Growth dilutes existing norms
Aliases: community scaling · norm dilution · governance capacity · newcomer ratio
What it is
Norm dilution at scale occurs when growth means newcomers no longer learn by repeated familiar observation and veterans cannot remind or repair individually. Scale changes how norms spread and are enforced, not only headcount.
Why it happens
Small groups rely on repeated encounter, reputation, and immediate correction. Growth adds strangers, content, and handling load; violations are harder to see and exceptions become mistaken as rules.
The second-order mechanism clarifies a commonly misread point: what actually drives dilution is not total headcount but the share of active members who are newcomers at any given moment. Norms propagate through observation and imitation. A community that is large in absolute size but grows slowly keeps newcomers as a small share of any given moment's interaction, so most exchanges still happen among norm-fluent veterans, giving newcomers ample observational examples and letting norm transmission roughly keep pace. Conversely, a community that is not large in absolute terms can dilute quickly if it gains a large number of members in a short window — a sudden referral spike from an external platform, a marketing push — because the newcomer share spikes sharply and veteran modelling and correction cannot cover the surge in new interaction, even though total size stays modest. This is why the intuition "the community got bigger so it got worse" is often wrong: the metric that actually matters is the newcomer-share curve over time, not a single static headcount.
Studying it
- Paradigm: trace interaction, violation, handling, and retention around growth episodes; interview entry cohorts; specifically compute the share of active members who are newcomers within rolling time windows and correlate that time series with contemporaneous violation rate and rule comprehension, rather than correlating violation rate with total size alone.
- Variables: total size, newcomer share and its rate of change, repeated interaction, rule understanding, violation rate, governance capacity, retention, and belonging.
- Methodological caution: attributing every change to scale ignores product change and external events. Also avoid conflating "large total size" with "high newcomer share" — they act through different mechanisms, the former mainly stressing whether governance capacity is adequate, the latter directly determining whether the model-and-imitate chain can keep up.
Where it stops holding
Growth can add diversity, resources, and mutual help; dilution is not inevitably decline. Closed entry preserves norms but can exclude newcomers and new views. The newcomer-share mechanism matters most in communities that rely on tacit norms with little codification, transmitted mainly through veteran modelling — a sudden influx overwhelms them quickly. In communities where norms are already heavily codified and conveyed mainly by the system rather than by interpersonal observation (extremely detailed written rules with mandatory reading and testing at entry), a sudden spike in newcomer share does relatively less damage, because norm transmission there never depended on "are there enough veterans to demonstrate it" in the first place.
Applying it
- Embed key norms into entry, feedback, and handling flows rather than relying on oral tradition, reducing how much norm transmission depends on whether veterans have time to model it.
- Monitor the newcomer-share curve over time, not just total size, and expand governance capacity and onboarding resources ahead of a known sharp rise (an anticipated large referral influx, say).
- Publish which habits are adaptable and which boundaries are non-negotiable, so a sudden wave of newcomers can locate the boundary themselves rather than waiting for veterans to correct them one by one.
- Verification: test scenario-rule understanding and actual treatment consistency by entry cohort; compare that consistency against the newcomer share at the time to check whether dilution actually tracks the share curve rather than total size.