Most marketing budgets in small schools are set the same way. The owner looks at the bank balance, looks at last month's ads bill, and decides whether it "felt worth it". A month with lots of enquiries feels worth it. A quiet month does not, so the budget gets cut, and three months later there are fewer students and the budget gets cut again.
The problem is not that the owner is careless. It is that the question "was this month's spend worth it?" cannot be answered in a month. A student who enrols in October and pays for eight months is worth eight months of fees, not one. Judge the ad by October alone and almost every channel looks too expensive.
This guide builds the number that should anchor the budget instead: how much a new student is worth over the time they actually stay, and from that, how much you can afford to spend to find one. We will use one illustrative school throughout. Its figures are invented; the method works with yours.
Step 1: find out how long students really stay
You need to know, for students who started in a given month, how many are still enrolled one, two, three and more months later. This is called a cohort: a group of people who started at the same time, followed forward.
Here is the illustrative school's January intake of 40 students:
| Month of study | Still enrolled | Share of starters |
|---|---|---|
| 1 | 40 | 100% |
| 2 | 34 | 85% |
| 3 | 30 | 75% |
| 4 | 27 | 68% |
| 5 | 25 | 63% |
| 6 | 22 | 55% |
| 7 | 20 | 50% |
| 8 | 19 | 48% |
| 9 | 17 | 43% |
| 10 | 16 | 40% |
| 11 | 14 | 35% |
| 12 | 13 | 33% |
Two things stand out in this example. Half the cohort is gone by month seven. And a quarter of it is gone by month three, before most schools have had a single structured conversation with the family about how things are going.
To turn the table into one number, add up the "still enrolled" column: 40 + 34 + 30 + ... + 13 = 277 student-months. Divided by the 40 who started, that is 6.9 months of fees per new student in their first year.
We stop at twelve months on purpose. Some students will stay for years, and they matter, but a budget built on a three-year average is a budget built on students who have not stayed three years yet. The first-year figure is conservative, and conservative is what you want when the money leaves your account before the fees arrive.
Step 2: work out what one student-month is worth
Revenue is not the right number here, because some of each fee is already spoken for by the cost of teaching the student. What you want is contribution: the fee minus the cost of delivering the lessons.
In the illustrative school a student pays about 80 a month for two hours a week. Using the cost model from what one class hour really costs, a place in a class of six costs about 5.24 per hour to deliver (the student's share of teacher, room and admin, plus materials and payment fees). Two hours a week is about 8.7 hours a month, so delivery costs roughly 45.
Each student-month contributes about 35.
Note what is not in that 45: the cost of finding the student. That is the thing we are trying to size, so it stays out.
Step 3: lifetime contribution and a spending ceiling
Put the two together:
6.9 months x 35 = about 243 of contribution from a typical new student in their first year.
That is the most a student is worth to you. It is not what you should spend to get one, for three reasons:
- The estimate is uncertain. One cohort of 40 is a small sample, and next January's intake may behave differently.
- The money moves in the wrong order. You pay for the ad now and collect the contribution over seven months. Spending the full amount means financing every new student yourself.
- Contribution is not profit. The owner's pay, accounting, software and everything above the level of a class still have to come out of it.
A cautious ceiling is about a third of first-year contribution. In this example that is about 81 per enrolled student. There is nothing magic about a third; it is a margin for being wrong. If your retention data covers several years and is stable, you can justify a higher share. If it is one term old, go lower.
The illustrative school, by the way, currently spends about 100 per new student. The ceiling says that is too much, which is useful to know before cutting the budget: it is only too much at the current retention and conversion rates, as the next sections show.
Step 4: translate the ceiling into enquiries and trials
Nobody buys "enrolled students" directly. You buy clicks, leaflets, open days and referral rewards, and those produce enquiries, some of which become trial lessons, some of which become students. To judge a channel, you need the ceiling expressed at the stage that channel produces.
The illustrative school's last three months:
| Stage | Count | Rate from previous stage |
|---|---|---|
| Enquiries | 100 | - |
| Trial lessons attended | 45 | 45% |
| Enrolled | 18 | 40% |
So 18% of enquiries become students. If an enrolled student is worth up to 81 of acquisition spend, then:
- An enquiry is worth up to about 14.60 (81 x 18%).
- A trial lesson attended is worth up to about 32 (81 x 40%).
Now a channel can be judged in its own terms. A campaign that brings enquiries at 9 each is comfortably inside the ceiling even if this month's enrollments look thin. A partnership that costs 50 per trial lesson is not, however busy it makes the front desk feel.
What actually moves the ceiling
Here is the part that is easiest to miss. The ceiling is not fixed. It depends on how long students stay and how well enquiries convert, and both of those are things the school controls without spending anything on ads.
| Scenario | Average months (year 1) | Lifetime contribution | Ceiling per student | Ceiling per enquiry |
|---|---|---|---|---|
| As today | 6.9 | 243 | 81 | 14.60 |
| 4 more of 40 still enrolled at month 3 | 7.6 | 266 | 89 | 16.00 |
| Trial-to-enrollment from 40% to 50% | 6.9 | 243 | 81 | 18.20 |
| Both | 7.6 | 266 | 89 | 20.00 |
The retention row assumes the four extra students who stay past month three then behave like the rest of the cohort. The conversion row does not change what a student is worth, but it means more students per enquiry, so each enquiry is worth more.
Put together, the two improvements raise what the school can afford to pay for an enquiry by about 37%. That is the same effect as a 37% cheaper advertising channel, and it comes from things that happen inside the building: the trial lesson, the first thirty days, the phone call after the second absence.
This is why retention and conversion belong in the same conversation as the marketing budget. A school that loses a quarter of its students in the first ninety days is not short of marketing money. It is spending its marketing money on students it does not keep. Students who quietly leave and the trial lesson that converts cover the two biggest levers in more detail.
Six ways to get this wrong
Counting paused students as enrolled. A student on a two-month break is on the roster and not in your revenue. Count months that were billed.
Using one small cohort. If a month's intake is under about 20 students, combine two or three neighbouring months. One family moving away should not change your marketing budget.
Mixing seasons. Students who join a summer course often leave in September by design. Measure them separately, or they will make your year-round retention look worse than it is.
Ignoring where students came from. Different channels can bring students who stay for very different lengths of time. A heavily discounted promotion may produce many enrollments that leave when the discount ends. If you record the source of each enquiry, run the cohort table per source at least once.
Forgetting siblings and referrals. A family that stays often brings a second child or a friend. That value is real but hard to measure, which is one more reason to keep the ceiling conservative rather than to inflate it.
Letting the ceiling become the target. The ceiling is the most you should pay. If a channel brings students well below it, that is good news, not room to spend more on that channel until it stops being cheap.
Getting the numbers from your own records
You need three things for each student: the date they started, the date they stopped (or that they are still enrolled), and where they came from. Invoices can stand in for the first two if enrollment records are patchy: the first and last month a student was billed tells you the same story.
For the funnel you need enquiries, trial lessons attended and enrollments, ideally by source. If enquiries currently live in a chat app, the first job is simply to count them in one place for a month; you will learn something from the count alone.
Do this once a term. It takes an hour when the data is in one place, and it replaces the monthly argument about whether marketing "worked" with a number you can check a channel against.
When enrollments, attendance and invoices already sit in the same system, the cohort table is a report rather than a weekend with three spreadsheets. That is the part Schoolory is built for.