Cancellations Are Up. Here is How to Find the Leak

By OSJJ

Open Source Jiu-Jitsu

Cancellations Are Up. Here Is How to Find the Leak.

A working diagnostic for the owner whose cancellations have climbed and who does not yet know why. Where the leak usually is, how to tell which one it is from numbers you already have, and what to change first.

Justin Hall / Co-Founder, Open Source Jiu-Jitsu

For a long stretch your academy lost about five percent of its members a month, maybe a little more some months, and the new sign-ups covered it with room to spare. Then at some point this year the cancellations started climbing, and now you are at ten, eleven, twelve percent, and the sign-ups are not covering it anymore. Revenue is flat or sliding while the front desk is as busy as it has ever been. You have a feeling about why. The feeling is probably wrong, or at least incomplete, because the feeling is almost always the most recent complaint you heard rather than the pattern underneath it.

This piece is a diagnostic. Not a theory of retention, and not a list of twenty things to try. It is the sequence of questions I run when an owner tells me their churn, the share of members who cancel in a month, has doubled, in the order I run them, with the number that answers each one. Nearly all of those numbers you already have, sitting in your membership software, unread.

One thing before the sequence. None of it works if you are not already tracking the numbers, and in a lot of academies nobody is. The owner does not know churn is high because the owner has never measured churn. The piece on The 11 KPIs lays out the full set; for this problem the two that matter are cancellations and holds, counted every month against paying members. Track them and a retention problem stops being sudden. Churn does not go from five to ten in two months. It goes five, six, six and a half, seven, and an owner who is watching catches it at six and a half, when the cause is still one thing and the fix is still small. An owner who is not watching finds out at ten, from the bank account, and by then the cause has had a year to compound.

First, make sure it is real

Churn is a rate, and a rate has a numerator and a denominator, and owners get both wrong. The numerator is cancellations in the month, not holds. A hold is a warning, and we will come back to it, but it is not a loss yet, and mixing the two inflates the number. The denominator is active members at the start of the month, not whoever happens to be in the system. If your software counts a member who has not paid in three months as active, your denominator is lying to you in the other direction.

That is not a hypothetical. An academy I pulled the numbers on recently showed 273 active members in its software. One hundred and fifty-five of them were paying. Another 110 were on freeze. Of the 91 people who had stopped paying since spring, 48 were sitting in freeze, 7 were recorded as cancellations, and the rest had simply lapsed with no status at all. The owner's cancellation report said he had lost seven members. Most of those freezes were not members who had asked for a pause. They were front desk staff, since gone, filing cancellations as freezes and leaving lapsed accounts active, for reasons nobody there can fully reconstruct now. Run churn against 273 and it looks like a healthy school with a small problem. Run it against the 155 who were paying and it is nine percent, and the problem is the business. At five percent churn instead of nine over the same stretch he would have had about seventy-five more paying members, which at his dues is close to ten thousand dollars a month. None of that was visible from the number he was looking at, because the system underneath it was dirty.

That is more common than it should be. Memberships on the wrong plan, people frozen who should have been cancelled, people cancelled who are still training, accounts that have not billed in months and nobody noticed. Clean it first. Then run it clean for the last six months. If the honest number is seven or eight, you have a problem worth attention but not an emergency, and the cause is usually one thing. If it is ten or above and has been for three months, something structural has changed, and the rest of this piece is for you. One month at ten is noise. Three is a trend.

Second, look in the mirror

Before you pull another number, decide whose problem this is. Every owner I have had this conversation with starts in the same place: the economy, the new gym down the road, a bad run of injuries, people moving away, the time of year. Some of that is real. A move, an illness, a job that changed, a life that changed: those happen, you will recognize them when you see them, and they are outside your control. But they are a point on your churn, maybe two. They do not take an academy from five percent to ten. Nothing outside your four walls doubles your cancellations. Something inside them does.

So if the number is real, the first step is not a tactic. It is accepting that a member who signed up and then left did so for a reason that lives in your building, and that you are the one who has to find it and fix it. That is harder than it sounds, because the owner built the building, and most of what is wrong with it was once a decision the owner made or a decision the owner never got around to making. The owners who get out of ten percent are the ones who stop explaining it and start investigating it. The ones who keep blaming the outside stay there, and nothing in this piece or any other will help them, because every finding will be filed under someone else's fault.

What follows assumes you have made that decision. Each section tells you where inside the four walls to look, depending on when people are leaving.

