My Time at Spotify as a PM: Learning the DNA of Spotify

Before Spotify, I was working on moonshot projects. Big, ambitious, slightly crazy bets on what technology could do. I loved that energy, but I wanted to learn how to do it at the place I believed did it better than anyone in my world. So I moved to Stockholm. Not for the city, not for the weather (anyone who has done a Swedish winter knows what I mean), but primarily to work at Spotify.

I spent three and a half years there as a product manager building tools for music creators. I led collaboration experiences on Soundtrap, worked on AI Music Making with Spotify's AI Labs, and co-invented Basic Pitch. But honestly, the features are not the story. The story is what those years did to the way I think. I walked in believing I was a strong PM. I walked out realizing I had only just started.

This is what I learned about the DNA of Spotify.

DIBB: the framework that runs the company

If there is one thing I carry from Spotify into everything I build now, it is DIBB.

DIBB stands for Data, Insight, Belief, Bet. It is the backbone of how Spotify makes decisions, from a tiny feature all the way up to company strategy. It sounds simple. It is brutal in practice.

Here is how it actually works.

Data is the raw signal. Numbers, user behaviour, research, everything you can observe. Not what you feel, what you can see.

Insight is where it gets hard. An insight is the non obvious thing the data is actually telling you. Most people stop at the data and call that an insight. It is not. The insight is the "oh, that is why" moment hiding underneath the numbers.

Belief is the conviction you form from that insight. It is your point of view on how the world works, or where it is going. This is where you are allowed to be bold, as long as the belief is earned by the insight.

Bet is the decision. The thing you actually go and do, resourced and sized to match your conviction. Small belief, small bet. Big belief, big bet.

What DIBB does to you as a PM is take away your right to just have opinions. You cannot walk into a room and say "I think we should build this." You have to walk the whole chain. Here is the data, here is the insight, here is the belief it gives me, so here is the bet I want to make. Half my ideas used to die at the insight step, because I realised I had a vibe and not a reason. The ones that survived, I would defend in a fire. I still think in DIBB today, every single week.

An obsession with data and experimentation

DIBB only works because Spotify is genuinely obsessed with data, and that obsession is wired into the tools.

At Spotify you do not win an argument by being senior or loud. You win it with an experiment. Almost everything ships as a test first. You roll a change out to a small slice of users, you watch what actually happens, and you let the numbers tell you whether you were right. Opinions are cheap. Evidence is the currency.

What made this real was the internal tooling. Spotify invested heavily in experimentation platforms that let any PM design an A/B test, ship it to a controlled group, and read clean results without having to beg an engineer or a data scientist for weeks. That sounds like a small thing. It is not. When running an experiment is easy, you run more of them. When you run more of them, you learn faster. When you learn faster, you stop guessing and start knowing. The whole company moves on evidence because the tooling makes evidence cheap to get.

I learned to fall in love with being proven wrong quickly. A failed experiment that cost a week is a gift. It is a guess you did not have to bet the roadmap on.

Winning an industry it had no business winning

Here is the thing that quietly reset my ambition.

Spotify should not have won. Apple basically invented the modern music business, the iPod, iTunes, then Apple Music sitting pre installed on a billion phones. Amazon had endless money and Prime to bundle music into. Google had YouTube, the biggest audio and video library on earth, with free distribution to match. Every structural advantage, hardware, capital, reach, belonged to someone else.

Spotify won on the one thing the giants treated as a feature and Spotify treated as the whole company. The product. Not out spending Apple, not out bundling Amazon, but out caring about the listening experience, week after week, at a level of craft the others could not be bothered to match.

The lesson went deep. Resources are not destiny. Focus and obsession beat money and reach far more often than the spreadsheet says they should. I think about that every day now as a founder going up against companies a thousand times my size.

The obsession with the top 0.01% outcome

The bar inside the building was almost irrational. "Good" was treated as an insult. The teams I worked with were not chasing a solid outcome, they were chasing the best possible one, the top 0.01%, the version that would still feel great in five years.

That obsession is contagious, and it is permanent. Once you have watched a team refuse to ship something merely good, you cannot unsee it. You carry the higher bar into every room after that. That, more than any single skill, is what changed me the most.

SVPG and a real product culture

Spotify treated product management as a craft worth investing in, not an afterthought.

