Oscar Hoglund on Value Chains, AI Tools, and Building Epidemic Sound Without Losing Your Life

Marina Mogilko introduces Oscar Hoglund, cofounder and CEO of Epidemic Sound, with the observation that the company’s music sits behind billions of daily video views. She asks whether that perspective reveals patterns in content that breaks through. Hoglund starts with the purpose behind the metrics: creators are trying to make meaningful human connections, and music helps them create the emotion that supports those connections.

Mogilko asks whether a familiar song is a reliable way to achieve that effect. Hoglund argues that commercial familiarity can also reduce a storyteller’s control. Listeners bring their own memories of a famous track, so the emotion a creator intends may differ from the one an audience experiences. Choosing music means considering that existing emotional baggage as well as the song itself.

He recalls a post-pandemic pattern in which listeners gravitated toward reassuring, warm, calming music. Classical music rose within that trend. He compares it with comfort food: when the world feels low, people look for something that helps them feel that things will be all right. The choice of soundtrack reflects the audience’s emotional circumstances, not only a platform trend.

When Mogilko asks about creators whose results changed after adding music, Hoglund says striking connections between audio and visuals matter. Sometimes the two create strong harmony; sometimes the attraction is a deliberate mismatch. What stands out is a relationship the audience finds interesting, rather than the simple presence of a track beneath the footage.

He calls those surprising combinations crosses and tells a story about a cofounder’s lunch with a tabloid editor. A famous athlete talking about sporting success is predictable. The same athlete speaking about addiction creates a different kind of attention because it contradicts the expected public story. The editor’s lesson was to notice combinations that produce this cognitive dissonance.

Years later, the cofounder was working in commercial television and needed a small broadcaster to break through. The company committed heavily to a Saturday-evening entertainment format designed to appeal across ages and genders. It combined competition, danger, humor, and the possibility of elimination, but it still needed a public face that would make the show distinctive.

The proposed host was a beloved cross-country skier associated with a very different world from commercial entertainment. Hoglund describes the logic as choosing the opposite of what such a highly commercial show would normally suggest. The broadcaster initially resisted, but the cofounder insisted. In his telling, the unexpected pairing helped create an enduring hit.

Music can perform the same role. A creator may tell one story visually and another through sound, using either resonance or contrast to make the result memorable. Hoglund sees this as useful beyond individual videos: businesses and podcasts can also combine elements that would not normally appear together, provided the connection creates a compelling experience.

Mogilko asks whether the company is building tools to help creators decide when music should develop, stop, or change. Hoglund describes an ambition to make professional creative support available to everyone, from a large editing team to a solo entrepreneur. That includes the kinds of recommendations a music supervisor, editor, or sound specialist might otherwise provide.

Historically, assembling that support has been expensive. Hoglund believes Epidemic now has a useful combination of a music catalog, artists, customers, platform knowledge, and performance data. AI can help bring those resources together for an individual creator. He does not see the purpose as replacing human connection, but as making the craft that enables it more accessible.

The company can observe which music has been available, how it has been used, and where engagement and feedback appeared over time. Hoglund wants to return those insights to creators in a practical form: how their own videos perform, what peers use, and which platform patterns might matter. The aim is to provide assistance that once required a much larger budget.

Asked when these capabilities will arrive, he says the work is underway and some tools have already been released. He names AI voice and tools that alter the length of a music track. The latter changes an awkward editing constraint: creators previously often had to reshape a video to fit a piece of music rather than adapt the track to their story.

He describes a creator who has finished an edit and found the right music, except that the track is too long or short. The newer tools let that recording be adapted to the desired duration while working with the artist’s track and intent. The important change is that the story can remain the organizing decision instead of being recut around a timing mismatch.

For a creator or small business seeking a stronger emotional connection, Hoglund recommends first deciding what it wants to sound like, which values it expresses, and which stories it tells. Recommendations become useful once that direction is clear. More interaction with the platform can then help it learn the creator’s preferences and make its suggestions more relevant.

