← All notes Retention

Member Retention for Studios: Why Members Leave

A churn rate tells you how much. Never who.

Every studio owner can produce a churn rate. Divide the month's cancellations by the members you started it with, and there it is: 5.0%. A number, on a slide, in a color.

Then what?

That is the trouble with churn as a measure of member retention. It reports how many people left after every one of them is gone, weeks after each of them decided, and says nothing about which ones you could have kept. You cannot phone a percentage.

The member retention rate formula shows how to work the number out honestly, but the useful question is who is fading, and why. One invented studio runs through this piece: Alturra Studios, 240 members on March 1 at $169 a month. In March, 12 of them canceled: the 5.0%. Read back through the booking data, 11 of the 12 showed weeks before the email. The twelfth moved away, and no list would have kept her.

What the report says5.0%12 canceled ÷ 240 members on March 1how many, never who
The same 12, read back one by one
6frequency fades
3never started
2one room, after a change
1moved away, not a fade
11 of the 12 showed weeks early
Alturra · March · the same 12 peoplethe rate says how many; the calendar says who
worked example on an invented studio · 6 + 3 + 2 + 1 = 12 · 12 ÷ 240 = 5.0% · 12 × $169 = $2,028 a month

Member retention in one minute

  • Nobody quits cold. The decision shows in the booking calendar weeks before the email; the churn rate reports it weeks after.
  • Four fades, four fixes. A regular's visits thin, a card fails, a new member never starts, or one room changes and several people drift.
  • Compare each member with herself, every Monday. A gap twice her usual is the signal; a thirty-minute review gives each name a person and a day.
The fadeHow it showsWhat it meansWhat to doAlturra's March
Frequency fadea regular's visits thin; the gap since her last class doublesthe habit is losing to something elsea note from a person, naming her class6 of 12
Failed paymentnothing in attendance; a payment did not go througha card problem, not a decisiona kind note the same day0 of 12 · 9 cards failed
Never startedno week yet with two classesno routine formedbook the second class before she leaves3 of 12
Quiet dissatisfactionseveral regulars thin out of one slot or teachersomething changed in the roomfix the room, not the people2 of 12
Not a fadenothing, until she tells youa real decision: a movelet her go kindly1 of 12

Why do members quit?

Rarely because of one event, and rarely on the day they cancel. The cancellation email feels like the event. It is the paperwork.

The decision came weeks earlier, and usually not as a decision at all: a new job with an earlier start, a knee that twinged, two weeks away that never quite restarted. The habit weakened, the visits thinned, and coming back began to feel like starting over.

A member decides with her calendar long before she decides with her email.

That calendar is data you already hold. Rachel, one of Alturra's twelve, took three classes a week for two years. Here are her last twelve weeks.

Rachel · classes each week, her last twelveUnlimited · $169 a month
3
1
3
2
3
3
2
4
2
5
1
6
1
7
0
8
1
9
0
10
0
11
✕
12
before the flagfrom week 6: a gap past twice her usual✕week 12: the cancellation email
Six weeks between the first sign and the email. The churn report counted her in week 12.
worked example on an invented studio · classes a week, bars drawn from zero · 3 + 3 + 3 + 2 + 2 + 1 + 1 + 0 + 1 + 0 + 0 = 16 in eleven weeks

The four fades, and what each one is telling you

Different reasons leave different fingerprints. Each of these four can be seen, and answered, while she is still a member.

1. The frequency fade

What it looks like: a three-times-a-week regular at two, then one, then two weeks gone. No complaint, and nothing in a report of totals.

What it means: the habit is losing to new hours, travel or a hard season. She has not decided against you; she has stopped turning up.

What to do: reach her while the gap is fresh, from a person, naming the class she used to take. Six of Alturra's twelve were frequency fades, each visible four to eight weeks before she canceled.

2. The failed payment

What it looks like: nothing in attendance at all. A happy member stops being one because a card expired or a bank reissued it.

What it means: not a decision, which is why it gets its own section below. The fix is a kind note the same day: no persuasion, just a working card.

