· Gustav Söderström

How Spotify Thinks — Gustav Söderström on Invest Like the Best

Spotify survives technology shifts by prototyping + stack-ranked bets, running a fully synchronized leadership team, and demanding explanations (not pattern recognition) for why anything works — even A/B winners.

spotifyai-or-diebets-boardsuper-appbundlingproduct-strategymusic-industrydeutschgood-explanationsynchronized-orgfree-tiermarginal-cost-aipodcast-exclusivitymeasure-inputs95% confidence

Why this is in the corpus

Rare operator-dense view into how a 700M-user super-app allocates capital (bets board), runs product (E-Team), embraces AI without overfitting to the current moment, and rebuilt its business model (free shuffle tier) from first principles.

Summary for skimmers

Gustav Söderström walks through Spotify's operating system: a VC-style "bets board" where ~44 bets from 14 VPs are stack-ranked every 6 months; a 3-hour Tuesday E-Team meeting where no topic goes "offline" and direct reports are banned so VPs must know their own details; prototyping the next 6 months in Figma/AI tools before committing to synchronize the super-app org; David Deutsch's "good explanation" bar — falsifiable, has reach, hard to vary — applied to product decisions (no launch without a theory); the macro-wind / "AI or Die" framing; generative AI flipping consumer products from asymmetric downlink to symmetric conversation; admitting podcast exclusivity was a bad bet and reversing quickly; the shuffle-mode free tier as a first-principles answer to YouTube's foreground ad model; Spotify as the de facto R&D department of the music industry (15 years unprofitable, labels profitable throughout); Bezos-style "measure inputs, not outputs" culture that lets Gustav survive failed launches like the Moments UI.

Briefing

What survives the editorial filter

This page should feel like a smart colleague already listened for you and left only the operating logic worth keeping. Not everything said in the episode makes it through.

Trust signal

Direct episode extraction

Best used for

Decision-grade retrieval metadata not yet added for this episode.

Hold lightly

No explicit downgrade reason stored yet for this episode.

Principles

Durable claims that survive beyond the speaker's biography — each with explicit limits, transferability judgment, and evidence.

Principle

Gradual succession — hand over the full P&L years before the title

The right way to hand over a CEO seat is to transfer the full operational P&L years before the title.

Daniel Ek made Gustav and Alex co-presidents running the full P&L and balance sheet for three years before the CEO handover, so the operational job was known "by heart" and only PR/government/public-face work remained new.

Don't hand over the title and the job at once — transfer the operational job first, the title last.

already about three years ago he asked me and Alex Nordstrom to step up and become co-presidents and sort of start to run the, the day-to-day of Spotify for him. He was still the CEO... So when it actually came to sort of taking over CEOs, we already knew that part like by heart. And we had ran it for three years, the full, the full p and a and balance sheetGustav Söderström

Principle

Optimize hard for the one thing that matters, accept being average at the rest

Choose the dimension your org is excellent at to match what wins the category; be average elsewhere.

"You can't win, you can only not lose as much" — Spotify chose to be excellent at a single unified experience and accept being average at parallel-team velocity, because the unified experience is what wins.

Pick the one capability that wins your category and deliberately under-invest in the rest.

the best outcome is that you re optimized for the thing that is important for the company and you suck at the thing that is not as important. The worst outcome is you re really good at the non-important thing and you suck at the important thing. So like figure out what s important and then optimize for that and just accept that you re gonna be average at the others.Gustav Söderström

Principle

Stack-rank every bet globally — equal priority is a decision punted to the org

Always stack-rank; refusing to rank is how leaders unknowingly set their orgs up for political fighting.

Very few people manage to say this is actually more important than that. They're just saying these things are very important, both of them... If you as a leader don't bring clarity, you're going to set your org up for fighting.

Principle

Admit bad strategy and reverse — defending past decisions is the real cost

Two ways to be right: always guess right, or change your mind when wrong. The second is cheaper.

There are two ways to always be right. One is to always guess right. The other is to just change your mind whenever you're wrong... The real cost is when you try to defend your past decisions.

Principle

Never launch an A/B winner without a theory of why it works

Require a causal explanation, not just an A/B lift, before launch — explanations scale across the org, pattern recognition doesn't.

I don't want to launch it until you have a good theory of why it works... if you figure out the why, it's the difference between pattern recognition and actually understanding something.

Principle

Measure inputs, not outputs — good ideas that fail should still be rewarded

Judging outputs promotes the lucky; judging inputs gives good reasoners more at-bats until they hit.

