Moderator burnout has a predictable curve

Moderator burnout has a predictable curve

Communities track member churn obsessively and mod churn almost never. The data argues that's backwards.

What the data shows
In studies of volunteer moderation, mod attrition clusters around identifiable stress thresholds: reports-per-mod per day, exposure to a single high-conflict incident, and the ratio of thankless enforcement to positive interaction. Servers that lost a mod frequently lost others within weeks — a contagion pattern, since remaining mods absorb the vacated load and hit their own thresholds faster.

Why it happens
Moderation is unpaid emotional labor with negativity bias built in — mods mostly touch the worst 1% of interactions, so their lived experience of the community is far darker than the average member's. Past a load threshold, the cost-benefit flips and they quit, raising load on the rest. Telegram's lighter native moderation and heavier reliance on anti-spam bots shifts more of this load onto automation, for better and worse.

The caveat
Mod psychology is hard to instrument and most evidence is qualitative or self-reported. "Burnout" lumps together very different exits (boredom, conflict, life changes).

Open question: what's the right reports-per-mod ceiling, and would simply capping each mod's queue do more for community survival than any member-retention tactic?