0:00
We always try to reproduce a small part of your neocortex on our servers.
Search this video
0:02
Can you describe how you think this will play out?
0:06
That's why startups are ahead.
0:06
They don't have to rebuild.
0:07
They don't have 15 years of data.
0:09
We are the R&D department of the music industry.
0:10
Talk is cheap, so we should do a lot of it.
0:24
Maybe a fun place to begin is the obvious place.
0:28
Everyone is facing this giant shift in technology.
0:31
My friend Ravi Gupta calls this imperative AI or die.
0:36
That companies, even the big sexy established technology companies, need to find ways to embrace and use this new technology as it unfolds or face elimination.
0:46
I would love to hear how you and Spotify are thinking about this challenge.
0:52
I know you've embraced it very quickly and you were very early to using machine learning and and data and data science all over the product.
0:58
But this is a big shift and you and Daniel and the team are some of the most thoughtful people about addressing shifts like this and you've done it before.
1:10
Walk us through in some detail how how you first felt it, what you did about it, what it's like to be at a big company and process something like this.
1:19
Yeah, it's it's a great question.
1:19
I think it is the right description, at least in the longer term, I think it is AI or die.
1:24
Just like it was smartphone or die and before that internet or die, computer die.
1:30
This is one of those shifts that, you know, it's not your choice whether you adopt it or not.
1:34
It's going to happen to you.
1:36
It's it's the the epitome of a macro wind, I would say.
1:40
And usually when these macro winds come, we have a saying internally that, you know, you can have the macro wind blowing your face and it's not going to change its direction.
1:48
So you basically need to reposition yourself so you get the wind at your back and you can sort of surf this this macro wind or if, you know, some people call it macro wave that you surf.
1:58
So we've been through a few of these.
1:58
Um the first one was really the smartphone when that came along.
2:03
Spotify was really well positioned for for the internet before the smartphone where we had a free tier on desktop and that's where we sort of acquired users and they built a playlist and they they retained themselves.
2:17
And then mobility to listen on the go was a paid feature on Spotify.
2:19
And that was fine when the majority was computers and the minority was was smartphones.
2:27
And then when smartphones took off, we faced an existential crisis where there started to be consumers who didn't have a desktop.
2:34
They only had a phone, so they had no free experience and our entire model died.
2:37
So so that was one of those moments we had to reposition the entire business model actually and figure out how do we do a free tier on mobile that doesn't cannibalize the paid feature which was mobility and we we can talk about how we figured that out later.
2:52
But that was one of those examples and I think this is a similar one.
2:54
The big question to me is does this require a business model change or is it {quote} just a product change?
3:04
The other thing that is different, I think, about AI is that it's not it's not going to touch one thing.
3:09
It touches the consumer product, but it also touches your productivity and competitiveness as a company.
3:13
So there there are lots of um different angles to start, but as you said, we were quite early with machine learning and the journey we had was, you know, we saw users coming on Spotify and then they started playlisting and that retained themselves.
3:30
But it was only a certain amount of people who were good at playlisting because you have to know the catalog in your head, the new releases, the back catalog.
3:37
So some people retained themselves really well.
3:41
And then we tried to scale that behavior by having editors who created playlists for people who couldn't playlist that well.
3:46
And we saw people using social to find inspiration.
3:48
Eventually machine learning started happening and we saw this opportunity of sort of building a music friend for everyone.
3:55
So so that's where we started.
3:57
We started investing in that and got quite good at that.
4:00
I think some people say that AI is just machine learning, it's just a new word.
4:06
And it's an interesting question, what is the difference?
4:08
I think the difference between what people used to call machine learning and what we call generative AI is that the statistical machine learning was sort of an output mechanism and I think the epitome of that age is the full screen TikTok feed.
4:23
It is like, you know, UIs shape themselves after technology that powers them to maximize metrics and I think that is the UI that maximizes the statistical explore exploit paradigm of of old school machine learning.
4:33
What happens with generative AI, I think the big shift is that you can take natural language input.
4:41
And so even if technically they're both machine learning, I think of generative AI as a as a new age.
4:45
And the big shift is that it's two-way.
4:48
If you think about Spotify, for example, as consumer product the way it looks, it's it's almost like a old school um broadband.
4:59
You know, the broadband where you had like maybe 1 megabit down link but only like 150 kilobit up link.
5:04
So a lot of bandwidth down, but not a lot of uh feedback.
5:09
This is what most consumer services look like.
5:11
You have um you know, streaming video on the down link, a lot of information per second.
5:19
But the up link is only like a few clicks and swipes.
5:20
It's very very narrow signal.
5:22
And this is what the previous machine learning age focused on.
5:24
I think what changes in the age of generative AI is that the up link can now be English language.
5:30
It can be almost as rich as the down link.
5:33
And I think that requires all of us consumer companies to in the limit totally rethink the product.
5:37
So if you just do the deduction of what I said if the full screen TikTok feed is the epitome of the ML paradigm the the asymmetric down link up link paradigm what are the chances that that is also the epitome of this generative AI age? I don't think so.
5:52
I think consumer products are going to change fundamentally.
5:55
I can't predict exactly how.
5:58
I think they're going to be much more symmetric in terms of information you receive versus information you give.
6:01
And I think if you fast forward 5 to 10 years, almost all big consumer products are going to be a conversation to some extent rather than this service that you use.
6:12
So really the job of us for us on the product side is to try to figure out what is the next paradigm.
6:18
And I don't know exactly what it is yet.
6:19
We're experimenting and if I did know, I probably wouldn't tell you right now. I'd sit on it for a bit.
6:23
But this is where on the on the product side.
6:27
And then we can talk a bit about um the productivity side as well where there are the obvious the obvious gains in terms of uh coding productivity where we are using all the tools that everyone else is doing.
6:41
But as a big company, there are few differences from the startups because so far um generative AI in coding has had the most impact when you write net new code which is a lot of what you do as a startup and a tiny bit of what you do as a big company.
6:55
Most of it is just refactoring, etc.
6:58
And I think I saw some statistic that you know, in a big company you basically code one out of every 8 hours in a day.
7:07
So not only is coding only 1/8 of the time, of that 1/8 of the time net new code is very small.
7:14
So I actually think the biggest impact is yet to come uh when it comes to coding. That's two things.
7:17
These models are getting big enough to understand really large and complex code bases like Spotify's.
7:23
And we're not quite there where these things can refactor our code base.
7:28
You know, it doesn't have quite a deep but it will.
7:31
And that will be a big shift.
7:34
The other is doing automatic peer review which is like just on the verge of that working.
7:40
It's not quite good enough that you can trust it.
7:41
So a lot of developers sit and wait for their code to to be in review and come back.
7:47
So I think we're we're seeing that ramp.
7:50
I think we're going to see it ramp a lot in the next few years.
7:52
But then the really interesting side is these other 7 hours what a developer does, which is a lot of communication, planning, working with designers, prototyping meetings.
8:04
Those things I think will actually have as big or even more impact than the coding itself.
8:09
I want to start with this down link up link part.
8:10
What have you learned about consumers' willingness to put a lot of effort into the up link?
8:16
It seems like the the chat interfaces, the the GPTs of the world, we know that people are willing to do a lot of back and forth.
8:25
And because it's the native interface.
8:27
Like you're going there expecting to write a lot of stuff.
8:29
Copy prompts from Twitter or whatever.
8:32
In an app like Spotify, how willing are people to get not lazy and really descriptive about what they actually want?
8:39
What have you learned about like the nature of people's laziness versus willing to put a lot of work in to get the thing that they want via that more rich up link?
8:47
Yeah, so that's that's probably the most exciting thing for us of this generative AI age and the and the dual up link paradigm.
8:55
So previously we mostly relied on some explicit input when you playlist.
9:00
That's like high value information.
9:02
You are sitting there thinking like this song goes really well with this song and that song.
9:06
So it's if you think about it as labeling even though you're playlisting for yourself, you're sort of labeling these tracks in terms at least in relation to each other.
9:15
And you're putting a lot of effort in it.
9:16
And that that was our was and is our big advantage in music recommendations even though generative recommendation systems are starting to take over from these more old school collaborative systems.
9:27
So we had some really strong signal like that where you quite seldomly invested a lot of time in producing a data set that described you about playlisting.
9:36
But most of the time we just had skips and the challenge for us is the phone is in the pocket.
9:40
So even if we had like a thumbs up down, you're not going to take out the phone every time and say like I didn't like this because of that or even do a thumbs up thumbs down requires you to take out your phone, unlock it, open Spotify.
9:52
What you can do from your earphone is to skip.
9:54
So we have the skip signal, but that is a very blunt signal.
9:56
So, we play your song and you skip it.
9:59
That could be because you absolutely hated it.
10:01
Could be because you love it, but it's a hundred times, you're tired of it.
10:05
It could be that you love it, you're not tired of it, but you're at the gym.
10:09
So, jazz is not the right thing.
10:11
Like, all of those just look like a skip to us.
10:14
So, that's a very blunt We have a lot of that signal, but it's blunt.
10:16
And you will never get to like perfect personalization through that.
10:21
Now, what we find with generative AI, one of the first services that we've launched that is live now in like 40 countries is something called AI playlisting.
10:29
We literally use an LLM that is trained on on your your listening data and world knowledge and so forth.
10:34
And you can literally tell us in English what kind of playlist you want.
10:37
Previously, you could playlist songs and maybe you put a title on it and we could guess like this is probably a running playlist.
10:43
So, we could do something.
10:44
Now, you can say I want a running playlist that is EDM. I want big drops. I want it to be 160 BPM.
10:51
And then you get a suggestion from the LLM and then you can keep a few tracks and say these were good, these were not good. Now, refine it.
10:55
I I don't like these artists, but I want more of that.
10:58
So, for us it's the first time that we get that kind of fidelity of what is actually in the user's mind.
11:02
One one way to think about Spotify is we always try to reproduce a small part of your neocortex on our servers.
11:09
It was just very hard with a clickstream of skips.
11:11
Now, when you tell us what is in your mind, it gets easier to approximate uh you as a person.
11:18
So, so this is uh this is really the first time that we have that signal.
11:22
Now, one way I like to think about it is when we do user research, we do two things.
11:27
We do quantitative testing, AB test, but before that, we do qualitative testing.
11:32
We interview a few people deeply to understand the need.
11:34
Then we build a product and then we AB test to see if we're right.
11:38
The promise of generative AI is AI is really a deep, ongoing qualitative user research with almost 700 million users all the time.
11:48
And, you know, it sounds big, but if you squint at it, that's kind of what it is.
11:51
It's interesting how many different ways you could take the product with this new technology.
11:55
I would be really curious to know the apparatus inside of Spotify, like the leadership team, the product team, and literally how you run the process of deciding what to do with your You have a lot of a big team, obviously, but no matter what, you have limited effort, limited units of energy you can apply.