Daniel was like, 'I understand. I agreed with the thoughts and the ideas. What was the mistake?'... I judged you by the inputs you had, not the outputs. That made me actually take more risk instead of scaling down.

Principle

Technology is necessary but not sufficient — change needs a contrarian business model

Durable change requires marrying a new technology to a new, often contrarian, business model.

Gustav cites access-vs-ownership (streaming) and the Kindle's whisper-sync (fixed-cost bundled data) — in both cases the business-model innovation, not the technology alone, created the change.

When evaluating a tech shift, ask what contrarian business model would convert it into durable advantage.

I believe that technology is necessary ingredient for change but not sufficient. I think you can cause havoc with technology like piracy, but when things really change is when you take a new technology and marry it with a new often contrarian business model. This is the Spotify story of access versus versus ownershipGustav Söderström

Principle

No direct reports in the executive meeting — force VPs to know their own details

Executives should be able to defend their own work without backup; rotating participants kills candor.

You're not allowed to bring anyone else in to explain your thing. You have to be on top of it enough to explain it to yourself. Over time these groups get very tight... People can be honest, no one is afraid.

Principle

Prototype the next 6 months before committing — synchronize disagreement early

Render the future visually before committing so alignment is forced while changes are still cheap.

What I've tried to do now, together with Alex Nordstrom, we synchronized the entire company... we prototype everything up front. So all this so-called fighting happens before you actually commit to doing something.

Principle

Ban "offline" and "later" in executive meetings — resolve in the room

Real-time resolution compounds; deferral compounds faster. With all decision-makers present, deferral is a choice, not a necessity.

You're not allowed to say the word 'offline' or 'later' because that person is in the room... Very simple in theory, but incredibly powerful in practice.

Principle

Always be first into the change — market share only moves during discontinuity

Market share is only winnable during technology discontinuities, so always adopt the change first.

Gustav notes Spotify grew most during periods of change (broadband→streaming, then mobile) and stagnates when markets are stable — so his rule is to accept that the world will change and get ahead of the curve.

Bet aggressively at inflection points; coast and you forfeit the only window where share moves.

Guess when Spotify grew the most... periods of change when things are stable, market share, stay stable, you don t eat market share. So when there s change, there s risk. You can lose market share. But that is also when you have the most opportunity to, to eat market share if you adopt a change. So my principle is just always be first. Be first and adopt it first.Gustav Söderström

Principle

Subscription revenue buys you the freedom to optimize for user value over engagement

Subscription aligns revenue with perceived value, freeing you from the engagement-maximization trap.

~90% of Spotify revenue is subscription, so they can make anti-engagement decisions (like letting users turn off video) because people pay for value felt, not time spent.

If you want license to optimize for user value, build a business model that monetizes value, not attention.

most of our revenue is from subscribers paying for their experience. You know, like almost 90% of the revenue. And I think that what you pay for when you vote with your wallet every month, you re not gonna pay for your engagement or time spent. You re gonna pay for the value you feel... it s a luxury for us because most of our revenue come from subscription to not have that pressure.Gustav Söderström

Frameworks

Reusable systems and operating models — including when they help and when they break.

Framework

The Bets Board (6-month VC-style stack rank)

A structured ritual that combines bottoms-up idea generation with global top-down prioritization, replacing political allocation with a transparent rank.

  1. VPs pitch bets as if they were startups pitching a VC
  2. Co-presidents stack-rank all bets 1..N globally
  3. Orgs resource from the top down until capacity is exhausted
  4. Orgs COMMIT to what they can deliver (bottoms-up commitment)
  5. Execute for 6 months
  6. Prototyping phase for NEXT 6 months runs in parallel
Use when: Large, multi-team product orgs that need to allocate scarce engineering time across many competing bets without letting VP politics decide.
Skip when: Small teams (<50) where a single roadmap works; or pure research orgs where scheduled commitment destroys serendipity. Also fails if planning tooling is weak — overhead exceeds execution.
Every six months, these VPs, they pitch, literally pitch, as if we were a VC and they were a startup... This time we have 44 bets. We stack rank them from 1 to 44.

Framework

Three-lens product evaluation — strategic, business, and emotional

Evaluate every product bet against three lenses: strategic (distribution), business, and emotional (time well spent).

For podcasts and audiobooks, Spotify applied all three: distribution favored a single app, the business case was real, and the content was "time well spent" — only moves clearing all three got built.

Run new product bets through strategic, business, and emotional lenses; require all three to pass.

If you look at what we focus on and what we optimize for, you have a strategic lens, then you have sort of a, a, a business lens and then an emotional lens. So the strategic L lens is really that we saw a long time ago that the biggest challenge in the world is going to be distribution.Gustav Söderström

Framework

Good-Calories Litmus (nutrition test for product)

Subscription model frees you from engagement-at-any-cost; pick verticals that produce "good calories" and you compound retention instead of guilt.

