Table of contents
- The 150 threshold: when informal systems fail
- The social brain hypothesis: where does Dunbar's number come from?
- Dunbar's concentric circles: a fractal geometry
- Channel proliferation: the mathematics of disconnection
- Model limitations: biological determinism versus managerial engineering
- How to structure growing companies without bureaucracy
- In conclusion: making scale a deliberate choice
- Structuring human scale with FluidOps
- Frequently asked questions
The 150 threshold: when informal systems fail
In the early 1970s, Wilbert "Bill" Gore, the engineer and entrepreneur behind W. L. Gore & Associates, noticed a striking empirical pattern. As long as a manufacturing facility employed fewer than a hundred people, strategic alignment occurred almost through osmosis. Everyone knew their colleagues by name, understood the machinery they operated, tracked order statuses, and recognized each worker's level of dedication. Frictions were settled organically by the coffee machine, without executive memos or rigid approval chains.
Then, almost imperceptibly, as headcounts crossed the 150-employee threshold, the workplace dynamic shifted. Silos emerged, anonymous complaints replaced direct conversations, absenteeism climbed, and workers began demanding formal employee handbooks, detailed job descriptions, and supervisory layers. To curb this bureaucratic drift, Bill Gore instituted a radical operational rule: whenever a facility approached 150 employees—signaled quite tangibly by a parking lot filled with 150 vehicles—the company stopped expanding the building. Instead, it purchased adjacent land and built an independent, identical sister plant (Gladwell, 2000).
This dynamic was neither an industrial quirk nor a reality unique to manufacturing. It illustrates a universal neurobiological constraint that leaders of growing businesses and scale-ups routinely encounter. As your organization expands, an unfamiliar reality sets in: you pass unfamiliar faces in the hallways, decisions stall in endless sign-off loops, and spontaneous trust is replaced by procedural oversight. You have run straight into Dunbar’s number.
The social brain hypothesis: where does Dunbar's number come from?
To grasp the origin of this invisible barrier, one must look back to 1992 and the foundational research of Robin Dunbar, an evolutionary anthropologist and biologist at University College London. Studying non-human primates, Dunbar noticed that the time spent on social grooming—a vital mechanism for forging alliances and resolving conflict—scaled in direct proportion to troop size. However, keeping track of these shifting social alliances requires substantial cognitive processing power (Dunbar, 1992).
The neocortex equation and anthropological prediction
Dunbar formulated what became known as the "Social Brain Hypothesis." He identified a strong statistical correlation between the relative volume of the neocortex—the brain region responsible for executive function and conscious deliberation—and average social group size across 38 primate genera. By applying the dimensions of the modern Homo sapiens neocortex to his regression formula, Dunbar made a remarkable theoretical prediction: the maximum sustainable social group size for humans is approximately 147.8 individuals, with a 95% statistical confidence interval ranging from 100 to 230 people (Dunbar, 1992). This figure was swiftly rounded in common usage to 150.
In subsequent papers, Dunbar demonstrated that articulated language eventually replaced physical grooming: conversing, sharing updates, and storytelling allow individuals to maintain bonds with multiple peers simultaneously, drastically optimizing social time budgets (Dunbar, 1993, 1996). Yet even with the efficiency of speech, our neocortex faces hard biological limits. Beyond 150 people, we lack the working cognitive memory needed to track who is reliable, who holds a grudge, and what social obligations exist between any two peers.
Historical and anthropological convergence
Far from being a sterile laboratory abstraction, the 150-person ceiling recurs across human history with remarkable consistency:
- Traditional hunter-gatherer societies: Anthropological studies across Africa, Amazonia, and Australia indicate that while nomadic hunting bands often count around thirty adults, extended clans that share survival rituals and mutual defense almost invariably hover between 140 and 150 members (Dunbar, 1993).
- Hutterite agricultural colonies: For centuries across North America, these pacifist Anabaptist farming communities have adhered to an unyielding rule: whenever a colony surpasses 150 individuals, it splits to establish a new farm. Elders observed that below this threshold, work ethics depend on peer accountability and fraternal solidarity; above it, anonymity takes root, necessitating formal rules, moral policing, and explicit penalties (Dunbar, 1993; Gladwell, 2000).