12:15
there's this huge space of stuff you could do with this technology and the exciting advantage that you have of all these 700 million users, what is the pro- the literal like meeting-by-meetings process, the setup process look like?
12:25
And the reason I'm asking this question is so many companies face this same challenge.
12:28
It's exciting, but also scary that they need to get the innovation before somebody else does and and and disrupts them.
12:34
So, what is the background process for how you arrive at the things you might try?
12:41
So, there there are really two things.
12:42
We have a very structured process um that that I want to talk through how it works, but there's also there are also some um concepts that we use.
12:51
Uh and over the years I've introduced some strategic frameworks to the company, some of which I know you're passionate about, like Seven Powers from Hamilton Helmer.
13:00
Um bundling framework that Tushar wrote and wrote.
13:04
He's on the board, right? He's on the board.
13:05
And secret secret weapon. Exactly. Very good secret weapon.
13:09
And also I found um Better, Simpler Strategy by Felix Oberholzer to be very good. What's that one? I don't know that.
13:18
It's a concept of the value stick where you have um willingness to pay, which is very important for us.
13:23
Like, the the way we measure value is the willingness to pay.
13:26
But what it introduces is also the willingness to sell.
13:28
If you think about your your staff, like what is their willingness to sell their services to you?
13:33
And everyone focuses on increasing the willingness to pay, but you can also sort of increase or depending on how you think about decrease the willingness to to sell.
13:41
And it turns out like the best companies in the world are not necessarily the ones that actually pay the most, it's the ones with the with the most interesting mission, the best culture, et cetera.
13:48
So, so because we're a bundled service where we try to just give users more and more value all the time.
13:56
You know, we put in lots of music, that's value.
13:57
Then we put in more podcast, uh that's value.
13:59
Now we put in books, it's more value.
14:02
This framework of willingness to pay and willingness to sell is very useful for It just fits our business really well.
14:10
And the job of us is to keep the willingness to pay quite far from the actual price.
14:14
Like, that gap is how much consumer surplus you're giving.
14:18
And our goal as a service is to make sure that this Spotify is just an amazing deal.
14:21
You're always going to feel like the willingness to pay, the actual value you perceive is way over the price that we have.
14:27
So, we use that framework quite a lot.
14:29
So, introducing these frameworks, not just in the business org, but also in the product and technology org, makes people think in more structured ways.
14:38
It makes people have a vocabulary.
14:39
We can talk about network effects, amortization, um you know, brand power, all of these things.
14:47
So, I spend a lot of time getting the teams to use these frameworks so that we have structured strategic thinking.
14:54
And the other thing I try to push as a thesis is that you know, I'm a big fan of sort of Socratic debate.
14:59
I'm amazed, like many other people, at how far you know, the Greeks and the Ro- and the Romans came with just discussion, even though they didn't have science, just reasoning.
15:09
Strong reasoning is very useful.
15:10
So, I try to push this sort of provocative line of talk is cheap, so we should do a lot of it.
15:16
Sort of as a counter to like moving fast and breaking things. Yeah.
15:21
It's like it's so cheap to talk, so we should actually do a bit more of it.
15:23
If you you know, sometimes it takes you to like, you know, it took the Greeks to the concept of the atom, you know?
15:31
So, we do a lot of talking and and ideation that is quite structured.
15:33
And I do that with my leadership team and often the leadership team sort of plus one.
15:39
Which means the the VP layer and sort of the director plus.
15:43
And I have a lot of time just for discussing concepts.
15:46
And so, um back to one of my my heroes in life, David Deutsch, uh his book The Beginning of Infinity and The Fabric of Reality shaped me quite a lot.
15:57
And he talks about something called good explanations.
16:00
And he has a list of what a good explanation is.
16:02
It is obviously needs to be falsifiable and so forth, but it also needs to be to have reach. It needs to to scale.
16:09
So, an okay explanation explains this phenomena, but it doesn't scale to other phenomenon, right?
16:14
Doesn't scale up and down.
16:14
Really good explanation scales, you know, from from explaining how the Earth works to the solar system to the planets.
16:20
But it's also very hard to vary.
16:23
Which I think is often underestimated.
16:26
If you have an explanation and you can switch it out for another explanation that explains the same thing, like, you know, if you explain the weather using gods, you can switch out this god for that god, it's probably not a good explanation.
16:35
It needs to be very hard to vary.
16:37
If you vary it, it doesn't explain it anymore.
16:40
Um the last thing he says is that explanations should not just be predictive.
16:45
That's not an explanation, that's a model.
16:46
An explanation needs to explain why.
16:48
So, I try to push my teams even if something works in an AB test, I tend to say like I don't want to launch it until you have a good theory of why it works.
16:56
Because if you figure out the why, i- i- if it's the difference between pattern recognition and actually understanding something.
17:06
Pattern recognition is useful, that's called seniority.
17:10
You know, I love people good pattern recognition, but if they can explain why it works, it scales to the entire org.
17:15
Other people can use that knowledge.
17:16
It's it's much much more valuable.
17:16
So, those are some of the concepts that I've tried to put into the org over time.
17:20
So, so I think that's important because that shapes the culture.
17:24
Then we have the structured process, which is we execute for 6 months at a time.
17:27
We have something called a bets process where all the VPs, which is about 14.
17:35
So, one of the benefits of Spotify is they're so small that all the VPs, the entire company can fit in one room.
17:41
And we meet and we meet 3 hours every Tuesday.
17:43
The entire So, so the company is completely synchronized, for good and bad, and we can talk about that later.
17:50
But so, every 6 months, these VPs they they pitch, literally pitch, as if we were a VC and they were a startup.
18:00
Um the bets that they think the company should do and why.
18:04
And it's very much like a startup process.
18:06
You know, you don't get to use the fact that, you know, um Gustav or Alex or Daniel may like you.
18:11
Like, you know, this is like a VC meeting. You have to convince us. So, they pitch.
18:16
Then um me and the other co-president, Alex Nordstrom, we decide based on these pitches a stack rank, a global stack rank.
18:25
This this time we have 44 bets.
18:25
Happens as usual between 30 and maybe 50.
18:31
We stack rank them from 1 to 44.
18:34
Then we go out to the org and say, "Now, try to resource this."
18:35
And they start from the top and then maybe they get to 30 and say, "This is what we can do in the next 6 months."
18:40
And then they commit to those things. And we start executing.
18:45
And it's a good mix of sort of bottoms-up innovation where you leverage like not just Daniel, not just me and Alex, but all the VPs and the layers below to come up with good ideas because they're closest to the user.
18:58
But then there's global synchronization.
19:00
We stack rank them, make sure that they fit a single strategy, then it's back to the org and they commit.
19:03
And as I think you know, you're going to be much better at delivering something if you were the one who said I can do this than if your boss said you can do this, right?
19:09
So, so that's the the process.
19:12
But leading up to that, we have something called a prototyping phase.
19:16
So, the previous 6 months, we prototype in the combination of Figma and increasingly gen AI tools what Spotify should look like after the next 6 months or could look like.
19:28
And this prototype also helps synchronize the entire company.
19:30
What I found previously was that when people submitted these bets, everyone had in their mind what their great feature would be.
19:39
You start building and then down the line you realize that you were not actually aligned.
19:44
And then you get a lot of fighting towards the end of the cycle where, you know, this thing doesn't work with that thing and, you know, things don't don't work out so well.
19:51
What I've tried to do now together with Alex Nordstrom uh we synchronize the entire company.
19:58
Alex and I don't have our direct reports team.
20:00
We meet as a single team 3 hours every Tuesday.
20:03
And we try to use the fact that we're small as an advantage instead of as a as a dis- disadvantage versus our competitors who are, you know, very very large companies.
20:12
And so we we um prototype everything up front.
20:16
So, all the so- so-called you know, quote-unquote fighting um happens before you actually commit to doing something and you have something you can hold in your hand say, this is what Spotify would look like if we pull this off.
20:29
So, that's a combination of sort of cultural input and then a very structured process for actually making it work.
20:35
I I have so many questions about process.
20:37
The first is how that 3-hour meeting on Tuesday works.
20:39
Like, what is the structure of that meeting?
20:42
It's called the E-team, execution team.
20:43
So, it's very focused on execution of the company.
20:47
And the idea is that, you know, if you have five five 5-day working weeks there's never on average more than 2 and 1/2 days before you, if you're if you're blocked on something, can escalate to me Alex and all the other VPs.
20:59
So, the idea is you should never be blocked more than max 2 and 1/2 days.
21:06
Because we run this synchronized ship if you're blocked, it gets very expensive because everyone else is downstream of you.
21:12
So, if you're running a synchronized operation the way we're doing um escalation processes are very important and resolution is very important.
21:20
So, a big part of that meeting is people say, you know, we're we're off track here.
21:25
I'm um dependent on this um man or woman over there who hasn't done what they said.
21:32
And the the beautiful thing about ha- being able to have all the VPs in the same room is you know, I've met so many meetings that I'm sure you've been in people say like, okay, we'll we'll take that offline. I'll talk to you later.
21:42
And what we said is you're not allowed to say the word offline or later because that person is in the room.
21:46
So, it's like, you know, I'm dependent on maybe Anna over there for this, but then Anna is actually there.
21:52
And then Anna can say, you know, okay, I didn't know that or I'm going to solve that.
21:56
So, it's like real-time resolution.
21:59
Very simple in theory, but incredibly powerful in practice.
22:01
Most companies don't do it.
22:03
So, this notion of like, no, taking it offline, taking it later, real-time resolution. That's why it's 3 hours.
22:10
So, that's one thing of this meeting.
22:10
Uh another principle we have in that meeting except um nothing goes offline is you actually can't bring uh your direct reports for good and bad.
22:22
The idea is that if you bring in a lot of direct reports two things are going to happen.
22:27
One is the VP is not going to get forced to know the details as much.
22:33
So, I'm trying to literally force the VPs to to solve it themselves cuz I want them to be in the details.
22:37
So, you're not allowed to bring anyone else in to explain your thing.
22:40
You have to be on top of it enough to explain it to yourself.
22:44
The other benefit of that is over time this group gets very tight because you don't switch people out all the time.
22:49
So, you build stronger rapport. People can be honest. No one is afraid.
22:51
It's a it's a very strong and high-functioning team.
22:56
Uh so, that's a lot of what we do.
22:56
The other part is uh strategy and looking forward.
23:00
So, let's say that, you know, something we've been working on for some time now that's public.
23:04
We wanted to introduce music videos.
23:06
And the team goes off and says, what does that take in terms of licensing, product?
23:09
What is the the cost implications for the company, for the P&L?
23:12
They come and present to the E-team like, this is what we want to do.
23:15
This is what how long we think it should take.
23:17
So, it's a combination of keeping the engine running and never stopping and also planning for for the future.
23:25
But we don't really plan in that.