  1. Test 1: Post-hour feeling — energized vs. guilty
  2. Test 2: Parental-time-transfer — do parents push kids INTO it or OUT of it
  3. Test 3: Is it in line with the existing "nutritious" mission?
  4. Green-light if it passes all three
Use when: Choosing new bundle additions, vetoing feature ideas that would optimize short-term engagement at the cost of user regret.
Skip when: Ad-supported businesses where regret isn't penalized by the business model; discovery features where "junk food" engagement is the whole product.
If you lose an hour on Spotify, how do you come out feeling versus if you lose an hour doom scrolling in the bathroom... We see parents restricting screen time for their kids and saying, 'Go to Spotify instead.'

Framework

The extrapolation-vs-discontinuity test — which era are we in?

Before forecasting, decide whether you're in an extrapolation era or a discontinuity era — they demand opposite strategies.

2015-2025 rewarded extrapolation (more mobile/subscription/ads); 2005-2015 punished it (missed the smartphone). Gustav believes AI puts us in a discontinuity era where "everything changes."

Classify your era first; only then choose between optimizing the curve and betting on a new one.

if I was an analyst and I try to predict the world between 2015 to 2025... I would ve done really well to just extrapolate more mobile, more subscription, more ads, just more of everything... But if you would ve shifted that 10 years earlier, 2005 to 2015, you would ve extrapolated PCs and internet. You would ve missed a smart, you would ve missed everything... So the question is, which era are we in? Are we in the con borders of extrapolation or in the macro change?Gustav Söderström

Framework

Willingness-to-Pay vs Willingness-to-Sell value stick (Oberholzer-Gee)

Bundling + keeping price far below WTP is how Spotify manufactures consumer surplus; mission + culture lower willingness-to-sell so talent stays below market wage.

  1. Increase willingness-to-pay (stack value: music + podcast + books + video)
  2. Keep actual price far below willingness-to-pay
  3. Decrease willingness-to-sell via mission + culture, not just wages
  4. Capture value only where the gap is widest
Use when: Bundled consumer subscriptions that need to justify price raises over time; talent markets where cash alone won't win.
Skip when: Zero-margin commodity businesses where there is no surplus to divide; early-stage startups that can't afford mission-over-cash hiring.
Our goal as a service is to make sure that Spotify is just an amazing deal. You're always going to feel the willingness to pay the actual value you perceive is way over the price.

Framework

Deutsch's Good Explanation bar

An explanation you can swap characters in (like a conspiracy theory or Thor-causes-thunder) is too easy to vary; a theory where parameters are load-bearing is close to truth.

  1. Test 1: Is it falsifiable?
  2. Test 2: Does it have reach? (works at multiple scales / domains)
  3. Test 3: Is it hard to vary? (swap a parameter → prediction breaks)
  4. Reject: pattern recognition dressed as reasoning
  5. Accept: a theory that survives parameter perturbation
Use when: Evaluating competing product/strategy explanations, filtering plausible-sounding rationales from durable ones, onboarding senior hires into structured reasoning.
Skip when: Pure exploratory brainstorming where premature falsification kills options too early.
A good explanation has to be hard to vary. If you move one of the parameters, the entire thing is not predictive anymore, then you're probably close to the truth.

Signals

What appears to be shifting, for whom it matters, and what happens if you ignore it.

Signal

AI is a dual-use technology — the same models can be the most addictive algorithm or hand control back

Generative AI is dual-use — it can be the most addictive algorithm ever or the tool that returns control to users.

Spotify's bet is to let users tell the system in plain English who they want to be musically (e.g., "I want classical, less EDM") — turning AI's deep understanding toward user agency rather than capture.

Expect the AI wave to split into engagement-maximizers and control-returners; choose your side deliberately.

There is a promise for generative AI to be the most addictive algorithm you have ever heard of. Because now we can understand you so deeply there is potential for for darkness over there... You could choose to do something else with generative ai. And so what we re choosing to do, for example, is to give back users control of the algorithmGustav Söderström

Signal

Non-developers are starting to use Cursor via MCP

The bottleneck for AI inside big companies is no longer AI engineering — it's boring old-school API exposure. Once data is real-time and MCP-wrapped, the user base of AI-native tools explodes past developers.

I had one of my PMs who doesn't speak Swedish, she did her taxes in Cursor, managed to wrap the Swedish tax authority in an MCP, not a developer. So I think it's going to grow outside of developers.

