0:00
The internet sort of started with curation of the user curation.
Search this video
0:01
So you you took something some good like people or books or music and you digitize it and you put it online and then you ask users to curate it.
0:09
And that was your Facebook, Spotify and so forth.
0:10
And then after a while the world switched from curation to recommendation where instead of people doing that work you had algorithms.
0:18
And that was a big change that required us and others to actually rethink the entire user experience and and sometimes the the business model as well.
0:25
And I think what we're entering now is we're going from your curation to recommendation to generation.
0:29
And I suspect it will be as big of a shift that you will eventually have to rethink your products.
0:36
We have to rethink the user interface and the experience for recommendation first era.
0:39
And so what what does that mean in the generative area?
0:42
No no one really knows yet.
0:47
Welcome to Lenny's podcast where I interview world-class product leaders and growth experts to learn from their hard-won experiences building and growing today's most successful products.
0:54
Today my guest is Gustav Söderström.
0:56
Gustav is a product legend and he's now the co-president, chief product, and chief technology officer at Spotify where he's responsible for Spotify's global product and technology strategy and oversees the product, design, data, and engineering teams at the company.
1:11
I've had Gustav on my wish list of dream guests to have on this podcast since the day I launched the podcast and I'm so happy we made it happen.
1:18
happen. In our conversation we dig into what Gustav has learned about taking big bets and what to do when they don't work out, how Spotify moved away from squads and how they structure their teams now, how AI is already impacting the product,
1:31
and also the future of music generated by AI, also why all great products need to pull some kind of magic trick, how accurately Succession represents Swedish business culture, and his hilarious analogy of being in your pants. Enjoy
1:42
Enjoy this episode with Gustav Söderström after a short word from our sponsors.
1:50
This episode is brought to you by Microsoft Clarity, a free, easy-to-use tool that captures how real people are actually using your site.
1:57
You can watch live session replays to discover where users are breezing through your flow and where they struggle.
2:02
You can view instant heat maps to see what parts of your page users are engaging with and what content they're ignoring.
2:07
You can also pinpoint what's bothering your users with really cool frustration metrics like rage clicks and dead clicks and much more.
2:15
If you listen to this podcast, you know how often we talk about the importance of knowing your users.
2:20
And by seeing how your users truly experience your product, you can identify product opportunities, conversion wins, and find big gaps between how you imagine people using your product and how they actually use it.
2:33
Microsoft Clarity makes it all possible with a simple, yet incredibly powerful set of features.
2:37
You'll be blown away by how easy Clarity is to use and it's completely free forever.
2:41
You will never run into traffic limits or be forced upgrade to a paid version.
2:46
It also works across both apps and websites.
2:50
Stop guessing, get Clarity.
2:53
Check out Clarity at clarity. microsoft. com.
2:58
This episode is brought to you by Eppo.
3:01
Eppo is a next-generation AB testing platform built by Airbnb alums for modern growth teams.
3:05
Companies like DraftKings, Zapier, ClickUp, Twitch, and Cameo rely on Eppo to power their experiments.
3:11
Wherever you work, running experiments is increasingly essential, but there are no commercial tools that integrate with a modern growth team stack.
3:19
This leads to wasted time building internal tools or trying to run your own experiments through a clunky marketing tool.
3:24
When I was at Airbnb, one of the things that I loved most about working there was our experimentation platform where I was able to slice and dice data by device types, country, user stage.
3:35
Eppo does all that and more delivering results quickly, avoiding annoying prolonged analytic cycles, and helping you easily get to the root cause of any issue you discover.
3:42
Eppo lets you go beyond basic click-through metrics and instead use your north star metrics like activation, retention, subscription, and payments.
3:51
Eppo supports test on front end, on back end, email marketing, even machine learning clients.
3:56
Check out Eppo at geteppo. com. That's geteppo.
4:01
com and 10x your experiment velocity.
4:08
Gustav, welcome to the podcast.
4:08
Thanks for having me, Lenny. Pleasure to be here.
4:12
It's my pleasure to have you on.
4:15
So at this point you've been at Spotify for over 14 years, which is a rare feat in the tech world.
4:20
And you've held a lot of different roles while you've been at Spotify.
4:24
Can you just start off by giving us a sense of what these various roles and what you've done over the years at Spotify and then just what do you What do you have to do these days?
4:32
What are you responsible for now?
4:32
So I came into Spotify in early 2009, late 2008.
4:41
And my job then, I had been an entrepreneur, started some of my own companies in the back then very very early sort of feature phone smartphone space.
4:49
So I had a bunch of knowledge there.
4:52
I had sold a company to to Yahoo in the mobile space.
4:56
I worked there for a while. I came back to Sweden.
4:59
And then I met through a mutual friend Daniel Ek, the the CEO and co-founder of Spotify.
5:06
And they had built the desktop product already, the free streaming desktop product.
5:10
And it was amazing and I could try it.
5:13
But they needed someone to figure out what to do with mobile.
5:15
And because I had been an entrepreneur in that space I got that job.
5:19
So my job was to to head up mobile for Spotify and figure out what the mobile offering would be, which was a challenge because obviously Spotify desktop was a free on-demand streaming application and back then specifically with edge networks you couldn't really stream at all in real time.
5:39
The performance wasn't there.
5:41
And also you could you couldn't fund that with an ads model.
5:43
So it was a product and business model innovation that was a lot of fun. So that's how I started.
5:49
Then after a few years I took on all of product development for Spotify.
5:55
Then a few years later I actually took on the technology responsibility, sort of the CTO role for Spotify as well.
6:03
And recently my official title is co-president of Spotify together with Alex Nordstrom.
6:10
So we kind of run half of the company each.
6:12
I run the product and technology side and he runs sort of the business and content side.
6:17
So that's the super fast version.
6:17
Aside from getting more responsibilities like taking on the technology department it has been sort of the same job by title.
6:27
I've always reported to Daniel.
6:31
But because Spotify has grown so much, every 6 to 12 months it's been like starting at a new company.
6:37
First it was sort of a Swedish Nordic challenge.
6:41
And then it was an a European challenge and then it was you know, getting into the US and then we became a public company.
6:48
So it's sort of as if I had jumped around between a lot of jobs actually even though it was largely the same title and role.
6:55
Your story makes me think of the classic be careful what you're good at because you end up taking on more and more.
6:59
And clearly uh you've been given more and more responsibility over the years and so um clearly things are going well and you're doing well. Shifting a little bit.
7:09
So you're on my podcast currently.
7:11
You actually have your own podcast which uh was kind of this limited series on this the product story of Spotify which I listened to and loved and it's kind of surreal to listen to your voice in real time cuz I've been listening to that recently in preparation for this conversation. Uh two questions.
7:24
Just what made you decide to launch your own podcast knowing you had a full-time job and a lot going on and the production value for your podcast was very high for what I could tell.
7:32
And then two, just what did you learn from that experience in terms of the product you ended up building and just like empathizing with the podcast creator side?
7:39
There were a bunch of different reasons why I did that.
7:44
Uh one is uh and and not a small one.
7:46
I think like you I I love writing and I have this this secret creator dream in me. Mhm. Yeah.
7:55
I used to write blog posts a long time ago and I write internally a lot.
7:56
You can't write that much externally when you work at a company like this. Yeah.
8:02
But I love writing and talking and presenting.
8:05
So there was certainly that.
8:07
And then no small part was to to from a product point of view to empathize with one of our main constituents the podcast creator.
8:15
I'm unfortunately not a great musician.
8:17
I try to play instruments and so forth, but I don't I don't have any records. I don't sing very well.
8:23
But I decided to make a podcast.
8:25
And uh that taught me a huge amount about what it's like to be a creator.
8:31
How you know, creating different styles of podcast.
8:33
For example we wanted to do a sort of higher um production cost podcast with music.
8:40
And then right away you run into a bunch of problems.
8:43
As Spotify is actually pretty well positioned to solve, but still like it's really hard to have music in a podcast from a rights perspective.
8:50
So you get you understand all these problems that podcasters have.
8:54
And and you can be better at solving them.
8:56
But the the biggest benefit and and the real reason for for um for doing the public podcast was that I had actually done an internal podcast through sort of a hack where we could gate the podcast only employees. Mhm.
9:12
And um I tried to figure out internally how to how to build more culture around Spotify and sort of help um define for new employees and existing employees who we are, the mistakes we did, the successes we had.
9:23
And and how we think about strategy specifically in product strategy because we were quite well known externally for for technology and the squads and all of these things.
9:32
Not so much for for for product strategy.
9:35
And because I I love storytelling more than Google Docs I decided to do an internal podcast.
9:40
And I went around and I interviewed actually Daniel's direct reports.
9:44
So the CMO, the CHRO, and and um CFO and so forth.
9:50
And and just asked them about a bunch of stuff.
9:52
And the idea was to make them more approachable for employees because I felt listening to podcast, you know, even these people that have no idea who I am because I've never met them. I feel like I know them.
10:04
I feel like I know how they think and I just like them much more.
10:07
So, the the secret idea was what if you could get to know your leaders much better than you do through occasional meetings or or or some town hall.
10:16
So, I did that internally and because I'm a product person we ended up talking a lot about product product strategy.