23:25
It's too big to have detailed planning.
23:26
That happens in focus rooms with smaller groups with experts.
23:31
And then they come and present to that team.
23:33
It's been we're not the only company that does this.
23:38
I know um Airbnb does something similar.
23:40
I spoke a lot to Brian Chesky about it.
23:43
Um I think um Netflix may have have had something similar at a time, but they're now sort of divided into content and business and product.
23:49
What What I think is important about this is it's both the business and the product side.
23:53
And we talk a lot of product there.
23:55
So, the business people in Spotify, they know an awful lot about AI, about what a mono repo is.
23:59
They're there for the participation on technology, but but on the flip side all my product and people and engineers, they know exactly what the P&L looks like. They know our goals.
24:11
They know, you know, they know what gross margin is, what operating they know everything.
24:14
So, so that's quite unique and that gives them sort of a CEO perspective that I think disappears in many companies because we put on these roles of like, you're a product person, so you're not supposed to understand finance. That's not true.
24:25
If you're the CEO, you have to understand all of it.
24:28
If people are listening and are curious about this BETs board process where you can submit projects and seems like a really elegant way to allocate capital.
24:37
Um what what advice would you give them about the pros and cons of this process?
24:43
And I know you've been doing it a long time, how it's changed over time to reflect the learnings of what makes it work or fail.
24:50
So, the concept itself is actually really straightforward.
24:52
It comes from the from the Kanban board.
24:54
It kind of comes from the developer uh community.
24:57
And it actually which actually comes from car manufacturing eventually.
25:00
It's just like it's really the concept of stack ranking which is very easy in theory and very hard in practice.
25:06
Very few people manage to say this is actually more important than that.
25:10
They're just saying like, these things are very important, both of them.
25:13
And when you press them, they say like, no, they're equally important.
25:14
But then they're not ranked.
25:17
So, the real secret is to stack rank and say like, you have your two your two darlings, but if you have to kill one of them, which do you kill first in reverse order?
25:26
So, very easy, but hard to do across the entire company to agree on that.
25:28
But once you have it it removes it gives so much clarity to the org.
25:34
Cuz what happens when you say like, these three things are equally important, but they're not really they never are.
25:38
You're going to have to choose.
25:39
Is you just push the decision down the org.
25:42
And this VP who's on the hook for that thing is going to start fighting.
25:45
This VP who's on the hook for the other thing.
25:47
And if you as a leader don't bring clarity you're going to set your org up for for fighting.
25:50
And people are very nice.
25:52
They're going to think that they don't like each other.
25:54
So, just the stack ranking and being completely transparent across the entire company means that if I come to you and I say like, you know, I need you to do this and you say, yeah, but I'm doing this.
26:04
We look at the board and say like, oh right, we should do this.
26:08
It's very simple, but very effective.
26:10
Um And and when you do that, there are a bunch of things you run into, uh theoretical questions of, okay, once you have this BETs board done, do you resource it globally?
26:20
Do you go through every developer and say it's like, oh, let's just try to get as far as we can?
26:26
That planning process is hell if everyone is up for grabs.
26:28
None of my my um VPs have any any uh estimate of what resources they will have.
26:35
You basically disempower your entire VPs.
26:38
And uh it's effective in a sense cuz you do perfect globally perfect resourcing, but it's incredibly inefficient to do the planning.
26:45
So, then the question is, how do you divide it, you know, into blocks?
26:47
And then we have so, the the structure we have is we have a platform organization uh that does, you know, they they work with GCP in the cloud and the developer tools and so forth and security and all of that.
27:01
Then we have an experience organization which is responsible for the entire consumer product across mobile, car desktop, etc.
27:06
Then a personalization organization because that's so important to us um that does all the uh the AI and the recommendations and balances between, you know, books, music, podcast, video, etc.
27:19
And then we have three business verticals, music, podcast and books.
27:24
So, they have their own resources.
27:26
And what we do is we start by asking them to resource as far as they can with the resources they have without sort of stealing from each other.
27:35
And then we get as far as we can because you need to give them predictability for them to be able to plan their own work.
27:40
Um and then at the end of that process you may move some people around globally to make sure that you don't, you know, you have something really important and there are two people missing.
27:48
That's not optimal for the company.
27:50
So, you may move some people around, but largely we try to let people keep the resourcing.
27:55
So, lots of those problems that you run into.
27:58
But I would say the biggest uh risk with this model, it sounds nice if you're perfectly synchronized.
28:04
The drawback of that model is that the planning is very expensive.
28:08
So, you have to be really good at planning.
28:11
And we've had to build our own tooling.
28:13
We tried some external tooling for planning. That wasn't good enough.
28:15
And if the planning doesn't work, the overhead just grows very quickly versus execution.
28:22
And you know, we execute for 6 months in order for the overhead to not get too big.
28:27
But we can't go to a year. Then you can't react. A quarter is too short.
28:30
It's too much planning overhead versus execution.
28:31
So, the planning is the thing you have to get really good at.
28:34
And I'm not going to say we're really good.
28:35
But we're getting better all the time.
28:37
It's the thing that I sort of care the most about making sure that the planning is reasonably big.
28:43
If you can do it for for us, it's critical because the whole whole of Spotify's product strategy is that we have large distribution closing in on 700 million MAUs for a single application.
28:56
And our entire strategy is basically we decided this many years ago before it was popular, but you saw the Chinese app the Chinese starting to build super apps.
29:06
Whereas the Western world built one app per use case.
29:07
We kind of adopted the Chinese super app idea and said the hardest thing is going to be to get installs.
29:14
You could see the average number of installs from the App Store dropping like below one on average.
29:19
So, distribution became the most important thing.
29:22
And then we chose, you know, when we when we did podcast and later books and videos, we're going to build it in the same application because then we can leverage our own distribution.
29:29
Um but that has drawbacks.
29:32
You have to have a an organization because then everything is dependent on each other.
29:35
You're going to ship one app to the App Store and everyone is a stakeholder.
29:39
So, you cannot divide and conquer.
29:40
You cannot say like, well, the book team, you can run ahead or the music team, you do this.
29:44
No, everyone has to wait for everyone.
29:48
So, because of our consumer strategy, the company needed to be synchronized.
29:52
And because it needed to be synchronized, we needed a really strong planning process.
29:55
So, it's kind of an outcome of of our consumer strategy.
29:59
And what I would say is it's not the right one.
30:00
It's the right one for us.
30:02
We're good at doing global changes, like changing the entire UI, because we're synchronized.
30:07
Um but we're probably much slower than other companies at, you know, quickly trying something.
30:10
Because it needs to go through like a lot of planning and so.
30:15
I think we're both I I don't think you can win in planning.
30:16
The best you can hope for is to be quite good at the important things and not so good at the less important things.
30:22
I'm going to come back to something very interesting you said around the adoption of some of the tooling that's at the most cutting edge.
30:29
So, let's take Cursor as an example of a company that now everyone's familiar with, $10 billion valuation.
30:32
Seems like every software engineer is using Cursor to make themselves better. Yeah.
30:37
But the way you framed it was so cool that yeah, sure, but uh that's new code primarily.
30:41
That's a fraction of 1/8 of their time.
30:44
And so, there's all in the pie chart, there's it's a very small sliver that's being addressed by Cursor at at big companies.
30:50
Can you describe how you think this will play out because it feels like the public markets, especially, are are very curious.
30:56
I guess private markets, too.
30:58
Very curious about how AI companies and products and tools will address this, you know, this bigger much bigger part of the pie that sounds like really hasn't been hit too directly yet.
31:10
There are a couple of things that are interesting that I don't think are super obvious.
31:15
One is used to be that, you know, every developer started using Cursor.
31:20
But now, I'm starting to see a lot more non-developers using Cursor.
31:22
And that's partially because the industry is starting to agree on this protocol called MCP model context protocol.
31:31
Which means that if you take your internal services and you wrap them in an MCP, you can speak English to your to your infrastructure.
31:38
So, if you're a developer now, you can sit in Cursor and say like, you know, I'm going to Or or if you're a designer, for example, or a product person, let's say you want to prototype a feature in Spotify.
31:51
One workflow is you take the existing Spotify, you double click and screenshot it, you upload that into Cursor and say, "Wire this up clickable in HTML."
32:00
And then if your services are wrapped in MCP, you could theoretically say, "Now, wire this up to, you know, our my liked music my my liked songs feed or or something."
32:09
And you can prototype even though you're not a developer, because the infrastructure is wrapped in in English language now through MCP.
32:15
I think that's an important thing.
32:17
And so, I think you're going to see many more people using Cursor than just developers.
32:22
I'm starting to see I had one of my PMs who is uh in Sweden. She's from New Zealand.
32:27
She doesn't speak Swedish.
32:28
She did her taxes in Cursor.
32:30
Managed to wrap the Swedish tax authority in an MCP.
32:32
Not not a developer, right?
32:34
So, I think it's going to grow outside of of developers.
32:36
But I think this points to what is actually happening in many of these big companies, which is why the startups can move faster.
32:43
So, if you think of a company like Spotify, it has tons of infrastructure.
32:46
You know, you have the database with play history going 15 years back.
32:50
You have, you know, who is in the family plan.
32:51
That's that's one server.
32:53
This one's data set somewhere.
32:56
Your taste graph is a data set and so forth.
32:59
Now, you know, here comes the the big AI companies and they give you this reasoning engine.
33:02
You know, some of them are open source.
33:04
So, basically for free, you get what is getting close to AGI.
33:08
So, now you have this thing that you thought would be incredibly expensive and you get it almost for free. It's a gift. You start using it.
33:16
What is the first problem you run into?
33:18
You say like, you know, uh how has my music listening changed over the last year? Nee.
33:25
There is that's not exposed as an API, because in the previous machine learning world, that data, the listening data 15 years back, it's on cold storage somewhere.
33:33
And an engineer would have had to do like an SQL, you know, job that may have taken a week to pull it up.
33:38
Then you would have trained a model, then you would have put put it back in cold storage.
33:41
If now you want to be able to reason over that in real time, you need to expose all your data as APIs in real time.
33:50
And actually So, actually, my biggest job to enable AI is not AI engineering.
33:54
It's old school engineering, exposing all this these this data that we have so that you can have a reasoning engine reason for you as a product person, or actually for me as a consumer potentially over my own data in real time.
34:09
So, I think that's what's happening.
34:09
So, the combination of now there's a standard MCPs that you can wrap these APIs in, and many of these com At least us trying to expose all of this data means that a business person, a lawyer, a product person, a designer will be able to use Cursor without having to code.
34:27
And they can actually at least prototype or talk to real services.
34:31
So, that that I think is the journey that we're on.
34:33
It started with developers.