- Fundamental military units: From the Roman centuria commanded by a centurion to modern infantry companies (typically numbering between 120 and 180 soldiers), military doctrines have historically calibrated autonomous units so that a commanding officer can intimately know the strengths and frailties of every soldier without bureaucratic intermediaries (Dunbar, 1993).
- The holiday greeting card test: Examining postal greeting cards sent in the United Kingdom, researchers found that the median active network mobilized by modern households centers around 153 individuals, including roughly 125 regular direct contacts (Hill & Dunbar, 2003).
Dunbar's concentric circles: a fractal geometry
Dunbar's thesis is often oversimplified into a single static figure. In practice, social relationships do not follow a binary split between confidants and strangers. Instead, they organize into concentric layers of decreasing intimacy. Empirical research across diverse cohorts reveals that these circles scale according to a fractal ratio of approximately 3 (Zhou et al., 2005):
The support clique (~5 people) and the sympathy group (~15 people)
The innermost circle consists of 4 to 5 people. This is your intimate core—the ultimate confidants you turn to during severe personal crises. In an organizational setting, this represents the founding core or a dedicated crisis unit, where communication operates with complete candor and zero latency.
The next layer encompasses roughly 15 individuals: the "sympathy group." This is the upper limit for an operational team capable of profound cohesion and instinctive synchronization. Within this boundary, members understand one another's working styles and spontaneously compensate for temporary performance dips. This explains why agile frameworks consistently advocate for compact development squads, rarely exceeding 8 or 9 contributors.
The band (~50 people) and the community of trust (~150 people)
At approximately 50 people lies the "coherent working group." At this scale, you still know exactly what everyone in the organization does, even if you do not interact with them daily. High-level alignment remains achievable in a single gathering.
Finally, the 150-person tier defines the "tribe": the limit beyond which personal relationships based on direct, unmediated reputation become untenable. Beyond this threshold lie casual acquaintances (roughly 500 people) and recognizable faces (around 1,500 people), where emotional connection evaporates in favor of surface-level recognition (Zhou et al., 2005).
Channel proliferation: the mathematics of disconnection
Why do team dynamics deteriorate so rapidly once these cognitive limits are breached? The answer lies in combinatorics. In a group of n individuals, the number of unique, bidirectional communication channels is calculated as n(n-1)/2 (Brooks, 1975). This relationship does not grow arithmetically; it expands quadratically.
| Group size | Potential communication channels [n(n-1)/2] | Primary coordination model | Critical friction point |
|---|---|---|---|
| 5 people | 10 channels | Informal, direct trust, verbal alignment | Negligible: instantaneous synchronization |
| 15 people | 105 channels | Lightweight rituals, daily stand-ups | Emergence of smaller sub-cliques |
| 50 people | 1,225 channels | Departmental pods, written documentation, formal briefings | Departmental silos and information drift |
| 150 people | 11,175 channels | Formalized processes, shared tooling, middle management | Cognitive overload: impossible to know all members |
| 300 people | 44,850 channels | Strict bureaucracy, steering committees, contractual governance | Anonymity, social loafing, employee disengagement |
Comparing a team of 15 people (105 channels) to an organization of 150 employees (11,175 channels) reveals a 10x increase in headcount alongside a more than 100x explosion in relational complexity. This flood of potential interactions overwhelms human attention. As computer scientist Frederick Brooks highlighted in his seminal analysis of software projects, coordination overhead inevitably devours the energy meant for execution (Brooks, 1975). For a deeper look into this coordination trap, explore our analysis of Brooks' Law and the pitfalls of late-stage hiring.
The collapse of personal proximity also triggers an insidious social dynamic: as accountability disperses across a crowd of unfamiliar faces, personal ownership declines. This collective drift is thoroughly documented as the Ringelmann effect and social loafing in teams. Modern digital networks offer no escape from this biological baseline: a large-scale computational study analyzing 1.7 million Twitter users proved that regardless of total follower counts, the volume of sustained, bidirectional interactions reliably plateaus between 100 and 200 contacts (Gonçalves, Perra & Vespignani, 2011).