10:22
And people internally really liked that.
10:25
Uh so, next time the question was what if people that don't even work at Spotify yet could feel as if they knew people at Spotify?
10:35
That'd be great because most leaders in most companies are very uh opaque and appear as some sort of otherworldly creatures that aren't really real, I think, when you see them in like business papers or something.
10:47
So, what if you have heard them talk for an hour or so?
10:49
So, that was general idea.
10:49
So, combination of um recruitment tool uh sharing more about how we think about product strategy and just because I think it was a lot of fun.
11:02
I got to interview a bunch of smart and interesting people both externally and and internally.
11:09
Did it have the effect that you're hoping after looking back? I think it did.
11:12
Uh the podcast did well and and no, we did not give it our our own sort of promotion.
11:17
I had to compete as everyone else, which also gives you a lot of empathy for the problem of like, okay, now you have a product, what about user how do you actually get people to listen to it?
11:29
So, it did achieve what I wanted in the sense that uh we have we have this thing called intro days where especially in the in the past few years when we hired a lot, we actually fly people to Stockholm for certain onboarding session to learn about Spotify.
11:44
And and the leadership is on stage talking about what they do and their departments and strategy and so forth.
11:50
And um it's very common that people come and tell me that, oh, you know, I listened to this podcast or this in the episode and it's it's at least one of the key reasons why I joined or sometimes the reason why I joined.
12:03
So, it's sort of anecdotal, but it it may be in the many tens of people at least have said it. So, that seems to work.
12:12
That's really interesting.
12:12
Just again, and this comes up a few times in the podcast, is just the power of content in all these different ways for hiring, for culture building, and it sounds like internally it was the original goal is just internally build this company culture around strategy.
12:24
That that was the original goal.
12:24
Make make a senior leadership more approachable and um so, reduce the distance and then also share more of the thinking in an entertaining or uh way rather than just through docs that people end up not reading. I love that.
12:38
So, I was listening to it as I said and what was really interesting is uh I think episode four was actually all about AI and I think your first kind of attempts at leveraging machine learning and AI within Spotify and I think that's what led to discover weekly and a few other tools.
12:53
And that was like years ago, but it's interesting listening to it now where AI is again like, you know, a huge deal.
12:59
And so, I'm curious very tactically on the product team what you advise product managers and product teams uh on how to think about AI in their product thinking and also just in their day-to-day work.
13:11
I can give a few a few examples there.
13:13
Um I I don't know that we're more more sophisticated than anyone else, but but it what we've been doing at least the traditional machine learning for quite a long time.
13:23
And I think in the podcast, I think I talk about the the journey of the internet in sort of stages and one way to think about it is that the internet sort of started with curation of the user curation.
13:35
So, you you you took something, some good like people or books or music and you digitized it and you put it online and then you asked users to curate it.
13:44
And that was your Facebook, Spotify and so forth.
13:48
And then after a while the world switched from curation to recommendation where instead of people doing that work, you had algorithms and that was a big change that required us and others to actually rethink the entire user experience and and sometimes the the business model as well.
14:04
And I think what we're what we're entering now is we're going from your curation to recommendation to generation.
14:10
And I suspect it will be as big of a shift that you will eventually have to rethink your products. So, so that's one lens.
14:18
So, I tend to talk to my teams about even though it's all machine learning, I asked them to think of this as something completely different.
14:29
The recommendation era was one type of machine learning.
14:32
The generation era is a different type.
14:33
So, don't think of it as just more of the same, think of it as something actually completely new instead.
14:40
And what we learned in well, a few things.
14:43
So, if you look at this new era of large language models and diffusion models and so forth, uh there are two types of of applications.
14:50
As I said for the recommendation era, we had to rethink the user interface and the experience for recommendation first era.
14:57
And so, what what does that mean in the generative era?
14:58
No no one really knows yet.
15:02
Uh there are a bunch as usual there are a bunch of um iterative improvements.
15:05
So, you know, we use these large language models to improve our recommendations.
15:09
You can have bigger vectors, they can have more cultural knowledge, you can use it for safety classification on podcasts that no one has listened to yet and so forth.
15:18
So, there's lots of obvious improvements and we're doing those.
15:22
But so far, we've only really done one sort of real generative product in in the hard definition, which is a product that couldn't have existed without generative AI and that is uh the AI DJ.
15:36
So, that's a concept that we've been thinking about for a very long time.
15:41
And the AI DJ is you you press a button, a person uh of of a digitized person, there's a real person named X and we digitized X.
15:47
So, he's now an AI comes on and talks to you about music that you like and suggests music and you can listen to it and if you don't like it, you can kind of kind of call him back and he says, okay, now let's listen to something maybe from a few summers ago or here's some new stuff that, you know, were trending yesterday in the last, you know, last of us episode or something like that.
16:10
So, that product couldn't have existed without generative AI both generating the voice and generating what uh the the content of of what the the voice says.
16:17
So, you can have individualized personalized voice at the scale of, you know, half a billion uh people.
16:26
And so, we we had the use case we had seen for many many years.
16:31
Sometimes people call it the radio use case, we call it the Siri intent use case internally when you actually don't know what you want to listen to at all.
16:40
Spotify wasn't that good.
16:40
Spotify was good when you knew at least roughly, you knew the use case of what you wanted to if it was a workout or dinner, like we had lots of lots of options for all of those, but if you really didn't know at all, it was hard to open Spotify and just stare at it.
16:56
And people used to say longingly, you know, that this was the one thing that radio was good at.
17:00
Radio was quite bad to be honest.
17:02
I mean, it's not personalized to you at all. It's not on demand.
17:05
You come in in the middle of things.
17:07
It's actually terrible in many ways.
17:10
But people still often say that there was something good about it and I think that something was the fact that you had a knob and you could just switch between contexts.
17:18
It's like, no, boring, boring, boring, boring. Okay, this is good.
17:23
And and Spotify never had that mode of like, I don't know what I want, but I want to sort of cycle through things until I find something that I like.
17:31
And I think with AI DJ, that's actually the use case we managed to solve.
17:34
So, X comes on and says, I'm going to suggest something to you that you can listen to.
17:39
And if you like it, you can keep listening, but if you don't like it, you kind of bring him back again and you change genre.
17:43
And for one reason or another, we tried to solve that for many times for a long time, but just starting to play a random song without any context as to why you would hear this just didn't never worked.
17:57
So, so that was our first sort of foray into a product that couldn't exist before and I think to your question of principles around that, there are a few pretty distinct principles that we've learned.
18:09
One that I really like that is not my principle at all.
18:12
I think I think it is straight from Chris Dixon is the principle of fault tolerant user interfaces.
18:20
So, I can't say how many times during the early machine learning era when we said, you know, we're moving from curation to recommendation, I saw a design sketch that was a single big play button.
18:32
Cuz clearly, that is the simplest user interface you can do.
18:35
But if you don't understand the performance of your machine learning, you can't design for it.
18:39
The quality of your machine learning, if you're going to have a single play button, needs to be literally 100% or zero prediction error.
18:48
And that's never the case, right?
18:48
So, let's say that you have, you know, a one in five hits.
18:53
Four out of five things are dud.
18:55
Then you need a UI that probably at least shows five things at the same time on screen.
18:59
So, you have a one in five of something being relevant on screen.
19:02
So, you need to understand the performance of your machine learning to design for it.
19:06
It needs to be fault tolerant.
19:08
And often you need an escape hatch for the user.
19:11
If you were so, you make a prediction, but if you were wrong, it needs to be super easy for the user to say, no, you're wrong.
19:16
I want to go to my library or to this or to that.
19:21
So, so we have that principle of having a fault tolerant user interface and a user interface that corresponds to the current performance of your of your algorithms.
19:31
And I think that is going to be true for generative machine learning as well.
19:35
I think a very clear example actually is Midjourney.
19:39
You think about the early Midjourney user interface inside the Discord channel, actually generating an image was very very slow.
19:47
It took a long time to generate high quality image and they could have built silver button thing where you put in a prompt, you wait for minutes, you get an image, and I think one out of four times is going to be bad.
20:02
So, you would have been disappointed three out of four times, and it's a minute each.
20:04
So, like 4 minutes later, you'd be this is a shitty product.
20:06
What they did was they generated four simultaneous low-res images very quickly.
20:15
And you could say like So, So, apparently their performance was probably one in four.
20:17
That's why they four showed four and not six.
20:21
And so, one in four was obviously was usually pretty good.
20:23
You click that one and either continue to iterate or scale it up.
20:26
So, that's also an example of I think people understanding where the performance of generative AI was when they built the UI.
20:32
So, that's something that, you know, I would be inspired by.
20:37
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 act talk and talk and talk and talk.
20:46
You have to remember that people came there for the music.
20:50
So, the principle for the AI DJ coming from the team, by the way, this was a bottoms-up product, actually.
20:56
It It required a lot of support.
20:56
We actually required big companies and so forth to be able to build it.
21:01
But the idea has been have been built by by teams bottom-up.
21:04
So, the principle there was literally to do as little as possible and get out of the way.
21:09
And I think that was really helpful.
21:11
You know, it's it's not telling you what the weather is and what happened in the news and going on and on and on about this band.
