Gym Member Retention: The Benchmark Guide

Most operators carry a retention number in their head. Ask where it came from and the trail goes cold.

We went looking for the benchmarks that hold up: who published them, what year the data covers, how big the sample was, and whether anyone outside the publisher can check them. Several of the most repeated figures in this industry do not survive that.

What is the current average gym member retention rate?

The current industry average is 66.4 percent annual retention, published in the Health & Fitness Association's 2025 Fitness Industry Benchmarking Report using 2024 data. The sample covers 175 companies and more than 17,000 facilities across 27 countries, including 114 US respondents. Roughly one member in three leaves within a year. The same report puts median revenue growth at 9.9 percent and net membership growth at 5.5 percent.

That figure replaced 71.4 percent, which reports 2015 data and still circulates. If your target came from the older number, you are measuring against an industry that no longer exists.

Your retention rate (the share of members active at the start of a period who are still active at the end of it) is only comparable to a benchmark built the same way. That is the first thing most published numbers get wrong, and the reason two operators can quote wildly different figures for the same business.

Which retention benchmarks are real?

These are the figures we could trace to a named publisher, a stated sample and a data year.

The Association of Fitness Studios breaks its figure down to 80 percent for personal training studios and 73 percent for group exercise. These are the most flattering numbers here and the least documented, so treat them as directional only.

Why is a 2015 number still the industry standard?

Because nothing replaced it for years. IHRSA published Profiles of Success annually through 2018 and 2019, then the series stopped. The 2025 Benchmarking Report is the association's first in-depth study since. That six-year gap is where the old figure kept circulating, copied from post to post with the date quietly dropped.

Note what the shift says. Retention fell about five points between those samples while membership grew: the industry takes in more people and keeps a smaller share. That squeeze is why revenue per member became the lever operators reach for when member growth slows.

Are you reading operator data or consumer data?

This is the distinction that causes the most confusion, and almost nobody flags it. The Benchmarking Report is operator-side: 175 companies reporting their own books. HFA's 2026 US Health & Fitness Consumer Report is consumer-side: roughly 18,000 Americans surveyed about their own behaviour. The two get quoted side by side as though they measure the same thing. They do not.

The consumer data is also newer, covering 2025. It found 81 million Americans held a membership, an all-time high and 5.2 percent up on the year. The most useful line for operators is that 4.6 percent of members did not use their membership at all, an all-time low. The sleeping-member problem is smaller than the folklore assumes, which puts the revenue at risk with people who are showing up and leaving anyway.

The full report sits behind a $299 paywall. Treat any figure attributed to it that you cannot find in the public release as unverified.

Should you trust the claim that half of members quit in the first 90 days?

Treat it as unproven. It is among the most repeated statistics in fitness marketing, and we could not trace it to a primary source. Every citation we followed pointed to another blog post, which pointed to another. No study, no sample size, no year. Early churn is real and worth managing, but you should not set an onboarding budget against a number nobody can source.

The best-documented finding about early behaviour is duller and more useful: contract length predicts survival better than almost anything else. One Million Strong, which analysed over a million membership records, found roughly 75 percent of members on 12-month agreements still active at 24 months, against about 35 percent of month-to-month members. That is a 40-point spread on a single structural decision.

Be careful reading causation into it. Members who commit to a year are different from members who will not, so the agreement partly measures intent rather than creating it. The operators who get the most from it build the habit early instead of leaving the contract to do the work, which is the pattern behind the South Korean result where habit beat discounting, and why an app that keeps 3 percent of users at day 30 is not a retention plan.

What should you actually measure?

Compute your own number before comparing it to anyone else's. Take the members active on day one, count how many are still active on the last day, and exclude everyone who joined in between. Including new joins is the most common error, and it inflates the result by quietly measuring growth instead of retention.

Then split the leavers. Members who cancelled deliberately and members whose payment failed are two different problems with two different fixes, and a blended number hides both. Track them apart, hold the method steady, and compare yourself to 66.4 percent rather than to a figure from 2015.