Running a school
What one student actually costs a language school
Published 16 August 2026
Every school knows what a student pays. Far fewer know what one costs to keep, and the gap between those two numbers is where a full timetable quietly turns into a thin year.
The four costs behind one seat
A student is not a price with nothing behind it. Four costs sit under every seat, and most schools measure only the first one.
- Delivery - the teaching hours the student sits in. In a group this cost is shared between everyone in the room; one-to-one, it is carried by a single person.
- Preparation and marking - the hours nobody is in the room for. Building the lesson, reading what came back, writing the feedback that makes the next lesson worth attending.
- Acquisition - what it took to bring this student to a first paid lesson: the ad, the free trial, the hour someone spent answering questions before any money changed hands.
- Administration - scheduling, rescheduling, invoicing, chasing the payment that did not arrive, and putting the same information into two systems that do not talk to each other.
The unit that pays the bills is the seat-hour
Per-student revenue is the wrong lens, because the cost that matters does not scale with students - it scales with hours taught. A teacher paid for ninety minutes costs the same whether four people or eight are in the room. So the question is never “is this student profitable?” but “what does this hour of teaching earn once everything under it is paid?”
That reframes a lot. A group with two empty seats is not a group that earns slightly less; it is often a group that earns nothing at all, taught by someone who is doing the work perfectly well.
The half of the job nobody bills for
Preparation and marking are real hours with a real price, and they are the ones schools account for last. A teacher who spends forty minutes around each ninety-minute lesson is working a third again as long as the timetable says.
This cost does not go away by being ignored - it comes out of someone. Either the school pays for it, or the teacher absorbs it, and a teacher absorbing it is a teacher who leaves in eighteen months and takes their groups' goodwill with them.
A student who leaves in March costs more than one who never joined
Acquisition is paid once and recovered slowly. Spend sixty euros to win a student who stays eight months, and that cost lands as under two euros on every lesson they attend. The same student leaving after three months carries five euros a lesson instead, and the lessons they never took carry the rest.
Retention is therefore not a satisfaction metric. It is the single cheapest way to change unit economics, because nothing else lowers a cost you have already paid.
Where the money actually leaks
- Seats sold once and then left empty when a student pauses, while the group runs at full cost.
- Late cancellations that free nobody: the teacher is paid, the room is booked, the hour is gone.
- Trial lessons that are taught properly, cost a full hour, and are never counted against the students they win.
- Hours taught but never invoiced, usually the ones rescheduled by hand into a different week.
- The same lesson recorded in a spreadsheet, a calendar and a chat thread - three places to update and three chances to be wrong.
One group, one lesson, in numbers
Take one ninety-minute group lesson with six seats filled, at twelve euros a seat. The figures below are illustrative - the shape is what matters, and it is the shape most schools recognise.
| Per lesson | Amount |
|---|---|
| Six seats at €12 | €72.00 |
| Teacher, 1.5 h at €22 | −€33.00 |
| Preparation and marking, 30 min | −€11.00 |
| Platform, admin and overhead | −€8.00 |
| Acquisition, spread over an eight-month stay | −€11.25 |
| What the hour leaves | €8.75 |
Illustrative figures for one 90-minute group lesson, six seats filled, a €60 cost to win each student and an average stay of eight months.
Eight euros seventy-five, on seventy-two euros of revenue. Now take two students out of that room and change nothing else: the same lesson, taught just as well, loses eleven euros fifty. Two seats are the whole margin.
And lengthen the average stay from eight months to twelve, with everything else untouched: acquisition drops to seven euros fifty a lesson and the hour leaves twelve fifty instead of eight seventy-five. Keeping students longer paid more than any price rise on this list.
What to change first
- Measure the hour, not the head. Work out what one taught hour leaves after teacher pay, preparation, overhead and acquisition, and you will know which groups are worth running before the term does.
- Cut the minutes nobody is paid for. Every re-entry of a lesson into a second system is money leaving quietly, and it is the cheapest cost in the whole list to remove.
- Defend the first three months. Students leave early for reasons a school can see - a lesson that did not land, homework nobody read, a schedule that kept moving - and each of those is fixable long before the notice arrives.
Where the numbers live
None of this is an argument for one more spreadsheet. It is an argument for the timetable, the homework and the invoices sitting in one place, so a taught hour cannot be untracked and a payment cannot go unrecorded. That is what EdTechCore is: the lesson on a shared board, the tasks that check themselves, and a double-entry ledger behind the school - one record, not three.