21:16
It is trying to get you to the music, and I think that's that's why it's working because it it it is working very well for us.
21:24
I love this distinction between recommendation and uh generation.
21:26
And this kind of begs the question of there's this trend that I imagine you're you're seeing of people auto-generating music using artists', you know, catalog.
21:37
Like there's this Drake and The Weeknd thing that came out a week or two ago.
21:41
Where do you think this ends up going, and how do you think artists adjust to this world where music can just be auto-generated, you know, this play button is like all of it is generated versus just like the DJ in between the songs?
21:52
First big caveat this, this is just super early.
21:56
No one No one knows anything, you know, about how this is going to play out or or the legal landscape and so forth.
22:02
But I think it's going to be have a lot of impact.
22:06
And I I think if we talk about two things, one is what it could do for music.
22:10
The other is the rights situation.
22:13
And if if rights holders are getting compensated and so forth.
22:15
So, we talk about the first thing in isolation.
22:20
I think an interesting example is right right about when I grew up, Avicii came along.
22:25
And it's interesting to think about because Avicii was not really considered by the existing music industry as a real artist cuz he couldn't really play an instrument and he couldn't sing.
22:36
And he was just sitting with this computer in this DAW, digital audio workstation.
22:41
And so, it wasn't really considered real music.
22:43
And I think now all of us consider it very real music and that he had tremendous real musical talent.
22:50
So, I think right now we're probably in the phase where people say this isn't real music and it's it's somehow fake.
22:54
I think the way to think about these these diffusion models if and when they get good enough at at generating music is probably the same, like an instrument.
23:07
It's just a much more powerful instrument, and we'll probably see a new type of creator that wasn't proficient at an instrument and they, you know, they they couldn't they couldn't assemble a full orchestra and and do the thing that they had in their head.
23:21
And they can now uh generate very very new things.
23:24
I also think, by the way, that there is this distinction between AI music and real music uh that that doesn't exist.
23:30
For sure, very talented real musicians are using AI to get better and to help create new ideas.
23:38
So, that distinction is doesn't really exist. It's all going to be AI.
23:42
The question is what percentage.
23:46
Which makes the problem harder cuz you can't talk about if it should exist or not.
23:49
You have to talk about what percentage should exist and and who gets to use it or not.
23:55
But I think the way to think about it is probably as an instrument uh that could help create a huge amount of of art.
24:03
And I think uh this is not news to you who probably use these things a lot, but I think if you don't use these generative models, there is the perception that you tell it to create a hit and you will get that.
24:16
That's That's not how it works.
24:18
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.
24:28
Actually being original is very hard.
24:31
And from one point of view, as it now gets easier to create more generic music, it will actually be more difficult than ever to be truly unique.
24:37
So, I still think there will be tremendous skill in creating something truly unique.
24:43
And my hope would be that what happened with the DAW and that technology jump was you got a whole new genre like EDM that didn't you couldn't really produce it with an orchestra or live.
24:55
And maybe we'll see completely new music styles with these technologies.
25:00
I think that would be very exciting.
25:02
So, that's on the positive side.
25:02
But then you have the the rights issue, which I have a lot of empathy for.
25:08
And and Spotify specifically has seen this before.
25:10
So, we had a different technology shift like this, which was the technology shift online downloads of music and piracy and peer-to-peer.
25:19
So, first it was the technology shift in peer-to-peer, and it was it was exciting for consumers.
25:24
More consumers started listening to more music than ever.
25:28
And I think that's where we are now with generative AI.
25:29
There's a new technology, but it also required a new business model before creators and industry could actually participate and benefit from this.
25:38
And if that's obviously self-serving to say because we were a big part of innovating that business model, but I still think that's what's necessary.
25:46
And I hope that that's what I and we could be part of.
25:49
So, I think we've seen that first part, the technology shift.
25:53
And there would probably be a lot of discussion and and chaos here, which I have a lot of empathy for.
25:59
But I think um I think we haven't seen the second part yet.
26:03
What is a model where this could be a benefit?
26:05
What would actually happen after piracy is that the music industry got bigger than ever.
26:11
Not not just as big, but bigger than ever.
26:14
And I think that could happen with this technology as well.
26:17
But we're right in the beginning.
26:20
So, along the same lines, something else you teach is this idea of all truly great products have to pull some kind of magic trick.
26:26
This comes up in your podcast a lot, and I think you mentioned this other places.
26:31
And thinking about all the stuff you're talking about here, it feels like in a sense, everything's going to feel like magic cuz AI is kind of baked into it.
26:38
I think when we did the the AI DJ, we did a small version of that.
26:42
When people first listened to it, we could see that reaction in in user testing.
26:47
When they when they like So, So, the magic trick there was that how could they record this person saying so many different things because it's talking about my music.
26:56
So, the magic trick was obviously didn't record a person saying it's it's generated.
27:01
And and that magic trick wears off.
27:01
You hear it all the time now and so forth, but it was one of those magic tricks.
27:08
So, I still think that concept is important and it seems to correlate with products sort of going viral and and taking off.
27:15
And I think it was the same using something like um DALL-E or Stable Diffusion or Midjourney the first time, it completely seemed like a magic trick.
27:26
And And obviously there is no magic.
27:27
It's just data and statistics.
27:29
But I think getting to that point and iterating a product to the point where it feels like magic the first time is is very helpful.
27:36
And it's often a question of just getting the performance to certain levels, scoping down, removing things.
27:45
It There's a lot of fine-tuning, I think, that makes you cross that that uh line from it's cool and impressive, but not magic to it it feels like magic.
27:55
I don't I don't understand how this how this could be done.
28:00
Yeah, it reminds me of the launch of GPT, which ended up being the biggest, most, fastest-growing product in history, and it's like the epitome of a magic trick.
28:07
It's like feels like actual magic. Absolutely. Absolutely.
28:09
And And to most people, it is still very Actually, to a lot of us and even to researchers, it's a little bit magical.
28:17
No one really understands fully.
28:19
So, I guess there's maybe some magic left in the world. Absolutely.
28:22
And I think a lot of people are worried about not understanding what's going on there.
28:27
Shifting to the way you all build product at Spotify.
28:29
So, Spotify is kind of famous for popularizing this idea of squads and tribes.
28:33
And correct me if I'm wrong, but you guys have kind of moved away from that approach. Yeah, that's right. Okay.
28:40
So, I'd love to understand just like why you shifted and what you kind of learned from that approach to building product.
28:46
And then just like how do you organize the teams now?
28:47
What do you What do you do now?
28:51
This was something that we focused a lot on early.
28:54
And it turned out to be smart of us to name these things into squads and chapters and so forth.
28:59
It wasn't It wasn't really Well, maybe it was sort of deliberately branding.
29:05
Uh but it wasn't for purposes of branding that we made it up.
29:07
We made it up because we thought it was a good structure to to use.
29:09
And we needed names for things.
29:11
And And this was the early internet era, so you were allowed to like make things up.
29:17
And so, it was very good for where we were at the time, and it certainly helped us in recruiting.
29:20
It's It's become a little bit of a of a cost to us because people still think that we organize that way.
29:26
And it's not a very efficient way of being organized at this scale or or maybe even if you started over right now because we've learned more.
29:34
But I I think the big difference is is the idea with the with the squad specifically was uh twofold.
29:42
They were supposed to be small and sort of full stacks.
29:44
A a should be about seven people and it should have, you know, front and back end, mobile, QA, um Agile coaches and so forth.
29:52
And And it should be very autonomous was the idea.
29:56
And And that's really what we shifted.
29:58
So So, first of all, as you grow the company, scaling in increments of of seven engineers just creates a ton of overhead.
30:06
So, obviously our teams now tend to be much bigger, maybe maybe two, three times that at least per like manager.
30:12
So, maybe have like 14 or something instead of seven.
30:16
And just less less overhead roles. And So So, that's one.
30:20
It It looks more traditional as as you learn more and it's reasonable as you scale.
30:26
The second big thing I think we struggled with was back then when I joined, the average age at Spotify was I mean, I was the oldest and this This was 14 years ago.
30:40
I think the average age was probably under 30 or something.
30:41
And it was in most uh tech companies.
30:45
And so, we had coming from Sweden, which is a it's a different culture than than the US.
30:53
And I love a lot of things about Swedish culture and I think we managed to keep the best parts.
30:57
But Sweden is a very sort of bottoms-up uh autonomous culture.
31:01
There's this famous drawing of how you make decisions in Sweden and in the US, I I think it's just a a hierarchy and in Sweden, it's kind of a circle.
31:09
You sit in a circle, no one is in the middle.
31:11
There is no leader and so forth. Interesting.
31:12
So So, I think by by sort of culture, we were very inspired by this super autonomous thing.
31:18
And I think the idea with with autonomy is uh is very reasonable and and the right one, which is we we were and we are hiring the smartest people we can find.
31:31
And And we pay high salaries for that.
31:33
So, if you're hiring smart people, one way to think about it is you're you're renting brain power.
31:39
So, if you're renting all of this expensive brain power and then you give them no room to think for themselves, that doesn't sound smart.
31:44
Then you should actually hire less smart people and like keep your costs down or something.
