Spent two hours listening to Spotify’s CEO talk about building a $146B company
I went down a rabbit hole yesterday and ended up listening to Daniel Ek, the CEO (He’s stepping down next year) of Spotify, talk for two hours straight.
Here’s what stuck with me.
The whole thing started with one painful problem: make legal music faster and easier than piracy. That’s it. Not “revolutionize music” or some grand vision. Just solve one annoying thing people actually deal with.
Start narrow, then expand carefully
What really hit me was how disciplined they were about focus. They started with music. Only when that worked did they add podcasts. Then audiobooks. Each time using the same distribution engine they already built. It’s like they had one machine that just kept getting fed new formats.
Most of us (myself included) love chasing shiny new features.
But Ek said something that hurt to hear: don’t hunt for miracle features. Just ship lots of small improvements that compound. Stack wins. The math works out better.
Shipping beats planning
The bias toward shipping really resonated. Launch rough, measure, iterate. Momentum beats committees. I’ve sat in too many meetings trying to perfect something before anyone’s even seen it.
Also this: kill half-hearted side projects. If it’s not important enough for senior people to work on, it’s probably not important. Protect focus like your life depends on it.
Two things I’m stealing
First, the “dogfood like crazy” rule. If your own team doesn’t love using it, users definitely won’t.
Seems obvious but how many products do we build that we’d never actually use ourselves?
Second, designing for two modes: lean-forward (I know what I want) and lean-back (decide for me). Every interface needs both. Spotify gets this. Most apps pick one and forget the other exists.
The stuff that makes you uncomfortable
Some harder truths: your platform will be gamed. People will misuse it. Don’t fight human nature, design for it. Use the misuse as input for better policies and ranking.
Also, price is a product decision. You need to adjust plans when value changes. Can’t just set it once and hope.
And as you scale, your job changes. You stop deciding features and start allocating capital and setting strategy. That transition sounds brutal but necessary.
What I’m thinking about
The north star thing stuck with me. Spotify’s is simple: help more creators make a living and help more users find what they love. Everything maps back to that.
I keep asking myself if I have something that clear. Or if I’m just collecting features and hoping they add up to something.
Anyway, these are my notes. Probably jumbled. But sometimes the best ideas come from just listening to someone who’s actually done it talk for a while about what worked and what didn’t.
I had this idea before Spotify made their AI DJ: A concept: Spotify – now with 180 million DJs