34:36
But I think um as you as you as you uh expose the infrastructure and wrap it in in APIs, I think it's going to go outside It's going to have to go outside of Is there a way to say that and and lose a lot, but summarize it, that what we've seen happen with developers is going to happen maybe even it's I'm surprised that it's Cursor that they're using.
34:57
That's quite interesting.
34:58
But with other similar tools. Yeah.
35:00
And that just more of our work is going to be um it's going to feel like we're working with a team, speaking to a team, um using natural language to prototype things, to try things.
35:10
And that will diffuse slowly through the entire Not only the the hours of the software developer, but the hours of each of the other functional areas. Yeah, I think so.
35:21
It's hard to see where it where it's going to land, because you're you're somewhere right now, but we're pretty certain that that somewhere is on this curve.
35:27
So, you can be pretty certain that the workflows you see right now are not going to be the same.
35:31
And that's actually one of the problems.
35:32
Like, how much are we going to build for what we see right now, when you know the models are going to be more capable, they're going to be different tooling very soon.
35:42
So, you don't want to overfit too much to to the moment.
35:44
A reasonable view of a modern company is that all of its data is exposed in real time, and you have some tool on top like Cursor or something else.
35:55
Maybe different tools for different skills.
35:56
Maybe a tool more the licensing team at Spotify may have a different tool to reason over all the contracts and quickly say like, "Do we think we can do this in that market?
36:05
And what do we need to license to do this?"
36:06
But also the product team could ask that you know, licensing engine, "We have like 15 years of contracts, both current and previous."
36:14
So, so this AI has a lot of insight into what music licensing looks like more than any single person in Spotify if you if you train it that way.
36:24
So, there there will probably be slightly custom interfaces for different skills.
36:28
Uh I'm not sure which is going to win out, but I think it's going to look something like that.
36:32
Right now, what we see people doing is they're they're sharing examples of prompts they used for the workflows.
36:40
And then prototypes that they've used.
36:43
And that feels like very much a point in time.
36:45
It's kind of hacky and, you know, different things.
36:49
If you were to calibrate the world out there, so few people have the inside view that you do, where you're excited by this technology, you're trying to embrace it, you're only able to embrace it so fast in the ways that we've described.
36:59
Like on a 1 to 10 point scale or something like this, what what score would you give how much this is impacting you so far?
37:10
And and like how crazy this might get.
37:13
Like, people are very excited that this is going to literally change everything.
37:17
Um and there's some people that are actually worried about how much how powerful it might be.
37:22
From a practical real world standpoint, could you calibrate us a little bit as someone one of the few people that like actually is both excited about it and and also faces reality on a daily basis?
37:32
If you want to be as realistic as possible about it, you take the developer use case.
37:36
I've seen studies from other big companies that if you actually measure out of a developer's time, the the speed up is like 7% or something, which sounds very disappointing because of all these things.
37:46
Like, the net new coding is a small part.
37:48
Net new is a small part of that and so forth.
37:51
So, I think right now, it's a bit overhyped in terms of actual impact, at least for these big companies.
37:57
But I think it's going to turn it into the opposite.
37:59
And I think right now, people are overexcited versus the actual impact.
38:04
Um but I think the opposite is going to happen.
38:05
I think it's going to have tremendous impact uh over the longer term.
38:10
Um what I see people doing right now, it depends.
38:13
I mean, I use it a lot personally.
38:15
I see a lot of my developers and product people and designers um use it all the time for for just productivity purposes.
38:25
Uh fact, you know, putting things into an engine, you know, asking it for the summary and so forth.
38:29
Those things happen all the time.
38:30
It's hard for me to estimate how much that speeds them up already. It certainly does.
38:36
But I think the really big impact comes as you reshape these companies from this technology.
38:41
Right now, we're just tacking it on top.
38:42
But as I said, you have to reshape it and rebuild it for this work where where reasoning engine can reason in real time over the entire company's data.
38:50
But that requires actually a lot of retooling.
38:53
That's why startups are ahead.
38:53
They don't have to rebuild.
38:55
They don't have 15 years of data.
38:56
So, they're probably in the future, which is they why they feel like, "No, no, Gustav is wrong.
39:00
The impact is really big already."
39:02
And I think it is for a startup.
39:07
I think it takes a bit longer for big companies.
39:08
And big companies like us, we have to shape up and accelerate in order to not uh be behind.
39:14
I'm sure they would all like to have 700 million monthly active users to experiment with, though.
39:18
Yeah, that's that's one one of the one of the benefits.
39:21
Um on that topic, you mentioned going through mobile and the experience of not only was everything the everything was changing as a result of mobile, but actually the business model also needed to change.
39:31
We've really talked about product so far, and there's more to ask about product.
39:35
But talk about business model.
39:37
Like, what would be the world in which as a result of this technology, Spotify's whole business model need needs to change.
39:42
And how do you how do you go about evaluating something like that?
39:46
It's It's a great question.
39:48
And we I we've seen a few of those examples of business models.
39:54
And I tend to tell my product teams that everyone says that you know, the world is disrupted and changed by technology.
40:01
And I think that's true in the sense that the underlying force is technology itself.
40:06
And it's this gift that keeps on giving.
40:07
It gives you computers, internet, smartphones, ML, AI, quantum computing.
40:12
And these gifts keep coming almost on a schedule, and they actually come closer and closer.
40:16
Previously, technology companies were not called technology companies.
40:19
As as a side note, they were they were called car companies.
40:21
But it was a technology companies or or you know, pharmaceutical that that was the state of the art technology right then.
40:27
But because these microwaves came so far apart, they call themselves a car company.
40:31
They never became ubiquitous technology companies.
40:34
They kind of over fitted to that.
40:38
I think somewhere in the '90s, around Google, Amazon, etc.
40:40
, these microwaves started coming so fast that people tried to pin them down as a, you know, Amazon is a books company.
40:46
And they were like, "No, not really.
40:47
We're doing books, but here's other stuff we're selling."
40:50
And then they're like, "Okay, you're the everything store company."
40:52
It's like, "No, not really.
40:53
Now we're selling uh you know, Amazon Web Services over here."
40:56
So, I think these companies are the first set of companies to have technology as the strategy.
41:01
The previous ones took you know, one wave as the strategy.
41:05
And then, you know, IBM comes along and does, you know, computers as a strategy or first memory and so forth.
41:12
I think we're see This is the first wave of general technology companies.
41:16
Which, interestingly, might mean that they could be I mean, companies almost always die after a while.
41:21
These could be the first companies that never die because they're ubiquitous technology companies.
41:26
Whatever the technology gift is, just try to have a company that can quickly wrap around it, figure out the product and business model.
41:32
So, so I think that's interesting.
41:32
And that's how I think about Spotify.
41:34
Yes, we're music company, and then a podcast company, and then a book company, and then a video company.
41:41
But it's really about trying to anticipate technology, figure out what it can do, and then adapt the product and often the business model.
41:48
So, I said that mobile was one of these things where we needed to change the business model.
41:51
And I think what happens when these one of these technology gifts comes along is there is a big change when the technology happens.
41:58
Like, you know, uh piracy. Big havoc.
42:03
But the real change happens when someone also figures out the business model.
42:04
So, I tell my product teams like technology can do good things.
42:09
Technology and a new business model can really change the world.
42:13
But without a business model, there's seldom like large-scale change.
42:17
You can you can destroy a lot of things, but you never really create value.
42:20
So, mobile was the first where we needed to figure out the free tier on mobile without cannibalizing our paid tier.
42:26
And and what we did there was we looked at our data and saw that 50% of premium users were listening in shuffle mode.
42:33
So, we said, "What if we take shuffle as a feature, give that away for free.
42:38
It It It should be 50% of premium consumption is very valuable, but it's not going to be 100% of anyone's premium consumption, so no cannibalization."
42:46
And we managed to create a tier where you could playlist all your favorite songs in a playlist, press play, put the phone in your pocket, and listen forever for free in the background.
42:55
So, that was a business model innovation along with technology.
42:59
The the most previous one was audiobooks where, you know, there were audiobooks in the US a la carte.
43:05
Um and sure, we did some nice innovation around being able to stream that book.
43:09
But the real thing is not stream a book.
43:12
You You You've been able to stream audio for a long time.
43:14
The real innovation that was the business model, to be able to bundle audiobooks into Spotify Premium and take it from a sort of It's almost like music.
43:21
Music was also a la carte and quite niche.
43:23
And once we made it an access model, you know, with no marginal cost, it got way larger.
43:29
And that's how we think about audiobooks as well.
43:32
So, we've seen a few of those and managed to adopt them.
43:33
To your question, is AI going to do that?
43:35
Do we need to change the business model? I'm not sure.
43:40
I think there's one glaring thing that is different, which is the previous sort of VC model coming all the way back from from chips and silicon was you make a big upfront investment and then you amortize and you get to almost zero marginal cost.
43:58
That's how software worked.
43:58
It's not how AI AI works.
44:00
The marginal cost is high, and you need to cover it.
44:04
So, you could say that that should change everyone's business model.
44:05
You're going to need to somehow either monetize very effectively through ads or charge users.
44:10
And you see Open AI being a subscription product.
44:13
And I think you're going to see more of those to cover the So, the marginal cost is a net new thing.
44:18
For Spotify, it's interesting because we're like the one technology company that always had a marginal cost.
44:23
One more stream was a marginal cost to labels.
44:26
So, we grew up in a world that where if we were too successful on the free tier, we could go bankrupt overnight.
44:32
Which was never true for Twitter or Facebook, which is why VCs said like, "Just go crazy.
44:37
Worry about, you know, monetization later."
44:39
Spotify could never do that because we could go bankrupt overnight.
44:43
So, we always had to worry about monetization and the balance between free tier and paid tier conversion and free tier monetization.
44:48
So, the good thing for us is we're fairly used to marginal cost in our business model.
44:54
Uh so, I don't think I think you're going to see those things.
44:58
It's very likely that some consumers are going to want tons and tons of inference.
45:01
And because that's a marginal cost, you're probably going to have to pay somehow for that as a consumer.
45:04
So, I think you're going to see more tiering of consumer products based on how much inference you want.
45:11
But for us, that's not that new.
45:11
We've had several tiers already.
45:15
I'm curious cuz um I'm an investor in a company called Etch that's going to be one of these companies that pushes down that inference cost.
45:20
And And like the history of compute, you're going to see this incredible, you know, consumer surplus that And And consumer benefit that comes from cheaper and cheaper unit by unit inference cost.
45:29
But the countervailing force is that we would just use more of it, you know, more reasoning tokens, more whatever.
45:34
So, it makes me wonder how much more you can imagine better models being useful to you.
45:41
Like, it seems like if we just froze reasoning and and model capabilities today, we probably still have decade plus of digestion to do of how we could use these models to make better products, better features, whatever.
45:54
Can you imagine like a 10 times better, you know, another couple orders of magnitude better models opening up lots of features that you can't currently do?
46:03
Like, is that a Is Is that a thing?