31:51
So, I think you have to give a bunch of autonomy to actually maximize the value of the investment you're making.
31:59
So So, that's very reasonable that you would give a lot of space for people to use as much of their their their talent and capacity as possible.
32:06
But the problem with that is if you put autonomy very far towards the leaves of the organization.
32:10
And And 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.
32:20
You're going to have 100 squads with 100 strategies running in 100 directions.
32:25
And And And, you know, Spotify has been there in that camp.
32:27
I mean, we managed to get somewhere for sure in spite of this, but I'd be I'd struggle to say we were like efficient in doing that.
32:36
So, we've done a few things.
32:38
Uh the team structure is more traditional, larger teams, uh less overhead.
32:41
And we've been specifically working with where in the org do we put the autonomy?
32:49
Cuz the extremes are at the leaves and we were there.
32:52
The other extreme may be at the top.
32:54
Let's say maybe something like Twitter.
32:56
There's There's one person. Both have problems.
32:58
If you have it at the leaves, you're going to produce a lot of heat.
33:02
If you have it at the top, you you need someone with a lot of capacity and he alone has a lot of capacity, but you are by definition going to bottleneck.
33:11
All decisions have to go through there.
33:13
And And Daniel just it's not his personality that he even wants to make all the decisions.
33:19
He wants to maximize throughput rather than to bottleneck the throughput.
33:22
So, the question is if it's not at the top and not at the very bottom, where do you put it?
33:28
And And what we found, which I I don't think is very contrarian at all.
33:32
I think this is the case in most companies is around sort of the VP level.
33:35
So, if you have Daniel, then you have the C-level, myself and others, then you have sort of the VP level.
33:42
That is a good mix of instead of having one person in the company think, so only Daniel then and and the rest just do, you have on the VP level in a company like this many tens to maybe hundreds of people that have a lot of autonomy to think.
33:57
So, you get you get a good amount of freedom of thought and and people thinking in different directions, but it's it's not like 8,000 people. Right?
34:06
And the these people on the VP level are both quite quite a lot of them, but they're also usually quite senior.
34:13
They have a lot of pattern recognition.
34:16
So, I think that solves for It's like a good if if you think of it as a as an optimization problem, it's kind of a good optimization space.
34:23
So, the the autonomy level in Spotify now tends to be quite high at the VP level and then lower around those levels.
34:33
And when you say autonomy, what does that actually mean?
34:35
Is it the VP of say the podcasting product has a lot of say over what happens and there's not a ton of I don't know.
34:42
Like how how involved are people above?
34:44
And I know Maya's the VP of product, I believe, for the podcast product. Exactly.
34:48
Who I think is going to come on the podcast someday.
34:50
Um what does that mean in terms of autonomy for her? What practically?
34:53
So, it means that um I would ask Maya to define a strategy for what we do in podcasting, how are we going to be different, why would a podcaster want to be here?
35:07
Whereas another company, I would make that strategy or in another company, Daniel would make that strategy. Mhm.
35:12
Same with the The idea, for example, came from uh one of from our personalization team.
35:18
And so, that was a bet that they made.
35:20
So, they have autonomy to make those kinds of of bets and define strategies.
35:26
Same with the user interface.
35:26
We have an experience team.
35:27
Can talk about the org structure uh later.
35:31
But I put a lot of autonomy on the VP of experience to to define and suggest what it is that we want to do.
35:40
And in other companies, I I would define all of that myself, for example.
35:44
Just going even a little bit further here.
35:46
I know you have just like strong opinions on the way to organize teams and how the organization kind of helps you optimize for specific things.
35:53
What are your kind of just thoughts along those lines and what have you learned about how the impact of organization and what you're optimizing for?
36:02
Yeah, so I I talk about um sort of an um an idealized spectrum or or um maybe not idealized, but exaggerated spectrum.
36:12
It's not really not nothing is really true.
36:15
But you you create uh extremes to make a point, right?
36:18
So, on one spectrum, you have something like Amazon, which is known for two-pizza teams. Um no dependencies.
36:26
You try to minimize dependencies so you can run in parallel.
36:30
Teams compete with each other uh on even on the same project and so forth.
36:34
But they have direct access to the user.
36:36
And so, the benefit here is if you have an idea, the time to get to user is very low.
36:40
And And it it has worked for them.
36:42
It's produced, you know, it's produced um Kindle, it produced Alexa, it's produced a lot of very novel things.
36:52
There are a few interesting um downsides here.
36:56
One downside that I'm extremely impressed with Jeff Bezos for seeing is if you have teams that compete with each other, the incentives are to to hide your results, hide your code.
37:09
And that should make for an organization that gets no platform leverage cuz no one is cooperating.
37:15
And I think this Either he had that insight or because he saw this, he he had to do this, but he's well known for pushing extremely hard on hard APIs.
37:23
Like if you don't create hard APIs to your technology, you're out.
37:27
And if you think about it, it has to be that way cuz otherwise no one would do it.
37:31
And a hard API is essentially a like everyone knows how to use this API and connect to this team to interface with. Exactly.
37:37
You have to expose your technology to others.
37:38
You have to maintain those APIs and they have to be very structured cuz otherwise the whole thing would would collapse as everyone's supposed to compete cuz there are no incentives.
37:46
You have to centrally force that.
37:49
And interestingly, even though theoretically then they're the worst position to have a structured platform, I think because they forced it so hard, they were the ones who did Amazon Web Services because they had such hard-defined APIs because of this rule that it was easier for them to turn it inside out and expose it the rest of the world.
38:05
Whereas if you look at something like Google, I think they struggled more with externalizing their APIs maybe because it is so friendly and soft, so they didn't need as hard APIs on the inside.
38:15
Because there was no competition.
38:15
People could just go into each other's code.
38:18
So, it's an interesting uh anecdote around it.
38:20
But the main point is you're faster there, but it's going to be hard to cooperate.
38:27
And so, you will see something like maybe exaggerating a bit, sometimes you'll see multiple search boxes on the same page from different teams.
38:36
And this has been true in Spotify, by the way, as well.
38:37
You've seen like multiple toasters on the now playing view coming up from different teams because they're working when we were in the autonomous mode. Everyone running.
38:46
Um And then So So, 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.
38:56
But clearly that's been the right choice for for Amazon cuz they're they're a trillion-dollar company.
39:03
But then on the other spectrum, you have something like Apple, who's also a trillion-dollar company.
39:08
So, clearly both models work where you would never see two search boxes from the same team popping up on an iPhone.
39:13
That is centrally organized by, you know, something that is close to single individual.
39:20
So, they they are instead in a in what's probably the world's biggest largest functional org. They're doing as much.
39:27
If you think about what goes into Apple, it's I mean, they certainly do everything we do.
39:31
They have music service, podcast service, audiobooks, and they have a billion other services.
39:36
So, it's not like they have an easier problem.
39:40
And And yet they they um built something that feels more like it was built by a a single developer for a single user.
39:47
So they centralize and they have this bottlenecking function that everything has to go through and be decided how it fits with everything else.
39:55
And so that has the benefit of the user experience being simpler and not shipping the org chart and increasing complexity, but it also has the drawback of speed.
40:05
Without having facts on it, I've heard people working at Apple said like, yeah, took 7 years to get that thing to market because it just had to wait in the in the pipeline.
40:17
So you have these extremes and and I think that the most interesting example I think to think about is uh when you double click the power button on an iPhone, that Apple Pay comes up.
40:28
Like that decision, how did that happen?
40:30
You can imagine that all the services team would like to pop up when you double click that button.
40:36
And so someone had to decide should should I should music come up?
40:40
Should Apple Payments come up?
40:40
Should something else come up?
40:45
Um And so they have a different structure there.
40:47
And and on that spectrum of centralized versus decentralized, because of our strategy, which is we're single application, trying to add or not trying to, we have added multiple types of content with actually very different business models on the back end, you know, rev shares and royalties and book deals and so forth, into a single user experience. That is our strategy.
41:08
We think the user experience and keeping that simple is the most important thing.
41:11
So we've chosen more of the sensor centralized model, where these different sort of vertical businesses, if you think about it, the music business, podcast, audiobooks business, they have to go through a single recommendation organization. That's another problem.
41:26
You know, which one do you recommend to which user?
41:27
Should be a book or podcast or music?
41:29
And how do you weigh them against each other?
41:31
And also the user interface could easily get incredibly complicated if everyone built their own UI.
41:38
The music team built their UI and then someone had added features on top.
41:42
So that's how we chose to to optimize.
41:44
But it is based on our strategy and I think both models work.
41:49
This episode is brought to you by Eco.
41:52
Last month, Eco users earned an average of $84 in cash back rewards. How?
41:54
With Eco, the future of personal finance.
41:57
Eco is the update to a misaligned financial system, providing an app that works just like your bank, but removes almost all of the middleman, helping even the best money optimizers optimize in less time automatically.
42:10
What if you earned rewards for paying your rent?
42:12
Or got rewarded for ordering food and shopping online?
42:16
Or even earn rewards for saving each month?
42:20
And then imagine if you got rewarded again, just for getting rewarded.
42:21
With Eco, you can spend at some of your favorite merchants and automatically get 5% cash back.
42:27
Plus, Eco's APY rewards look more like $80, not 80 cents.
