The Metrics to Check Once a Week

The last lesson of the Academy is deliberately small. Not a dashboard of forty charts — seven numbers, checked once a week on a fixed day, fifteen minutes total. Threads of this ritual have run through the whole course: the first-month audit demanded a five-number row, the Money track needed real averages, every marketing lesson asked you to log. This lesson assembles it all into the operating rhythm of the business.

Why weekly, and why so few

Daily numbers at your scale are noise — one client cancelling reads as catastrophe on a Tuesday and as rounding by Friday. Quarterly numbers are archaeology: whatever went wrong is three months old. Weekly is where signal lives, fast enough to act, slow enough to be true.

And seven numbers, not forty, because the review must survive busy weeks. A metrics habit you skip when things get hectic is not a habit; it is a decoration. Everything below comes from two places you already have open: your service's partner panel (clients, payments, statuses) and your outreach log.

The funnel four: is new business coming?

1. Outreach touches sent — cold messages, community answers that reached a prospect, demos booked from any channel. Fully under your control, which is exactly why it is first: in a bad week, this is the number you can always move by deciding to.
2. Conversations started — replies, demo bookings, inbound questions. Against touches, this is your reply rate: the health check of your message and list.
3. Trials started (or first workspaces created, if you run manual onboarding). Against conversations — your demo's conversion, the number the selling lesson exists to move.
4. New paying clients. The line that matters. Against trials — your onboarding and follow-up quality from the trial playbook.

The power is not in any single value; it is in the ratios between adjacent lines, because they tell you which lesson to reread. Touches high, conversations low → positioning or message. Conversations high, trials low → demo. Trials high, payers low → onboarding and follow-up. A funnel that names the broken stage replaces a month of "marketing isn't working" mood with one targeted fix.

The health three: is the business you built holding?

5. MRR — monthly recurring revenue, the sum of live subscriptions. Watch direction, not drama: MRR that rose even slightly in a week when everything felt chaotic is the fact that keeps you going; MRR flat for six weeks while "new clients" trickle in means churn is eating your acquisition — see number six.
6. Churned clients this week, with the stated reason — from the exit-interview habit. Zero most weeks at your scale; the log matters precisely because patterns only appear when every loss is written down. Involuntary losses (card failures) are the platform's dunning job plus your one personal note; voluntary ones each get a category and, monthly, a look for clusters.
7. Activation: clients with zero activity in 21 days. The early-warning radar from the retention lesson, formalised. Every name on this list gets the five-minute personal email this week — it is the single highest-yield item in the entire review.

The fifteen-minute ritual

Same day, same time, recurring calendar event — Friday afternoon and Monday morning both work; chosen-and-kept beats optimal. The sequence:

1. Fill the seven numbers from panel and log (five minutes).
2. Compare each with last week and eyeball the three funnel ratios (three minutes).
3. Write one sentence: the most important thing these numbers say. "Replies died when I changed the opener." "Two restaurant churns citing closures — seasonal?" Forced compression is the analysis; a number-filled sheet with no sentence is stamp collecting.
4. Pick one action for next week, attached to the weakest ratio or the risk list: revert the opener; send the three activation emails; offer yearly to the two clients past month three (seven minutes, including doing the small ones on the spot).

One sentence, one action. Not five actions — the week only has room for one deliberate marketing improvement alongside running the business, and a single completed fix compounds; five intended ones don't.

What not to measure, on purpose

Equal and opposite discipline: the numbers that feel like progress and predict nothing. Total clicks across all client links — that is your clients' success metric (they should watch it in their analytics; you only care at plan-limit scale). Site visits and social followers — inputs at best; they earn attention only the week you pay for traffic and need cost-per-trial. Rankings, impressions, likes — weather. If a vanity number insists on being watched, give it a monthly glance, never a weekly row. The seven earn their places because each has a lesson-shaped action attached; a metric with no action attached is entertainment.

The course, closed

That is the Academy. You can explain the product to anyone, launch it in an evening, read its economics without illusions, and market it one deliberate week at a time. The sheet of paper from the first Foundations assignment — niche, plans, break-even, first channel — is now a business plan you have the skills to execute. The numbers ahead are yours to produce, not ours to promise; the pricing page is where the fourteen-day trial starts whenever you are ready.

The ritual, on one card

• [ ] Create the recurring calendar event and the seven-column sheet (touches, conversations, trials, new payers, MRR, churn+reason, silent clients).
• [ ] Run the first review now with whatever history you can reconstruct — an imperfect baseline beats a perfect blank.
• [ ] Write this week's one sentence and one action; do the action before the next review.
• [ ] Monthly, on the first review of the month: scan the churn log for clusters and the funnel ratios for a four-week trend.
• [ ] After eight weeks, act on the channel verdicts your log has quietly accumulated: keep the two producers, drop the rest — and go back to any lesson the weakest ratio names.