How Spotify Scaled Engineering Across 4 Countries — Through Dowling’s IHRM Lens

When a Stockholm startup opens an engineering hub in New York, the question is not whether Swedish management culture will survive the Atlantic crossing. The question is whether anyone planned for what happens when it doesn't.

In the spring of 2013, Spotify had a problem that most startups would envy. The Swedish music streaming company -- founded by Daniel Ek and Martin Lorentzon in Stockholm in 2006 -- had grown to roughly 1,200 employees and was scaling at a rate the Swedish labor market could not sustain. Stockholm, for all its strengths as a tech hub (Klarna, King, iZettle), has a metropolitan population of about 2.4 million. The pool of senior machine learning engineers, infrastructure architects, and product leaders with experience scaling consumer products to hundreds of millions of users was finite. Those people lived in New York, San Francisco, London, and Tokyo -- not Kungsholmen.

The company had already opened a small New York office focused on advertising and business development. But in 2013-2014, Spotify made a decisive strategic move: it would build a full engineering hub in New York City. Not a satellite support team. Not a liaison office. A second center of gravity for R&D that would own product development.

The stakes were enormous. Spotify's entire engineering culture was built on a model that had become famous in the agile community: squads, tribes, chapters, and guilds. Henrik Kniberg and Anders Ivarsson had published their influential "Scaling Agile @ Spotify" paper in 2012, and the model was already being imitated by companies from ING to Telstra. Squads were small, autonomous cross-functional teams of 6-12 people. Tribes were collections of squads working in related areas, capped at roughly 150 people (the Dunbar number). Chapters were functional groupings -- all backend engineers across squads in a tribe -- and guilds were company-wide communities of interest.

This model had been designed by Swedes, in Sweden, for a Swedish engineering culture. It assumed high trust, low hierarchy, intrinsic motivation, and a consensus-oriented approach to decision-making. It assumed that a squad could self-organize without a traditional project manager. It assumed that engineers would speak up when they disagreed with a direction, even to senior leaders. It assumed, in other words, a set of cultural norms that Sweden's national culture supplied naturally and that other countries did not.

Over the next decade, Spotify would expand to engineering offices in Stockholm, Gothenburg, New York, Boston, London, and Tokyo, among others. It would grow from 1,200 employees to nearly 10,000 by 2022. It would go public via a direct listing on the NYSE in April 2018 -- characteristically unconventional, skipping the traditional IPO process. Then, in 2023, CEO Daniel Ek would announce three rounds of layoffs totaling approximately 2,300 jobs, the largest of which cut 17% of the global workforce in December.

This arc -- from 1,200 in one city to nearly 10,000 across four continents and back down to 7,000 -- is one of the most instructive case studies in international staffing of the past decade. And it maps with surprising precision onto a framework published in a textbook before Spotify streamed its first song.

In International Human Resource Management (8th edition, 2017), Peter Dowling, Marion Festing, and Allen Engle lay out the foundational model for how multinational companies approach international staffing. The framework identifies four orientations, each with distinct assumptions about where talent should come from and who should lead:

Ethnocentric: The home country knows best. Key leadership and technical positions abroad are filled by parent-country nationals -- people from headquarters. The assumption is that HQ culture, processes, and standards must be maintained, and the most reliable way to do that is to send your own people. This approach ensures tight control and cultural consistency but limits local responsiveness, breeds resentment among host-country talent, and is expensive (expatriate packages typically cost 2-3x local hire compensation).

Polycentric: The locals know best. Host-country nationals manage local operations with significant autonomy. Subsidiaries develop their own cultures, management styles, and operating rhythms. This approach is cheaper, builds local goodwill, and leverages deep market knowledge -- but it can create isolated fiefdoms with weak coordination, inconsistent standards, and limited knowledge transfer back to headquarters.

Geocentric: The best person for the job, regardless of passport. Staffing decisions are based on competence, not nationality. Leadership roles are filled by the most qualified candidate whether they come from Stockholm, New York, Mumbai, or Sao Paulo. This is the orientation most multinationals aspire to and the one fewest achieve. It requires sophisticated talent identification systems, willingness to invest in global mobility, and -- critically -- a corporate culture that genuinely does not privilege one national culture over others.