He also encourages creators to be ferocious learners. Studying people whose work produces a desired feeling can be as important as enthusiasm for one’s own story. Those influences give both the creator and a recommendation system more information about the emotional direction. The conversation then moves past a sponsor segment to how a business develops beyond its first useful capability.

Mogilko asks about the progression from a feature to a product and then a platform. Hoglund credits the framework to others and applies it to Epidemic’s history. Its first offering was useful but narrow: a solution customers might appreciate without yet treating it as an indispensable part of their work. That initial feature addressed a problem the founders knew from television production.

The founders expected online media to become increasingly video-centered. They also knew that music rights were unusually difficult to clear, especially for content consumed instantly around the world. Traditional arrangements involved many participants with different interests—artists, producers, songwriters, labels, publishers, and collecting organizations—without one party able to authorize every necessary use.

Hoglund describes competing claims and veto rights that could prevent a recording from being used in a video. Epidemic’s initial innovation was to move from representing fragmented rights toward owning the rights needed for its catalog. He says that required paying artists well upfront and assembling music specifically suited to bringing stories and visual content to life.

With control of the necessary rights, the company could offer creators a much simpler proposition across platforms and territories. Hoglund describes immediate product-market fit, but still calls that first offering a feature. His analogy is a vitamin: useful to take, but something a customer can forget and continue without. The next ambition was to become a necessity in the creative process.

The second phase added software to the rights innovation. Epidemic wanted to understand creators well enough that editing without its tools would become difficult to imagine. Workflow integration, recommendations, and creative support turned the service into a more complete product. Hoglund cites its use by millions of creators and many leading YouTubers as evidence of that deeper role.

The third phase asks how more people can provide and receive value through the infrastructure the company has built. Hoglund compares the moment with Amazon recognizing that its internal systems could serve other businesses. The analogy is about opening a capability developed for one company’s own growth so that other participants can build and scale on top of it.

For Epidemic, that infrastructure includes its data stack and its understanding of how music travels online, creates emotions, and affects creators commercially. Hoglund says the company decided to open more of it through an initiative called Antidote. The platform ambition is to connect storytellers, artists, and holders of intellectual property rather than keep all the infrastructure inside one service.

He then explains why understanding value chains was one of the company’s most consequential decisions. In its early Nordic broadcasting market, Epidemic estimated that Sweden had roughly 5,000 freelance editors. The team could not see a practical way to reach and persuade every editor to switch from the music sources they already used.

Moving one step upstream revealed around 50 production companies responsible for much of the television output. That was a smaller group, but still difficult for a young company to influence. Another step revealed four major broadcasters commissioning work from those production companies and, indirectly, the freelance editors. Four serious meetings became a plausible sales plan.

The company understood the broadcasters’ problems and signed multiyear agreements, despite having a relatively small catalog. The broadcasters then brought the production companies into meetings with Epidemic, and the production companies in turn brought in their editors. Instead of persuading thousands of people independently, the team used the relationships through which work and revenue already flowed.

That is Hoglund’s definition of leverage: a small amount of input produces a large amount of output. The lesson is to understand the ecosystem before choosing the level at which to sell or intervene. An organization that appears to have thousands of separate customers may contain a few relationships that determine how those customers behave.

Mogilko asks whether a similar principle explains Epidemic’s hiring. Hoglund says he wants to hire people smarter than himself so they can tell him what to do. Being the smartest person in the room would be a warning, not a triumph. Different forms of intelligence and experience create the productive combinations he has described elsewhere in the conversation.

The discussion turns to hustle culture after Mogilko describes spending a relaxed weekend with the team. Hoglund resists reducing his view to simply supporting or rejecting hustle. He admires aspects of American culture, studied at Wharton, and values intense learning. He tries to combine useful ideas from different places rather than treat one national style as a complete formula.

He begins his account of entrepreneurial success with talent, but considers it the least important component. Talent is a ticket that allows someone to participate, and the relevant question is whether it creates an edge relative to the competition. It is a starting condition rather than a guarantee that a person will build a successful company or creative career.

