ voice synthesis to ChatGPT's instant viral traction. | |
| Moving Away from the Spotify Squads and Tribes Model | 6 | 8 | 3 | 2 | Lenny brings up Spotify's renowned squads and tribes model. Gustav dispels the mythology, explaining how hyper-autonomous junior squads created organizational friction and heat, necessitating larger teams and VP-level autonomy. | |
| Organizational Spectrum: Amazon, Apple, and Spotify's Strategy | 6 | 8 | 2 | 2 | Gustav contrasts Amazon's decentralized hard-API two-pizza teams with Apple's centrally organized functional hierarchy. He explains why Spotify chose a centralized UX model to integrate diverse content verticals. | |
| Sponsor Segment: Eco Cash Back and Rewards App | 5 | 9 | 2 | 2 | Following the Eco sponsor break, Lenny asks about the controversial home feed redesign. Gustav provides an in-depth retrospective on misdiagnosing user intent by flipping the 90% recall / 10% discovery balance, explaining A/B testing psychology under public criticism. | |
| The 10% Planning Rule and Socratic Decision-Making | 5 | 7 | 2 | 1 | Gustav explains the 10% planning time rule and discusses pushing teams toward Socratic debate, arguing that product strategy should be treated as 100% science rather than inexplicable art or magic. | |
| Product Analogies, Short-Termism, and Succession | 4 | 3 | 1 | 1 | Lenny brings up the Swedish 'peeing in your pants' idiom for short-term fixes and asks about Swedish business representation in Succession. Gustav shares humorous cultural context. | |
| The Future Roadmap for Spotify and Podcasting | 5 | 5 | 1 | 1 | Lenny asks what lies ahead for Spotify podcasting. Gustav highlights audience discovery and monetization infrastructure for creators, alongside cross-device ubiquitous playback for listeners. | |
| Lightning Round: Book Recommendations and Product Rituals | 5 | 5 | 1 | 1 | In the lightning round, Gustav shares an expansive reading list covering strategy, physics, and information theory, and explains the origin of the 'Think It, Build It, Ship It, Tweak It' product lifecycle framework. |
Statements from this episode (21)
Söderström: Spotify gave no special platform promotion to his podcast
“The podcast did well and no, we did not give it our own sort of promotion. I had to compete as everyone else”
Söderström: Generative AI shift will force companies to rethink products
“And I think what we're entering now is we're going from your curation to recommendation to generation. And I suspect it will be as big of a shift that you will eventually have to rethink your products.”
Söderström: Product interfaces must be designed around ML error rates
“The quality of your machine learning, if you're going to have a single play button needs to be literally a hundred percent or zero prediction error. And that's never the case, right? So let's say that you have, you know, a one in five hits, four out of five th…”
Söderström: Spotify's AI DJ succeeded by minimizing voice chatter
“And for the AI DJ specifically, Another principle is to try to avoid this urge of just wanting to show off the technology and have this voice, like talk and talk and talk and talk. You have to remember that people came there for the music. So the principle for…”
Söderström: The boundary between AI music and real music does not exist
“There is this distinction between AI music and real music that, that doesn't exist for sure. Very talented real musicians are using AI. To get better and to help create new ideas. So that distinction doesn't really exist. It's all going to be AI. The question …”
Söderström: AI makes generic music easy, making true originality harder than ever
“Actually, what these models do is Because they, because they've been listening to a lot of music, they are very good at doing something that sounds very similar to what already exists. Actually being original is very hard. And from one point of view, as it now…”
Söderström: Generative AI could make the music industry bigger than ever
“What would actually happen after piracy is that the music industry got bigger than ever, not, not just as big, but bigger than ever. And I think that could happen with this technology as well, but we're right in the beginning.”
Söderström: Making products feel magical requires scoping down and performance tuning
“But I think getting to that point and iterating a product to the point where it feels like magic the first time is very helpful. And it's often a question of, Just getting the performance to certain levels, scoping down, removing things. There's a lot of fine …”
Söderström: Spotify's famous "squads" organizational model is inefficient at scale
“It's become a little bit of a cost to us because people still think that we organize that way. And it's not a very efficient way of being organized at this scale, or maybe even if you started over right now, because we've learned more.”
Söderström: Spotify abandoned 7-person agile squads for 14-engineer teams
“As you grow the company, scaling in increments of seven engineers just creates a ton of overhead. So obviously our teams now tend to be much bigger, maybe, maybe two, three times that at least per like manager to maybe have like 14 or something instead of seve…”
Söderström: Pushing autonomy to junior leaf nodes produces conflicting strategies
“If you put autonomy very far towards the leaves of the organization, and then also if you combine that with having a very junior organization, which we did back then, there's a fair chance that you're just going to produce heat. You're going to have a hundred …”
Söderström: Spotify concentrates product and strategic autonomy at VP level
“The autonomy level in Spotify now tends to be quite high at the VP level. And then lower around those levels.”
Söderström: Decentralized team autonomy risks shipping org chart complexity to users
“You get the benefit of speed. But you get the drawback of kind of shipping your org chart and shipping complexity to the end user.”
Söderström: Spotify routes music, podcasts, and audiobooks through one recommendation org
“So we've chosen more of the central centralized model. Where these different sort of vertical businesses, if you think about it, the music business, podcasts, audiobooks business, they have to go through a single recommendation organization, because that's ano…”
Söderström: Spotify homepage activity is 90% recall and 10% discovery
“If you look at what people do on Spotify's homepage, the current one,
It is almost 90% what we call recall. So it is either getting to a session that you're already in or a specific playlist that you know you want to get to, or at least a specific use case. S…”
Söderström: Content recall requires dense UI while discovery requires large real estate
“When you do recall, you want the dense UI. With many items on screen, because you know what it is you're looking for, so you don't need a lot of real estate. When you're doing discovery of new things, you want a lot of user interface, a lot of pixels, and you …”
Söderström: Redesigns carry higher risk than features because participation is involuntary
“There are two fundamentally different types of product development. One is designing a new feature. It is hard to make it, but it's voluntary for people to use. So you do the AI DJ. Some people love it. That's fine. If you don't like it didn't make it worse fo…”
Söderström: Novel product paradigms require large MVPs to avoid false negatives
“The hardest thing about this is if you're trying something completely new, The MVP needs to be very big so you can build a new UI, but if you didn't do algorithms for a single item feed, you can't tell if it was the right idea, but poor machine learning, right…”
Söderström: Spotify runs on six-month execution cycles with two weeks of planning
“We work in, in sort of six month increment. So we try to spend two weeks planning and are roughly successful.”
Söderström: Short-term fixes are like peeing in your pants in winter
“The saying is that's like peeing in your pants in, you know, in, in cold weather. It feels really warm and nice to begin with. And then after a while you start to regret it. It's about being, being short term, basically.”
Söderström: Breaking habit bubbles is harder in podcasting than music
“We talked about it mostly for music, music discovery, but the same problem, it's the same problem and even harder for podcasts. So we're still focused very heavily on helping spot, on helping podcast creators find more audience. This is, like I said, it's even…”