Kicking lurkers to "clean up" the member list is recurring advice. The participation and discovery data argues it's usually self-harm.
What the data shows
Inactive members aren't inert. A share of them are dormant — returning seasonally, on a relevant topic, or via search months later — and "inactive" snapshots badly overstate true churn because they ignore long re-engagement tails. Purges also frequently sweep out people who'd have re-joined the conversation, while doing nothing for the actual health metrics (reply latency, tie density, retained-core size), which a smaller member list doesn't move.
Why it happens
The member count is mistaken for the unit of health. It isn't — active-poster ratio is. Lowering the denominator flatters the ratio cosmetically without improving anything real, and discards optionality (dormant members reactivate; kicked ones rarely return).
The caveat
There are legitimate reasons to prune: bot accounts, ban evasion, regulatory limits, or genuine security hygiene. Cleaning out confirmed bots is not the myth. Purging humans for the crime of lurking is.
Open question: what fraction of "inactive" members actually reactivate over a 6–12 month window, and does that long tail differ between Discord's notification-driven return and Telegram's quieter, search-and-forward-driven one?
Server Signal
@ServerSignal
Myth: purging inactive members makes a server healthier
This post was published in the Telegram channel Server Signal. Subscribe here: @ServerSignal.