Model limitations: biological determinism versus managerial engineering
While compelling and intuitively resonant, Dunbar's number should not be treated as an immutable absolute. Several prominent researchers have highlighted methodological weaknesses when directly transposing non-human primate metrics to modern knowledge work.
The statistical critique from Stockholm
In 2021, a research team at Stockholm University led by Patrik Lindenfors re-evaluated the original datasets using updated Bayesian phylogenetic methods. Their findings urged caution: depending on the statistical models applied, the point estimate for human group size fluctuated wildly (between 16 and 109 individuals), but more importantly, generated 95% confidence intervals ranging from 2 to 520 individuals (Lindenfors, Wartel & Lind, 2021). According to these authors, deriving a precise, rigid limit for human cooperation solely from primate cranial ratios represents an oversimplification of complex evolutionary data.
Furthermore, sociologists deploying alternative network estimation tools (such as the scale-up method) have suggested that the average social network of a modern adult hovers closer to 290 identifiable acquaintances (McCarty et al., 2001). The result hinges largely on how one defines a "relationship"—whether it denotes a close friend, a colleague seen once a month, or an external contractor.
Cultural ingenuity versus biological constraints
The primary flaw in applying Dunbar's number mechanically to organizational leadership lies in underestimating human structural ingenuity. Unlike baboons or chimpanzees, humans do not rely exclusively on neurological bandwidth to collaborate. We create cultural artifacts and operating systems designed specifically to offset cognitive limits:
- Explicit roles and clear mandates: You do not need to understand a finance director's personality quirks to process an invoice; you only need clarity on their domain boundaries and documented procedures.
- Governance rituals: Transparent operating practices allow organizations of 500 people to navigate changes smoothly without requiring every individual to deduce colleagues' personal intentions.
- Organizational architecture: As articulated by Conway's Law on organizational structure, the way you architect your teams shapes the structure of what you produce. With modular boundaries, large organizations can scale without hitting cognitive burnout.
- The strength of weak ties: High-performing companies do not require universal fraternity; they need clear, reliable information bridges between specialized groups.
How to structure growing companies without bureaucracy
For founders and project portfolio leaders, the takeaway from Dunbar's work is not to artificially halt growth, but to consciously design an operating model that respects human physiological limits. Here are four guardrails to safeguard agility and engagement as you scale:
1. Divide business units before reaching cognitive saturation (50 to 150 people)
Do not wait for commercial or engineering branches to collapse under the weight of endless status meetings. When a department approaches 50 people, begin partitioning it into self-contained functional pods with autonomous operating and budget scope. Embrace a cellular model: rather than managing a centralized 120-person marketing division, build three focused internal agencies dedicated to distinct segments or product lines.
2. Protect compact execution teams (5 to 9 people)
At the tactical execution level, keep teams aligned with Dunbar's sympathy group. Avoid launching strategic initiatives with sprawling 25-person cross-functional committees. Break major initiatives into modular components entrusted to small squads of 5 to 8 people with explicit ownership and deliverable scopes. This mirrors the proven fundamentals of agile frameworks and focused sprint cycles.
3. Replace grapevine noise with visual management
Spoken updates work when you are 12 people; at 60, they breed confusion. To protect your team from drowning in endless email threads and unread meeting notes, make work visible. Implementing Kanban boards to map workflows brings bottlenecks into plain view, eliminating the need for daily administrative syncs whose sole purpose is status reporting.
4. Build an asynchronous, documentation-first culture
Beyond 50 people, synchronous communications (unmanaged chat channels, emergency meetings) quickly become disruptive. Document project scopes, operating principles, and architectural decisions within a central, searchable knowledge repository. This gives team members the autonomy to act decisively without constantly pinging colleagues for procedural guidance.
En conclusion : faire de la taille un choix délibéré
Dunbar's number is not an inevitable ceiling forcing growing businesses to remain under 150 people. It serves as an early warning indicator on your operational dashboard: it signals that the informal social glue that drove your early momentum is evaporating under the pressure of mathematical combinatorics.
Attempting to run a 200-person organization with the conversational, centralized reflexes of a 15-person startup is a direct recipe for disengagement and executive burnout. Sustainable growth demands working with human biology rather than against it: maintaining small execution units, decoupling departments before they stagnate, and supporting teams with transparent visual systems that make trust observable and shared.