42:30
And then there are Eco points, the world's first open rewards system.
42:35
You earn them whenever you do almost anything in the Eco app.
42:39
Eco is working to make these points the most rewarding points ever, so it pays to be early.
42:43
Sound too good to be true? Go to eco.
42:45
com/lenny, sign up for an onboarding, and find out why it isn't.
42:50
Lenny's podcast listeners who attend an Eco welcome session will get an exclusive 4% APY on deposits over $1,000. Learn more at eco. com/lenny. That's eco. com/lenny.
43:05
It's interesting these two examples you gave, Apple and Amazon.
43:06
They're two of the biggest companies in the world and they're like at the extremes of these two into the spectrum.
43:12
And it's interesting most companies are somewhere in the middle.
43:16
I wonder if there's just like a benefit to being in an extreme and that ends up being really important. I think so.
43:20
In in almost all industries, you have this smiling curve concept, right?
43:24
Where you want to be at the extremes of the smiling curve.
43:26
And that's what big business opportunities are, but but not in the middle.
43:30
So it's probably true terms of organizational models as well.
43:34
Speaking of um extremes, I want to talk a bit about taking big bets.
43:38
So you guys had this big launch event recently where you basically redesigned the whole primary feed of Spotify to make it feel more like where kind of apps are going, like TikTok Reels feel of just, you know, stream and you start hearing videos and music starts playing and some people loved it, some people did not.
43:55
And I'm curious as a product leader, how you think about thinking long term and dealing with people that are just like, what the hell's change? I hate change. Stop changing things.
44:03
How do you think about that? Who do you listen to? Who do you ignore?
44:06
How do you know what to stay the course?
44:08
How do you approach that?
44:10
Yeah, you're being very kind.
44:10
There there was a lot of of negative feedback on Twitter on on some of that.
44:14
So let let me actually kind of dig into some detail because I think this is a really for product people listening to this, this is an interesting lesson that I think few people few companies talk about.
44:29
Cuz cuz you don't really want to talk about uh you want to talk about everything that went exactly as you thought they would and you don't want to talk about the things that didn't go exactly as you thought they would.
44:39
So I'll I'll go through kind of um what we are trying to achieve and and and what we learned.
44:46
So Spotify is mainly a background application and for a long time we've been considered very good at background music and podcast recommendation when the phone is in your pocket and you're listening to like a an EDM playlist or, you know, a pop playlist or something.
45:01
We're really good at inserting another EDM track there or another pop track there or something like that in the background.
45:09
What we hear from users again and again though is that you know, they say that they get trapped in a taste bubble.
45:15
So, you know, I love my Spotify, I love this, but I am a little bit bored with EDM now and Spotify is not suggesting something completely new.
45:26
And if you think about that problem, it may sound similar to the recommendation problem.
45:30
It's just another recommendation problem.
45:33
But it's actually fundamentally different.
45:36
Because when you're recommending another EDM track inside the EDM playlist, you have a lot of signal from that user that they like EDM.
45:44
But if you're going to recommend a completely new genre, by definition, you have no idea.
45:50
Because if you had no idea, it wasn't new to them.
45:52
So you can't know anything.
45:55
So back to hit rate, your hit rate is going to be incredibly low when you suggest something completely new to the user.
46:03
So this problem of helping people get out of their taste bubble isn't easy as it sounds.
46:09
And we can't really take some, you know, some genre that maybe isn't typical.
46:11
So I'm a big fan fan of reggaeton, for example.
46:16
It's not typically it's not that common in Sweden and if you would look the rest of my profile, it's kind of EDM heaviness, you probably wouldn't have guessed it.
46:22
And Spotify wouldn't have guessed it.
46:26
So if I'm listening to my favorite EDM playlist in the background or maybe my metal playlist, metal is very big in Sweden, it's really hard for us to just insert a reggaeton track in the middle of that.
46:38
You know, most people are going to think Spotify's broken.
46:40
What the hell are they thinking, right?
46:44
So that doesn't really work.
46:44
So in order to help people break out of their taste bubbles, you need something different.
46:52
You need something where your hit ratio can be very low.
46:56
And you need people to expect it to be very low.
46:59
So when we recommend things in the background, our hit ratio needs to be at least nine out of 10.
47:04
Maybe one dud is okay, but if you get, you know, five duds, you're going to think we broke your playlist and your session.
47:12
We need something where one out of 10 is is a success.
47:15
If you find one gem out of 10 tries, you're very happy.
47:20
So so you need a completely different paradigm.
47:22
And and you also need to be able to go through many candidates quickly, right?
47:28
Because the hit rate is so low.
47:29
You can't take 3 minutes per item.
47:29
It's like, okay, I didn't like this and it's still like 2 minutes left before the next one comes on.
47:36
You need to quickly say, no, no, no.
47:38
So the the obvious candidates for this are these uh feed type experience, where you can go through lots of content.
47:45
You're expecting the hit ratio to be much lower and if you don't like it, the cost is very low. You just swipe.
47:48
And this is the reason why people have been when they want to break out of their of their taste bubbles or when they come into Spotify and listen to something completely new, it is usually because they found it on one of these services, like a TikTok or YouTube or something, where they get exposed to lots of of new content.
48:07
So people were asking us for these tools.
48:09
And so that's what we wanted to solve for.
48:13
And so we built a bunch of features, um feed like structures where you can go through either a shot a new genre with many tracks uh or a podcast channel with with channel with many episodes or or even full playlists.
48:28
And and we implemented those and we put them in something called sub feeds.
48:32
So in the current experience, and this is rolled out worldwide, if you click the podcast sub feed, you get a feed of podcast episodes.
48:40
Click the music sub feeds, you get a feed of of playlists, where you can quickly, you know, you can go through many playlists and if you don't understand the name, you can quickly hear what they sound like and check out a few tracks and understand if this is for you.
48:51
And if you go to the search and browse pace page, you can find completely new genres that you can quickly go through.
48:58
And so those are working as we intended.
49:00
People go through them, go to them when they want to find new music.
49:03
They browse through them and they save new songs.
49:08
So they're working as we intended.
49:10
Uh the thing that didn't work as we intended was when users asked us for this uh again and again, we took sort of the sum of these things and we put it on home because people ask so much about discovery and we can see clearly how how correlated discovery is with retention on Spotify and so forth.
49:27
But what we what we misjudged or or failed to to uh or rather learned about our own homepage is that the way it works right now and and this is what you can see in the Twitter comments if you if you remove the angry voices and sort of try to see
49:45
what they're saying they're saying the following which is I see quite clearly in the quantitative data as well that 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
50:00
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 so you come in with a high intent you actually knew what you wanted and maybe only 10% of the time is it
50:13
true discovery like I don't know what I want so if you think about that it's 90% recall and 10% discovery when we tested that the design so so the sub feeds were working and are working but when we tested the some of them on home we kind of switched it from 9010 to
50:29
1090 so 10% recall 90% discovery and while people want to discover they don't they probably don't want 90% discovery instead of 90% recall so if you then look at the comments on Twitter what they're saying is like hey I can't find
50:44
my playlist anymore like where are these things they're not really complaining about the discovery they're complaining about the things they don't get anymore and we could see this in the quant data as well then you can see traffic
50:54
shifting from home into search and into library which is a clear sign people are trying to find find the things they can't find anymore and and you can even see people then trying to use these discovery tools which are optimized for understand
51:07
quickly understanding new things to do the recall like where is that workout playlist I know I want and it's actually very bad UI for recall it's kind of like a slot machine right very unpredictable if you ever get to that workout playlist it was optimized
51:21
for finding new things not for recall of existing things 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
51:33
want a lot a lot you want a lot of user interface a lot of pixels and you probably want sound because you don't know what it is so what kind of what we learned about our UI and I think there's maybe maybe a little bit of you know
51:46
product jealousy here you always look at other experiences and if you look around you could be forgiven for thinking that most other products if you look at something like YouTube for example their homepage is exactly that it's a
52:00
huge single item discovery feed with only new items and people don't seem to tweet angrily about about how angry they are due to say they love YouTube and it's a big product and I think what we discovered was that we actually did something really well on
52:15
our homepage which was supporting you being inside of multiple sessions at the same time so you could be in the middle of two podcasts and an audiobook and also then actually I just want to get to that workout playlist I don't
52:29
remember the name of it but I know it's workout we actually did that part really well I would venture to say much better than the other experiences where you literally have to go to your to some tab and into library and start browsing to get back to where you were
52:43
and so maybe it's path dependent if we had you know because we have done recall pretty well people got I think reasonably upset when they couldn't find it when they couldn't do the recall anymore and we really don't want we didn't want
52:58
to lose that cuz it was one of the things we did well and underestimated and my takeaway is actually we do it better than other experiences and so we certainly want to keep that so what we did was now we're just updating the hypothesis to achieve the
53:12
same goal which is these things are working and when people want to discover they use them and they seem to work they can they can also get better you know you're on this like um hill climbing journey from machine learning point of view
53:25
but the question is how do you make sure that whenever people feel that they are in that I'm trapped in my taste bubble they understand that these things are there and they're easy to use so now we have a version of of home that