Regiocentric: A pragmatic middle ground. Talent moves within defined regions (within Europe, within Asia-Pacific) but not necessarily across the entire organization. Regional hubs develop their own leadership cadres, and cross-border mobility happens within geographic clusters rather than globally. This balances the geocentric ideal against the practical barriers of language, time zones, visa regulations, and cultural proximity.

Dowling, Festing, and Engle make a critical observation that practitioners frequently miss: most companies do not choose one orientation and stick with it. They evolve through orientations as they mature internationally, and they often operate with different orientations in different functions simultaneously. A company might be geocentric in engineering but ethnocentric in finance. The staffing approach is not a corporate policy. It is an emergent pattern shaped by stage, culture, competitive pressure, and the specific talent markets a company enters.

That observation is exactly what makes Spotify's case so instructive.

Spotify's early years were unmistakably ethnocentric -- not by deliberate policy, but by the organic logic of a company built in one city by a culturally homogeneous founding team. The company's management culture reflected Swedish workplace norms with remarkable fidelity.

Swedish management culture has distinctive characteristics that cross-cultural researchers have mapped extensively. Erin Meyer, in The Culture Map, places Sweden at the extreme consensual end of the Deciding scale and in the egalitarian zone on the Leading scale. Richard Gesteland's Cross-Cultural Business Behavior classifies Scandinavian business culture as moderately Deal-Focused, Informal, and Reserved. These combine to produce workplaces where hierarchy is flat, first names are universal, conflict is handled through patient deliberation rather than confrontation, and there is a deep cultural expectation of lagom -- a Swedish concept meaning "just the right amount" or "not too much, not too little."

Lagom is not merely a word. It is a cultural operating system. In a lagom workplace, no one should stand out too much. Decisions should feel balanced. Credit is shared. Self-promotion is uncomfortable. The ideal colleague contributes expertise without dominating the room.

Daniel Ek embodied this style. Despite being the CEO and co-founder of one of the world's most valuable tech startups, Ek was known for a low-key, deliberate leadership presence. He did not deliver the performative keynotes associated with American tech CEOs. He led by consensus, trusted engineering teams with radical autonomy, and resisted the command-and-control instincts that characterized many fast-scaling Silicon Valley companies.

The squad model was lagom encoded as organizational architecture. Autonomous squads meant no single leader dominated. Consensus within the squad meant decisions felt balanced. The absence of traditional project managers meant individual contributors were trusted to self-organize. This worked beautifully in Stockholm, where cultural norms aligned perfectly with organizational design.

But when Spotify sent its first engineering managers to New York -- and they were, overwhelmingly, Swedish nationals or people who had been deeply acculturated in the Stockholm office -- Dowling's framework would classify the move as textbook ethnocentrism. Key positions abroad, filled by parent-country nationals carrying HQ culture in their operating assumptions.

This was not malicious. It was natural. When you need to open an office fast and maintain cultural consistency, you send people who already know how things work. But it set a pattern: Spotify's international offices would be led by people who had internalized Swedish norms, and those norms would become the invisible baseline against which all other approaches were measured.

The New York engineering hub was where the ethnocentric model first cracked under pressure.

American engineering culture -- particularly New York tech culture in the mid-2010s -- operated on fundamentally different assumptions. Where Swedish culture valued consensus, American culture valued speed. Where lagom discouraged individual prominence, American engineering culture rewarded visible impact. Where Swedish engineers were comfortable with ambiguity and trusted the group process, American engineers recruited from Google, Facebook, and Amazon expected clear ownership, explicit career ladders, and visible recognition for individual contributions.

The friction showed up in specific, predictable ways.

Decision-making speed. In Stockholm, a squad might spend days or weeks reaching consensus on a technical direction. The Swedish engineers were comfortable with this -- the process felt thorough and respectful. American engineers, accustomed to a "disagree and commit" culture where a tech lead makes a call and the team executes, found the consensus process agonizingly slow. They interpreted it not as thoroughness but as indecisiveness. Meyer's Culture Map illuminates this precisely: on the Deciding scale, Sweden sits firmly in the consensual zone (decisions made by group agreement, take longer, but execute fast because everyone is aligned). American tech culture sits in the top-down zone (a leader calls it, the team executes, course corrections happen in real time). Neither is superior. But importing one into the other without adaptation creates organizational whiplash.

