How to improve retention in a Skool community

By Christopher Ward, Co-founder, SkoolgradesPublished

Short answer

Retention improves when you treat it as three separate jobs rather than one. Get new members to a first interaction inside two weeks, give mid-tenure members a ritual and a relationship, and give long-tenured members visible progression or a next role. Measuring retention by tenure band tells you which of the three is actually leaking.

Retention is the least glamorous number in community ownership and the only one that compounds. A community that keeps 90 percent of its members each month and one that keeps 80 percent look almost identical in week four and look like different businesses in year two.

The trouble is that "improve retention" is not a task. It is three different tasks that happen to share a metric, and owners who treat it as one thing end up doing the work that fits their temperament rather than the work their community needs.

How should I be measuring retention?

Retention should be measured by tenure band rather than as one blended number, because splitting it by how long people have been members is the single change that makes the work tractable.

Skoolgrades resolves member lifecycle across six lanes and three tenure bands — newly joined at 14 days or less, mid at 15 to 90 days, and long-tenured beyond 90 days. Those cut points are not arbitrary. They mark the three distinct ways membership fails.

BandThe jobFailure looks like
Newly joined, 0 to 14 daysGet them to a first interactionJoined, read a bit, never typed, gone
Mid, 15 to 90 daysBuild a habit and one relationshipActive for a fortnight, then thinning out
Long, over 90 daysProgression, or a next roleSteady for months, then quietly stops

Run your own numbers this way once and there is a good chance one band is doing most of the damage. That is the band to work. Everything else is optional until it is fixed.

A blended retention number tells you that you have a problem. A tenure-split retention number tells you which problem.

Why do new members leave in the first two weeks?

New members leave in the first two weeks because they never got to a first interaction, which makes most of what owners call a retention problem a first-contact problem.

A new member arrives with a burst of motivation that decays fast. If that burst is spent reading a welcome document and watching a module, it is gone and nothing anchored them. If it is spent typing something and getting a reply from a human, they are in.

The mechanics that matter:

  1. One obvious first action, and only one. Introduce yourself in this thread. Anything more elaborate gets skipped.
  2. A reply by name within 24 hours, referencing what they actually wrote. This is boring and unscalable and it is the highest-return work in the whole job.
  3. A named person to talk to. "Ask Priya about the ad tracking thing, she's been through it" beats any onboarding sequence.
  4. A second reason to return inside three days, inside the community rather than in an email.

The list of new joiners who have not yet posted is short, urgent and finite. It is the one list worth checking more than once a week.

Why do members drift away after a month or two?

Members who make it past two weeks but drift out before three months usually did not find either a rhythm or a person.

Rhythm means something happens on a schedule and does not depend on your mood. A Monday wins thread. A Thursday question. A monthly call at the same time. The specific ritual matters much less than its predictability — people build their attention around things they can anticipate. A brilliant post that arrives at random is worse for retention than a mediocre one that arrives every Monday.

Relationship means at least one other human here knows who they are. You can force the conditions for this without forcing the outcome:

  • Introduce two members to each other by name, in public, with a reason.
  • Reply to the same person's comments more than once so they know you register them.
  • Ask a specific member a question they are qualified to answer, publicly.

The third one does double duty, because it also pulls people out of lurking. See activating lurkers for more on that.

There is one more mid-tenure pattern worth understanding. Some members are quiet with you and enormously active elsewhere on Skool. Skoolgrades classifies these as untapped with concrete thresholds: at least 25 posts in other Skool communities, active in two or more communities, seen in the last 14 days, joined 30 or more days ago, and silent in your community for 30 or more days. That combination says something precise. They have the habit. They are not disengaged people. They simply have not built the habit here. That is a very different fix from a member who has gone quiet everywhere.

Why do my long-time members quietly stop showing up?

Long-tenured members stop showing up for two reasons that both look like drift: they finished, or they faded.

They finished. They came to solve a problem, they solved it, and the community became a thing they pay for out of inertia. There is a limit to what you can do here, and some of this churn is healthy. What helps is a visible second act: an advanced tier of conversation, a harder problem, a peer group at their level.

They faded. Life got busy, the community got noisier, and their attention went elsewhere without a decision ever being made. This one is very recoverable and almost always missed, because the member looks fine right up until they are gone.

The other move for long-tenured members is to give them a role. Members who welcome new joiners, run a thread, or moderate a channel churn far less, because they now have a responsibility and an identity here. It is also the only way onboarding stays good as the community grows past what one person can personally reply to.