53:39
that we're also testing obviously where these things are very available but but voluntary and you can still do all of the recall and so from my point of view this is the reason we AB test because you know you want to be scientific about it and
53:55
you know you want to learn as much as possible about your own product and your users and now I'm sharing a lot of the learnings maybe we should keep them to ourselves but uh my hunch is that it's going to make it a much better product but what I told my
54:10
teams when we went into this cuz I've done this a few times like redesigning I think there is a there are two fundamentally different types of product development one is designing a new feature it is hard to make it but it's
54:26
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 it didn't make it worse for you but when you redesign it is much more tricky because it's not voluntary to participate in the redesign
54:42
so there is there is a cost even you know what for people who don't like it you have a very tricky problem here which is there are going to be two types of feedback one is you did something and it was right but people are upset because you changed stuff
55:02
the other is you did something and it wasn't right and people are also upset but for good reasons and so how do you separate these two because I I think I explained this um to um when we talked through this with my my teams I think the analogy to think
55:18
about is you have you know your desktop your physical desktop you have your computer in one place you have your pencil over here you have your notebook over there and I come in and I just rearrange all of it and you have spent in our case maybe 12
55:31
years with that setup it doesn't matter if I have a lot of of quantitative data that my new setup is better you're going to get upset because you are effective in this old setup and it's hard to tell those apart the
55:46
most classic use case is the the Facebook newsfeed which people are very upset about when it became a single newsfeed but it turned out to solve a lot of user problems that you didn't have to run around all of Facebook collecting events yourself
55:58
so there are some ways of of understanding if if if you made it better but people's habits are broken or if it's not better and one thing is for example to look at new user cohorts that don't have that behavior versus old user cohorts and so forth
56:12
so we went through all of this with the teams before we did it I said this is going to be painful probably going to be a lot of tweets cuz chances that we get it exactly right are very low so for that reason it hasn't been you know very hard on the
56:25
team um it is hard you know you want to respond to people but the right way to do it is to listen understand try new hypothesis to really figure out what's going on so I've I think I've done it um maybe three or four times now one
56:40
three maybe one unsuccessfully two successfully so kind of knew what I was getting into uh so it's it's almost like uh you punish yourself very painful but also the most exciting things and I think any product person knows that the the
56:57
easiest and most straightforward thing to do is to iterate around where you are there's no risk you're not going to get fired no users going to get angry but everyone also knows that eventually if you don't adopt new technologies new paradigms etc. you're going to get
57:09
paradigms etc. you're going to get replaced you have to find this balance of trying new things that's you know when you work in software you have this tool of AB testing and being scientific about it when you when you build hardware it's worse if you're if you're wrong you're wrong you can't update I love this story so appreciate you
57:28
sharing it I imagine also with a big launch like this you can't actually AB test it ahead of time cuz of the press season they're like oh my god look what Spotify's doing and so you're kind of limited there I imagine right you couldn't really test this ahead of time the the hardest thing about this is if
57:42
you're trying something completely new the 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 but poor machine learning right you have poor machine learning or the other you have to build a lot and it gets quite
58:00
expensive that's actually the biggest way it's painful is not really the feedback from the outside it is the cost you have to take on the inside you know you incur a lot of cost and you're really hoping you're right and and in our cases the the changes on the homepage aren't that hard for us to do the important
58:21
thing is that the underlying hypothesis of can we help you break out of your taste bubble actually works and then you know you update the the acquisition funnels into that experience but I think the problem is that you need to get so many things in place to be able to say if you get you know you you might get a false negative
58:43
just because like you didn't do it well enough that's the biggest chance I think with these big rewrites where everyone has to update everything before you can know if you're right or wrong what was that process like of helping understand what is not working and what is working and what you wanted to change like I imagine there's like a bunch of
58:59
data you're looking at some tweets things like that like what was kind of like the tactical oh shoot something's not going the way we expected here's what we should do well but feeds we tested but the home feed uh we we rolled out and tested afterwards and we tested out on users few different variants of it and then we got the data back and we we
59:19
looked more at the quantitative data and we do a lot of user research where people sit and use the the feeds to understand and and build like our own theory of mind of what is working and what is not working and then obviously, you you look at you look at user feedback, of course. And some users are very good at
59:38
And some users are very good at expressing what is that that isn't working.
59:42
Others are not as good as expressing what isn't working.
59:45
So, it can be hard to parse that.
59:47
But certainly, that's a factor as well.
59:50
And so then, what once you do that, then you have quantitative data to look at.
59:56
And then you sit and reason through what you think is right and wrong, what are the different hypotheses, what is working, what is not working.
1:00:02
And then just update and test again and again until you find uh until you prove or sort of disprove your hypothesis.
1:00:09
Trying to be as scientific as possible about it.
1:00:13
And and also, I think the biggest risk also when you've invested so much time in something is you you know, getting precious about things.
1:00:21
You have to just be brutal.
1:00:24
You have to believe in things 100% until the data says no.
1:00:26
And then you believe in something else 100%. That's That sounds easy. It's very hard to do.
1:00:32
It it to to the extent that people get upset when you do it because for some reason people don't like when people change their mind.
1:00:42
It is what we should want from everyone.
1:00:43
I would love a politician who said, "I've looked at the data and I realized actually this is right and now I believe this."
1:00:49
But we hate politicians that do that.
1:00:52
You know, they feel untrustworthy and like we ridicule them.
1:00:54
So, I think that's the biggest risk with anyone.
1:00:59
You just have to be like unemotional.
1:01:02
And just just look at the proof in the data.
1:01:07
And then, you know, if you do that, you just move on and then you get to where you want to be.
1:01:10
You solve the same problem, but you adapt.
1:01:13
I really like that philosophy.
1:01:15
Essentially, it's the idea of strong opinions loosely held, right? Is that Exactly. Exactly what it is.
1:01:19
And it sounds so easy, but it it's hard.
1:01:21
Right, cuz to your point, people don't like don't respect someone changing their mind.
1:01:27
They're like, "Oh, I see.
1:01:27
They were wrong the whole time and they were so confident about being wrong." Yeah, exactly. And it's unclear why.
1:01:34
It is what we should want.
1:01:37
But uh I think it I think it has something to do with human human psychology.
1:01:42
We actually tend to love prophets and people who hold very strong opinions with very little data.
1:01:48
Those are the people we like.
1:01:51
People who look at a lot of data and actually adapt, we don't like. Not sure why. We're flawed. Flawed creatures. For sure.
1:01:59
Is there something that you recently changed your mind about along these same lines that maybe comes to mind of like, "Oh, yeah."
1:02:06
No, I think these learnings about the what our own uh design system and homepage does really well, maybe better than others, that we don't sort of what want to wash out with the bathwater or whatever the other thing expression is.
1:02:20
I think that's the biggest um current learning.
1:02:23
I'm actually very very happy about.
1:02:25
Yeah, I love learning that uh we're doing something really well that we didn't really realize necessarily and maybe we should lean into that more. Exactly.
1:02:34
Going in a somewhat different direction, Shirish Murarka suggested I ask you something.
1:02:39
He's on your board, I believe.
1:02:40
And he suggested I ask you about your 10% planning time.
1:02:43
What is What is that about?
1:02:45
This is a concept that I think Shirish has used for a long time ever since he worked at uh YouTube.
1:02:52
And the idea is that roughly, you shouldn't be spending more than 10% of your time planning versus uh executing or or building.
1:03:01
Uh which means that if you work in quarterly sort of 10 weeks, you should spend 1 week planning.
1:03:07
Since we we work in in sort of 6-month increment, so we try to spend 2 weeks planning.
1:03:13
And we're roughly successful.
1:03:13
And this is um actually, when we talk about org models, uh give a shout-out to to Brian Chesky at Airbnb, who is who is actually one of the first, I think, to have these more contrarian org models.
1:03:26
He's much more Apple-esque than most of Silicon Valley.
1:03:30
He also works in 6-month uh increments.
1:03:32
So, he has a lot of experience in that as well.
1:03:36
Uh so, that's what the 10% planning time is.
1:03:38
And I think if you find yourself planning much more than that, you're either planning too much or your execution period is just too short for that amount of planning.
1:03:47
It's a It's a rule of thumb, but I find that it it works.
1:03:53
I asked uh a few PMs what I should ask you, PMs that work at Spotify, actually, that I haven't told you.
1:03:57
And someone pointed out that you you always bring a lot of energy and clarity to a room.
1:04:02
That's something they see you as really strong at.
1:04:04
What have you learned about just importance of that or just how to do that well as a as a leader?
1:04:10
Well, that's great to hear. Um I didn't know that.
1:04:12
So, I'm trying to figure out what to answer.
1:04:15
Uh I think that uh the energy, I don't know.
1:04:20
I guess I'm just excited about what I do.
1:04:25
Uh I've always been excited about technology.
1:04:27
I love um seeing new things.
1:04:29
I My My core drive is still this notion of, you know, you you see something, which I think you'll you'll empathize with, that doesn't exist yet.
1:04:38
And you're like, "Wow, I wonder if that could exist. That would be so cool."
1:04:42
And then in order to get people to do it, you try to share that excitement.
1:04:46
So, I don't think I can be bring a lot of energy for something I'm not excited about.