Career progression signals. Swedish Spotify had minimal formal hierarchy in its early years. There was no elaborate leveling system. Compensation was relatively compressed by American standards, reflecting Sweden's broader social norms around income equality. American engineers, coming from companies with clearly defined L3-through-L8 levels, stock refreshes tied to performance ratings, and promotion cycles as cultural rituals, felt unmoored. Without explicit signals of progression, they could not tell if they were succeeding.

The *lagom* problem. An American engineer who shipped a major feature and wanted recognition -- in a team meeting, in a performance review, in a promotion -- was operating within normal American professional norms. In a lagom culture, that same behavior registered as self-promotion, which was uncomfortable for Swedish colleagues and managers. The reverse was equally corrosive: Swedish engineers who quietly did excellent work without advocating for themselves were overlooked by American managers who interpreted silence as mediocrity.

Spotify's response was instructive -- and, through Dowling's lens, represented a genuine shift. Rather than forcing the Swedish model on NYC, the company began allowing and eventually encouraging local adaptation. The New York office developed its own subculture: slightly more hierarchical, faster-paced, with more explicit performance feedback and career laddering. Swedish engineering leaders who had been sent to establish the office gradually handed off leadership to locally hired Americans.

Spotify's CHRO, Katarina Berg, who joined in 2013 and became one of the longest-tenured executives in the company, advocated for what she called "being the same but different" -- maintaining core values while allowing regional interpretation of how those values were lived. In Dowling's terms, this was a shift from ethnocentric toward polycentric: host-country norms began shaping how the global model was implemented locally.

But the tension was real. Stockholm-based leaders sometimes perceived the NYC office as "not doing Spotify right." New York-based engineers sometimes felt like second-class citizens in a Swedish company. Cross-office collaboration required navigating not just time zones but fundamentally different assumptions about how work should happen.

Spotify launched in Japan in 2016, and Tokyo became the ultimate stress test for the staffing model.

Japan occupies the opposite end of nearly every cultural dimension from Sweden. Gesteland classifies Japanese business culture as Relationship-Focused, Formal, and Reserved. Meyer places Japan at the extreme high-context end of the Communicating scale, the indirect end of the Evaluating scale, and the hierarchical end of the Leading scale. And while Japan is also consensual on the Deciding scale, Japanese consensus (nemawashi -- the practice of building agreement through careful behind-the-scenes individual consultations) operates through fundamentally different social mechanics than Swedish consensus. In Sweden, consensus emerges through open group discussion. In Japan, consensus is manufactured privately before the group convenes, and the meeting is a formalization of what has already been decided in the hallways.

Spotify's open, real-time squad discussions mapped onto neither model. They were too informal for Japanese hierarchy norms and too public for Japanese consensus-building norms.

Recruiting in Tokyo exposed the first cracks. Senior engineering talent in Japan expected formal titles, clear reporting lines, and structured career progression. Spotify's flat org chart, where a "squad lead" had influence but no formal authority, was legible in Stockholm and tolerable in New York. In Tokyo, it was confusing. Candidates asked questions Spotify's recruiters had not encountered: "Who will be my direct superior?" "What is my rank?" "How is promotion decided?"

Cultural integration was harder still. The radical transparency that worked in Sweden -- where anyone could see any team's OKRs, roadmaps, and performance data -- collided with Japanese norms around information sharing and face-saving. A Japanese engineer receiving blunt feedback in a public Slack channel experienced it very differently from a Swedish engineer. The concept of mentsu (face) meant that public critique, however well-intentioned, could register as humiliation. And in a culture where silence signals disagreement rather than consent, the standard Spotify practice of interpreting a lack of objections as approval produced systematically wrong readings of team sentiment.

Dowling's framework predicts exactly this collision. Ethnocentric staffing in culturally distant markets fails because the exported culture carries assumptions that do not translate. But geocentric staffing also fails when "global culture" is actually the home-country culture wearing a universal disguise. What Tokyo needed -- and what Spotify eventually moved toward -- was a genuinely polycentric approach: local leaders empowered to adapt the organizational model to Japanese norms while maintaining strategic alignment with Stockholm through shared goals rather than shared processes.