Third, find out when they leave

This is the single most useful thing you can do, and it is rarely done. Pull every cancellation from the last six months and, for each one, write down how many months the member had been training when they left. Then sort. You are looking for a cluster, and it is usually in one of four places: the first two months, months two to six, six months to about two years, or after two years, which for most rooms means the blue belts. Each is a different door, with a different reason behind it and a different fix. The month does not tell you whether the problem is inside; you have already settled that. It tells you which part of the inside to look at, and the sections after the holds take the three doors one at a time.

If your cancellations are spread evenly across all three, you do not have one leak, you have a general drift, and the answer is the whole system rather than any one part of it. But that is uncommon. Usually there is a cluster, and the cluster tells you which door people are walking out of.

Fourth, read the holds

Holds are the leading indicator. A member who asks to pause has already half-decided, and a large share of real holds become cancellations within ninety days. So the hold count three months ago predicts the cancellation count today, and the hold count today is telling you what the next quarter looks like. This assumes the holds are real, members who asked to pause, which is why the system has to be clean before this number means anything.

Two things to do with that. First, chart holds against cancellations with a three-month offset and see whether they move together. If they do, you have a reliable early warning, and you also know that your hold process is where the retention work should happen, because that is the last moment you have the member's attention. Second, call every person currently on hold. Not text. Call. Ask what would bring them back, and listen for the pattern across the calls rather than the content of any one. Ten hold calls will tell you more about your academy than ten cancellation forms, because the person on hold still thinks they might come back, so they tell you the real reason instead of the polite one.

Here is the honest shape of the work. Tracking the numbers is easy. Finding which window people are leaving in is easy once you track them. Finding the reason inside that window is not, and there is no formula for it. That is the investigative part, and it is the part that takes real time and some real willingness to look at your own building without flinching. What follows for each window is not the answer. It is the list of what I have seen most often, in that window, in the academies I have been inside, offered as the places to look first and the questions to ask. Your answer is in there somewhere. Nobody can tell you which one it is from outside.

The four doors. Which window the cluster falls in tells you where inside the four walls to look.

If they leave in the first two months

They came, they did a few weeks, they stopped. Whatever the form says, this is a gap between what they were told and what they got, and that gap opened during the sale or the first month. The billing was not explained properly and the second charge surprised them. They were never really qualified and the price was outside their budget from the start. The schedule was not walked through and the classes they could make turned out not to be the ones they needed. The first classes were the wrong classes, nobody learned their name, or they were put in a room that was too far ahead of them. Different details, same failure: the expectation set at the front desk did not match the reality on the mat or on the bank statement, and they left before anyone noticed.

The test: pick five people who cancelled inside two months and, for each, write down what they were told about price, billing, schedule and what their first month would look like, and then what actually happened. Then write down who was responsible for their first month and what that person did. If the answers are vague on the first half and nobody and nothing on the second, you have found it.

The fix is a process, and it works fast. A sales conversation that qualifies honestly on budget and schedule, explains the billing before the card is run, and walks the member through which classes are theirs. A written intro process. A named person responsible for every new member's first month. A check-in at day seven and day thirty. The right first classes, in a room where they can keep up. New members leave when the promise and the product do not match and nobody is watching closely enough to catch it. Make the promise accurate and make sure someone is watching.

If they leave between month two and month six

They stuck around long enough to be a real member, then drifted. This is the window many owners are living in when they call me, and the cause is the program itself: what they are being taught, in what order, toward what, and whether they can see and reach the next level from where they stand. It shows up as "schedule" or "money" on the cancellation form, because nobody writes "I did not know what I was working toward" on a form. Check these in order.

There is no pathway to the next level. The curriculum may exist, but the member cannot see where they are on it. Nobody has told them what a first stripe requires, how far off they are, or what they should be able to do by month six. Without that, progress is invisible, and when progress is invisible the only reason left to come is that they enjoy it, which does not survive a busy month at work. The test: are the requirements for each stripe and each belt written down, and does the member know where to find them? If the only person who knows the criteria is you, there is no pathway, there is only your opinion.

The next level is behind too many barriers. The pathway exists and the member still cannot get there. The fundamentals class they are required to pass through runs at one time slot and it is one they cannot make. Promotion needs an attendance count that a twice-a-week member cannot hit. The test is held twice a year and they missed it. Each rule made sense on its own; together they make the next level unreachable for an adult with a job. The test: take a member who trains twice a week at six in the evening and map, on paper, whether they can actually reach the next level on your schedule. If the honest answer is no, that is a schedule problem wearing a curriculum's clothes.