What actually keeps members from leaving?

Members stay for five things — being known, predictability, visible progress, peers other than you, and fit — and everything else is downstream of those:

  • Being known. Named, replied to, remembered. Almost nothing beats this.
  • Predictability. Rituals over quality.
  • Visible progress. Levels, streaks, a wins thread, anything showing movement.
  • Peers, not just you. A community where all value flows through the owner does not retain past the owner's energy.
  • Fit. Retention that comes from clear promises at the join point never appears in a retention tactic list, and it is upstream of all of the above.

And what does not hold people: more content. There is always more content elsewhere, free. Posting harder is the most common response to a retention problem and the least effective.

What should I do every week to hold retention steady?

Holding retention steady takes twenty to thirty minutes once a week, done in this order:

  1. New joiners who have not posted. Reply, welcome, ask one question. Highest priority, shortest window.
  2. Members whose activity has fallen off relative to their own past. Two or three genuine messages. Not everyone who is quiet — only the people who have changed.
  3. One public introduction between two members who should know each other.
  4. One member asked a public question they are qualified to answer.
  5. Log the numbers. Retention by band, and cancellations by band. Five minutes.

That is it. It is unglamorous and it works better than any campaign, because retention is an accumulation of small acts of noticing rather than a project.

Do I need a tool for any of this?

You can run all of this on a spreadsheet and a calendar reminder, and for a community under about a hundred members you probably should. Tools earn their place when the roster gets past what you can hold in your head.

Skoolgrades exists for that point: it reads your community through a Chrome extension on your own logged-in Skool session, places every member into a lifecycle state through the Cohort Blender, resolves lifecycle across the six lanes and three tenure bands, keeps a churned-member archive that freezes a member's full state at the moment they left, and hands you drafted messages you review and send with one human click.

What it does not do is predict who will churn. There is no predictive model and no backtest behind it — snapshot history is shallow and we would rather say that plainly. It surfaces behavioural signals of drift, which is a smaller and more honest claim, and in practice the more useful one anyway. Also see how to reduce churn and how to build a churn baseline.

Frequently asked

What is a good retention rate for a Skool community?
There is no published benchmark for Skool communities, and any single figure would ignore price point, niche, free versus paid entry, and community age, which are the four variables that move retention most. Track your own rate monthly for three to six months to build a range, then measure changes against your own history rather than a borrowed number.
Why should I measure retention by tenure band?
Because membership fails in three different ways. New members leave because they never had a first interaction, mid-tenure members leave because they never built a habit or a relationship, and long-tenured members leave because they finished or faded. A single blended number tells you something is wrong but never which of the three to fix.
Does posting more content improve retention?
Usually not. There is always more content available elsewhere for free, so content alone is a weak retention lever. What holds members is being known by name, predictable rituals they can build attention around, visible progress, and relationships with other members rather than only with the owner. Posting harder is the most common ineffective response.
What does untapped mean in Skoolgrades?
Untapped members are prolific elsewhere on Skool but silent with you. The definition is concrete: at least 25 posts in other Skool communities, active in two or more communities, seen in the last 14 days, joined 30 or more days ago, and silent in your community for 30 or more days. They have the habit, just not here.
How much time should retention work take each week?
Twenty to thirty minutes is enough for most communities. Reply to new joiners who have not posted yet, message two or three members whose activity has dropped relative to their own past, publicly introduce two members to each other, and log your numbers. Consistency each week beats a large occasional push by a wide margin.
What happens to member data when someone leaves my community?
Inside Skoolgrades, a churned-member archive freezes that member's full state at the moment they churned, so their history does not disappear from the roster the way it otherwise would. That means you can still see what a departed member looked like before they left, which is what makes patterns across several departures readable at all.
What owners say

I got a little obsessed.

I got a little obsessed. It's one of those tools that you open and immediately understand what you're supposed to do. You see exactly where your members are, no more guessing who needs attention and who's doing fine. That clarity matters more than people realize. I didn't expect to love this as much as I do. I'll be using it going forward.

Silla Jonsdottir

See this in your own community

Skoolgrades reads your Skool community through your own browser and turns it into a report you can act on: participation measured on real interactions, who is drifting, where your members go when they leave your room, and a ranked list of what to do next — with the draft already written.

Get your report

Keep reading

Last updated 2026-07-29. Skoolgrades is an independent tool and is not affiliated with or endorsed by Skool.