Epidemic Sound CEO on AI, value chains, and building a $1.4B company

Executive overview

Oscar Hoglund, co-founder and CEO of Epidemic Sound, built a $1.4 billion music licensing platform that soundtracks three billion video views daily. The company's growth came from a legal innovation — acquiring full music rights rather than licensing fractionally — and from understanding that leverage in business means achieving massive output from small, targeted input. This interview covers the frameworks behind that growth: value chain analysis, the feature-to-platform arc, and a four-part success model. The personal cost was near-total: a family crisis forced a complete rethink of role, schedule, and priorities, leading to a discipline-based operating system that replaced hustle. The core insight is that disciplined structure, not relentless availability, is what sustains both the company and the life around it.

The "crosses" principle: cognitive dissonance drives virality

• Music creates emotional connection; using only famous tracks surrenders control over what emotion the audience feels.
• Crosses are combinations where two elements are polar opposites — a mismatch that creates cognitive dissonance and compels attention.
• Example: pairing Sweden's most commercially unpalatable cross-country skier with the most engineered-for-mass-appeal TV format produced the most popular show in Swedish TV history, still running a decade later.
• The same principle applies to music: a visual story told against a contradictory audio story generates the engagement spike associated with virality.
• Post-Covid data showed a macro-level comfort food effect — classical, warm, reassuring music spiked globally because audiences were emotionally depleted.

AI tooling: the platform moment

• Epidemic has moved from feature (legal innovation) → product (soundtracking platform) → platform (Antidote), following the same arc as Amazon's internal infrastructure becoming AWS.
• Six months before this interview, Epidemic recognised it held more proprietary data on how music travels online than any other player — making it the right moment to open that stack externally.
• Existing AI releases: AI Voice, and a track-length adaptation tool that lets editors fit a track to a finished edit rather than re-cutting the edit to fit the track.
• Roadmap: personalised music supervision at scale — the system learns a creator's channel, style, and peer set, then serves targeted recommendations.
• Target: give solo creators access to the same music supervision tooling previously only available to large production budgets.
• The biggest near-term opportunity identified: vibe coding. Currently ~99% of people cannot code; vibe coding expands that population, but everything being built is muted — Epidemic wants to be the default music layer for that world.

Value chain analysis: the leverage framework

• Facing 5,000 freelance editors in Sweden with no realistic way to reach them, the team stepped up the value chain.
• 5,000 editors → 50 production companies → 4 major broadcasters: the broadcasters were reachable.
• Closed deals with all four broadcasters; the broadcasters then summoned the production companies, who summoned the editors — three layers solved by four meetings.
• The framework: map who depends on whom, find the smallest set of decision-makers whose approval cascades down, target them with deep empathy for their specific pain points.
• Leverage = small amount of input producing a huge amount of output; understanding the value chain is how you find the leverage point.

Feature → product → platform

• Epidemic launched as a feature: a one-trick vitamin. Nice to have, but forgettable if skipped.
• The legal innovation (full rights ownership, artist paid handsomely upfront, complete indemnification for users) created instant product-market fit.
• Became a product by layering software on top: deep tooling, 70% of the world's largest YouTubers using the platform.
• Became a platform (Antidote) by opening the proprietary data stack to artists, storytellers, and IP holders — connecting supply and demand through infrastructure others can build on.

Four components of success

• Talent: least important; just a ticket to play. Matters only relative to your competitor set, not in absolute terms.
• Grit / relentlessness: more important than talent because it is chosen. Expect 70–80 hour weeks for years before lift-off.
• Being a net giver: the most overlooked component. Tastemakers, gatekeepers, and facilitators are everywhere and mostly invisible — you never know who will have the chance to help or block you. Conduct yourself so others want you to succeed.
  • Practical method: warm introductions, done rarely (~once a week) but with full research — two paragraphs that humanise both parties with a serious note and a fun fact.
  • Default opener with new contacts: "How can I help?" — primes the relationship before asking anything.
• Discipline: the unsexy answer. Structures that protect non-negotiable priorities create the long-term sustainability that hustle alone destroys.

Discipline as the operating system

• During Epidemic's first decade: no meetings before 9 a.m. (school run was non-negotiable); work hard 9–6; phone off 6–9 p.m. for family dinner and bedtime; second work window 9 p.m.–1 a.m. for US/Asia calls.
• No weekend work. Proper summer holidays protected deliberately.
• The crisis: after a year of near-continuous travel, his wife said "our life is better when you're not here." He called his co-founder the next day to quit.
• That moment forced a restructuring of his role and the business. He stayed, but only by making structural changes — not motivational ones.
• The guardrails were not about working less; they were about making the hours outside work genuinely unreachable, so those hours counted.

The "why three times" analysis framework

• From BCG: an analysis only qualifies as analysis when you have asked why at least three times.
• Surface observation (revenue not growing) → first why (costs up) → second why (unit costs up, not volume) → third why (supplier cartel formed after bankruptcy and merger) → actionable solution (run a tender, introduce competition).
• In the AI era: artificial intelligence is abundant; the scarce resource is rigorous root-cause thinking.
• Framework for AI-assisted building: ask why three times to locate the real problem or opportunity, then deploy AI tools to answer and act on those questions fast.
• Rushing to an idea without this depth produces solutions to the wrong problem.

Hiring above yourself

• Only hire people smarter than you; if you are the smartest person in the room, the company is in trouble.
• Epidemic's executive team includes the person who built the original Amazon app and a former co-CEO of Klarna.
• The same "crosses" logic applies to team building: pair people who seem incongruous but whose combination creates something more powerful than either alone.

AI tools currently in use

• Primary tools: Gemini and ChatGPT — used consistently enough that the models now understand the user's context and return increasingly useful output.
• The value of AI tools compounds with use: the longer you interact with a specific model, the more calibrated its suggestions become.
• No single app is transformative alone; the discipline of using a small set deeply outperforms sampling many tools shallowly.