Everyone in this market publishes claims. Almost nobody publishes evidence. This is eight years of one music school’s operating data, measured from the records rather than gathered from a survey.
| Finding | Result |
|---|---|
| Students still enrolled after 12 months | 42.9% |
| Lost enquiries that were lost to no reply | 81.0% |
| Tuition invoices paid late | 22.8% |
| Average class capacity actually filled | 55.8% |
| Student exits recorded as “personal reasons” | 78.0% |
School owners assume they lose prospects on cost. Price accounts for roughly one lost enquiry in sixty.
| Reason the enquiry was lost | Share |
|---|---|
| No reply, the conversation stopped | 81.0% |
| Other | 15.8% |
| Price | 1.6% |
| No space available | 1.0% |
| Too far away | 0.6% |
Conversion by how the enquiry arrived says the same thing from the other side:
| Channel | Share of enquiries | Converted to enrolment |
|---|---|---|
| Walk-in | 4.5% | 48.9% |
| Phone | 3.6% | 42.8% |
| Text message | 63.7% | 15.8% |
| Web form | 14.7% | 10.0% |
| 2.5% | 9.8% | |
| Social media | 10.8% | 6.4% |
The two channels where somebody answers while the person is still there convert around three times better than the channel carrying almost two thirds of the volume. Text messaging is where this market actually enquires, and it converts at a fifth of the rate of a walk-in.
That gap is not about intent. A parent who walks in and a parent who messages at 10pm want the same thing. One of them gets an answer.
| Months since enrolment | Still enrolled |
|---|---|
| 1 | 95.2% |
| 2 | 88.9% |
| 3 | 79.9% |
| 4 | 71.5% |
| 6 | 61.2% |
| 9 | 53.0% |
| 12 | 42.9% |
One in five is gone by month three. Two in five by month six. The steepest single drop falls between months two and three: nine percentage points in one month, more than any other. That is where the novelty wears off, the first billing cycle completes, and the student discovers that learning an instrument is work.
A school keeping 43% of its intake through year one is not failing. That is what the curve looks like.
These figures are corrected for censoring: a student is only counted at month N if they enrolled at least N months ago. Without that correction the numbers look worse than they are, because recent enrolments have not had the chance to stay.
Of all paid monthly invoices, 22.8% arrived late, a median of five days and nine on average when late. But only 1.2% went past thirty days.
The money almost always arrives. It arrives about a week after it should, every month, for one invoice in five. That is a permanent low-grade administrative tax rather than a bad-debt problem, which is exactly why it never gets examined.
Across every open class, the overall fill rate is 55.8%. Four in ten classes are at or below half full, and each one runs with a teacher paid the same as a full one.
Of exits with a recorded reason, 78% say “personal reasons.” Only about a fifth carry anything actionable: motivation 9.3%, money 5.3%, class format 4.3%.
This is the least satisfying finding and possibly the most useful one. Any vendor claiming to tell you why your students churn is claiming more than school records contain.
A multi-site music school running alinaflow as its operating system. Records span eight years; meaningful volume begins in 2023, so every rate above is restricted to 2023 onward. All figures are aggregates computed directly against the production database. No student, family or staff member is identifiable, and no individual record appears in this report.
If you run a school and want to compare your own numbers against these, or you would like the method in more detail, get in touch. We intend to update this annually and to widen the sample.