Signal

Big-company coding speedup from AI is ~7% today — but the unlock is yet to come

Public-market expectations of AI productivity gains at large companies are temporarily inflated; the durable gains will come from refactor-capable models + non-coding workflows, not Cursor-style autocomplete.

I've seen studies from other big companies that if you actually measure out of a developer's time, the speed up is 7% or something, which sounds very disappointing... but I think it's going to turn into the opposite. I think it's going to have tremendous impact over the longer term.

Signal

Natural language is the new universal interface — every user becomes a "developer"

Natural language turns every one of your users into someone who can directly instruct your product.

Gustav reframes generative AI as "computers finally understand English" — replacing 10-user research panels with continuous high-fidelity dialogue across all 761M users.

Build for a world where every user can express intent in plain English — and design for that scale of dialogue.

The best way I think to describe generative AI is that finally computers understand English. Like it used to be a small population of about 1 million developers on GitHub who could talk to computers. Now we all can. And so I think consumer companies could and should give everyone access to talk to them in plain English.Gustav Söderström

Signal

AI has non-zero marginal cost — business models will tier by inference consumption

The next wave of consumer pricing will look more like Spotify's label-royalty model (per-use cost must be recovered) than Twitter's 2010s model (worry about monetization later).

The previous VC model was you make a big upfront investment and you get to almost zero marginal cost. That's how software worked, that's not how AI works. The marginal cost is high and you need to cover it... you're probably going to see more tiering of consumer products based on how much inference you want.

Signal

Media habits will change because of AI — sitting still is not an option

A top media operator is directionally certain AI will reshape media habits, even though the specifics are unknowable.

Gustav says the AI question is what keeps him up at night — not because he knows the answer, but because the certainty of change makes standing still the most dangerous option.

Treat directional certainty of disruption as sufficient reason to move, even without knowing the endpoint.

the thing that keeps me up at night is not surprising what, what AI truly means in the limit. I m pretty sure that media habits are gonna change because of ai. And so I don t think it s an option to just sit still... I think we re one of these where everything changes. So I can t tell you what s gonna change, but I m pretty sure that things are gonna change.Gustav Söderström

Opportunities

Only included where there is a buyer, a real wedge, and a plausible revenue path — not vague idea theater.

Opportunity

Product-overhang exploitation — ship two years of features on today's models

Aggressive product refactoring on today's GPT/Claude-class models: rebuild core workflows as two-way conversation, not downlink-heavy UIs.

Wedge: Mid-stage consumer products with large user bases — Notion, Duolingo, any media subscription — not AI-native startups already doing this.
Why now: Gustav explicitly subscribes to product overhang; inference cost is dropping fast enough that the economics already work.
I subscribe to the product overhang idea that there's a huge product overhang and if we froze, I think we would see products shipped that look amazing for several years before we exhausted what we have.

Opportunity

Wrap legacy enterprise data in MCP so the non-engineer 80% can reason over it

Boring-but-critical infra work: API-ify every cold dataset, wrap in MCP, ship an internal AI workbench per skill-group.

Wedge: Start with one workflow (e.g. "query 15 years of contracts") at orgs with deep structured data — banks, pharma, telcos, media.
Why now: LLMs cheap enough to be a reasoning substrate + MCP stabilizing as the standard + non-engineers demanding it. Three-way convergence that didn't exist 12 months ago.
Actually my biggest job to enable AI is not AI engineering, it's old-school engineering exposing all this data.

Opportunity

Mainstreaming audiobooks via subscription bundling (à la Nordics)

Bundle audiobooks into Premium with a generous monthly cap + top-up — exactly Spotify's playbook.

Wedge: Audio/content bundles that can license publisher catalogs — not just Spotify; also niche literary apps (e.g. Substack + audio).
Why now: Consumer willingness to stack subscriptions has peaked; bundled audiobooks land inside an existing subscription.
Audiobook was a very niche behavior. 10, 11 million or something... you can see in the Nordics where you have the access model that it's getting very mainstream.

Opportunity

"Premeditated media" — agents that filter your feed against your stated future self

Agents that let users pre-commit their content filters resolve the gap between their reflective and in-the-moment selves.

Gustav built a personal agent that filters X for rage-bait/clickbait/politics and surfaces only what people he trusts are discussing — a "premeditated media" category he thinks everyone should be able to access.

Build for the reflective self that pre-commits, not the in-the-moment self that gets captured.

