Your Students Start Leaving Before They Cancel: Building an Attendance Early-Warning System
The cancellation email is not the moment a student decides to leave. New data from a studio-software company says the decision shows up in the attendance record three to eight months earlier, while the student is still training and still paying. We reported the figures in this week’s news roundup. This article is about what to do with them.
What the data says
Mariana Tek, which sells booking and membership software to boutique fitness studios, published an analysis of its own client data in late October. Its director of marketing, Julie Sippy, put the finding in one sentence: “The membership change is often the end of a slow, subtle disengagement trajectory, not the beginning of one.”
| Member group | What happened before they cancelled or downgraded |
|---|---|
| Members who cancelled | Average monthly visits fell from 6.2 three months out to 5.4 in the final month; monthly spend fell 11% |
| Members under one year | Visits fell 14% in the three months before a downgrade |
| Members over one year | Visits peaked at 7.4 a month eight months before a downgrade, then fell 39% to 4.5 by the month before it; spend fell 24% |
| After the downgrade (over-one-year members) | Visits fell a further 58% and revenue a further 57% in the following month |
Visit frequency was the strongest warning sign in the data, and spend the second. The analysis does not say why members disengaged.
Why a martial arts school should care
A boutique cycling studio is not a karate school, and nobody should quote these numbers as martial arts benchmarks. But the mechanism carries over directly, and if anything it should be stronger in our business. A martial arts student is on a belt path with a known cadence of classes, a known testing calendar and an instructor who knows their name. A school is better placed than almost any fitness operator to notice when a student who trained three times a week starts training twice.
The finding that should make owners uncomfortable is the long-tenure curve. The member most at risk of a quiet exit was the loyal one, the person who had been around long enough that nobody was watching them anymore. In a school, that is the brown belt who has stopped coming on Saturdays, or the adult who has been “busy with work” for six weeks.
A weekly attendance-intervention routine
- Run the trend, not the snapshot. Once a week, pull each active student’s attendance for the last 90 days and compare each month with the one before. Most school software can do this already; check before you buy anything new.
- Set a simple trigger. Flag any student whose attendance has fallen two months running, or who has missed a full week they would normally train. Flag any student who has not booked or attended in ten days.
- Assign the call, by name. Each flagged student goes to a specific instructor, with a date. A phone call or a conversation on the mat, not an automated text.
- Talk about the next goal, not the absence. The conversation is about the next stripe, the next test date, the next thing they are working toward. If something has changed at home or at work, that is when you find out.
- Split the approach by tenure. Newer students usually need a short-term goal and a reason to come back this week. Long-tenured students usually need recognition: a role, a leadership position, a visible milestone.
- Record the outcome. Note who was contacted, what was said, and whether attendance recovered within 30 days. After a quarter you will know which conversations work in your school.
- Review the list every week at the staff meeting. The point is that somebody is looking, every week, and that the list never belongs to nobody.
What a school owner should take from this
The time to save a student is while they are still paying you. If your school only learns that a student is unhappy when the cancellation arrives, you have given away months in which a call, a goal and a stripe could have changed the outcome. Make the attendance trend a weekly report with a name next to every flagged student.
Sources, and what we could not confirm
James Burke, “Mariana Tek: churn signals surface 8 months early”, Gym Business News, October 27, 2026, drawing on Athletech News, October 23, 2026.
This is vendor data from Mariana Tek’s own boutique-fitness clients, not an independent study, and the percentages should be read as directional. We have not seen the underlying analysis: the number of studios and members, the time period, and how “downgrade” was defined are not stated in the reporting. None of the studios in the dataset is identified as a martial arts school, and we know of no published martial arts dataset that measures the same thing.
