Dunbar's Number and Community Scale — What Changes After 150 Members

What Is the “150-Member Wall”?

Community operators often notice a particular shift at a certain scale. “We used to talk freely as a group, but lately things feel strangely formal.” “Only certain people speak up now.” “Nothing the team does generates the same response as before.”

In most cases this shift is not caused by a drop in operational quality or by members losing interest. The community has crossed a structural threshold tied to its scale.

Anthropologist Robin Dunbar studied the relationship between primate brain size (specifically neocortex volume) and group size, and concluded in his 1992 paper:

“The size of a typical human group is around 150 — the number of individuals with whom stable relationships can be maintained.”

This is known as Dunbar’s Number. The figure of 150 is not a hard value; it is a midpoint in a distribution spanning roughly 100–250. What matters is that this cognitive limit has direct consequences for how a community must be structured.

Up to 150 Members: “Village” Mode — Everyone Knows Everyone

A community of fewer than 150 members operates within human cognitive reach. Participants recognize one another by name and intuitively know who has expertise in what. This is “village mode.”

The defining characteristic is that trust grows from relationships. When someone posts, everyone knows whose voice it is, so replies carry context. New members can find their footing within the existing network of relationships relatively quickly.

Village-mode operation works on human effort alone

At this scale, even without formal systems, the personal commitment of moderators and founders fills the gap.

• Unwritten norms function as an implicit code of conduct
• Handwritten welcome messages to new members remain manageable
• When a disruptive post appears, the community as a whole can handle it

The governance “OS” is essentially personal and informal — and that is sufficient.

Village-mode KPIs

At this stage, intuition often outperforms dashboards.

• Are 10–20% of all members “core members” (posting at least once per week)?
• How long does it take a new member to make their first post? (Under 48 hours is ideal.)
• Can the operator tell who has been less active recently — and reach out?

If you want to quantify, the three-tier KPI framework of leading, intermediate, and lagging indicators applies here too, with leading indicators (core-member engagement) carrying the most weight.

150–1,500 Members: Transition to “Town” Mode

The community does not collapse the moment it crosses 150, but operational limits begin to surface gradually.

Specific changes that emerge:

• “Wait, who is that?” becomes more common (face–name matching starts to break down)
• Core-member voices dominate the feed; new members find it harder to speak up
• The operator can no longer track every member; isolated individuals begin to appear
• More channels dilute traffic, causing spatial density ρ to drop

What is needed at this point is developing a middle layer and beginning structural design.

Building the middle layer (hub layer)

In village mode, the founder or operator is the hub of all relationships. In town mode, that role must be distributed across a middle layer.

The middle layer consists of people who stand between core members and general participants — those who are knowledgeable about specific themes, who welcome newcomers, and who embody the community’s culture.

• Appoint moderator roles (with explicit authority and responsibilities)
• Create sub-group leads organized by topic or geography
• Hold regular gatherings just for core members (meetups, video calls)

Structural channel design

At this scale, channels need deliberate design — otherwise uneven traffic distribution causes a density mismatch that is hard to recover from.

A practical guideline is to target roughly √N channels for N members. For 150 members, that is around 12–13 channels. Keeping channel scope consistent is critical — a base of #general, #random, and #questions supplemented by function-specific channels tends to work well in early town mode.

Town-mode KPIs

Quantitative tracking becomes more important.

• Middle-layer ratio: Are 5–15% of all members actively engaged on a weekly basis?
• New-member first-post rate: What percentage make their first post within 7 days of joining?
• Channel-level density (ρ): Does daily post volume in key channels stay at 50–80% of $v_{max}$ (guideline: 30 posts/day)?

1,500+ Members: Governance Redesign for “City” Mode

Beyond 1,500 members, matching faces to names stops being functionally possible. The community shifts from being held together by personal bonds toward being maintained by institutions, culture, and structure.

At this scale, an operator trying to “manage” the whole community will inevitably burn out. What is needed instead is designing the infrastructure that enables the community to operate autonomously.

The √N organization model

Trying to run a 1,500-person community from the center is unsustainable. A useful framework is the √N organization model.

√N means the square root of N. For example, if N = 1,500 then √1,500 ≈ 38.7.

For a community of N members, aim for roughly √N moderators or leaders.

Why √N? One leader can maintain village-like relationships with roughly 30–50 people — within Dunbar’s range. Dividing N members into groups of ~40 gives N/40 leaders, but cross-group coordination and overlap push the practical number slightly higher, toward √N.

Each leader owns a specific area (theme, channel, sub-group) and maintains village-like relationships within that area. City-level governance is, in effect, a federation of villages.