3. The member who never started

What it looks like: she joins, comes once or twice, and is rarely seen again. She may not cancel for weeks, until the next payment makes her think about it.

What it means: onboarding, not retention. There was no habit to lose, and a new member's first week decides whether one forms.

What to do: give the first six weeks their own routine: the second class booked before she leaves the first, and a call if nothing is booked. Three of Alturra's twelve joined in January or February and never had a week with two classes.

4. The quiet dissatisfaction

What it looks like: several regulars drift away from one slot, teacher or time at once. Read one member at a time, it is invisible.

What it means: something changed in the room: a schedule move, a teacher who moved on, a class that got too crowded. It is the fade most likely to be about you, and the one an exit survey misses.

What to do: look across the list, not down it. Alturra's Wednesday 7pm changed teachers in January. Two of March's twelve were its regulars, and three more are fading. That is one room, not five decisions, and no note fixes a room.

Wednesday 7pm · attended averagea room of 20 · new teacher from January
December
16 of 20
January
15 of 20
February
13 of 20
March
11 of 20
December's nine Wednesday regulars, on March 31
4still most Wednesdays
3fading now
2canceled in March
Across about 260 classes a month the studio average barely moves off 12. Read slot by slot, the room shows it.
worked example on an invented studio · bars drawn from zero on a room of 20 · 4 + 3 + 2 = 9 regulars

Talk to the regulars who still come, then decide on the room: its time, its format, who teaches it. Reading fill by slot is how you find rooms like this one.

The fades wear different clothes in different rooms. Yoga students drift with classes left, a barre studio's cliff comes weeks after a challenge, a CrossFit gym's on-ramp decides the year, and an access gym has only check-ins to go on.

How do you spot a member who is about to cancel?

Compare her with herself, and watch the gap since her last visit. When it reaches twice her usual gap, the fade has started, whatever the rest of the studio is doing.

One studio-wide rule cannot see that. "Anyone quiet for two weeks" is days late for a three-times-a-week regular. Here are three Alturra members on one Monday, each nine days since the last class.

One Monday · nine days since each last classdays quiet, on a 14-day scale
Priyausually 3 a week · usual gap 2.3 days3.9× her usual
Claireusually 2 a week · usual gap 3.5 days2.6× her usual
Jordanusually once a week · usual gap 7 days1.3× his usual
twice the usual gappast it: fadinginside it: normal
A two-week rule · flags nobody todaymisses 2
Each one's own pattern · flags Priya and Clairecatches 2
worked example on an invented studio · usual gap = 7 ÷ classes a week · 9 ÷ 2.3 = 3.9 · 9 ÷ 3.5 = 2.6 · 9 ÷ 7 = 1.3

Watch the gap, not the count. In a monthly total, four classes in the first week and silence since looks like one a week. Priya passed twice her usual gap four days ago; a two-week rule would find her five days from now.

Write down each member's usual once. Take a normal stretch before any fade: classes a week, and the gap between them. The rest is arithmetic.

Give every name a reason. "Priya, 78% risk" tells the desk nothing. "Priya: three a week for two years, nine days quiet, the Monday 6pm" tells them what to say.

Is a failed payment churn?

No. Count it in its own column. A card that did not go through is not a member deciding to leave, and counting the two together flatters your billing and slanders your classes.

It is also the most fixable fade, when the chase takes days rather than weeks. On March 1, nine of Alturra's membership payments did not go through.

Chose to leave12canceled in March · $2,028 a month
frequency fades6
never started3
one room, after a change2
moved away · Lena1
Card did not go through9on March 1 · $1,521 of dues at stake
collected by day 217
paused until April 15 · Marcus1
moved away · Lena, counted on the left1
Alturra · Marchmembers lost to a card: none
worked example on an invented studio · 6 + 3 + 2 + 1 = 12 · 7 + 1 + 1 = 9 · 9 × $169 = $1,521 · 12 × $169 = $2,028

A kind note went to each that day, and seven were collected within 21 days. Lena was moving and canceled; her card failing that week was a coincidence, so she counts once, as a cancellation. Kept apart, the columns answer two questions one rate blurs: are members choosing to leave, and is the card chase working? The failed payment email templates have the words, and the seven-day failed-payment playbook has the schedule.