46:03
Or Or do you think we kind of have what we need?
46:06
And therefore, inference we could expect to be really cheap.
46:08
So, I both subscribe to the product overhang idea that there's a huge product overhang if we froze.
46:14
I think we would see product shipped that look amazing for several years before we exhausted uh what we have. So, I subscribe to that.
46:22
But I also subscribe to that there is no limit for for compute.
46:27
You eventually you get to computronium.
46:31
But if you look at this the physics of computronium What's computronium?
46:32
It's the smallest It's the most computation a universe could do, you know, theoretically. Yeah.
46:39
Um we're very far from that limit.
46:39
Uh so, I think we're going to go all the way there before we stop.
46:47
And I think we're going to be very inventive of It's uh there is a nice analogy that I think uh I don't know who came up with it, but I think Ben Evans talks about it quite often.
46:59
You know, when the spreadsheet came along, the idea was the same, you know, now all the accountants are going to sort of go out of business.
47:08
What happened was we could just not imagine if if computa- if if calculation or or basically, you know, yeah, if calculation cost of calculation went to zero, what's going to happen is you you could imagine that the value of doing that is going to go to zero because there were so many accountants in the world.
47:24
So, what happened was we just started doing massively more accounting.
47:28
When when there's no cost to spreadsheeting, you're going to start uh you're going to start do models to predict the futures of this this, you know, uh asset or good or something, you know, into the future forever.
47:39
We just came up with so much more spreadsheeting that you could do.
47:43
And it's bigger than ever.
47:46
And I think we're going to see exact I I think from a financial point of view, um when the cost of something drops, the demand usually increases more than the cost than than the drop.
47:56
And I think that's bound to happen with intelligence.
47:59
It is like the ultimate thing.
48:02
And to say like, "No, I have I have enough intelligence."
48:04
It's not interesting to I think we're going to be ashamed of how mundane things we spend inference on.
48:11
It's like, you know, "Could my coffee be like 1° warmer tomorrow?"
48:15
If it's truly no cost asking the questions, I think people will.
48:19
Maybe now's the time to ask you about sitting in the back garden with David Deutsch and talking to him about this concept of The Beginning of Infinity.
48:27
Computronium made me think of your interest in this topic that like and and your answer there that No, no, there's no there's no end point here.
48:34
Like, we're going to It's just going to keep going.
48:35
We're going to keep learning, keep deploying our new technology.
48:38
Can you talk about him, that book, why it influenced you, your conversation with him?
48:42
Yeah, so David Deutsch has been a hero of mine since since I read The Beginning of Infinity.
48:49
And then he wrote another book called The Fabric of Reality.
48:50
He's considered the father of quantum computing.
48:54
And um obviously, quantum computing is is one of these gifts that uh technology is going to give us, and it's about to get very real, I think, very soon.
49:04
Um so, I've always been interested because quantum computing is the or or quantum mechanics is the most insane thing on this planet.
49:12
You know, we we live in what we consider this reality, but if you go to to the bottom layer, this is not reality.
49:18
It's it's just some sort of three-dimensional projection that we live in.
49:24
So, that that The Fabric of Reality had a big impact on me.
49:26
And he he's a he's a believer in is an Everettian.
49:29
He believes in in in uh multiple worlds worlds scenario.
49:34
Uh so, yeah, know, that book, and it blew my mind.
49:37
Then, Beginning of Infinity is um is maybe his most famous book, whereas Fabric of Reality is really about quantum computing um and how quantum quantum computer works.
49:48
Beginning of Infinity is is very philosophical and he has he he has a bunch of big ideas there.
49:54
He's a very positive person.
49:56
And I've now at 70-plus, uh I finally got to interview him in his garden in Oxford.
50:01
He's uh he's not of great health, so had to be outdoors, you know, distanced.
50:06
And uh everyone is very negative on the future, you know, there's so many problems that could go wrong and, you know, all these could go wrong, climate.
50:12
He's actually very positive about the future.
50:15
He's he's clear that there are risks.
50:17
But, you know, he sees us going out there into the stars and like I asked him like, where do you think we are in a in a million years?
50:23
And he's like, well, maybe we're, you know, this far outside of the solar system, but not quite there.
50:27
He's like very certain we're going to get there, so.
50:30
He's a very positive person and when I asked him about his life, he's very content with his life. He's very very happy.
50:36
So, he's just an inspiring person still at this age, you know, I wish I will be like him in that age.
50:44
But, this book uh has a few concepts that I've tried to apply at Spotify and one of them is he talks about the power of explanations.
50:54
And he thinks the human mind is infinitely scalable.
50:56
He does not think there's a limit to what we can understand because of explanations.
51:00
And I think this is something that a lot of people uh I agree with that, but a lot of people disagree.
51:05
Certainly, there are things we could never understand.
51:06
His view is no, there is no limit to what we can understand.
51:11
We are the only species who who broke that barrier because we have explanations.
51:16
Other species have pattern recognition.
51:18
They they can do things and learn the pattern that this works.
51:20
There's some uh cultural transfer maybe of looking at someone else doing that pattern.
51:24
A bird can see another bird.
51:27
Some, you know, species can teach their kids, but they never produce explanations.
51:32
And um the uh he has a definition of a good explanation.
51:37
He's very inspired by by Karl Popper as a philosopher. It's his house god.
51:40
Um so, he takes a bit from Popper and he takes a bit from science.
51:45
Uh so, he says obviously that a good explanation has to be falsifiable.
51:49
But, he says a few other things that I think are obvious in retrospect, but not before.
51:53
He says that a good explanation has to scale. Uh has to have reach.
51:57
What does he mean with that?
51:59
He says that some explanations uh explain something quite locally, you know, you can have an explanation about um the for example, the the sun revolving around the earth, which explains a bunch of stuff, but it doesn't scale, right? To other planets.
52:14
It's a better explanation is to have the earth revolving around the sun.
52:18
It just scales better to different scales.
52:20
So, a good explanation has to scale up and down.
52:24
A good explanation has to be compatible with all the previous explanations, but most interestingly, he says that a good explanation has to be hard to vary.
52:33
This I find very obvious, but also very non-obvious to people.
52:36
So, what does he mean with a good explanation has to be hard to vary?
52:40
He means that for example, if your explanation for the weather on the planet is that now Thor is angry, so there's thunder there.
52:51
You know, it's an explanation, but it's too easy to vary.
52:53
You can say like, well, now someone else is angry. They also had a hammer.
52:57
It's too easy to vary and get the same result.
52:58
A good explanation, if you if you move one of the parameters, the entire thing is not predictive anymore.
53:04
Then you're probably close to the truth.
53:06
And I think this is so interesting because the problem with most conspiracy theories that people love is they're so easy to vary.
53:12
You can just exchange that character for other crazy person did something crazy and still going to produce the same thing.
53:17
So, if it's too easy to to change people in a conspiracy theory, it's probably not true.
53:24
So, I think that's something very powerful.
53:25
Like good explanations need to be very hard to vary.
53:29
So, this is something I've tried to to to instill in my org.
53:32
And I think there's an interesting meta point here, which is people ask me as a product person, you know, how much of product development is magic and how much is science.
53:43
And I try to be provocative in saying I think it's exactly 100% science and 0% magic.
53:46
And people get provoked because it kind of implies that there's no skill.
53:54
Um so, what I mean with that, I say it to provoke.
53:58
What I mean is that um certainly, people are going to have pattern recognition in this neural network.
54:04
They've seen a lot of examples.
54:06
That's what we call seniority.
54:06
And people have seen a lot of things.
54:09
They're going to get instinctively to the right conclusion faster than others. So, that is valuable.
54:15
And and I want lots of seniority.
54:17
So, I don't I don't discard seniority.
54:20
Uh and it brings a lot of value.
54:21
You can save a lot of time and a lot of uh mistakes.
54:24
But, the reason you call it magic is because that person can't explain it. It isn't actually magic. It's just science.
54:30
It's just you are not smart enough to explain yourself.
54:34
If you could think even further and explain it and come up with an explanation for what you see, the way David Deutsch does, it's so much more valuable for the company.
54:41
If you have a theory, instead of saying like, no, patching my intuition is this, you're not smart enough to understand it, so I'm not going to tell you. Just do what I say. Maybe I'm right.
54:51
Maybe I'm wrong, but it's not very helpful for you, you know, when I leave the company, you're going to take over, you're like, I have no idea why they did that.
54:57
You have to develop your own intuition and your own pattern recognition.
54:59
But, if I can come up with an explanation, which is, you know, I think the psychological behavior of people, you know, it's like um Kahneman's, you know, loss adversity uh or prospect theory.
55:10
I think people value losing something one and a half times the the value of getting it, so therefore, we should not just launch feature and test it because it's 1.
55:18
5x hard more expensive to remove it.
55:22
Then you have a theory and it can spread across the company in like a week and now everyone has that.
55:27
So, I really want to force people in my company to try to even if we see something working in a AB test, I try to tell them I don't want to launch it until you at least have a theory of why it works.
55:36
Even if it's super clear, there's a lot of pressure to launch it cuz there's like engagement value and monetization.
55:43
But, if you I want you to at least have a theory.
55:46
Cuz then over time, the company builds up a theory a consumer theory.
55:48
And if you have a strong consumer theory, then you can predict things that were very unlikely.
55:53
The what what David Deutsch also says is that pattern recognition will iteratively get you more on the same path, but it's never going to jump all the way from the geocentric to sort of the heliocentric model.
56:07
Only an explanation can take you to quantum physics. Entirely unintuitive.
56:12
No pattern recognition gets you to like maybe it's a wave and a particle at the same What what's an example internally of a great explanation that then led to some that the geo to heliocentric type of jump?
56:24
Like how did how what's an example of how that actually played out?
56:25
I think a good example that is sort of public is uh the free tier that I told you about.
56:30
When we when we faced the prospect of we only had a paid mobile tier.
56:36
You actually paid to get mobility on Spotify.
56:40
Now, smartphones are scaling.
56:40
Users don't have a computer. We need a free tier.
56:44
The competition there was YouTube.
56:44
They were foreground, on demand with video.
56:50
And so, the you know, the the pattern recognition, the obvious thing would have been to say like, let's do that. It's it's proven.
56:58
Um but um what we did instead and the specific specifically attributed a person named Charlie Hellman, was to reason around it from sort of first principles and say, okay, let's look at our usage of Spotify.
57:09
How much of it if we limited our license to the same thing?
57:12
It only works in the foreground.
57:13
As soon as you lock the screen, the music stops.
57:15
How much of the listening is in the foreground?
57:18
Turns out back then, it was like 9% or something.
57:19
So, you have 91% of the use case being in the background.
57:23
We probably want to get something else.
57:25
Uh the user need there is probably background listening.
57:27
And you then you look at the App Store.
