Health Metrics

Communitas surfaces seven health patterns that indicate structural problems in a community. Each pattern is detectable from the community graph and maps to specific interventions a community leader can take.

These are not vanity metrics. They measure structural conditions that research links to community resilience, retention, and collaboration quality.

Bridge scarcity

What it measures. The number and distribution of members who connect otherwise separate clusters. When too few people bridge different groups, the community has single points of failure — individuals whose departure would fragment the network.

Why it matters. Bridges carry information across groups and hold the network together. Research on social networks shows that losing bridge nodes increases shortest-path distances and can isolate subgraphs entirely. Communities rarely notice bridge scarcity until someone burns out and leaves.

What to do about it. Identify overloaded bridges and reduce their broker load. Create new cross-cluster connections through opt-in introductions. Design events and projects that bring different clusters together around shared interests.

Newcomer drop-off

What it measures. The rate at which new members become inactive after their first week. Specifically: what proportion of members who join in a given period are still participating two weeks later?

Why it matters. Most communities lose the majority of newcomers before they form a single meaningful connection. Early drop-off is often invisible because the people who leave never become visible enough to miss. High drop-off rates mean the community is constantly losing potential contributors, mentors, and bridges.

What to do about it. Provide personalized onboarding paths — a clear first step, a second step, and a person to talk to. Match newcomers with mentors or buddies. Track time-to-first-reply and time-to-first-contribution as leading indicators. See the newcomer onboarding experiment.

Clique lock-in

What it measures. The degree to which tightly connected clusters interact only with themselves. High clustering coefficient within a group combined with few or no edges to other groups indicates lock-in.

Why it matters. Cliques are natural and often healthy — people form strong bonds around shared work or interests. Lock-in becomes a problem when clusters stop exchanging information, perspectives, or members with the rest of the community. Innovation, learning, and resilience all depend on cross-cluster flow.

What to do about it. Surface shared interests across clusters. Suggest cross-cluster collaborations on artifacts or events. Use bridge invitations to create low-risk connections between clique members and outsiders.

Moderation hotspots

What it measures. Concentrated patterns of conflict, escalation, or moderation activity. This includes threads with disproportionate moderator intervention, members who appear repeatedly in conflict edges, and topics that consistently generate escalation.

Why it matters. Moderation load that concentrates in a few threads, topics, or relationships signals unresolved structural problems — not just “difficult people.” Hotspots drain moderator energy, discourage participation from bystanders, and can drive away the members least willing to tolerate conflict.

What to do about it. Distinguish between content-driven and relationship-driven conflict. For recurring interpersonal conflicts, consider conflict scaffolds. For topic-driven escalation, consider facilitation prompts and clearer norms. Distribute moderation load across more stewards.

Fragmentation risk

What it measures. The tendency of the community’s clusters to drift apart over time. Indicators include a shrinking giant component (the largest connected subgraph), increasing average path length between members, and declining cross-cluster interaction frequency.

Why it matters. Fragmentation is usually gradual and silent. Two active subcommunities can drift apart over months without anyone noticing — until a crisis reveals they no longer share context, norms, or trust. By the time fragmentation is visible, reconnection is expensive.

What to do about it. Monitor cross-cluster interaction trends over time. When interaction between specific clusters declines, investigate whether it reflects natural specialization (acceptable) or loss of shared context (concerning). Use cross-cluster sharing and shared events to maintain connective tissue.

Participation concentration

What it measures. How evenly participation is distributed across members. High concentration means a small number of people carry most of the community’s activity — answering questions, reviewing contributions, organizing events, moderating discussions.

Why it matters. This is the community equivalent of bus factor. When three people do everything, the community is fragile. Those people burn out, and when they leave, institutional knowledge and relational infrastructure leave with them. Concentration also discourages broader participation: newcomers see that “the regulars” handle everything and don’t step in.

What to do about it. Identify members who are ready to take on more responsibility and create pathways for them. Use recognition rituals to make diverse contributions visible. Distribute recurring tasks across more people. Track whether concentration is decreasing over time.

Information flow gaps

What it measures. Messages, ideas, or decisions that circulate within one cluster but never reach others. Detected by comparing topic presence across clusters and identifying cases where relevant information stays local.

Why it matters. When information doesn’t flow, clusters make decisions without full context, duplicate work, or develop incompatible norms. Members in peripheral positions miss announcements, discussions, and opportunities that affect them. Flow gaps are a structural problem, not a communication skills problem.

What to do about it. Surface relevant threads and decisions to affected clusters through cross-cluster sharing. Create digest mechanisms that summarize activity across groups. Ensure that governance decisions reach everyone they affect, not just the cluster where the discussion happened.

---

These metrics work together. Bridge scarcity and fragmentation risk are related but distinct — you can have enough bridges but still see clusters drifting apart if those bridges aren’t active. Newcomer drop-off and participation concentration reinforce each other: when newcomers leave, the same people keep carrying the load.

Use these metrics as a starting point for your own pilot. Pick one, measure it, and try an intervention.