1:04:53
So, I kind of have to work on things I actually believe in and that I'm excited about.
1:04:56
And so, maybe then the the energy comes more naturally.
1:05:00
Unfortunately for me so far, Spotify has been in this phase where a lot of innovation is allowed and I'm even asked to try to do new cool things.
1:05:09
Maybe I would have less energy for a pure optimization phase.
1:05:13
Uh on the clarity, I've always liked uh trying to explain things.
1:05:19
It's It's a you know, well-known fact that the best way to understand something is to try to explain it to someone else.
1:05:25
So, I try I go around explaining things to people who didn't ask for it.
1:05:29
And not to sound smart, but to see if I actually understood it.
1:05:32
And so, may maybe it's that practice.
1:05:35
And And on that note, I actually do ask my leaders that work for me and I ask them to ask their leaders to always explain themselves.
1:05:46
And I think when you we talked a little bit about autonomy and so forth, I don't think we don't promise everyone that they have to agree.
1:05:55
But I think the promise we should make to all employees is that even if they don't agree, they should be entitled to understand why you're making the decision.
1:06:03
What I don't think is acceptable is to say, "No, we're going to do it this way because I'm I'm more senior.
1:06:11
I've I've seen this a bunch of times.
1:06:14
Uh you're not smart enough."
1:06:15
Like all of those things.
1:06:18
I think you have to explain yourself.
1:06:20
So, you owe an explanation.
1:06:21
And I find that valuable.
1:06:21
Back to like the only way to understand something is to explain it because it usually turns out that if you can't explain it yourself, you probably don't really even understand it yourself.
1:06:33
Sometimes, I think it's possible that you can have product instincts that are good, but you can't express them.
1:06:42
But most often, when people say there's something there, you know, but they can't explain it, they actually don't understand themselves.
1:06:51
And many times there actually isn't anything there.
1:06:54
And also, if you can explain it as a product person, that knowledge is now shared.
1:06:58
So, it just becomes much more effective for the organization.
1:07:03
So, I sometimes I try to provoke people a little bit and say, you know, when people ask like how much is art versus science, you know, I say it's 0% art, 0% magic, and 100% science.
1:07:18
And that's because I want to force people to try to explain it.
1:07:24
I think we used the word art and magic we have historically used the word art and magic for anything that we couldn't yet explain.
1:07:31
You know, um genetics was was magic and art until it was science.
1:07:38
And uh you know, quantum physics was magic until it was science.
1:07:41
And most recently, actually, intelligence and creativity was art and magic until it was statistics and an LLM.
1:07:49
So, I I think um I try to push people to say, "Are you sure you can explain this?"
1:07:56
Because that forces people to think through.
1:07:57
So, that's um maybe I like it and I try to force it on people, so maybe that's why people think I sometimes bring clarity. I love that.
1:08:09
Question along those lines.
1:08:11
Is there a system or an approach to explaining that you recommend?
1:08:12
Is it just like write it out in a document?
1:08:13
Is it explain in a certain style?
1:08:15
Or is it just like however is natural to the person?
1:08:19
I used to write everything.
1:08:22
And then write and rewrite and make it more and more condensed.
1:08:26
So, that that worked for me.
1:08:29
I don't write as much anymore.
1:08:31
Now, I tend to like walk and talk in my head myself.
1:08:35
What What I actually do is I am And And I found this different for different people.
1:08:39
A lot of people want to bounce something with someone else. That's how they think.
1:08:45
You kind of repeat the same thing again and again and you get some feedback on it.
1:08:49
And so, I used to write a lot.
1:08:49
I sometimes do when it's an idea I want to understand better.
1:08:56
And at some point in my life, I would love to write something real like a book or something.
1:09:00
But what I do increasingly now is I do my one-on-ones with peers or people who report to me or something.
1:09:07
And I just put on AirPods and do like a distributed walk and talk.
1:09:11
Both people are walking, but in different locations.
1:09:15
And you spend an hour discussing something.
1:09:18
That has actually turned out to be very very fruitful.
1:09:21
So, then you get the power of you're not alone.
1:09:22
So, you get more brainpower than your own.
1:09:27
And I think I you know, I don't think there's strong evolutionary proof for this, but there's certainly indications that you're thinking better when you're walking, whether it's because you're oxygenating your brain or because it's evolutionary for some other reason, I'm not sure.
1:09:43
But I found that walking, talking, and thinking uh actually even if you're not in person, just over AirPods, is super effective.
1:09:51
It was the pandemic that kind of forced this.
1:09:53
I thought we would get less creative and that strategy strategizing would suffer during the pandemic.
1:10:01
You know, I have found the opposite.
1:10:01
We had more of this than ever and I started thinking about why and I think it's all that these walking talks that we did.
1:10:07
You can't have a router that you want to write a book someday.
1:10:08
What do you think your book would be about? I have no idea. No idea.
1:10:15
Uh statistically it's probably going to be about something that I did a lot, so it has to be about something with technology or product or something. Mhm.
1:10:22
But I would love to write something fictional.
1:10:24
That would be a lot of fun. Oh boy.
1:10:26
Uh I'll pre-order as soon as that's up.
1:10:29
Another concept I wanted to touch on that another PM suggested, which is he called it uh the pee in the pants analogy.
1:10:37
Does that ring a bell and is that interesting to talk about?
1:10:39
I don't know exactly which occasion this person is referring to, but I know I've used that analogy uh a few times. Okay, promising.
1:10:50
I uh the I don't know if it's like a Swedish analogy um because I I thought it was more more widely known.
1:10:57
But the idea is that you do something, so the the saying is that's like peeing in your pants uh in you know, in in cold weather.
1:11:04
It feels really warm and nice to begin with and then after a while you start to regret it.
1:11:11
It's about being being short-term, basically.
1:11:14
So now I just say that I just say like that's like peeing in the pants instead because people know what I mean. It's short-term thing.
1:11:19
That's a hilarious way of communicating that idea.
1:11:24
Uh must be a Swedish thing.
1:11:25
Yes, I think yeah, Swedish people do it for some reason. Apparently others don't.
1:11:30
Maybe because it's cold a lot of times a year. Yes, that's probably it.
1:11:31
This is a saying in cold climates.
1:11:33
In the warm it doesn't doesn't help.
1:11:35
No one understands what you mean.
1:11:38
Speaking of Sweden, uh do you watch Succession? Yes, I do.
1:11:42
Okay, so Sweden has become a big part of the show, uh specifically the company trying to I guess I don't want to spoil, but there's a character that's really important.
1:11:50
Yes, exactly, that is Swedish.
1:11:52
And so I'm curious just what what do you think of the way they portray the Swedish culture and Swedish business dealings?
1:11:59
It's it's super fun to see this as a Swede.
1:12:02
And and I guess for first and foremost like anyone or any person or any uh country that gets represented by super tall, well-built, great-looking Alexander Skarsgård should probably be pretty happy. So So I that's good.
1:12:19
Then I think um this is episode where they are in Norway without giving away too much. Yeah.
1:12:25
It's it's uh there are elements that are authentic.
1:12:27
There's a lot of uh of uh I think um paid brand positioning from a Swedish brand name Fjällräven, which I think means arc- arctic fox. Mhm.
1:12:36
Um which is actually very popular outdoor brand in Sweden, so that's kind of that's kind of authentic.
1:12:45
The the sauna things and so forth are authentic.
1:12:48
So it's like it's real, but it's it's exaggerated.
1:12:50
Actually the thing that isn't very authentic is his uh negotiation style. Mhm.
1:12:55
Uh Swedish people tend to be serious, cautious, and and this guy is more of a player.
1:13:05
So it's he's not the typical Swedish businessman from a negotiation tactic point of view, I think.
1:13:10
Yeah, it doesn't make me think of the way you described it where in Sweden people sit in a circle and no one's in the center. No, exactly.
1:13:16
He's very much in the center.
1:13:19
And then when people go saunas, they're just like a chant, sauna, sauna. Like that. Exactly. The last episode. It is a great show. I love it. I love it. This season is insane.
1:13:27
I'm so curious where it all goes.
1:13:30
Maybe just the last question before very exciting lightning round.
1:13:32
Spotify is at this point the biggest podcasting platform for me specifically and I think globally. And I love using it. It works great.
1:13:42
I'm curious just what's next for Spotify and specifically Spotify podcasting.
1:13:48
There are two sides to it.
1:13:48
It's for Spotify creators and for Spotify listeners.
1:13:53
For Spotify creators, there are two things.
1:13:55
One is and this is what we talked about at Stream On.
1:13:59
We talked about it mostly for music music discovery, but the same problem it's the same problem and even harder for podcast.
1:14:07
So we're still focused very heavily on helping spot on helping uh podcast creators find more audience.
1:14:15
Uh this is like I said, it's even a bigger problem to break up break out of your habits and your bubbles in podcasting cuz it's such a big investment to to um find a new podcast.
1:14:24
And so that is something I think we could and should do really well.
1:14:29
So we keep investing a lot there.
1:14:33
Um and as I said, you'll see more as we roll out more features now.
1:14:37
The other big need for for creators is monetization.