I literally asked my agent filter for rage bait filter for for click bait filter for you know, politics... can you give people the chance to decide their own future ahead of time? This is the idea behind like the taste profile behind these agents... Premeditated Media... is something that, that I think everyone should at least have access to if they want to.Gustav Söderström

Opportunity

AI-composed adaptive fitness audio — beat-matched to your cadence and taste

AI-composed, cadence-matched fitness audio is an unfilled gap sitting on a large, engaged, low-regret user base.

~70% of Spotify users already exercise with the app; AI can now generate pace-matched, taste-personalized, beat-mixed running audio with coaching overlays — a product impossible before generative composition.

Look for AI-unlockable products that sit on top of an existing low-regret, high-engagement use case.

you say to Spotify, you know, I want running playlist for like an eight minute mile and I want it to be in my taste, but I want it to be on the downbeat either on my, on both my feet or just one of my feet... then you actually want to speed em up or slow them down to exactly 160 or exactly 80. And then you actually want to mix them together with perfect beat match transitions, right? And then you actually wanna overlay commentary on topGustav Söderström

Lessons still worth keeping

Useful takeaways that did not fully clear the bar for durable principle status.

Lesson

Podcast exclusivity — betting on celebrity content in a low-production-cost medium

Exclusivity is powerful when content is capital-intensive and content-picking skill is rare. In podcasts, neither held — they should have followed the YouTube model from the start.

Before copying a content-strategy from another medium, check whether the underlying economics (production cost, talent supply) match. If they don't, the strategy inverts.

The macro trend for podcasts was that the production cost was so low... to go in and do exclusivities on top of that is counter-purpose. The whole point is more like YouTube... We also betted a lot on celebrities and they are celebrities, but they're not always good podcast hosts.

Lesson

Buying voice tech (Sonantic) ahead of the curve to intercept, not wait for, the AI window

If you believe in an exponential, acquire the enabling pieces early to intercept the curve, not chase it.

Spotify bought Sonantic for cheap-per-minute voice before LLMs could even write the scripts — betting on intercepting the AI curve rather than waiting and building once it arrived.

When you believe a curve is coming, buy or build the bottleneck capability before it's obviously needed.

We bought this company called Semantic to be able to produce voice... if you wanna serve 700 million users with like a few minutes of voice per day, you re gonna go bankrupt. We bought this company producing voice very, very cheaply with the different technology even before the lms were smart enough to produce the script for that voice because we just bet on, you know, intercepting that curve instead of waiting for it and then then building for it.Gustav Söderström

Lesson

The Moments UI — shipped ahead of the underlying ML

Great product vision + weak underlying technology = premature launch. Even a clean A/B result can hide an instrumentation bug when the UI is radically new.

Don't ship a UI paradigm that requires capability your stack doesn't yet have — and treat "A/B looks okay" on a novel surface with extreme skepticism.

The idea was far ahead of where the technology was and it costed a lot of money. We actually announced it. We had A/B tested it and it looked okay, which is what we launched. Then we discovered there was a bug in the A/B test when it was live. And it actually underperformed drastically what we had.

Lesson

Three deliberate counter-positions to Apple — freemium, personalization, ubiquity

Spotify's three founding bets were deliberate counter-positions Apple structurally could not match.

Freemium (Apple ad-averse), personalization (Apple data-averse), and ubiquity (Apple ecosystem-locked) all paid off because each exploited a constraint inherent to Apple's identity — and against a much larger, more respected product company.

Pick strategic positions your dominant competitor is structurally forbidden from occupying.

when we created our strategy, we bet on three things. And they were all basically counter positions to Apple deliberately. One was freemium... The second was personalization... The third was ubiquity. We bet that they were gonna prefer their own products and never get be good on an Samsung TV or an Android phone. So these were our three bets. Freemium, personalization, and ubiquity. And I have to say they pan out really well.Gustav Söderström

Lesson

The free shuffle tier — first-principles reasoning beat pattern-matching YouTube

When the pattern-matched move exists, reason from underlying usage data instead. Foreground ads were a local optimum; 91% of actual listening was background.

Even inside a company, pattern-matching to a visible competitor feels safer than first-principles reasoning — but the first-principles answer is where durable differentiation lives.

Turns out back then it was 9% or something. So you have 91% of the use case being in the background... Even the people inside the company said that's a terrible idea. But we trusted the data... This is what made growth explode.

The Plays

Try these this week

Verb-first executable actions — each one tied to a stated outcome in the episode.

Launch a Free Shuffle-Only Tier Using Premium Engagement Data

Outcome: If premium users spend ~50% of time shuffling, offer shuffle-only access for free — it captures roughly half the value but never 100% of any user's need, minimizing cannibalization while driving top-of-funnel growth.