Structuring human scale with FluidOps
FluidOps helps growing businesses and distributed teams maintain operational clarity without adding procedural weight. From our free tier, you get access to a dynamic Kanban board to track task progress, an interactive Gantt chart to manage project timelines, and an intuitive Scrum view designed for sprint execution. Eisenhower matrices and integrated productivity frameworks are available at no additional cost to keep priorities sharp, complete with bidirectional calendar synchronization for Google Calendar and Outlook.
Our plans adapt to every stage of your company's growth: start risk-free on our Free plan at €0 per month. The Individual plan is available at €12 per month or €120 per year, while our Enterprise plan begins at €12 per seat per month on annual billing. Review the full breakdown of features on our pricing page.
Frequently asked questions
What is Dunbar's number exactly?
Dunbar's number is a theoretical cognitive threshold of approximately 150 individuals with whom a human can sustain stable, reciprocal social relationships. Formulated in 1992 by evolutionary anthropologist Robin Dunbar, it is derived from the mathematical correlation between neocortex volume and social group size in primates.
Why does company cohesion deteriorate beyond 150 employees?
Past 150 members, our neocortex can no longer naturally track the individual reputations, roles, and allegiances of every colleague, diluting organic trust. Anonymity takes root, necessitating formal administrative workflows, middle-management hierarchies, and written regulations to maintain operational coordination.
What is the optimal size for an operational project team?
Based on Dunbar's fractal layers and operational observations, high-performing tactical teams operate best with 5 to 9 individuals. This corresponds to the sympathy group, where team members communicate with minimal friction and naturally compensate for workload variations.
Can modern communication tools and social platforms bypass Dunbar's number?
No. Data from platforms such as Twitter show that despite having thousands of virtual followers, the number of stable, reciprocal relationships an individual maintains remains capped between 100 and 200 contacts. Technology improves information accessibility, but it does not increase the emotional processing bandwidth of the human brain.
How can a growing business scale past 150 people without becoming bureaucratic?
The most effective approach is adopting a cellular organizational model, dividing the company into semi-autonomous operational pods of 50 to 100 people. Combining this structure with an asynchronous, documentation-first culture and transparent visual task boards keeps cognitive overhead low while preserving agility.
Sources
- Brooks, F. P. (1975). The Mythical Man-Month: Essays on Software Engineering. Addison-Wesley.
- Dunbar, R. I. M. (1992). Neocortex size as a constraint on group size in primates. Journal of Human Evolution, 22(6), 469–493. doi:10.1016/0047-2484(92)90081-J
- Dunbar, R. I. M. (1993). Coevolution of neocortical size, group size and language in humans. Behavioral and Brain Sciences, 16(4), 681–694. doi:10.1017/S0140525X00032325
- Dunbar, R. I. M. (1996). Grooming, gossip, and the evolution of language. Harvard University Press.
- Gladwell, M. (2000). The tipping point: How little things can make a big difference. Little, Brown and Company.
- Gonçalves, B., Perra, N., & Vespignani, A. (2011). Modeling users' activity on Twitter networks: Validation of Dunbar's number. PLOS ONE, 6(8), e22656. doi:10.1371/journal.pone.0022656
- Hill, R. A., & Dunbar, R. I. M. (2003). Social network size in humans. Human Nature, 14(1), 53–72. doi:10.1007/s12110-003-1016-y
- Lindenfors, P., Wartel, A., & Lind, J. (2021). ‘Dunbar's number’ deconstructed. Biology Letters, 17(5), Article 20210158. doi:10.1098/rsbl.2021.0158
- McCarty, C., Killworth, P. D., Bernard, H. R., Johnsen, E. C., & Shelley, G. A. (2001). Comparing two methods for estimating network size. Human Organization, 60(1), 28–39. doi:10.17730/humo.60.1.efx5t9gjtgmga73y
- Zhou, W.-X., Sornette, D., Hill, R. A., & Dunbar, R. I. M. (2005). Discrete hierarchical organization of social group sizes. Proceedings of the Royal Society B: Biological Sciences, 272(1561), 439–444. doi:10.1098/rspb.2004.2970