Your churn and retention, and what a year keeps.
Put in last month’s member counts and see monthly churn, member retention, what a year at that pace keeps and the dues that left. The cohort tab follows one month’s new members to day 90. The numbers already in it are an invented example, Alturra Studios’ March; replace them with yours. No signup.
Last month, counted.
Count both ends of the month under the same definition of a member.
Left in, the 18 new members make it 242 ÷ 240 = 100.8%: that is growth, a net member change of +2, not retention. The starting numbers are Alturra Studios, an invented studio, and every figure is arithmetic on the counts in the boxes, not a measurement of your studio.
One month’s new members.
Count from each person’s join date, under the same freeze rule as the month.
Alturra’s 18 March joiners: 16 were members on day 30, 13 on day 60 and 12 on day 90.
A monthly rate is held up by long-standing members, so it hides how quickly new ones leave. Compare each month’s cohort with the ones before it, not with the studio-wide rate.
Wondering where the rest leaks? Try the revenue leak calculator →
How each rate is calculated.
Four rates and one dollar figure, each worked on Alturra Studios, an invented studio: 240 members on March 1 at $169 a month. During March 18 people joined, 12 members canceled and 4 froze, leaving 242 active members and 4 paused on March 31.
Member retention
(members at end − new members) ÷ members at start
(242 + 4 − 18) ÷ 240 = 95.0%
Member retention is the share of the members you started the month with who are still members at the end of it, so the 18 who joined in March come out: they were not members on March 1. Left in, they make it 242 ÷ 240 = 100.8%, which is growth, not retention: 2 more active members than on March 1, a net member change of +0.8%. The 4 paused members count at the end because Alturra’s written rule says a freeze under 60 days is a pause, not a loss; counted as losses, March is (242 − 18) ÷ 240 = 93.3%. Take both counts under the same written definition of a member, and settle how a freeze counts before you calculate, not after.
Monthly churn
canceled ÷ members at start
12 ÷ 240 = 5.0%
Monthly churn is the share of members who canceled during the month, measured against the members you started it with. Churn and retention split the same 240 people into left and stayed, so under the same rules they add up to 100: 5.0% + 95.0%. If your rule counts a freeze as a loss, it counts here too, (12 + 4) ÷ 240 = 6.7% beside 93.3%, and the freeze switch in the calculator moves both at once. A new member who joins and cancels inside the month belongs in neither count. When your pair misses 100, someone is counted by one and not the other, and the check line shows by how many.
A year at this pace
monthly retention to the 12th power
0.9512 = 54.0%
Each month’s losses come from a smaller group than the month before: 240 becomes 228, then about 217, then about 206, and after twelve months about 130 remain. Multiplying skips that. Five percent times twelve is 60% lost, which says 40% remain, or 96 of the 240, and the gap grows as churn rises. Give the calculator a typical month, or the average of your last three or six, never an unusual one.
Cohort retention
still members at day 90 ÷ that month’s joiners
12 ÷ 18 = 66.7%
Cohort retention is the share of one month’s joiners who are still members a set number of days after each person’s join date, here day 90: of Alturra’s 18 March joiners, 16 were members on day 30, 13 on day 60 and 12 on day 90. Over the same three months the studio-wide 95% a month compounds to about 86% (0.953 = 0.857). The monthly rate is held up by long-standing members and hides how quickly new ones leave, so compare each month’s cohort with the ones before it, not with the studio-wide rate.
The dues that left
members who left × monthly dues × 12
12 × $169 × 12 = $24,336
The 12 who canceled paid $2,028 a month between them, so a year of their dues is $24,336. It assumes each would have stayed the whole year, so read it as the size of one month’s leak, not a forecast. With freezes counted as leaving, the 4 paused members join the 12.
Every formula here, with the edge cases that move it, is worked through in the member retention rate formula guide, and the member definitions guide sets out who counts.
All of these rates are counted after members have gone, while the fade shows in attendance weeks earlier. Why members leave covers the four patterns to watch for, and Xyzios puts current members whose attendance is slipping against their own pattern on an at-risk list.
What owners ask about churn and retention.
How do you calculate churn rate and retention rate for a gym or studio?
Monthly churn is the members who canceled during the month divided by the members you had at the start of it. Member retention is the members at the end, minus the new members who joined during the month, divided by the members at the start. A studio that starts March with 240 members and loses 12 to cancellations has 12 ÷ 240 = 5.0% churn; if 18 joined and it ends with 242 active members plus 4 on a short freeze its rule counts as staying, retention is (242 + 4 − 18) ÷ 240 = 95.0%. Counted under the same rules, the two add up to 100%.
Why do new members come out of the retention rate?
Retention asks whether the people you started the month with stayed, and a new member was not one of them. Leave them in and a busy month can read as better than perfect: 242 active members on March 31 against 240 on March 1 is 100.8%, in a month when 16 of the original 240 canceled or froze. That end-over-start figure is real, but it measures growth: 2 more active members than on March 1, a net member change of +0.8%. Keep it beside retention, and follow each month’s joiners on their own as a cohort.
How do you turn monthly churn into an annual retention rate?
Compound it. Subtract monthly churn from 1 and raise the result to the twelfth power: at 5% a month, 0.95 to the twelfth is 0.540, so about 54% of today’s members would still be members in a year. Multiplying says 5% × 12 = 60% lost and 40% left, which is wrong, because each month’s 5% comes from a smaller group than the month before. Use a typical month, or the average of the last three or six, never an unusual one.
Do frozen or paused members count as retained?
Only if your written rule says so, and having one rule matters more than which way it goes. One workable rule: a freeze under 60 days is a pause, not a loss, and it becomes churn if it passes 60 days without a return. The calculator’s freeze switch applies your choice to retention and churn together: counted as staying, Alturra’s March is 95.0% retention and 5.0% churn; counted as leaving, 93.3% and 6.7%. Either way, count paused members as their own group beside active and canceled.
What is 90-day cohort retention, and how do you calculate it?
It follows one month’s joiners on their own and asks what share are still members 90 days after each person’s join date. If 18 people joined in March and 12 of them are members on their 90th day, March’s 90-day cohort retention is 12 ÷ 18 = 66.7%. Over the same three months a studio-wide 95% a month compounds to about 86%, because long-standing members hold the monthly rate up. Compare each month’s cohort with the ones before it, not with the studio-wide rate.
What is a good churn rate for a boutique fitness studio?
A little better than your own last six months, counted the same way each month; your studio against itself is the comparison that tells you something. For a rough sense of scale, in our observation, healthy studios lose about 5% of their members a month and the best about 3%. That is what we have seen, not a target: a rate counted under someone else’s definitions of a member and a freeze cannot be compared with yours line by line. At 240 members one cancellation moves the rate about 0.4 points, so one month is noise and three in a row is news.
Can Xyzios help with churn and retention?
Yes, in two ways: it counts members under your own rules, and it watches the people behind the rate. You decide what counts as a member and whether a freeze counts as churn, the boards follow, and paused members are counted as their own group, apart from active and canceled. Current members whose attendance is slipping against their own pattern land on an at-risk list, and a short save note is drafted for those at High risk; nothing reaches a member until you tap approve. On the Recovered Revenue Ledger, a save is checked at 60, 90 and 120 days and earns one month of dues at each check the member is still active, three months at most. Xyzios works with Mindbody and is an approved Mariana Tek integration.
Churn is a lagging number. The fade is a leading one.