Real metrics vs vanity counts

Every platform dashboard leads with the numbers that make the platform look good: total members, total posts, total reactions. Some of those are useful. Most are decoration. Your first job as a numbers-literate manager is knowing which is which before an owner asks you a hard question.

Start with active members: how many distinct members did something — posted, replied, reacted, attended, even just logged in — in the last 30 days, and how does that compare to the last 30? Raw activity can rise while the community rots, if three superusers are doing all of it, so pair it with posts per active member and a rough spread: what share of activity comes from the top ten members? A community where the top ten produce 80% of activity is a panel show with an audience, and it's one burnout away from silence.

The metric above all of them is retention: of the members who were paying last month, how many are paying this month? Retention is the community's health in one number because it aggregates everything — onboarding, rhythm, events, moderation, value — into the only verdict that's cast in money. Track it monthly, and by cohort when you can: members who joined in March versus April tells you whether your new onboarding is actually better or March was just lucky.

Now the vanity shelf: total member count (rises even while activity dies, because cancellations lag signups), follower-style reach numbers borrowed from social platforms, reaction totals with no denominator, and "engagement" screenshots of the single best thread of the month. These aren't lies, exactly — they're answers to questions nobody who pays you is asking. When an owner waves a growing member count at you while activity slides, your job is the respectful redirect: growth is the top of the bucket; let's look at whether the bucket is leaking. That's next lesson.

Churn math: the number your job hangs on

Churn is the percentage of paying members who cancel in a period, usually a month. It sounds abstract until you price it. Take the course's running example: a community of 500 members at $49 a month is $24,500 a month in revenue. At 8% monthly churn, it loses 40 members — $1,960 walking out every month — and has to recruit 480 new members a year just to stay at 500. Most of what owners call a growth problem is that leak.

Now the part that pays your rent: cutting churn from 8% to 6% keeps 10 more members a month — about $490 more every month, immediately, and those saved members keep paying, so the gain compounds month over month. No ad campaign a small business can buy returns money that reliably. This is why owners pay community managers management retainers in the $1,500–5,000 range for ongoing engagement work: if you hold even a fraction of that churn reduction, you're the cheapest revenue they have.

Work the calculator both ways with every owner you serve. First with their real numbers, so the leak has a price tag. Then at one and two points lower, so your retention work has a price tag too. When you negotiate, renew, or ask for budget, you won't be arguing that community "feels better lately" — you'll be pointing at the difference between $1,960 a month leaving and $1,470. Feelings get thanked. Math gets renewed.

Reporting owners actually read

Most community reports die unread because they're data dumps: forty metrics, three dashboards, no conclusion. Owners are busy people paying you partly so they don't have to study their own community. A report they actually read fits on one page and answers three questions: how healthy are we, what could hurt us, and what do you recommend?

Lead with health: the short list from this module — active members and its trend, posts per active member, retention or churn against last month, and one line of context for each ("activity dipped the week of the holiday, recovered the week after"). Four numbers with a sentence each beat forty numbers with none. If a metric moved, say why you believe it moved, and mark clearly when you're guessing. Owners don't need you to be certain; they need to know which parts are certain.

Then risks, stated while they're still cheap to fix: a superuser going quiet, a cohort from a launch starting to drift, a recurring complaint surfacing in DMs, moderation incidents trending up. This section is where you earn trust disproportionate to your fee, because you're the only person positioned to see these early. Never pad it — a report that cries risk monthly gets ignored the month it's real. "No new risks" is a perfectly good entry.

Close with one recommendation. Not five — one, with the expected effect in the owner's language: "Add a buddy match to onboarding; March cohort retained 9 points better in the community I ran last year, and our first-week silence is where our churn starts." Then next month, report what happened with it. That loop — recommend, act, report the result — is the visible shape of a manager who moves numbers. It's also your renewal file: when the retainer conversation comes, twelve months of these one-pagers are the case, already written.