57:30
Is there a way to listen to music for free in the background in the App Store?
57:34
The closest thing was Pandora. But, that was radio.
57:36
You could not listen to your favorite songs.
57:40
So, then we said we would like a consumer product where you can listen to your favorite songs with your phone in the pocket forever for free.
57:49
So, the problem with that is that's that's almost a premium use case.
57:51
If we just launch that, it's going to cannibalize our premium tier. So, what do we do?
57:58
Then we looked at the premium usage and we saw that premium users about 50% of the time, they were shuffling their playlist.
58:04
They were using on demand features, you know, searching and clicking and playing specific songs 50% of the time, but they were shuffling playlist 50%.
58:12
So, then we thought, what if we take this that seems to be something that even when you have have on demand, you voluntarily shuffle. It's a big use case.
58:19
We give that away for free.
58:21
That should mean that none of the premium users convert back to free cuz they still want their 50% on demand, but you're giving a lot of value away for free.
58:29
So, we tried to model a consumer need, reason around it, came up with this shuffle background tier.
58:33
It was very very very unintuitive.
58:35
Even the people inside the company said like, that's a terrible idea. It's a terrible idea.
58:41
But, we kind of trusted the data and I was even skeptical of it myself.
58:43
I was like, look look like look at this on demand.
58:47
Shouldn't we try like time caps or a lot of people just want us to try long free trials.
58:52
But, the problem with the free trial is even if you know, Nokia, I think, Nokia comes with music, they tried a year-long free trial.
59:01
But, even then, the user knew that if I start investing in playlist now, a year from now, my playlist investment is going to disappear.
59:06
So, they never started investing.
59:09
So, we went with this shuffle tier and and this is what made growth explode.
59:14
And to this day, that's our differentiation against the other services.
59:16
It's like it's the only way to listen to music for free forever with your phone in your Fascinating.
59:20
So, that's an example that's an example of like theorizing and explaining rather than pattern recognition.
59:26
I'd love to talk about the evolution of the relationship with the music industry.
59:30
It's a company that unquestionably has wholesale changed music, which is so interesting and so cool.
59:36
You know, thinking back to the You've been here a long time.
59:39
Thinking back to the early days, um it's amazing the impact that it's had.
59:41
And from a investor's perspective, one of the things that many were always keyed in on is just the gross margin of the business.
59:49
Just like how much transfer pricing problem are you always going to have that no matter how big you get, you know, you the the music industry that owns the IP is just going to always take their same cut cut of the meat.
59:59
Talk about how you've thought about that change over time.
1:00:04
It seems like it's been uh both a good relationship for them, but also a very patient path for Spotify.
1:00:10
Maybe just give us like the the insight into how it's worked and how you think about it. Yeah, for sure.
1:00:15
Um I mean, I I grew up as Spotify grew up in the era of piracy in Sweden, which was the worst market.
1:00:22
And there's this famous quote from a UK uh label exec to a Swedish label exec uh you know, around early 2000, where the Swedish label exec showed a P&L of one of these Swedish companies and said, "That's not a business. That's a hobby."
1:00:39
That's how broken it was.
1:00:39
And that's actually why Spotify could happen because the music industry was prepared to take risk in Sweden.
1:00:46
Uh and and I want to give a lot of credit to the music industry.
1:00:47
They took a lot of risk with Spotify.
1:00:49
Spotify took an enormous amount of risk enormous amount of capital risk. We MG'd a lot.
1:00:55
We ate a lot of the risk, but certainly they took a lot of risk.
1:00:59
So so you know, I think the music industry certainly deserves the uh the success. Um as does uh Spotify.
1:01:03
I've called my team since you know, I joined in 2000 2008 somewhere around 2012 or something, I started saying that, you know, my team the the the R&D team and and all of Spotify, we are the R&D department of the music industry.
1:01:18
And first people were like, "What do you mean?"
1:01:21
And I'm like, "Well, look at it.
1:01:22
It's an entire industry that doesn't have an R&D department."
1:01:24
Like, you know, mobile phones has an R&D department.
1:01:27
It's called, you know, Apple or Google.
1:01:29
Everyone else has a lot of R&D.
1:01:30
But there's no R&D spend in the music industry.
1:01:34
Um and I think that's turned out to be true.
1:01:36
And if you look at the trajectory, this year is sort of the first year of of of profitability for Spotify since its founding.
1:01:44
People say that, you know, there's there's a lot of talk about is Spotify sharing enough of the revenue.
1:01:49
You know, we share about 70%.
1:01:52
But the truth is the other 30% we haven't kept.
1:01:53
We've invested all of that in the music industry and then more.
1:01:57
So we were unprofitable um for 15 years.
1:01:59
We just invested invested invested. So a ton of patience.
1:02:04
And And at the same time actually the music industry has been profitable.
1:02:07
Spotify has been unprofitable.
1:02:07
So we've taken I think it's fair to say we are literally the R&D department of the music industry.
1:02:13
We invested and we're we're, you know, had had losses for 15 years and the music industry has been gaining um uh profit.
1:02:21
Now, that is not sustainable forever.
1:02:24
We need to get profitable.
1:02:24
We needed to get profitable.
1:02:26
So we can't be the music the the R&D department of the music industry unless we can, you know, have the best machine learning engineers, the best product people, developers, etc.
1:02:34
For that, you need to be profitable.
1:02:36
It turns out these people are expensive because they're sought after.
1:02:40
So we just, you know, we are very very patient and long-term company and we invested for a long time. But it was just time.
1:02:46
About 2 years ago we decided now it's time for us to become profitable to sort of take control of our own fate.
1:02:51
Um you know, in terms of be you know, being able to invest in ourselves.
1:02:55
So yes, we're profitable, but we're we're actually investing um almost all of that back into more people, more product, more AI.
1:03:03
So we're still Now we just have our own investment vehicle instead of having to ask private investors initially or the street for more money.
1:03:14
So that's how I think about it.
1:03:14
Really as the as the R&D department of the music industry.
1:03:18
And I think we've done a good job.
1:03:19
Uh you know, this year we paid out over 10 billion.
1:03:24
And that's, you know, up from 1 billion I think in you know, uh almost 10 years ago.
1:03:29
It's like just steadily increased.
1:03:32
The music industry is bigger than it ever was.
1:03:34
People still talk about the heyday of the CD ROMs. the the CD era. Yeah.
1:03:38
The truth is the music business is bigger than it was back then.
1:03:42
So this is the best it's it's ever been. It is better than ever. More money than ever.
1:03:46
The pie is both bigger and higher, but it's also getting sliced up.
1:03:50
But that's because more people take a shot.
1:03:53
And And it feels very wrong for us to say like, "No, the creators up until 2020, they were good, but no one should be able to try after 2020."
1:03:59
I you know, new creators should be able to try to do music. So that's the dynamic.
1:04:05
And And I think a way to think about this is people talk about the per stream payouts and so forth a lot.
1:04:13
And, you know, Spotify should should share more per stream.
1:04:15
There are two things that are happening when other companies say that they share more per stream.
1:04:19
The industry doesn't pay per stream. They pay per subscriber.
1:04:23
But we have twice more than twice the engagement of our competitive services.
1:04:27
So if you take the same $10 and you listen twice as much as Spotify, the per stream is half.
1:04:32
So these are the companies that higher per stream because they have a worse product.
1:04:36
We've learned from the labels that we have twice the engagement and half the churn of competing services.
1:04:42
So uh that's sort of a curse where the per stream model just the better we are as a product, the lower the per stream is going to look.
1:04:50
Um but we're we're looking at the aggregate number and we're leading everyone else there.
1:04:52
We're the vast majority of these of these payouts.
1:04:56
So I think if you look overall, the model is working.
1:04:57
We took a lot of investments and now the industry is getting a huge return.
1:05:01
And Spotify also is profitable now.
1:05:05
And uh the way to grow this pie is to you know, now we are closing in on 300 million paid subscribers, closing in on 700 million MAUs.
1:05:16
There's a about 500 million paid subscribers I think in the world.
1:05:19
And so we're we're almost 300 of those.
1:05:22
But that's like half a billion out of the world's population.
1:05:25
If you look at markets like Sweden, you know, on average you can just look at the public numbers convert about 40%.
1:05:31
But if you look at the mature markets, I won't give you the exact number, but it is much higher.
1:05:34
And if you look at the emerging, it's lower. So the average is 40.
1:05:39
But that's not that's not the average across the world.
1:05:40
That's a blend of low and high converting.
1:05:42
And so far throughout our history, everything starts to look more and more like Sweden the more time passes.
1:05:46
So the solution to this is just to scale it faster.
1:05:51
Better free tier that gets more people on that converts to premium.
1:05:56
We think there should be billions of people paying for music.
1:05:59
And that's how you make the pie truly bigger.
1:06:01
The The rev share is actually a red herring.
1:06:02
So let's say that we share, you know, 70% today, ish.
1:06:07
Or let's say 2/3 to make it easier.
1:06:10
Um even if we were a charity and we paid out 100%, that would only be, you know, 1. 5x what you get today.
1:06:15
So if you think like, you know, X pennies per stream is too little, even if we were a charity, it would be 1. 5.
1:06:21
The solution is not the rev share.
1:06:24
We're giving away the vast majority.
1:06:25
The solution is to quickly scale the amount of people paying for music.
1:06:29
And if you just look at the numbers, you just have to keep going and it's going to get to billions of users paying.
1:06:35
And then several billions.
1:06:35
Then the music industry is absolutely massive.
1:06:38
I just think the music industry is like undervalued.
1:06:43
Terminally it's going to be much bigger than it looks.
1:06:45
I'm curious how you're thinking about the podcasting world.
1:06:49
This is something that um we're sitting here doing right now.
1:06:51
Uh I've been doing for a long time.
1:06:53
Now it seems we've entered this interesting new era where um when I started doing this, I remember it it was quite I would call it like low status.
1:07:01
Like, when I told people about it in 2016, they either didn't know what it was or thought it was kind of silly.
1:07:08
And now, especially in the US with what happened around the election and the importance of podcasts in the election, it seems as though it it it's hit some tipping point where basically anybody that might make sense to have a podcast now has one or is launching one.
1:07:21
And it it it's it's the corporate marketing strategy is to get a podcast.
1:07:25
And it's the it's the communication strategy is to go on them.
1:07:29
Um so it's it's really exploded in importance and visibility.
1:07:35
What role has and will Spotify play in all this?
1:07:38
And just like what do you think about podcasting and and and its importance?
1:07:41
So the the reason we went into podcasting, one thing that I'm very precious about when it comes to Spotify and um so is Daniel and and uh the other co-person, Alex, who's my closest partner, is that uh there are many ways we could go um as a company.
1:07:59
And I think your business model to some extent steers you.
1:08:01
If you're an advertising business model mostly, you're going to be steered towards any additional engagement.