What does a weekly fade review look like?

Thirty minutes every Monday, four lists, and a person beside every name. The lists are the four fades; the person is whoever reaches that member, and by what day.

Weekly matches the speed of a fade. Rachel's first sign came six weeks before her email, and a monthly review would have caught it once at best. Here is Alturra's sheet on the first Monday in April.

Fade review · the first Monday in Aprilthirty minutes · nine names
Below their own pattern
Priyawas 3 a week · 9 days quiet
Sam · Tue · a note
Clairewas 2 a week · 9 days quiet
Sam · Tue · a note
Tomwas 2 a week · 3 classes in 30 days
Ellie · Wed · a note
New, not started
Leojoined March 16 · 1 class since
Sam · Wed · a call
Avajoined March 23 · 1 class, nothing booked
Sam · Wed · a call
Card
OmarApril 1 payment still open after the note
Sam · today · a call
One room
Rosa, Kate, BenWednesday 7pm regulars, all thinning
owner · this week · the room first
9 names · $1,521 a month of duesnext Monday: who came back?
worked example on an invented studio · 3 + 2 + 1 + 3 = 9 names · 9 × $169 = $1,521 a month

What to do this week:

  1. List the regulars below their own pattern: a gap twice the usual, or a last 30 days at half a usual month or less.
  2. Add the cards: every membership payment still open.
  3. Add the new members who have not started: joined in the last six weeks, with no week yet of two classes.
  4. Look across the list for one room: if several names share a slot or a teacher, look at the room before writing to anyone.
  5. Give every name a person and a day: a note naming a class, a call for a new member, a note or a call for a card, all by Friday.
  6. Next Monday, count first: mark who came back to a class before you build the new list.

Count modestly: some would have come back anyway, so claim only what you can confirm months later. What one member is worth shows why a careful count claims less.

What should you say to a member who is fading?

Name the class she used to take, invite her back to it, and ask nothing else. Two or three lines from a person, sent early, while coming back still feels like the next class. For Priya:

Hi Priya, Leah's Monday 6pm has space this week. Want me to put you in? Sam

No discount, which says the problem was the price, and nothing about her absence: no "we noticed you've been away", no "we miss you". For a new member who never settled in, call instead and book the next class while you are on the phone.

Once she has canceled, it is a different job. A call while her access is still running can offer a freeze with a return date or a smaller plan; after that, winning back a former member has its own timing.

Where this lives

A weekly review needs lists that stay true without anyone rebuilding them, which is the part a spreadsheet loses by the third Monday. It is what Xyzios, the studio operating system, does for this job. It works with Mindbody and is an approved Mariana Tek integration, and the booking platform stays your system of record.

It flags current members whose attendance is slipping against their own pattern, on an at-risk list with High and Medium bands, and drafts a short save note for those at High risk. When a membership payment fails, it sees the status, not the reason, and drafts a kind fix-up note that assumes an innocent cause; your platform's own retries carry on, and Xyzios never retries a card. Nothing reaches a member until you tap approve, and approved notes send from your studio's own email.

At risk · against their own patterncurrent members
Priya S.was 3× a week · 9 days since her last class
Highsave drafted · approve?
Tom R.was 2× a week · 3 classes in 30 days
Highsave drafted · approve?
Claire D.was 2× a week · 9 days since her last class
Mediumflagged · no draft
Rosa P.Wednesday regular · 1 Wednesday in 5 weeks
Mediumflagged · no draft
Every name here still has a membership. High gets a drafted save note; nothing goes until the owner taps approve.
recreation · demo data · the save happens while she is still a member · only High is drafted

The Front Desk board lists today's first-timers to greet by name, and the Programming board shows the attended average by slot, which is where to look for a room like the Wednesday 7pm. Members who canceled in the last 30 days without switching plans go on the Sales Desk call sheet for a person at your studio to call; Xyzios does not write to people who have already canceled. Each approved save is checked on the Recovered Revenue Ledger at 60, 90 and 120 days, crediting a month of dues at each check her membership is still active, three months at most. How the save works and how payment recovery works have the detail.