After the 2018 direct listing on the NYSE, Spotify continued expanding aggressively. The company grew from roughly 4,400 employees at IPO to 6,500 by 2020 and nearly 9,800 by late 2022. Engineering centers operated in Stockholm, Gothenburg, New York, Boston, London, and other locations. Podcast acquisitions (Gimlet, Anchor, Parcast, The Ringer) added hundreds of employees, primarily in the United States.

During this period, Spotify's staffing model evolved toward what Dowling would characterize as a blend of regiocentric and geocentric. European offices -- Stockholm, Gothenburg, London -- shared talent pools and management practices more easily with each other than with New York or Tokyo. The cultural distance between Stockholm and London was manageable. Between Stockholm and Tokyo, it required constant bridging.

The leadership team itself became more internationally diverse. While Gustav Soderstrom (Swedish) led R&D and the core product/engineering function remained Stockholm-anchored, the C-suite and VP ranks increasingly included Americans, Brits, and other non-Swedish leaders. Chief Content Officer Dawn Ostroff (American) oversaw the massive podcast expansion. The company introduced more formalized HR infrastructure -- structured performance reviews, leveling frameworks, compensation bands -- partly in response to the expectations of a large American workforce accustomed to these systems.

Spotify's "Work From Anywhere" policy, announced in 2021, was the geocentric aspiration made explicit: employees could work from any country where Spotify had a legal entity. The policy reflected a genuine belief that geography should not constrain talent. But it also dramatically increased the coordination costs of the squad model, as teams that were once distributed across two or three offices became distributed across dozens of locations and time zones.

Internal surveys referenced in Spotify's culture blog posts and HR communications revealed a persistent pattern: employees in non-Stockholm offices consistently rated "connection to company culture" lower than Stockholm-based employees. Dowling, Festing, and Engle would recognize this as the "HQ-subsidiary gap" -- the predictable result of a company whose cultural center of gravity remains in one city even as its headcount globalizes.

In January 2023, Spotify laid off approximately 600 employees -- about 6% of its workforce. In June, another 200 positions were cut, primarily in podcasting. Then in December 2023 came the largest reduction: roughly 1,500 employees, approximately 17% of the company. Daniel Ek described the necessity in a public memo: "Over the last two years, we've put significant effort into reducing costs and making our business more efficient. However, we still have too many people dedicated to supporting work and even doing work around the work rather than contributing to opportunities with real impact."

CHRO Katarina Berg addressed the organizational dimension publicly, acknowledging that Spotify had "over-hired during the growth period" and that the distributed model -- all those squads and tribes and chapters -- had created coordination overhead that was no longer sustainable.

The 2023 layoffs are revealing when read through Dowling's framework.

First, the restructuring re-centralized decision-making authority. Ek personally drove the layoff decisions and took public responsibility in a tone notably more direct and decisive than the consensus-oriented style associated with Swedish leadership. This was pragmatic -- you cannot build consensus around who to lay off -- but it signaled a shift in leadership style that had broader organizational implications.

Second, the restructuring disproportionately affected roles and functions created during the rapid expansion of 2020-2022, many of which were in non-Stockholm offices. While Spotify did not publish a geographic breakdown, reporting from multiple outlets indicated that the New York and London offices were significantly impacted. The podcast division, which had been heavily US-based, was hit hardest.

Third, post-layoff, Spotify consolidated its organizational structure. The company moved toward fewer, larger teams and reduced the number of autonomous squads. This was a partial retreat from the squad model's original design -- a model that depended on organizational slack (redundancy, overlapping mandates) to enable autonomy. When headcount drops 17%, there is no slack left. Autonomy gives way to coordination.

Dowling's framework predicts this pattern. When multinationals face financial pressure, they re-centralize: pulling decision-making authority back to headquarters, reducing local autonomy, reverting from geocentric aspirations toward ethnocentric control. By late 2024, Spotify had stabilized at roughly 7,000 employees and reported its first annual operating profit in company history. But the organizational model that emerged from contraction was leaner, more centralized, and -- through Dowling's lens -- more ethnocentric than the company's peak geocentric ambitions.