The schedule has a hole in it. Match your cancellations to the class times they attended. If a cluster trained at one slot, the problem may be that slot: the coach, the mix of levels in it, the time itself. This one is cheap to check and cheap to fix, and it gets missed because owners look at churn as a whole-academy number instead of a per-class one.

The fix is to build the pathway. Write down what a member learns in their first six months and what each stripe and belt requires, put both where the member can see them, and check that a member on a normal schedule can actually reach the next level under your rules. This is the slowest fix and the one with the biggest payoff, because it is the one that was costing you the most members for the longest time without showing up on any form.

If they leave between six months and about two years

They were committed. They had a locker, a regular partner, a spot on the mat they stood in every night, and most of them were still white belts, somewhere between their second stripe and the blue belt they could almost see. Then they left, and it was not a move or an injury. These are the hardest cancellations to take and the easiest to misread, because the owner assumes a committed member leaves for a big reason, and often it is a quiet one that has been there for months.

The fix here is less a system than a habit. Every hold gets a call. Every price change gets explained in person before it hits. Every promotion gets a conversation about what comes next, including for the people who did not get one. And once a quarter, look at the room and name who is on the outside of it, then do something about that before they do. These members wanted to stay. Something made it easy to leave.

If they leave after two years: the blue belt

Everyone in this industry has a saying about blue belts quitting, and part of the reason is fair. For most rooms the white belt took a year to eighteen months. The purple belt is going to take three more, sometimes four. Not everyone who signed up for a hobby is going to sign up for a five-year project, and the further up the belt system you go, the narrower the pyramid gets, on time and probability alone. You will always have more white and blue belts than anything else, and you should.

But there is a difference between a pyramid and a wall. If your academy has been open ten years and the room is still white and blue belts with a handful of purples and browns, that is not probability, that is a problem, and the problem is almost always the same one: nobody can see the path from blue to black. The member is told, in effect, to keep training for the next six to eight years and it will come. That does not fly for most adults, and eventually they move on. You cannot make most people stay five years. You can maximize how many do, and the way you do it is a path to purple, brown and black that is defined, attainable, reasonable, and written where the member can read it, the same thing the two-to-six-month section asked for, carried up the belts.

One more thing about this window, and it matters for the diagnostic. If your churn went from five to ten, it almost never came from here. You do not get a sudden exodus of people who have been with you three, four, five years unless something happened that you already know about, and when purples and browns leave in numbers there is always a story, and you are usually in it. A doubling comes from the earlier windows, from white belts and new blue belts, and there the story is the program. So read this section as a long-term retention problem to fix on its own timeline, not as the place to look for the leak that opened this year.

Fifth, reread the cancellation reasons knowing they are polite

An academy with a cancellation form tends to have a stack of them that says schedule, money, and moving. Those are real sometimes. Often they are the polite version. Schedule frequently means the classes I could make stopped being worth making. Money frequently means it stopped feeling worth it, which is a value judgment, not a bank balance. Treat the reasons as a weak signal that confirms the timing data, never as the diagnosis on their own. The diagnosis is in when they left and what was happening inside the four walls when they did.

Then wait

Change one thing, the thing the cluster points at, and wait three months before you judge it. An owner who changes five things at once will not know which one worked, and will usually stop all five when the number does not move in a month. Churn is a lagging number. The changes you make this month show up in the cancellations of the month after next, and the owner who panics at sixty days and reverses course has learned nothing and spent a quarter doing it.

And be honest with yourself about the timeline. Your cancellations did not double in a month. They doubled over six months or a year, a point at a time, while nobody was counting. So it is unlikely to be something you fix in a month either. Sometimes it is: a hole in the schedule, a billing conversation nobody was having, one coach, one time slot. If you find something that simple, change it and enjoy it. But go in prepared for the other case, which is that it takes close to as long to get from ten back to five as it took to get from five to ten, and that the number comes down the way it went up, a trickle rather than a switch. That is not a reason to wait. It is a reason to start now and to keep tracking, so that six months from now you can see the trickle going the right way.

The honest part

Ten percent is not a mystery and it is not bad luck. It is a specific group of people leaving through a specific door for a specific reason. The numbers to find the group and the door are already in your software, or will be a month after you start tracking them. The reason takes longer, because it is inside the building, the owner built the building, and looking in the mirror is the one step no diagnostic can do for you. Do that step first. Track the two numbers so it never surprises you again. Then go find the door. The members who left did not tell you. The numbers will.

Open Minds. Open Mats. Open Source.

Back to the Free Library