The next component is grit. Hoglund likes the word relentless because it reminds him that a serious venture can require years of persistence and sacrifice. He acknowledges having worked seventy- or eighty-hour weeks. The distinction he wants is between sustained effort and the dishonest meaning of hustle: claiming that something is good when it is not simply to sell it.

Persistence matters more to him than innate talent because people can choose how they respond to the work. But he thinks another component receives too little attention: every industry contains tastemakers, facilitators, and gatekeepers. A founder or creator will meet only some of them, while many more people can affect whether an opportunity becomes available.

His conclusion is that people should conduct themselves in a way that makes others want them to succeed. Others repeatedly have chances either to help or to hinder a career. Hoglund’s preferred way to build that support is to be a net provider, giving more value than he takes. He would like that phrase to describe his life as a whole.

Practical helpfulness feels good immediately and can also compound over time. When introduced to someone, he often begins by asking how he can help. Making an introduction, opening a door, or solving a small problem can create an indirect supporter of the person and the company. The future opportunity to reciprocate is uncertain, but the relationship need not depend on predicting it.

He gives introductions as a concrete example. He does not make them constantly, but tries to make them well: obtain both people’s agreement, do some homework, and write a short explanation of each person. Alongside their professional relevance, he includes a human detail that helps them find a point of connection rather than leaving them with a bare email address.

Connecting good people can make the world better and also reflect positively on the person who brought them together. Hoglund treats this as a repeated practice rather than a one-time networking maneuver. Shared experiences and useful actions can build goodwill even when nobody knows exactly how that support will matter later.

After a brief newsletter promotion, Mogilko asks how he stays sane while building in a competitive market. His one-word answer is discipline. During the company’s early years, he recognized that business success was uncertain. He was willing to take that risk, but did not want his relationship with his family to depend on the venture succeeding.

He therefore declined meetings before nine in the morning, even when finance professionals wanted an early slot. He used the time to take his children to school or daycare. During the working day he worked intensely, but he made other people adapt to that morning boundary rather than repeatedly trading it away for the company’s immediate demands.

At six in the evening, his phone stopped working for business purposes, and he remained unreachable until nine. Family dinner, bedtime, and reading to the children were protected. Afterward he worked another shift, sometimes until one in the morning, to overlap with the United States or Asia. He also kept travel to weekdays and protected weekends and proper summer vacations.

The system followed a much darker period. After extensive travel while the children were young, his wife told him that family life had become better when he was away. Hoglund calls that the darkest moment of his life. The next day he told a cofounder he intended to quit, and the company made major changes to the business and his role.

He says he has always put family first and works to live rather than living to work. The same idea of discipline applies to exercise, food, and other habits. It is an unglamorous answer, but he sees consistent boundaries and moderation as necessary conditions for sustaining the effort that a long entrepreneurial journey requires.

Asked which AI applications he uses most, Hoglund names Google’s Gemini tools and ChatGPT. He is investing time in understanding the systems and helping them understand his context so they become more useful. He also sees an opportunity to make music available to people building applications through AI, since much of that new software currently has no sound component.

His final advice to someone starting an AI business is a framework from his consulting years at BCG: ask why three times before calling the work an analysis. He illustrates it with a hypothetical company that has stopped making a profit. The first investigation finds that revenue is stable but costs have risen, establishing that the immediate problem sits on the cost side.

The second why asks what changed within those costs. The company buys the same amount of inputs, but the unit price has risen, so simply reducing the volume is not yet an explanation of the underlying change. The third why investigates the supply market rather than stopping at the observation that materials have become more expensive.

In the example, one of three suppliers failed and the other two merged, leaving a single supplier able to dictate prices. That diagnosis points to a different remedy: invite additional suppliers to bid and restore competition. Hoglund’s point is that the solution becomes clearer only after tracing the symptom through several levels of cause.

AI makes intelligence and execution more available, but it does not remove the need for curiosity, rigor, and discipline in choosing a problem. Hoglund recommends first understanding why something works or fails, then using AI to investigate and build around that cause. Mogilko closes by reinforcing the distinction: dig deeper before rushing into an idea, and keep asking why.