• R&D expenses in 2022: approximately EUR 1.8 billion (~16% of EUR 11.7B total revenue)
• Gross margins improved from ~25% in 2022 to ~28% in late 2023, partly due to workforce reductions
• December 2023 severance charges: estimated EUR 130-145 million
• First annual operating profit in 2024: approximately EUR 1.1 billion, a dramatic turnaround from years of losses

Dowling's model is typically presented as four parallel options: choose your orientation. Spotify's experience reveals something more useful for practitioners: for scaling companies, the four orientations are not choices. They are phases. And every growing company passes through them in a roughly predictable sequence. The challenge is not picking the right orientation. It is managing the transitions without losing talent, coherence, or cultural identity.

Stage 1: Ethnocentric Launch (first international office, 0-50 international employees)

Your first international hires will be acculturated HQ employees sent abroad, or carefully selected locals who are culturally aligned with headquarters. This is fine. It is arguably necessary. You are establishing a beachhead, not building a democracy. The risk at this stage is not ethnocentrism -- it is failing to recognize that you are being ethnocentric. That failure prevents you from planning for the transition ahead.

Spotify's version: Swedish engineering leaders sent to NYC in 2013-2014 to establish the squad model. Effective for speed. Set the ethnocentric pattern.

Stage 2: Polycentric Collision (50-500 international employees)

Local offices develop their own norms. Host-country hires push back on HQ practices that do not translate. You will feel this as friction: complaints about decision-making speed, feedback style, career progression, or meeting norms. This friction is not a problem. It is a signal that your international footprint has outgrown your HQ-centric operating model. The risk is either crushing local adaptation (retreating to ethnocentrism) or allowing so much local variation that the company fragments (excessive polycentrism).

Spotify's version: NYC develops a faster, more hierarchical subculture. Tokyo finds the squad model confusing. Berg advocates "same but different." Local leaders gradually replace Swedish expats.

Stage 3: Geocentric Aspiration (500+ international employees, deliberate effort)

You begin staffing leadership based on competence rather than nationality. You create global frameworks (leveling, compensation bands, performance systems) that standardize outcomes while allowing cultural variation in implementation. You monitor whether your corporate culture privileges one national group over others. The risk is confusing geocentric rhetoric with geocentric reality -- claiming to hire "the best person regardless of nationality" while maintaining HQ cultural norms as the invisible standard.

Spotify's version: International senior hires, "Work From Anywhere" policy, formalized HR infrastructure. But Swedish cultural norms (squad model, lagom, consensus) remained the default operating system. The aspiration was genuine. The execution was incomplete.

Stage 4: Contraction Reversion (any size, triggered by financial pressure)

When layoffs come -- and they come for every scaling company eventually -- the organization re-centralizes. Authority flows back to headquarters. Local autonomy shrinks. Long-tenured HQ-culture leaders regain influence. This is not a failure. It is a predictable stress response. The risk is treating reversion as permanent, allowing the crisis to undo years of cultural progress.

Spotify's version: 2023 layoffs re-centralize authority around Ek and Stockholm. Squad model consolidated. Geocentric ambition downshifted to regiocentric pragmatism.

For each factor, identify which orientation best matches your current reality -- not your aspiration. If four or more factors cluster in one column, that is likely where you are operating today, whether or not it is where you want to be.

1. What cultural norms are we exporting? Name them specifically. For Spotify: flat hierarchy, consensus decision-making, lagom modesty, radical autonomy, tolerance for ambiguity. These are not universal. They are Swedish.

1. Which are genuinely universal principles, and which are nationally specific practices? "Teams should have autonomy over how they deliver" is a principle. "Squads self-organize without project managers through open group consensus" is a Swedish practice. The principle can travel. The practice must be adapted.

1. Who are we sending first, and what cultural DNA do they carry? If your first five hires in the new office are all HQ transplants, you are planting an ethnocentric seed. That may be intentional. Name it.

1. What do local candidates expect from an employer? Research employment norms in the target country: career progression expectations, feedback styles, decision-making patterns, hierarchy norms, work-life boundaries. Not to accommodate everything, but to know where friction will emerge before it surprises you.

