An audit is not a dashboard. A dashboard runs continuously and you stop seeing it after a week. An audit is a deliberate, dated sit-down where you write numbers on a page, compare them to last time, and decide what to change.
Do it twice a year, or after anything big — a launch, a price change, a format change, a quiet stretch. Block two to three hours. Write everything down; the writing is the mechanism, because next time you need something to compare against.
Here is the checklist.
How do I set up a Skool community audit before I start?
Setting up a Skool community audit means creating one dated document with your member count at the top and nine empty headings underneath, so every number you write has somewhere to live and something to compare against next time.
Put the date, your member count, and one sentence on what has changed since your last audit. If this is your first, say so — the first one produces a baseline, not conclusions.
Then create nine headings, one per step below, each to be filled with a number and a sentence. Resist the urge to fix anything while auditing. Note it and move on.
Who is actually in my community, and is anyone talking?
Finding out who is actually in your community and whether anyone is talking takes two audit steps: a roster split by tenure, and a participation count of members who genuinely did something in the last 30 days.
Step 1 — Roster composition
You are answering: who is actually in here?
- Total members.
- Members who joined in the last 30 days.
- Members who joined more than 90 days ago and are still around.
- Paying members, if you charge.
The ratio that matters is the third number against the first. A community that is 80% recent joiners is a leaky bucket wearing a growth costume; one that is 90% long-tenured with almost no new joiners is stable and slowly dying. Neither is visible from a headline member count, which is why that count is the least useful number you own.
Write down the split. Next audit, the direction it moved tells you more than either number alone.
Step 2 — Real participation, not visits
Most audits get this step wrong, so be careful.
Count the number of distinct members who did something in the last 30 days — left a comment, replied to a comment, voted in a poll. Divide by total members. That is your participation rate.
Now write your dashboard's engagement figure next to it. Skoolgrades counts participation — comments, comment replies and poll votes — where a visit-based engagement rate counts anyone who loaded the page, which is why the two will not match (the full explanation). For the arithmetic, see how to calculate engagement rate on Skool.
Label which figure is which. The trap is not the gap between them, it is comparing this quarter's participation rate against last quarter's visit-based figure and concluding the community collapsed. Pick one definition, note it at the top of the sheet, and use it every time.
Which of my posts work, and are new members sticking around?
Judging which posts work and whether new members stick takes two audit steps: sorting your last twenty posts by comments to find the pattern, and measuring what share of recent joiners did anything in their first fortnight.
Step 3 — Content performance
Pull your last 20 posts. For each, note the number of comments. Then sort.
Look at the top five and the bottom five and ask one question of each group: what do these have in common? Format, topic, length, time of day, whether you asked a question, whether you replied in the thread yourself.
Two patterns show up constantly:
- Posts that ask something specific outperform posts that announce something. "Here is the new module" gets acknowledgement. "What is the one thing in module 3 you would cut?" gets a thread.
- Posts where the owner replied to the first comment within an hour outperform the rest. Your reply is permission for everyone else.
Write down the one format that works and the one that does not. If nothing separates them, the variable is probably timing — see the best time to post on Skool.
Step 4 — The first fourteen days
New members either build a habit in the first two weeks or they do not. So audit that window specifically.
Take everyone who joined 15 to 45 days ago — old enough to have had their chance, recent enough that onboarding has not changed. For each:
- Did they post or comment at least once?
- Are they still active now?
The percentage who did something in their first fortnight is your activation rate, the single most useful number in the audit. If it is low, nothing further down the funnel matters much.
Then walk your own onboarding as if you were new. Join flow, welcome message, first thing they see, first thing they are asked to do. Most communities ask a new member to consume something. The ones that stick ask them to say something.
Who is drifting away, and where do those members go instead?
Working out who is drifting and where they go instead takes two audit steps: sorting the roster into retention lanes so the fading members separate from the dormant ones, and looking at which other Skool communities your quiet members are active in.