Whatever you run it in, start this Monday. The churn rate calculator shows what a point of monthly churn costs over a year, and the two-minute studio check shows where retention sits among the numbers that decide growth.

Straight answers

Common questions.

Why do gym and studio members quit?

Rarely because of one event, and rarely on the day they cancel. Usually a habit weakens first: a new job with an earlier start, an injury, a schedule change or two weeks away that never quite restarted. The visits thin until coming back feels like starting over, and the cancellation email arrives weeks after the decision. Some members never built the habit at all, some drift away when a class they loved changes, and some are lost to a card that expired, which is not a decision at all.

How do you spot a member who is about to cancel?

Compare each member with her own pattern, not with the studio average. Work out her usual gap between visits, then watch the gap since her last one: when it reaches twice her usual, the fade has started. A three-times-a-week regular who has been quiet for nine days is fading; a once-a-week member quiet for nine days is not. A single studio-wide rule, such as anyone quiet for two weeks, is days late for the first and raises false alarms for a member who only ever came every other week.

Does a failed payment count as churn?

Count it separately. A membership payment that did not go through is usually an expired or replaced card, not a member choosing to leave, and counting the two together hides both problems: churn looks worse than the decisions members made, and nobody notices when the card chase is not happening. It is also the most fixable loss: a kind note the same day that assumes an innocent cause, and a call if it is still open a few days later. If a member with a failed payment then cancels for another reason, count her once, as a cancellation.

What is member retention, and how is it different from churn?

Member retention is the share of the members you started a period with who are still members at the end of it; churn is the share who left. Counted under the same rules they add up to 100%, so a studio that loses 12 of its 240 members in a month has 5.0% churn and 95.0% retention. Neither number says who is leaving or why. That is the weekly work: finding the members behind the rate while they are still members.

How can a fitness studio improve member retention?

Work the fades, not the rate. Once a week, list the regulars whose gap since their last visit is twice their usual, the membership payments that are still open, and the new members who have not yet had a week with two classes in it. Look across the list for a shared slot or teacher, then give every name a person and a day. Chase cards the day they fail, give new members a first week that books the second class, and count who came back before you build the next list.

What should you say to a member who has stopped coming?

Name the class she used to take, invite her back to it, and ask nothing else. Keep it to two or three lines from a person, signed with a name. Leave out any discount, which suggests the problem was the price, and any line about her absence, such as “we noticed you’ve been away” or “we miss you”. For a new member who never settled in, a two-minute call that books the next class does more than a note.

Is it worth contacting a member who has already canceled?

Yes, but it is a different job from reaching a member who is fading. In the first days, while her access may still be running, a call can offer a freeze with a return date or a smaller plan. After that, a win-back works best one to three months after the membership ends, sorted by how she left and naming what has changed. The cheap, repeatable saves happen before the cancellation, not after it.

Can Xyzios tell me which members are about to leave?

It flags them, and you decide what happens next. Xyzios works with Mindbody and is an approved Mariana Tek integration. It flags current members whose attendance is slipping against their own pattern, on an at-risk list with High and Medium bands, and drafts a short save note for those at High risk; when a membership payment fails, it drafts a kind fix-up note. Nothing reaches a member until you tap approve, and approved notes send from your studio’s own email. It does not write to people who have already canceled.

Try it on your own numbers

Every board in this piece is a screen in the product.

Shown here on demo data. Connect Mindbody or Mariana Tek and it fills with yours. The first month is $99: the whole OS, every board, hands-on onboarding.

No card today · billing starts when you're connected and say go

Every number, one screen. The studio grows.

Free templates

Nine studio templates,
free to your inbox.

Two policies, notes for members and leads, a price increase letter, and four sheets for the desk and the numbers, each cut down to the part you copy. After that, the occasional studio note. No spam, no drip campaign, unsubscribe any time.

Written by the people who build Xyzios — not a content farm. What each template is for →