Gustav Söderström — How Spotify Thinks — Gustav Söderström on Invest Like the Best
Gustav Söderström
  1. 1

    Analyze premium user behavior to identify the single highest-usage feature or mode.

    At Spotify, 50% of premium listening sessions were shuffle.

  2. 2

    Model cannibalization risk: confirm that the feature represents a large share of aggregate usage but not 100% of any individual's consumption.

    If it's 50% of sessions, no user relies on it exclusively, so free access won't fully replace premium.

  3. 3

    Launch a free tier restricted to that single mode (e.g., shuffle-only playback).

    At Spotify, this became the free mobile experience.

  4. 4

    Track both free-tier growth and premium conversion/churn to validate the no-cannibalization hypothesis.

    Spotify saw 'growth explode' without material premium erosion.

Stop or pivot when

  • If the feature accounts for ≥100% of any significant user cohort's usage, cannibalization risk is too high

Before you start

  • · Detailed usage telemetry by feature/mode
  • · Ability to gate features at the product level (e.g., shuffle vs. on-demand)
  • · Willingness to launch a lower-value tier despite internal skepticism
freemium-tier-designgrowthcannibalization-modeling1-1010-50

Run a multi-year co-president apprenticeship before handing over the CEO seat

Outcome: Transfer escalating operational control to your successor over years so the title handover is low-risk.

Context: Daniel Ek named co-presidents three years out, widened their rope from daily to weekly to monthly P&L ownership, and kept only external/ceremonial duties to transfer at the end.

already about three years ago he asked me and Alex Nordstrom to step up and become co-presidents and sort of start to run the, the day-to-day of Spotify for him. He was still the CEO and so he is been gradually like giving us more and more rope to, to run the, the business day to day and week to week and then month to month.
Gustav Söderström
2-3 years before handover per
  1. 1

  2. 2

  3. 3

  4. 4

  5. 5

  6. 6

  7. 7

Before you start

  • · A founder genuinely willing to delegate
  • · A successor with enough tenure to hold institutional trust
  • · Board alignment on the transition timeline

Take the distribution pain of a single app instead of shipping a separate one

Outcome: When distribution is the binding constraint, build into your existing app rather than shipping a standalone.

Context: Seeing 7+ good podcast apps stuck near 0% share against Apple's 98.5%, Spotify judged distribution (not product) the real problem and built podcasts into its 300M+ user app despite the engineering pain.

is that the biggest problem to design the experience or is the biggest problem actually getting distribution? ... I think Apple Podcast was still like 98.5%, all of them. So like the problem wasn t that there wasn t a good enough podcast product, the problem was that they didn t get a distribution. So we chose to take the pain of doing a single app with all the complexity that comes with that for the benefit of reaching what was then already 300 something million users
Gustav Söderström
per major product expansion per
  1. 1

  2. 2

  3. 3

  4. 4

  5. 5

  6. 6

  7. 7

Before you start

  • · A large existing distribution surface / installed base
  • · Engineering capacity to absorb multi-model backend complexity
  • · Willingness to take the harder build for the larger reach

Merge separate leadership meetings into one synchronized all-functions weekly

Outcome: Replace siloed leadership meetings with one weekly all-functions meeting so blockers resolve in real time.

Context: Spotify's co-presidents killed separate product and business leadership meetings in favor of a single 3-hour Tuesday "ET" meeting with ~14 SVPs across all functions, banning "take it offline."

We chose to say we re not gonna have our own leadership meetings. We have a single one with all of our SVPs. So we have a single meeting every Tuesday called ETE for three hours with all the SVPs of, of all the internal functions like you know, marketing ads, subs, but also, you know, all the product and technology functions in the same meeting.
Gustav Söderström
3 hours, weekly per
  1. 1

  2. 2

  3. 3

  4. 4

  5. 5

  6. 6

  7. 7

Before you start

  • · A single unified product/experience strategy that justifies the coordination cost
  • · Senior leaders willing to sit through non-immediately-relevant discussion
  • · Tolerance for high leadership time spend

Run an anonymous third-party "regret" survey across all competing platforms

Outcome: Measure how much time users regret — anonymously, third-party, across all platforms — to surface value vs capture.

Context: Spotify ran a blind third-party survey across Spotify, YouTube, Apple Music, Amazon, TikTok et al., asking both satisfaction and regret; it found Spotify lowest-regret and some platforms at 60%+ regret, which became the basis for the "time well spent" strategy.