1. What is our 24-month plan for local leadership? If HQ expats will run the office indefinitely, you are ethnocentric. If leadership transfers to local hires within 18-24 months, you are on the polycentric path. If you select leaders from a global pool, you are aiming geocentric. Whatever it is, name it and resource it accordingly.

Spotify's two-decade arc -- from a Stockholm apartment to 10,000 employees across four continents and back to 7,000 -- is not your company's story. But the pattern it reveals is universal enough to be actionable. Here are five things you can do this week, grounded in the adapted framework.

In your next leadership meeting, put Dowling's four categories on a whiteboard. Ask your team: "Which one are we?" Not which one you aspire to. Which one you are right now, based on evidence -- who holds leadership positions, where decisions actually get made, whose cultural norms are treated as the default. Most companies think they are geocentric. Most are ethnocentric with polycentric exceptions. Naming the reality is the precondition for managing the transition.

Spotify's squad model was a brilliant organizational innovation. It was also a Swedish cultural product -- designed for high-trust, low-hierarchy, consensus-comfortable teams. If you are scaling a management model internationally, make a two-column list. Left column: structural elements (roles, reporting lines, meeting cadences, decision-making processes). Right column: cultural assumptions baked into those elements (who speaks up in meetings, how disagreement is expressed, what "good performance" looks like, how credit is distributed). The structural elements can export. The cultural assumptions must be examined, named, and adapted. Spotify's mistake was treating the squad model as a structure when it was also -- inextricably -- a culture.

The single most consequential hiring decision in international expansion is not the first engineer or the first sales rep. It is the first local leader who can bridge cultures -- someone who (a) genuinely understands and values your company's core principles, (b) has deep local networks for recruiting, (c) can translate your culture into locally legible norms, and (d) has the standing and confidence to push back on headquarters when something will not work locally. This person is not an implementer. They are a cultural translator. They are rare, they are expensive, and they are worth every dollar. Spotify's smoothest expansions happened where this person existed. Its roughest happened where this person did not.

Every company that hires aggressively during growth will eventually contract. When it happens, the instinct is to re-centralize -- pulling authority back to the home office. If you have invested years building local leadership and distributed decision-making, a poorly managed contraction can undo that work in months. Before the next downturn, document which decisions should re-centralize under pressure (budget, headcount targets, strategic priorities) and which should stay local regardless (team management, cultural adaptation, client relationships, local compliance). Build the contraction playbook while you still have the luxury of time and headcount.

Most companies budget for salaries, office space, and maybe an EOR or entity setup. They do not budget for: cross-cultural management training for HQ leaders who will work with the new office; compensation benchmarking in the target market (which may differ dramatically from HQ); regular inter-office travel for relationship-building (not optional -- this is how trust is built across cultures, especially in relationship-focused markets like Japan, Brazil, and India); local legal counsel for employment law compliance; and the 6-12 month productivity dip while a new office finds its rhythm. For a 20-person international office, expect these hidden costs to add 15-25% on top of direct compensation. Plan for them now or be surprised later.

Dowling, Festing, and Engle wrote their framework to describe four strategic choices. Spotify's experience reveals that for most companies, it is not a choice at all. It is a lifecycle. You start ethnocentric because you must. You become polycentric because your international offices force you to. You aspire to geocentric because it is the right thing to do and because the best global talent demands it. And when the market turns, you re-centralize and have to fight to preserve whatever geocentric progress you have made.

The companies that navigate this lifecycle best are not the ones that skip phases. They are the ones that name each phase honestly, plan for the transition to the next one, and resist the temptation to confuse their home-country culture with a universal standard.

Spotify built one of the most innovative organizational models in modern business. They exported it across four continents. They learned -- sometimes painfully -- that a model designed for Swedish cultural soil does not transplant automatically into American, British, or Japanese ground. And then they had to make wrenching decisions about what to keep and what to cut when profitability demanded it.

That is not a failure story. That is what international growth actually looks like when you are honest about it. The framework gives you the map. Spotify gives you the terrain. What matters is whether you walk in with your eyes open.