1:08:06
Fortunately for us, we're a mostly a subscription-based business model.
1:08:12
Um so we focus more on retention.
1:08:12
And you're going to vote with your wallet every month if you want to keep paying for us.
1:08:17
So we don't we're not as steered towards again engagement at any cost.
1:08:20
So having been at Spotify, you know, for a long time when this happened, one of the things that made me feel very good about Spotify was that when people used it when they lost an hour on Spotify, they felt very good about it.
1:08:31
If you lost an hour on music, you come out feeling that was a good hour.
1:08:36
One of the reasons I really pushed I pushed quite hard for podcast in the company was that I was using it myself and a lot of our developers were using it.
1:08:43
And I saw it being hacked into the product at hack week every year.
1:08:48
It's like people wanted them there.
1:08:48
And we just said like our developers is like a small sample of the world.
1:08:53
What if they're What if they're a good sample of the world? So So that was one push.
1:08:56
We saw people using it internally and hacking it.
1:08:58
But what made us decide on it was that it was this format.
1:09:03
Everything in the world was getting more and more short form. People were bite-sized.
1:09:08
And, you know, attention spans were going down.
1:09:10
And there was this this counterforce, which was long-form discussions, deep.
1:09:13
People spoke in full sentences, you know, um about you know, quantum physics or whatever.
1:09:22
And that just felt like something very important and good for the world. And so we looked at it.
1:09:25
We saw that it seemed to be growing from a small base.
1:09:30
Um we saw the biggest competitors sort of being asleep at the wheel.
1:09:34
And we did basically the Peter Thiel idea of, you know, it's better to go after a small markets early and better on organic growth than to take try to take a small share of a mature market.
1:09:43
It looks more more like less risk in the mature market to get like 1% but the thing people miss is the cost my acquisition cost in a mature market is just massive. Right?
1:09:55
Where is the cost my acquisition cost in a new market is usually small.
1:09:58
So we decided to go for it because we thought it was something that was in line with music.
1:10:01
Like if you lose an hour in a deep podcast, you come out feeling like you learned something.
1:10:08
And this is the reason we also went into books.
1:10:11
Um because it's in line with that.
1:10:11
We want to be this nutritious service.
1:10:15
Uh there are two litmus tests for this.
1:10:16
One is if you lose an hour on Spotify, how do you come out feeling versus if you if you lose an hour doom scrolling in the in the bathroom, how do you feel about that?
1:10:24
In one case you feel like you ate a lot of candy.
1:10:26
Like you had a lot of energy in you but it's bad calories.
1:10:29
In the case of Spotify, you feel like you learned something.
1:10:31
The other litmus test that we have is see a lot of parents restricting screen time for their kids and saying go to go to Spotify instead.
1:10:41
Which means that expresses how they feel about it and how we feel about it.
1:10:44
So that was one of the reasons to go into podcast.
1:10:48
It was it was partially philosophical.
1:10:51
But we also saw the market opportunity of a small market that was poised to grow.
1:10:55
And we saw need in early adopters, you know, trying to hack it in and then we did this bet on leveraging our own distribution combining it with music saying that the market is this big right now but what if we could expose podcasts to people who listen to music?
1:11:08
Could we grow the market?
1:11:11
So that's the bet we did and the truth is audiobook is something similar.
1:11:14
In the US audiobooks was a very niche behavior.
1:11:19
I don't know, 10 11 million or something people who who paid for them a la carte and the idea was that's a limitation because of the business model.
1:11:28
When you pay a la carte you're not going to explore new books at, you know, $15 per book cost.
1:11:32
Just as in music, when you pay $0.
1:11:35
99 per song, you're not going to soundtrack your sleep.
1:11:38
It's too expensive at $0. 99 per 3 minutes.
1:11:42
But what if we had like a no no marginal cost model uh where you can just explore.
1:11:46
Is is audiobook much bigger than it looks?
1:11:48
Is the business model that is wrong?
1:11:53
So again, it was seeing a market that looked pretty small but you can see in the Nordics where you have the access model that it's getting very mainstream.
1:12:02
So I bet on the market but it was the same philosophical discussion.
1:12:03
Like is this are these good calories? Is this nutritious?
1:12:07
Is this in line with Spotify's mission of being like the place where you go when you want to feel good about yourself instead of when you want to feel bad about yourself.
1:12:14
I remember when the very first time I ever talked to Daniel walking along the Westside Highway here many years ago, he talked about this notion of Spotify needing to be better than free.
1:12:21
And it was kind of a cool idea.
1:12:26
If you think about podcasting, it's very different than music.
1:12:27
That you don't when someone listens to this show on Spotify, you don't owe me anything.
1:12:31
How do you think about the way that podcasting and then obviously books is a little bit maybe more like music and I'd like to hear how you think about it.
1:12:38
How do you think about if that's the supply of the stuff that people are listening to on Spotify or watching on Spotify, the ways in which that affects your business model and the and the bundle?
1:12:50
We didn't know before we started if podcasting and later audiobooks would be cannibalistic to the other media types or not.
1:12:58
But it turns out it's it's not.
1:12:58
So the easiest model to think about it is Spotify is a bundle now.
1:13:02
You you pay some price or or you have the advertising based here and you get a bunch of value.
1:13:09
And our job is to try to increase the value you get so you value it more.
1:13:11
And then over time maybe we can capture some of that value by price raising.
1:13:15
We price raised a few times which is part of why we're profitable now but that's because we had such user surplus in value.
1:13:23
That's because we kept just stacking value.
1:13:26
And value are two things.
1:13:26
Value are features like personalization and you know, just a really good product.
1:13:30
But the other value is different types of media.
1:13:35
So what we see is that have a user that uses music that has a certain amount of consumption when you add podcast, it's just more. It's not more.
1:13:43
It's not a fixed it's it's an infinite game looks like. At least for now.
1:13:47
We we haven't run out of time in the background yet.
1:13:48
Then when you add audiobooks, it's just more retention more time spent and and more willingness to pay.
1:13:56
So that's how we think about it as a business model.
1:13:57
Then on the back end, they have very different business model very different business model.
1:14:01
I think we may be one of the most complex companies in the world on the back end because we're you know, music is a sort of a pool based royalty model.
1:14:13
Um podcast as you know is advertising based largely.
1:14:18
Um but now we also have this Spotify partner program where you don't have Spotify ads in the premium tier uh if you're paying so you get more uninterrupted.
1:14:28
So that's another business model which is part of the premium bundle and then you have audiobooks which you know, the publishing industry works in a third way very different.
1:14:36
Uh where we also have certain amount of time included in the premium tier and then a top up if you run over that.
1:14:40
One of the really complicated things about Spotify I don't think is appreciated is on the front end it's one app, one consumer.
1:14:48
You just go between them.
1:14:51
But there are very different implications of where you click in that UI in terms of triggering different business models and so forth.
1:14:57
So to model a company financially is actually quite hard.
1:14:58
We have to predict your user behavior.
1:15:00
Where you click matters and you know we have the personalization uh that has different impacts in terms of of cost and so forth.
1:15:09
So we've had to build a system.
1:15:09
We call it the Spotify machine and that that's why I said I have one experience organization and the job of this experience organization is to make sure that all of this complexity, all of these teams who theoretically could be set up to compete with each other to fix their P&L that never ships to the user.
1:15:25
There's one person who is the responsible person for the consumer experience.
1:15:32
And that person's job is to make sure that as you go between mobile and desktop and car and speakers, the thing makes sense.
1:15:38
It's like the gatekeeper against, you know, the org holding them back from the user behind them protecting the user.
1:15:45
But it's also the same in in personalization.
1:15:46
I have a personalization organization because you have the same incentives of of, you know, programming music versus podcast versus books.
1:15:53
You know, everyone wants to take market share and so forth.
1:15:58
So it's the same problem.
1:15:58
We have to optimize for the user and sort of protect the user from the internal incentives of teams and business models.
1:16:07
So that's that makes Spotify a pretty unique company.
1:16:10
We're like one thing on the front end and we're many different things on the back end with different different products.
1:16:15
If you think about, let's say, 5 years from now and you dream as big as you can possibly dream for where Spotify might go from where it is today to where it will be in 5 years. Paint us that picture.
1:16:28
Certainly I hope we've cracked, you know, the the billion user um line but you know, as a subscription, I hope we're becoming one of the biggest media subscriptions in the world and we add more and more value to that.
1:16:41
So so hopefully music is bigger than it ever was.
1:16:45
I'm hoping that audiobooks is a mainstream phenomenon as it is in Scandinavia where, you know, it's almost as many people that listen to music listen to audiobooks.
1:16:54
Uh I think that would be a net good for the world.
1:16:57
Um but I also hope we've added a few more of these uh uh verticals.
1:17:03
I can't say what they what they are but they're, you know, the subscription model, the bundling model that we didn't talk so much about.
1:17:10
The way to differentiate we can differentiate on product or on content but largely we tend to license sort of commodity content.
1:17:15
We don't work with exclusivities at least on anymore.
1:17:16
We tried them in podcast for a while.
1:17:20
So you can you can differentiate on the product and consumption of the commodity content.
1:17:24
But you can also differ differentiate it on the offering.
1:17:28
So for example if you look at Spotify now versus other offerings uh some other offerings have the same music.
1:17:35
Some other offerings have some of the same podcast but you you can not really find the combination of music, podcast and audiobooks. That's a unique thing.
1:17:43
So to use bundling theory to to create more and more of a differentiated unique thing that is Spotify I think it's very very exciting and I think you will see more innovation on the bundling business model in in in addition to to to the I mean I'm the product guy but I'm very interested in business models.
1:18:03
I've been a a CEO myself.
1:18:03
So I think you will see a lot of innovation there.
1:18:07
What's the key to a good bundle?
1:18:07
And and I'm also curious, you know, you said you experimented with exclusive content that was only available on platform and less of that now.
1:18:15
You know, what drives a decision like that and and how do you think about other people that might want to create a bundle somewhere else?
1:18:21
When we looked at podcast you know, you you look at something like Netflix and it's this beautiful business model and and insanely good execution as well on top of that.
1:18:35
And it looked to us like that could be interesting. You know what?
1:18:39
I think when you're a product company that works with commodity content, you always had this envy of like what if we could differentiate through content?
1:18:44
You know, then life is going to be super easy.
1:18:46
You always think the other thing that someone else is doing is easy and your thing is hard and it's usually like very hard to do the other thing.
1:18:53
So we tried exclusivity in podcast as a way to differentiate the service but I think it was ultimately a bad bet because the macro trend for the whole thing with podcast was that the production cost was so low.
1:19:09
Joe Rogan was initially sitting in his trailer.
1:19:11
Like the production cost was low.
1:19:14
And then go in and do exclusivities on top of that is kind of counter purpose in a way.
1:19:17
The whole point is more like YouTube in that this is very cheap content so you can get a lot of it.
