Churn is usually discussed as a rate — a percentage per month, tracked on a chart, treated as weather. That framing hides the two things about leaving that are actually actionable: when it happens, and where those people go.
When do members leave?
Across 1,307 tracked departures, the median member who left had been in the community 35 days.
The full ladder, and one crucial instruction for reading it:
Read this table carefully, because it is easy to get backwards. The denominator is members who left, not members who joined. "47.2% by 30 days" means that of everyone who ever departed, 47.2% of those departures happened within the first month. It does not mean 47% of joiners leave in their first month. Most joiners are not in this table at all.
With that said, the shape is unambiguous. A quarter of all departures happen in the first week. Nearly half happen in the first month. Leaving is overwhelmingly an early event.
That has a blunt implication for where effort goes. Most retention work is aimed at members who have been around for months — the newsletter, the monthly call, the re-engagement campaign. The data says the decisive moment happened before any of that reached them. By day 30 the question has largely been settled.
Where do members go when they leave?
This is the finding that should change how you think about churn, and almost nobody has the data to see it.
86.8% of members who leave a community are still on Skool. And 70.2% were active somewhere on the platform in the following 30 days.
They did not decide that online communities are not for them. They did not get too busy. They went somewhere else on the same platform — and the overwhelming majority are still there now, participating in someone else's room.
Two things follow.
Churn is competitive, not attritional. You are not losing people to the general friction of life. You are losing them to other communities that are, at that moment, giving them more reason to show up. That reframes the problem from "how do I stop natural decay" to "why did someone else's room win the five minutes a day they had".
Almost everyone who left is reachable. They are not gone. They are active, on the same platform, most likely within the same interest space. That makes win-back a far more sensible investment than it usually is — see how to win back churned Skool members.
What predicts whether someone stays?
One signal separates the two populations cleanly, and it appears early.
Members who engaged in their first 14 days churned at 3.9%. Members who stayed silent for those 14 days churned at 7.5% — close to twice the rate.
And the context that makes it stark: 94% of members carry zero contribution points. The silent group is not a minority you can afford to treat as an edge case. It is nearly everyone.
The honest caveat: this is a correlation, not proven causation. Early engagement may mark intent rather than create retention — the people who were going to stay may simply be the people who spoke up. And given that 86.8% of leavers go on to be active elsewhere, early silence may mark people who found a better room rather than people you failed to activate.
That caveat does not make the signal useless. Even as a pure marker, it identifies who is at risk while there is still time to act, which is more than most retention metrics manage. It just means the intervention is "give this person a reason to engage in week one", not "make them click something so the number moves".
What should you actually do with this?
Four things, in descending order of how much the data supports them.
Move your effort to the first fourteen days. Half of all departures are decided inside the first month and a quarter inside the first week. Whatever you are doing at day 60 matters less than what happens at day 3. A welcome sequence that reaches a new member personally in their first few days is the highest-leverage work available.
Treat silence at day 14 as the trigger. Not day 30, not "when they stop logging in". Fourteen days of no contribution is the point where the risk roughly doubles, and it is early enough to act on.
Stop treating churned members as gone. Seven in ten were active on Skool in the last 30 days. A win-back message to that population is not a message into the void — it is a message to someone who is currently reading someone else's posts.
Watch for the quiet ones, not the loud ones. The member who complains is engaged. The member who has said nothing for two weeks is the one who leaves. Identifying them is a data problem before it is a retention problem — how to spot members about to leave is the practical version.
Is a 35-day median tenure bad?
It is not comparable to anything published, so treat it as a description rather than a grade.
What it does tell you is the timescale you are operating on. If your onboarding assumes members will find their way over a couple of months, it is calibrated for a member who, statistically, will not be there. The decision window is weeks.
And it tells you what to measure. Not a monthly churn percentage — that is a lagging summary of decisions already made. Measure the share of each joining cohort that contributes something in its first fourteen days. That number moves before the churn number does, and it is the one you can actually influence.
These figures come from The State of Skool 2026, an independent measurement of Skool communities including 1,307 tracked departures. See the full report for the full retention analysis and 21 other findings.