1:14:41
And you know, you can monetize today in many ways with uh uh DAI and so Spotify SAI and so forth, but we're working hard to to expand that and make it better cuz the the industry is starting to mature and I think this is one of the biggest needs and the biggest things we could do for for creators to help them monetize better.
1:15:02
Actually both free and paid.
1:15:02
We also have paid podcasts.
1:15:05
So that's on the creator side.
1:15:07
On the consumer side, I I don't want to share too much.
1:15:09
We've shown that we're investing a lot in discovery.
1:15:14
I want to keep some secrets for when they roll out, but we are investing a lot in the user experience itself.
1:15:21
I think it's far from optimal yet um what it could be.
1:15:27
One thing that I can share that we're investing a lot in is just the ubiquity and playback across different devices and in cars and all these things that we've done well for for music.
1:15:37
But I think the listening experience can get a lot more seamless.
1:15:41
Uh I think uh search can get better, the data about podcasts and well, I don't want to say too much, but looking at you know, AI and generative technology, there's there's a lot that can be done.
1:15:52
All right, well I'll take I'll take what I can get.
1:15:53
With that, we've reached our very exciting lightning round.
1:15:57
I've got six questions for you, Gustav. Are you ready? I think I am. Let's do it.
1:16:03
Okay, let's we'll find out.
1:16:03
What are two or three books that you've recommended most to other people?
1:16:08
Okay, this is why I tried to squeeze in seven into two and three.
1:16:12
So if we start with the on uh product, I think uh it's well known, but one that I would recommend for other people to read is Seven Powers by Hamilton Helmer, which um Netflix has used a lot, we use a lot.
1:16:23
It's just if you're starting out, great to have a strategy framework.
1:16:25
No strategy framework is right, but having one is better than none.
1:16:31
Another in sort of the space of mental models and frameworks, I think is the The Complete Investor by Charlie Munger.
1:16:38
So it it's yes, it's about investment, but really it's a bunch of mental models that he uses and I think the the key takeaway is you have a problem, you should always apply three different models to it because what models do is they they um simplify and and reduce dimensionality.
1:16:56
You know, the world has probably infinite dimensions and it reduces to maybe three or four.
1:16:59
And the risk with that is you happen to get rid of a really important dimension like, you know, maybe pandemic diseases or something.
1:17:08
But if you use three models that have different dimensions and diff- was reduced in different ways, statistically and and it comes to the same conclusion, even the second model you apply vastly increases your chances that you're right.
1:17:22
So that was a good book to read.
1:17:24
Then I think if we go outside of product, I'm very interested in just science and mathematics. So a few quick ones.
1:17:33
The Mystery of the Aleph, an amazing book.
1:17:37
Something Deeply Hidden by Sean Carroll on the Everettian interpretation of quantum mechanics.
1:17:42
Uh Helgoland by Carlo Rovelli on the relational interpretation of quantum mechanics.
1:17:49
The Beginning of of Infinity and Fabric of Reality by David Deutsch.
1:17:54
The Case Against Reality by Donald Hoffman on sort of evolution versus truth and that evolution doesn't optimize for seeing the truth, just for fitness.
1:18:03
Gödel's Proof, I think is an amazing book on his uh incompleteness theorem that in sort of any axiomatic systems, there will be true statements that can never be proven, which is a weird thing to to think about.
1:18:19
And then uh maybe one of my favorites is The Demon in the Machine by Paul Davies that I think is lesser known on how information is really just entropy and this concept of information engines that you can power something by just information and the exhaust is also information.
1:18:36
That was not a quick list.
1:18:36
No, but but uh no, I was just going to say you've set the record for the most number of books, but it also shows how you've become so insightful and wise is just, you know, reading books like these.
1:18:46
And so I think if people are looking to, you know, get to a place that you're at now, I think it I think there's a lesson.
1:18:54
I'll keep the others much shorter, I promise. And it's all good. We got we got time. Okay, next question.
1:18:59
What's your favorite recent movie or TV show?
1:19:03
So we talked about Succession and it is a recent favorite.
1:19:04
So I'll just previously take something that isn't recent, but is an absolute favorite, which is uh Halt and Catch Fire, which I think is on FX.
1:19:15
Amazing show if you ever worked in technology.
1:19:17
It kind of starts out in the Silicon Prairie in the '80s and follows up to present day. Amazing show. Halt and Catch Fire.
1:19:24
Yeah, I watched some of it.
1:19:25
I actually fell off of it, but I'm going to It's a good reminder to go check it out. Got to go back. I'm going to go back.
1:19:30
What's a favorite recent interview question you like to ask?
1:19:34
I don't ask it, but my favorite question is Lex Fridman's small ending question that is usually something like, "So, what's the meaning of it all?" I like that.
1:19:43
It's a tough question to get.
1:19:45
I'm so tempted to ask you, but uh No, don't. Okay. Let's move on.
1:19:50
That'll be another pod That'll be our second second take at this.
1:19:54
What is What are some favorite products you've recently discovered that you love?
1:19:58
The obvious one is is uh Chat GPT GPT-4 and just playing around with that, trying to create bots for yourself that do different things for you and so forth.
1:20:06
Uh but, I don't think that's um that's probably true for everyone.
1:20:10
The other really favorite is something you've written about and talked about, which is uh Duolingo, which I think is both very impressive from a product point of view, the the execution and and uh what they've done.
1:20:23
It is also insanely used in my family.
1:20:25
We have a family account, and everyone is, you know, using it and competing every day.
1:20:31
So, I'm both impressed by the product and also use the product quite a lot.
1:20:35
What languages are folks learning within your family?
1:20:38
In my family, it's Spanish right now. How's it going? Um bien. You got a gold star.
1:20:48
have like a few thousand XP. I'm not that good yet. So.
1:20:50
I don't know if that's good. That sounds pretty good.
1:20:54
Next question, what's something relatively minor you've changed in your product development process that's had a tremendous impact on your team's ability to execute?
1:21:00
I'm not sure I've done anything minor that had a tremendous uh impact.
1:21:08
Uh usually it takes something bigger to get big impact.
1:21:13
I think maybe one thing that I've tried to do back to like clarity and so forth is uh this thing I mentioned about I'm trying to push a lot for what I call Socratic debate, uh uh where the idea is obviously that the best idea wins, not the most um senior idea and so forth.
1:21:32
And and trying to push for this uh notion of uh having people explain themselves.
1:21:36
Not saying like, "I think there's something there. I have a feeling." or something like that.
1:21:43
And and apparently, as as you said, that has had some impact cuz people apparently say that about me. So.
1:21:50
That's probably the biggest thing.
1:21:53
Final question, what is one fun ritual of the Spotify product team, and uh is it saunas?
1:21:59
So, Spotify is so big now that uh we don't It's quite local, actually.
1:22:04
Different parts of Spotify have different product rituals.
1:22:07
I acci- sort of accidentally created one ritual many years ago, maybe 12 years ago, when we talked about um where where a product which phase a product is in.
1:22:17
And it was, you know, we needed some definition.
1:22:20
So, I think sort of off the cuff, I said like, "Well, you know, it's four phases.
1:22:23
It's uh think it, build it, ship it, tweak it.
1:22:28
And in the think it phase, you should you should it should be cheap, you know, not a lot of money spent.
1:22:34
In the build it phase, you're going to start spending a lot of money.
1:22:35
So, then you must have reduced the risk in the think it phase that you're right.
1:22:38
And then then you have the ship it phase, and then you go over and tweak it.
1:22:44
And it was something that wasn't that thought through.
1:22:47
Uh but, it's funny because I still hear it, sometimes even from other companies.
1:22:50
Like, "Oh, we're in the think it phase."
1:22:51
Or, "We're in the tweak it phase." So, it kind of stuck.
1:22:55
I don't know if it's very good, but it's stuck. It is catchy.
1:22:56
I think that the Anything getting stuck in people's head is a success.
1:23:02
Gustav, thank you so much for being here.
1:23:04
Uh we are two for two for Swedish people.
1:23:06
Gustav with an F Alstromer uh was on the podcast, who who is also an amazing person. Also an amazing person.
1:23:13
I am I feel very uh jealous of people that get to work with you and for you.
1:23:17
Thank you again for being here.
1:23:17
Two final questions, where can folks find you online if they want to learn more, maybe reach out, ask some questions? This is this @gustavs. Okay. Say it again. @gustavs. Awesome.
1:23:29
And then final question is just how can listeners be useful to you? Just reach out.
1:23:33
I do read feedback, and I try to remove the the the angry comments and understand what they're actually thinking and why, you know, why they're upset or what's not working.
1:23:44
And then the reaching out, would you recommend a an angry tweet at you or more of a email to that email address you shared?
1:23:49
Well, the the the the @gustavs is the Twitter handle. So, just tweet at me.
1:23:54
And you can be You can be nice as well. It's okay. Amazing.
1:23:58
Gustav, thank you so much for being here.
1:24:00
Thank you for having me, Lenny. It's been a pleasure. Bye, everyone.
1:24:05
Thank you so much for listening.
1:24:05
If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.
1:24:13
Also, please consider giving us a rating or leaving a review, as that really helps other listeners find the podcast.
1:24:19
You can find all past episodes or learn more about the show at lennyspodcast. com.
1:24:25
See you in the next episode.