1:19:24
You don't have to be right.
1:19:26
As soon as you go into exclusivity game, you have to you got to be right.
1:19:28
You got to be a content picker.
1:19:30
And that's a very hard skill that Netflix does extremely well. Right?
1:19:36
But we had this opportunity.
1:19:36
We didn't have to pick content.
1:19:37
We just get all of it and use machine learning to serve you what you wanted and me what me what I wanted and there wasn't this capital intensive need there that there is in in producing like costume dramas.
1:19:45
This is a bad strategic decision that we did.
1:19:52
We also betted a lot on on celebrities.
1:19:55
And they are celebrities but they're not always good podcast hosts.
1:19:59
And the podcast hosts that were really good, they grew up through this organic system.
1:20:03
So, you know, there there are two ways to be to always be right.
1:20:08
One is to to always guess right.
1:20:08
The other is to just change your mind whenever you're wrong.
1:20:13
So, we decided to change our mind and say this is the this looks like the age of syndication.
1:20:18
Creators actually want to be everywhere.
1:20:20
They create a video or they create music.
1:20:21
They want to be everywhere.
1:20:22
Okay, let's embrace that.
1:20:24
We're actually a platform.
1:20:24
In music we were always a platform.
1:20:26
We never played with exclusivity.
1:20:28
We said we want the maximum catalog.
1:20:30
Books we're doing maximum catalog.
1:20:33
Let's just embrace that in in podcast as well. So, we pivoted strategy.
1:20:39
And that saved us a lot of cost.
1:20:43
Um which is part of what we're doing well and it's also improved our our the catalog greatly and now we're on a really good trajectory with our podcast viewing.
1:20:50
So, it was an example of of a bad strategy and I think the the important thing is to admit it and change your mind.
1:20:56
The real cost is when you when you try to defend your past decisions.
1:21:03
What things do you do outside of Spotify in your life that most prepare you or make you capable to do the best job that you can in Spotify?
1:21:11
I think the world is moving very fast so a lot of my time is just trying to keep up with what is happening.
1:21:19
So, I I spent I spent a good deal I I was in a on a vacation in in Lisbon with my family recently and I spent a lot of time with them seeing Lisbon which is a beautiful city.
1:21:31
Then I tried to take I asked them for like one day off off from work and off from the family to just indulge myself.
1:21:37
This time it was, you know, going back to trying to code a bit use all these new tools, stay on top of what's happening.
1:21:44
Sometimes it's reading, you know, like physics or math or something.
1:21:49
It's a combination of keeping up with what is happening which is hard because it moves so fast but also stimulate myself mentally.
1:21:58
I have to have something that I'm excited about at any point in time.
1:22:03
And it can be new things like AI and what it would mean.
1:22:07
But it can be age-old things that I just didn't know like learning more about physics or math or something.
1:22:13
So, for a while yeah, I've read a lot of philosophy for a while because it's just an interesting area.
1:22:19
Um You know, you think through all the big questions of of intelligence and consciousness and all of those things and you can spend like 10 years there just reading all of that.
1:22:26
And now I feel like tapped out a little bit.
1:22:28
I I think I like when you start reading you're like, "Yeah, I'm going to crack this."
1:22:32
And then turns out I didn't crack it.
1:22:36
People have been trying to crack, you know, consciousness for a while but it's so deeply interesting.
1:22:39
It kept me like excited about life for for for um for a very long time.
1:22:44
I was never like a big math person in school.
1:22:46
I was okay but not great but I found myself getting very excited about math the older I got.
1:22:54
So, you as you start reading a bit of philosophy you get into like things like Gödel's incompleteness theorem and you know, constructive constructive mathematics and these things that, you know, they're loosely related to work but they keep you energized. keep me energized.
1:23:09
And I actually talked to it turns out that a lot of my product people and engineers are deeply interested in these things.
1:23:16
So, I just have something very interesting to talk to people around me about.
1:23:19
So, that's how I keep and then I do sports.
1:23:21
I do Brazilian Jiu-Jitsu with my kids which is very rewarding. What does that tell you? Humbleness. Humbleness.
1:23:30
You come in and you think you do something and you get absolutely smashed by someone half your size and they're not even sweating.
1:23:37
And you're like, "Okay, technique matters." It's technique. It's leverage.
1:23:41
Yeah, the beautiful thing about Brazilian Jiu-Jitsu is that the belt thing is the belt thing is real.
1:23:47
Um and it is there's There's a long story behind it but the net is that um Japanese person brought brought Jiu-Jitsu to a Brazilian family and there was there were a bunch of brothers there who fought a lot.
1:24:01
There was one brother that was just underdeveloped versus the others.
1:24:05
He was just not very strong.
1:24:05
So, he could not beat his brothers.
1:24:07
So, he started taking Japanese Jiu-Jitsu and figuring out how he could just use physics, just leverage.
1:24:15
And slowly slowly he started beating all his brothers and that became Brazilian Jiu-Jitsu.
1:24:18
So, it was literally like he had to solve the problem. He could not use power.
1:24:23
And then, you know, this family put up all these competitions to really I like it from a like evolutionary product point of view.
1:24:29
It's like they said like, "Okay, anyone come here.
1:24:31
Karate, kickboxing, just try it. Open, you know."
1:24:36
They fought in these basements just like evolving the sport, proving that it was real.
1:24:40
A lot of our martial arts it's like magic and secret and you know, they never test their skills.
1:24:44
Um So, it works very well in practice.
1:24:47
Um The other thing I like about it is that I've done a lot of other martial arts.
1:24:55
You know, boxing and Thai boxing and stuff.
1:24:58
Those things are great as exercise but for for like self-protection is not very good.
1:25:03
You cannot punch someone in the face.
1:25:05
You're going to get sued as protection.
1:25:07
The beautiful thing about martial arts which is called the gentle sport is you you control people. You constrain them.
1:25:15
And you can adapt the level of violence.
1:25:17
This is why police use Jiu-Jitsu and not like Thai boxing because you can you can regulate the violence to the other person and you can control them without hurting them.
1:25:27
And so, that's why I think everyone should try should use it and practice it.
1:25:31
It's it's it's good both for self-discipline because you get humble.
1:25:34
It's also actually useful and you can use it without harming other people.
1:25:39
People probably don't know this because how would they?
1:25:41
But Spotify, you and Daniel especially have been probably the most influential people and certainly the company on me and how I've thought about building our businesses over time.
1:25:53
And a lot of that comes back to the stuff that you don't see.
1:25:55
We've talked I've purposely talked about a bunch of it today with you.
1:25:58
The bets board, the the complexity that's hidden behind a beautiful consumer experience.
1:26:04
You were the first person years ago to describe the bets board concept to me and we've used that very effectively and so many lessons from Daniel on how to think about what matters to users and I think Spotify is not only a an incredible product but it's also one where the company, you know, the product is a reflection of the company behind it.
1:26:21
And I think it's one worth studying by listening to conversations like this one because it raises that for me it what it's done is raised the the bar of ambition and the standard for excellence of how a company should be constructed to mirror the needs that it has, that it's unique needs.
1:26:39
But also just like the the character and the discipline of the people running it.
1:26:42
So, it's been so it's been so fun to do this with you and and thank you so much Thank you.
1:26:45
for all the lessons over the years.
1:26:47
Well, you know that the closing question that I have for everyone.
1:26:48
What is the kindest thing that anyone's ever done for you?
1:26:52
The thing that made me really excel in my role was being allowed to take a lot of risk by Daniel.
1:27:00
So, I've actually screwed up a bunch of things in Spotify that didn't work.
1:27:04
And I never felt that I was going to get fired for it.
1:27:09
And and he actually encouraged that and I got like a second chance.
1:27:11
And that's that's what made me like have the higher ambition instead of holding back for risk of of failure.
1:27:20
So, I think it's a it's a series of those things like being allowed to to mess up things.
1:27:28
Um that has probably had the biggest impact on on my professional career.
1:27:32
Can you give an example of a bad mistake that you made and what it how he and the org made you feel through that process so that you could be re-emboldened to take more risk again?
1:27:42
I was interested in new user interfaces and um many years ago I took the company very hard on a journey for an interface that at the time was like very provocative.
1:27:56
The idea was like that Spotify just starts playing things.
1:28:01
You swipe up to get to the next genre and you swipe left or right to get other things within the same genre.
1:28:09
And now you would say that sounds almost like TikTok.
1:28:10
This was before Musically.
1:28:13
Um But two things happened is it was very provocative.
1:28:17
It started playing things without you asking.
1:28:19
So, people were upset but I pushed pretty hard because I was convinced that immediacy and the idea was um you just sound your way to what you want to hear in in a very low friction interface.
1:28:31
And it was maybe a decent idea but it was before machine learning.
1:28:34
It just did not work at all.
1:28:35
You could just not get there. And we built this. It was called Moments. Um the UI.
1:28:40
We used editors on the back end which is did not work at all.
1:28:45
So, the idea was like far far ahead of where the technology was.
1:28:51
And it costed a lot of money.
1:28:51
We actually announced it.
1:28:52
There is a video of like us presenting this user interface and so forth.
1:28:55
People luckily forgot it. But it just didn't work.
1:29:00
Um We had AB tested it and it looked okay which is what we launched.
1:29:06
Then we discovered there was a bug in the AB test when it was live and it actually underperformed drastically what we had.
1:29:13
So, we had to roll it back and I'd taken like the entire organization on this excursion that lost us like a year or something in a very competitive business.
1:29:23
That was a good opportunity to get fired. And and I didn't.
1:29:28
Daniel was like, you know, I understand the back to like explanations.
1:29:32
No, I understand I agreed with the I with the thoughts and the ideas. What was the mistake?
1:29:35
And the mistake was like the machine learning was not there.
1:29:39
We were not good enough to get you there in enough swipes.
1:29:42
And and he was more like Jeff Bezos, you know, he measures the inputs, not the outputs.
1:29:47
Like if the inputs are bad, if the ideas are fuzzy and stupid, that's a problem.
1:29:52
But but you're not going to be you're not going to be always right even with good good ideas.
1:29:56
So, and I heard him say this, Jeff Bezos quote of like I I focus on I judge you by the inputs you had, not the outputs, because the problem with judging the outputs is you could just get lucky and you get promoted even though you're not very good just by luck.
1:30:11
Whereas if you look at the inputs, you know, if you just give more chances, if they're structured ideas ideation and execution, eventually you're going to get right.
1:30:19
So, just got more chances.
1:30:21
And that made me actually take more risk instead of scaling down on the risk.
1:30:24
But I felt very very very burned for a long time.
1:30:28
There are jokes internally about moments, you know.
1:30:29
What a powerful story and mindset for us all to adopt.
1:30:35
Such a great closing story, Gustav.
1:30:37
Thanks so much for your time. Thanks for having me. Pleasure.