Step 5 — Retention lanes
Sort your roster into rough lanes and count each:
| Lane | Rough definition |
|---|---|
| Active | Did something in the last 30 days |
| Fading | Was active 30–90 days ago, quiet since |
| Dormant | No activity in 90+ days, still a member |
| Churned | Left or cancelled |
The fading lane is the actionable one. Dormant members have decided; fading members are mid-decision. If you have energy for one intervention this quarter, it belongs here.
Skoolgrades resolves this properly: lifecycle runs across six lanes and three tenure bands — newly joined at fourteen days or less, mid at 15 to 90 days, long-tenured beyond 90 — because three weeks of quiet means something different for a nine-day-old member than for a two-year veteran. The Cohort Blender also assigns each member a lifecycle state: star, rising, steady, fading, drifted, ghosted, untapped, or going dark.
There is no prediction happening. These are behavioural drift signals — what someone actually did — not a model forecasting who will cancel. Between audits, the shorter version is spotting members about to leave.
Step 6 — Where your members go instead
Step 6 is the one step you cannot do from inside your own admin area, and it is the one that reframes everything above.
Your members are in other Skool communities. Some are silent with you and prolific elsewhere. That is not a disengaged person — that is an engaged person who has not built the habit in your room. Completely different problem, completely different fix.
Manually, open a handful of your quietest members' profiles and see what else they are in. Slow, but do ten and you will learn something.
Skoolgrades collects this as one of its eleven data streams and ranks external rooms by how much of your members' attention they capture. One caution: those counts deliberately undercount. "Fourteen members active elsewhere" is a floor, not a total.
What do the money numbers say, and would I reply to my own posts?
The last two steps of an audit are the money — recurring revenue, new subscriptions, cancellations and net — and the uncomfortable one: reading your own recent posts as a member and asking whether you would have replied.
Step 7 — Money
If you charge: monthly recurring revenue, new subscriptions this month, cancellations this month, and net. If cancellations are outpacing joins, that is the audit's headline and everything else is a footnote.
If revenue lands outside Skool checkout — an off-platform course, a coaching offer — note that it is invisible to any Skool-based reporting, including Skoolgrades. Track it yourself.
Step 8 — The uncomfortable question
Read your last five posts as a member would, not as the person who wrote them.
Would you have replied? If not, the participation number in step 2 is explained, and you can stop looking for structural causes.
How do I turn an audit into changes I will actually make?
You turn a community audit into changes you will actually make by picking exactly three, writing each as an action with a day attached, and booking the next audit before you close the document. Go back through and choose three. Not ten.
Write them as actions with a day attached:
- "Every Tuesday, post one question-format post."
- "DM the fading lane this month, twenty-five at a time."
- "Rewrite the welcome so the first ask is an introduction post."
Then put the next audit in the calendar before you close the document. An audit with no rematch scheduled is a diary entry.
How do I make a Skool community audit repeatable?
A Skool community audit becomes repeatable when the counting steps stop being manual, so the afternoon goes on judgement rather than arithmetic. The manual version above is worth doing at least once, because you will notice things a dashboard never shows.
A tool helps the second and third time. Reports in Skoolgrades — participation rate, retention lanes, joins, timing, top posts, and a peer benchmark computed on posts matured 72 hours so it does not swing with fresh content — are included on every Skoolgrades plan, with revenue and acquisition sections on Pro ($99/mo). Grow is $19/mo.
The outreach action in step 9 is covered too. You build the list — pick a recommended group such as the fading lane, describe who you want in plain language, or paste a post URL and pull everyone who engaged — then Skoolgrades drafts each message verbatim or personalised per member, you approve every draft, and send from the extension panel one click at a time. Nothing goes out unattended, which keeps your account safe and the messages worth opening.
That turns steps 1, 2, 3, 5 and 7 from an afternoon of counting into reading a page, leaving your time for steps 4, 8 and 9, where the thinking lives. One caveat for large communities: above roughly 3,000 active members, collection samples rather than taking a full census, so those figures are estimates.
Once the audit is done, the follow-through is a weekly routine, not another audit.