We were trained on the SVPG model. SVPG is the Silicon Valley Product Group, the team founded by Marty Cagan, best known for the books Inspired and Empowered. Their whole philosophy is the idea of empowered product teams. You hire smart people, you give them a real problem to solve instead of a feature to build, you point them at outcomes instead of output, and then you trust them to figure out the how.

Spotify lived this. Teams were handed problems and real autonomy, not a list of tickets. And the Swedish culture made it work. Flat, low ego, high trust. Psychological safety was not a poster on the wall, it was the actual condition that made people brave enough to take the big bets that DIBB pointed them towards.

The smaller aha moments that rewired how I think

Some lessons just landed quietly and never left.

Aligned autonomy. Spotify's org itself was a product. Squads, tribes, chapters, guilds. The whole point was aligned autonomy, teams that were loosely coupled but tightly aligned on the why. The aha was that how you structure people is a product decision, not an HR one. You scale by making teams independent and pointing them all at the same north star, not by adding layers of approval.

Data informs, it does not decide. This one took me a while. Spotify is obsessed with data, but the best PMs there knew data tells you what is happening, never what to do. You cannot A/B test your way to soul. The hardest, highest leverage calls are still judgment and taste, and the experiment is there to check your taste, not replace it.

The value is in the tweak it. Spotify thinks in stages. Think it, build it, ship it, tweak it. Most teams everywhere else celebrate the ship. At Spotify the launch was treated as the starting line. The real value got created in the weeks of tweaking after, reading the data, fixing the rough edges, compounding small wins. Launch is the beginning of learning, not the end of work.

A roadmap is mostly a list of nos. The discipline that surprised me most was saying no. With unlimited ideas and limited bets, the roadmap is really a list of everything you consciously decided not to build this cycle. Focus was not a vibe, it was a daily act of killing good ideas to protect the great one.

Discovery never stops. Discovery and delivery ran in parallel, not in phases. You are always talking to users and testing assumptions while you build, so you are never six months into a thing before you find out it was wrong. The aha was that being slightly uncertain forever is healthier than being confidently wrong once.

You are building a two sided marketplace. Building creator tools taught me you are never serving one user. Every decision has a listener side and an artist side, and value has to flow both ways or the whole thing breaks. Once you see the marketplace, you stop optimizing one screen and start thinking about the loop.

The free product is the marketing. Spotify's freemium model reframed growth for me. The free tier is not a charity, it is the top of the funnel and the best ad the company runs. The aha was that giving the product away to the right person, with the right friction, is often a smarter acquisition strategy than a marketing budget.

Human plus machine beats either alone. The magic of Spotify's personalization, the Discover Weeklys of the world, was never pure algorithm. It was editorial taste and machine learning working together. The lesson I carry into AI products now is that the winning experience blends human judgment with the model, it does not hand everything to the model.

Play the long game. Spotify was patient in a way that looked almost reckless from outside. It optimized for where music was going, not the next quarter. Watching that taught me that conviction plus patience is a real competitive advantage, because most competitors cannot stomach the wait.

The best ideas come from the bottom. A lot of Spotify's biggest hits started as small bottom up bets during hack time, not as executive mandates. The aha was that your job as a PM is partly to create the slack and safety so the brilliant idea sitting with an engineer actually sees daylight.

Strong opinions, loosely held. The smartest people I worked with would argue hard for a position and then change their mind in a second when the data turned. No ego attached to being right. I learned that updating fast is a superpower, and that defending a dead idea to save face is the most expensive thing a PM can do.

Ship lovable, not just viable. The bar was never minimum viable, it was minimum lovable. Something people would actually feel something about, not just tolerate. That small reframe changes what you let out the door.

What I actually walked away with

None of this was a course I finished. It was an environment I got to stand inside, and environments are sticky. You absorb the standard whether you mean to or not.

I left Spotify a completely different operator than I arrived. Sharper about why before what. Allergic to good enough. Convinced that data beats opinion, and that craft and obsession can beat capital. Those are not lines on a CV. They are the operating system I run on now as a founder.

If you ever get the chance to spend real time around the top 0.01% of any craft, take it, even if the title looks like a step sideways. The skills you can pick up anywhere. The standard you cannot.