Your gym starts the month with 200 active members and finishes with 210. Membership grew, but that does not tell you how many of the original 200 stayed. New joins can hide departures, and freezes can change the answer without changing a single member record.
Gym member retention is the percentage of your opening members who meet your stated retained-member rule at the end of the period. Churn is the percentage of that same opening group who meet your loss rule. Count people first, then calculate percentages.
For an opening group of 200 people, if 185 remain active or on an approved freeze and 15 have left, relationship retention is 185 ÷ 200 × 100 = 92.5%. Churn is 15 ÷ 200 × 100 = 7.5%. If five of the retained members are frozen, active-only retention is lower: 180 ÷ 200 × 100 = 90%.
Those are different measures. Label them so the reader knows which one you mean.
Choose the period and member states first
Take an opening snapshot and a closing snapshot using the gym’s reporting timezone. For a June report, use the start of 1 June and the end of 30 June. Record the extraction date too, because a late correction can change an earlier payment or membership status.
Decide how each status affects the calculation before you view the result:
| Status at closing | Active-only retention | Relationship retention in this example | Separate reporting |
|---|---|---|---|
| Active opening member | Included | Included | Active |
| Opening member on an approved freeze | Excluded | Included | Frozen |
| Opening member whose cancellation is effective | Excluded | Excluded | Lost: cancellation |
| Opening member expired without renewal by the cutoff | Excluded | Excluded | Lost: expiry |
| New member who joined after the opening snapshot | Excluded | Excluded | New joins |
| Unresolved or conflicting record | Unresolved | Unresolved | Unknown |
Here, a freeze preserves the membership relationship but suspends active use. A gym may use another definition, but it should record that definition and keep it consistent between reporting periods. Do not quietly move frozen members out of the denominator to improve the percentage.
If your policy has a renewal grace period, state its length and whether the report is provisional until that period ends. Apply the same rule to all members. A grace period invented after seeing the result makes month-to-month comparisons unreliable.
Work through a complete month
This fictional example is not an industry benchmark. All groups below are mutually exclusive, and every person has one member ID.
| Movement | People |
|---|---|
| Active at the start of June | 200 |
| Opening members with an effective cancellation by closing | 12 |
| Opening members expired without renewal by closing | 3 |
| Opening members on an approved freeze at closing | 5 |
| Opening members still active at closing | 180 |
| New June joins still active at closing | 30 |
The opening group reconciles:
180 active + 5 frozen + 12 cancelled + 3 expired = 200 opening members.
The closing active headcount reconciles separately:
180 active opening members + 30 active new joins = 210 active members.
The closing active-or-frozen population is 215. It is not the same as the active headcount of 210.
| Measure | Calculation | Result |
|---|---|---|
| Active-only retention | 180 ÷ 200 × 100 | 90% |
| Relationship retention, including approved freezes | (180 + 5) ÷ 200 × 100 | 92.5% |
| Churn, cancellations plus expiries | (12 + 3) ÷ 200 × 100 | 7.5% |
| Freeze share of opening group | 5 ÷ 200 × 100 | 2.5% |
| Net active-member growth | (210 − 200) ÷ 200 × 100 | 5% |
Relationship retention plus churn equals 100% here. Active-only retention plus churn does not: the remaining 2.5% is frozen. This is why “churn is always 100 minus retention” is unsafe without matching definitions.
Copy the member-level worksheet
Keep identifiable working records in a restricted system. The owner report needs totals and reasons, not a public list of member names or health details.
| Member ID | In opening group? | Closing state | Relationship retained? | Active retained? | Loss reason |
|---|---|---|---|---|---|
| DEMO-001 | Yes | Active | Yes | Yes | None |
| DEMO-002 | Yes | Approved freeze | Yes | No | None |
| DEMO-003 | Yes | Cancelled | No | No | Member requested |
| DEMO-004 | Yes | Expired | No | No | Renewal not completed |
| DEMO-005 | No | Active | Excluded | Excluded | New join |
These rows illustrate classification, not the full 200-person example. In your complete worksheet, count each opening member exactly once. Reconcile the classified total against the opening snapshot before dividing.
If the opening group is zero, report retention and churn as “not applicable,” not 0% or 100%. If records are unknown, disclose the count and keep the result provisional. Unknown is not automatically retained or lost.
Handle the cases that distort the result
A person cancels and rejoins within the month. In a closing-state report, an opening member active at the closing cutoff is retained. Keep the cancellation and rejoin as separate events for operational analysis. If you instead measure every cancellation event, call that an event count; it is not the same as closing-state churn.
A new join leaves before month end. They do not belong in the opening-group calculation. They do belong in the new-join cohort report. Otherwise rapid losses among new members disappear behind a healthy opening-base percentage.
A member transfers branches. For a whole-business report, a transfer within the business is usually not a loss under a relationship-based definition. At branch level, record transfers separately so relocation is not mistaken for cancellation.
Two accounts represent one person. Reconcile duplicate identities before calculating. Preserve the evidence used to resolve the conflict rather than deleting records to make the totals agree.
An expiry is corrected later. Record a revised report with its cutoff and reason. Do not compare a corrected month with an uncorrected month without explaining the difference.
Keep member retention separate from revenue retention
Member retention counts people. Revenue retention follows revenue from the same opening customer group. A retained member moving to a lower-priced plan changes revenue retention without changing the member count.
For example, an opening group represents 20,000 currency units of normalized monthly recurring revenue. At closing, losses and downgrades reduce that to 18,000, while upgrades add 1,000. Under those stated assumptions, gross revenue retention is 18,000 ÷ 20,000 = 90%; net revenue retention is 19,000 ÷ 20,000 = 95%.
Do not include revenue from new members in either numerator. Do not substitute the cash collected during the month for normalized recurring revenue: an annual payment can make cash receipts jump even when membership is unchanged. Agree the revenue basis with the person responsible for your accounts.
Turn the result into a useful review
Break the losses down by tenure, plan and stated reason. Compare join cohorts at the same age; a group measured after 30 days is not comparable with one measured after 90 days.
Use the first-90-day cohort dashboard guide for new joins, the attendance-change guide to investigate earlier signals, and cancellation interview questions to understand stated reasons.
A later increase in retention is not automatic causal proof that a reminder or campaign worked. Plan changes, seasonality, member mix and service changes can also affect the result. Keep the counts and the reporting rule beside the percentage in your weekly owner scorecard.
Frequently asked questions
What is the formula for gym member retention?
Divide opening-group members who meet your retained-member rule at closing by all eligible opening-group members, then multiply by 100. State the period and treatment of freezes, expiries, rejoins and unresolved records.
Is churn always 100% minus retention?
Only when both measures use the same opening group and period, and retained and lost are the only possible states. If active retention excludes frozen members while churn also excludes them, the two percentages will not add to 100%.
Should new gym members be included in retention?
Exclude new joins from an opening-group retention calculation. Report them separately in net growth and in a join-cohort report measured at a consistent age.