We surveyed our users through a third party anonymously... We asked the question like, how do you feel about the time you spent... But then we also asked the opposite question, how much of your time did you regret afterwards? ... we were actually the lowest regret content sort of on the internet
Gustav Söderström
periodic (ongoing competitive tracking) per
  1. 1

  2. 2

  3. 3

  4. 4

  5. 5

  6. 6

  7. 7

Before you start

  • · Budget for third-party blind research
  • · Willingness to act on an unflattering finding about your own product

Run Six-Month VC-Style Bet Pitches with Stack-Ranked Resourcing

Outcome: Every six months, VPs pitch bets as if to a VC, the company stack-ranks them (e.g., 1–44), then resourcing teams work top-down until capacity is exhausted, ensuring only the highest-conviction initiatives get funded.

Gustav Söderström — How Spotify Thinks — Gustav Söderström on Invest Like the Best
Gustav Söderström
six months40 per 180 days
  1. 1

    Hold a pitch session every six months where VPs present their proposed bets as if pitching a VC.

    Treat the session formally; each VP makes the case for why the company should back their initiative.

  2. 2

    Stack-rank all submitted bets from highest to lowest priority.

    The example given was 44 bets ranked 1 through 44; typical range is 30–50 bets.

  3. 3

    Hand the ranked list to resource-allocation teams and start resourcing from the top.

    Teams work down the list until they hit capacity (e.g., 'maybe they get to 30' out of 44).

  4. 4

    Commit only the top bets that can be resourced over the next six months.

    Unfunded bets below the cut line are deferred or killed for the cycle.

Stop or pivot when

  • If a bet cannot be resourced after working top-down through the stack rank, it is not executed in that cycle

Before you start

  • · VP-level or equivalent leadership team with ownership of strategic initiatives
  • · Ability to estimate resourcing capacity for a six-month window
  • · Organizational willingness to kill or defer lower-ranked bets
roadmap-prioritizationresource-allocationportfolio-management10-5050+

Require a Falsifiable Theory Before Launching Winning A/B Tests

Outcome: Even when an A/B test shows positive results, insist that the team articulate a 'hard to vary' explanation for why it works before you ship — ensuring the learning scales across the organization.

Gustav Söderström — How Spotify Thinks — Gustav Söderström on Invest Like the Best
Gustav Söderström
  1. 1

    Run the A/B test and observe the metric lift.

    Standard experimentation; measure whether the treatment wins.

  2. 2

    Before greenlighting launch, ask the team for a theory that explains why the test worked.

    The theory must be falsifiable, have reach (scale), and be hard to vary (changing one parameter breaks the predictiveness).

  3. 3

    If no coherent theory emerges, hold the launch or run additional experiments to uncover the mechanism.

    Pattern recognition alone ('it worked in the test') is not sufficient justification.

  4. 4

    Document and share the theory so the entire org can apply the learning to future decisions.

    This is how insights compound across teams.

Stop or pivot when

  • If the team cannot articulate a hard-to-vary explanation, do not launch the winning variant

Before you start

  • · Active A/B testing infrastructure
  • · Organizational norm that metrics alone do not justify launches
  • · Leadership willing to delay or reject statistically significant wins without theory
experimentation-rigorknowledge-scalinglaunch-gating1-1010-5050+

Hold a Weekly All-VPs Escalation Meeting with No-Offline Rule

Outcome: Run a recurring all-VPs meeting (e.g., Tuesday) where blocked issues are escalated, 'offline' discussions are banned, and no direct reports attend — forcing senior leaders to resolve cross-functional blockers in real time.

Gustav Söderström — How Spotify Thinks — Gustav Söderström on Invest Like the Best
Gustav Söderström
ongoing weekly cadence1 per 7 days
  1. 1

    Schedule a weekly all-VPs escalation meeting (e.g., every Tuesday).

    With five-day work weeks, no one waits more than 2.5 days on average to escalate a blocker.

  2. 2

    Enforce a 'no offline' rule: when someone says 'let's take that offline' or 'I'll talk later,' immediately require resolution in the room.

    The goal is to eliminate deferred decisions and force closure on the spot.

  3. 3

    Ban direct reports from attending; only VPs may participate.

    This forces VPs to own details and resolve issues themselves rather than delegating on the fly.

Before you start

  • · VP-level or equivalent leadership cohort who own cross-functional outcomes
  • · Cultural buy-in to real-time decision-making and no deferral norm
escalation-hygienecross-functional-coordinationdecision-velocity10-5050+

Decision Moments

Actual decisions, real outcomes

Specific decisions narrated in the episode with their outcomes and transferable lessons.

Daniel Ek wanted to hand over the Spotify CEO role but knew most of the job is tacit, operational knowledge that cannot be transferred by briefing. A cold handover would put the company at risk during the most demanding part of the transition.