Systematizing thread redirection

At this scale, chronic traffic concentration in main channels becomes a persistent problem. The solution is formalizing and automating thread redirection.

• Bots that automatically prompt “please continue this in a thread”
• Documented moderator protocols for detecting traffic spikes and redirecting to threads
• A clear channel architecture that separates important announcements from everyday conversation

The optimal density framework is useful here: there is an upper limit ($v_{max}$, guideline: 30 posts/day/channel) to the volume of messages a participant can comfortably read in a single pass. When main-channel traffic consistently exceeds this limit, the excess cannot be managed without distributing it across threads or parallel channels.

Codifying culture for automatic inheritance

In city mode, “the vibe” can no longer be transmitted informally to new members. As a community grows, culture must be preserved through documentation and institutions rather than oral tradition.

• Community guidelines (written as “directions for behavior” rather than lists of prohibitions)
• FAQ and onboarding paths (a guide for what new members should do first)
• An archive of important past discussions and decisions

Communities where culture lives only in long-time members’ memories develop a deepening split between “people who remember how it used to be” and “people who don’t” as scale increases.

City-mode KPIs

• Tier-based participation rate: Track the ratio of core members (weekly activity), middle members (monthly activity), and mass members (read-only)
• New-member retention: Retention rate at 30 days and 90 days after joining
• Moderator load: Is each moderator responsible for an appropriate number of members? (Guideline: 30–50 members per moderator)
• Culture metrics: Trends in guideline violation reports; new-member first-post rate

The Case for Staying Small

“Bigger is always better” is not a principle that applies to community operation.

For communities where specific goals, culture, or quality standards must be maintained, deliberately keeping the community below 150 members can be a strategically sound choice.

Communities that should stay in village mode

• Communities where shared background is the source of value (e.g., practitioners in a specific industry, holders of a specific qualification, advanced users of a particular tool): The denser the shared assumptions, the richer the conversation — but as scale grows, variance in background increases and depth erodes. Examples: a healthcare-professionals information exchange, a senior-user group for a specialized software tool.
• Communities that go deep on a specific topic (a particular programming language, a craft technique, a music genre): Participants share strong common language; as membership grows, the mix of beginner-to-expert levels makes it harder to pitch discussion at the right level.
• Communities where psychological safety is paramount (health, parenting, career): As the number of strangers increases, the cost of disclosure rises

In these cases, managing inflow through waitlists or invitation systems — and onboarding new members only as existing ones depart — is a sustainable approach.

Trade-offs of capping scale

Deliberately constraining scale comes with costs:

• Without new-member inflow, the community ages and ossifies
• People who were turned away may generate negative word of mouth
• Fewer reference cases and operational resources exist for small communities (most published advice targets large-scale ones)

When choosing to cap size, it is important to communicate to current members why the community stays small, and to ensure shared understanding that value comes from quality, not quantity.

Summary — Swap the Operating “OS” as Scale Changes

There is no single correct model for community operation. But it is true that as scale changes, the assumptions underlying effective practice change with it.

The key insight is not “respond when scale grows” but “prepare the next OS before scale arrives.” Identify middle-layer candidates before hitting 150. Codify guidelines before approaching 1,500. The later preparation is delayed, the higher the transition cost.

And if growing the community is not the goal, intentionally staying small is a legitimate strategy. The question of scale should always follow — not precede — a clear answer to “what value does this community deliver, and for whom?”

Related articles

The "Just-Right Buzz" in Community — The 70% Rule

Why the impulse to "make it more lively" can backfire, explained through three independent theories — queuing theory, flow theory, and the community temperature function. Optimization, not maximization, is the real goal.

rokuse.com/en/blog/community/seventy-percent-rule

The Real Reason "Buzz" Disappears — Sparse and Crowded Are the Same Disease

A community feeling "flat" and a community where "the timeline moves too fast to follow" look like opposite symptoms, but they share the same root cause. This article explains how to diagnose your own community through the lens of spatial density (ρ).

rokuse.com/en/blog/community/density-mismatch-diagnosis

Chasing MAU Kills Communities — On Leading, Intermediate, and Lagging Indicators

If you only chase lagging indicators like MAU and post count, you'll notice the problem too late. This article explains how to catch community health early using a three-layer KPI tree of leading, intermediate, and lagging indicators.

rokuse.com/en/blog/community/leading-vs-lagging-indicators

References

• Dunbar, R. I. M. (1992). “Neocortex size as a constraint on group size in primates.” Journal of Human Evolution, 22(6), 469–493.
• Dunbar, R. I. M. (1998). “The social brain hypothesis.” Evolutionary Anthropology, 6(5), 178–190.