Did: Three years before the title handover, he named Gustav Söderström and Alex Norström co-presidents and gradually widened their rope from running the day-to-day to weekly to monthly ownership of the full P&L and balance sheet — keeping only the external/ceremonial duties (PR, government, public face) to transfer at the end.Outcome: When the CEO transition came, the co-presidents "already knew that part by heart" having run the full P&L for three years; only the externally-facing responsibilities remained new, so operational risk at handover was minimal.

Succession risk collapses when you transfer the operational job years before the title — master the tacit part live with the founder as backstop, and leave only the ceremonial part for the end.

Part of an emerging decision pattern across multiple episodes

Generative AI presented a fork: the same deep user-modeling capability could be pointed at building the most addictive algorithm ever, or at something contrarian. Meanwhile a blind cross-platform survey showed users regretted 60%+ of time on some big platforms — and Gustav judged AI to be a discontinuity era where media habits will change and standing still is the riskiest move.

Did: Repositioned Spotify around an "AI-or-die"-style imperative to move first into the change, but bet contrarian: instead of engagement-maximization, use AI to hand users explicit plain-English control of their own algorithm (taste profiles, "premeditated media" agents), anchored to a codified "time well spent / no regrets" strategy and protected by ~90% subscription revenue.Outcome: Spotify made anti-engagement decisions (e.g. letting any user turn off video) and is investing in user-control AI and no-regret verticals like fitness — a directional bet whose endpoint is unknown but whose direction (be first, return control) is committed.

In a discontinuity era you can be certain change is coming while unable to predict its form; the move is to bet first on a contrarian direction your business model can sustain, not to wait for clarity.

Part of an emerging decision pattern across multiple episodes

Tensions surfaced

Contradictions and trade-offs the episode raises — judgment calls a thoughtful operator has to navigate.

Tension

Synchronized super-app vs divide-and-conquer speed

Global changes at scale require synchronization. Rapid local experimentation requires decoupling. The same org cannot do both equally well.

We're good at doing global changes, like changing the entire UI because we're synchronized, but we're probably much slower than other companies at trying something. It's not the right one. It's the right one for us.

Tension

Build for today's AI workflows or wait for the next model

Ship velocity vs overfitting risk. Every feature you ship is effectively a bet on a snapshot of capability that will be obsolete before it pays back.

You're somewhere right now, but we're pretty certain that that somewhere is on this curve... you don't want to overfit too much to the moment.

Tension

Per-stream payout metric is lower when your product is BETTER

Engagement quality drives the per-stream metric down even as it drives aggregate label payouts up. Creator-facing transparency and shareholder-facing logic pull opposite directions.

These other companies have higher per stream because they have a worse product... we have twice the engagement and half the churn of competing services. So that's a curse.

Tension

Tenure's trust-and-efficiency vs the risk of becoming a closed group of old people

Tenure buys trust and efficiency but costs fresh blood — you must optimize one and engineer mitigations for the other.

Gustav contrasts Larry Ellison/Oracle (kept the core team for ~two decades) with Elon (wants churn and fresh blood) — both built trillion-dollar outcomes, so the answer is to pick tenure and mitigate, via Rising Stars programs and senior hires.

Optimize for tenure's trust, then deliberately counter-engineer for the fresh blood it costs you.

one benefit of, of tenure is the trust people keep you honest... the other benefit is this efficiency of like, people know you don t have to give as much context. There are some downsides to tenure, which is opposite of eons. You don t get fresh blood as much. And so you need to be careful. Maybe you re just a, a group of old people eventuallyGustav Söderström

Tension

User control as a product bet — power users want it, the majority want it done for them

Building user-control tooling pays off even though most users won't use it — the active minority's work improves everyone's defaults.

Gustav invokes the 1-9-90 rule and Spotify's ~10B user playlists: a small minority's curation already powers great recommendations for passive users, and AI control will work the same way.

Don't kill control features because most users skip them — their signal upgrades the automatic experience for all.

People like you and I will probably want more control, but like you re of 700 million users, like what percentage of them do you actually think? I, I feel like people just want you to do it for them.Patrick O Shaughnessy
some people engaging a lot talking to Spotify all the time saying that, no, that s wrong... They re doing a lot of work, they re making their Spotify better for themselves. But then we know that someone else who looks very similar to that but doesn t have the time or knowledge to do all of that actually probably want the same thingGustav Söderström

Corpus connection

Where this episode fits for retrieval

What kinds of decisions this briefing is best pulled into.

Primary decisions

  • product-strategy
  • capital-allocation
  • business-model
  • hiring-culture