Why Half Your Gym Members Quit in Year One (And the 6 Signals That Predict It)

Roughly half your members will be gone within a year.

That's not a scare stat invented by someone selling software. It comes from the largest retention dataset the fitness industry has: Dr Paul Bedford's analysis of 342,759 member records across 100 sites, which found 51.9% of members still holding a membership at twelve months. A separate ukactive study of 15,000 members put twelve-month retention at 47%. Different countries, different operators, same answer.

Here's the part that should bother you more than the number itself. Most of those members didn't decide to leave. They drifted, and the drift was visible in your data weeks before anyone cancelled.

This is a guide to reading that drift.

Why the exit reason you're given is almost never the real one

Ask a member why they're cancelling and you'll usually hear that they're moving, or money is tight.

Bedford tested this. In his qualitative study of 1,000 members, 32% told their club they were relocating. The actual annual relocation rate in the population is somewhere around 5%. People weren't moving house six times more often than the general public. They were being polite.

When the same members were asked properly, the real reasons came out differently: the club didn't meet expectations (17%), loss of motivation (15%), boredom with the environment or classes (11%), lack of support from staff (10%).

Notice what those four have in common. Every one of them is something you could have seen happening. None of them appear on a cancellation form.

So if you're running your gym on what members tell you at the door, you're working from a story they've constructed to avoid an awkward conversation. The behaviour data doesn't have that problem.

Signal 1: A gap of more than a week

The single best predictor of cancellation isn't how often someone comes. It's how long they've been away.

A peer-reviewed machine-learning study of 5,209 members tested which variables predicted dropout. Days of non-attendance came out on top by a distance, accounting for 35-54% of the model's predictive power. Membership duration managed 14%. Amount billed, under 18%.

The model's highest-risk decision rule kicked in at a gap of more than 7.5 days.

That's a useful, specific, actionable threshold. A member who normally comes twice a week and hasn't been in for nine days isn't taking a rest week. They're in the part of the curve where people quietly stop.

Most gym owners can't see this, not because they don't care but because a paper register doesn't sort by absence. You can flip through six months of sign-in sheets and still not notice who's missing. Digital check-ins make the question answerable in a glance: who hasn't scanned in for over a week.

Signal 2: A first month with almost nothing in it

The early window matters more than any other, and it's where most gyms are least attentive.

In Dr Melvyn Hillsdon's study tracking 10,000 members, around 30% of members made no visits at all in each of their first three months. Not few visits. None. By month twelve that figure was over 50%.

Those members are paying you and getting nothing, which feels like free money right up until they cancel and tell someone that your gym didn't work for them.

The same research found something more encouraging. In any month a member makes at least one visit, their likelihood of cancelling the following month drops by 20%. For members in their first six months, the effect is stronger: 27%.

One visit. Not a transformation, not a personal best. One.

That reframes what a new-member follow-up is for. You're not trying to get someone to commit to a routine in week two. You're trying to get them through the door once more, because the data says that single visit measurably changes what happens next.

Signal 3: Frequency drifting under once a week

Bedford's research points to roughly four sessions a month as the line where retention changes character, with members averaging about 1.2 sessions a week showing substantially lower cancellation risk.

An independent peer-reviewed study of 3,802 members across two gyms found the same shape from a different direction: training three or fewer times a month predicted dropout, and so did four to six times a month, in both gyms studied.

Two research groups, different countries, converging on the same rough threshold. Somewhere around four visits a month is where a membership stops being a habit and starts being a direct debit.

The practical version: sort your members by visits in the last 30 days and look at everyone under four. That list is your next quarter's cancellations, and you're seeing it early enough to do something.

Signal 4: Gym floor only, never a class

This one surprised me when I read it.

Hillsdon's analysis of 10,062 members compared people who attended group classes against those who only used the gym floor. Group exercisers retained at 88%, gym-only members at 82%. Adjusted for demographics, gym-only members were 26% more likely to cancel.

The detail that makes it interesting: there was very little difference in visit frequency between the two groups. So this isn't just a roundabout way of measuring who comes more often. Something about the class itself does the work, most likely that a class puts you in a room with the same faces every week and gives you a reason to be missed when you're not there.

Separate research with 601 facilities found average membership length rising from 23.4 months for members who never attended classes to 35.3 months for those attending three or more times weekly.

If you run classes and half your members have never tried one, that's not a scheduling problem. It's a retention problem wearing a scheduling costume.

Signal 5: Nobody has spoken to them

Hillsdon's data on this is blunt. Members who felt ignored by fitness staff were 50% more likely to cancel. Members who were spoken to by reception staff were 50% less likely. Over a quarter of members reported that fitness staff never spoke to them at all.

And the effect compounds with progress. Fewer than one in ten members who reported no staff interaction felt they'd made a lot of progress. Among members spoken to every visit, more than half did.

The most useful number in the whole body of research is this one: if every low-progress member were spoken to at least frequently, 10% of monthly cancellations would be avoided. That works out to roughly 16 saved memberships per 1,000 members, every month.

You don't need software to talk to people. What you need software for is knowing which twelve people to talk to on a given Tuesday, out of four hundred.

Signal 6: No visible progress toward anything

Hillsdon found that each goal a member reported progress on was associated with roughly 10% lower cancellation risk.

This is where most of the actual churn lives, and it's the hardest signal to see from the front desk, because a member who isn't progressing looks exactly like a member who is. They both walk in and walk out.

It's also where the reasons people give diverge most sharply from the truth. In YouGov's US polling, only 5% of people said they quit because they didn't get results, while 41% said cost. Set that against Bedford's finding that unmet expectations and lost motivation are the top real drivers, and a clearer picture emerges: people experience the absence of results as the membership no longer being worth the money. Then they report the price.

Two things make progress visible, and members can do both themselves if you give them the tools. Logging workouts turns four months of effort into a line that goes up. Tracking food closes the gap that quietly wrecks more results than any training programme — people train hard, eat unchanged, see nothing, and conclude the gym failed them.

There's a reason this matters more than it sounds. One well-known study on habit formation found that exercise took a median of around 91 days to become automatic, noticeably longer than diet or drinking habits. It's a small study and the exercise figure is partly extrapolated, so hold it loosely. But it lines up with what the retention data shows: the first three months are when members are most fragile, and they're fragile precisely because nothing feels automatic yet.

The ones who already left aren't necessarily gone

Worth knowing before you write off a lapsed member: a study following 5,242 members found that 38% of people who dropped out came back within twelve months. More than half of those returned in the first month after leaving.

That last detail is the actionable one. The window for winning someone back is far shorter than the window in which they might drift back on their own, and it's widest immediately after they stop coming, before they've replaced the habit with something else or joined somewhere closer to home.

The same study found that having been a member for more than six months was the strongest predictor of returning after a lapse. Which loops back to everything above: the work you do in the first ninety days doesn't just reduce the chance someone leaves. It raises the chance they come back if they do.

What this is worth

Retention isn't a soft metric. In the ukactive dataset, member lifetime value ranged from $378 to $720 depending on how engaged members were.

Hillsdon modelled the revenue effect directly. Across a 10,000-member base, a 10% increase in average attendance was worth an additional £43,343 over 24 months. Moving half of the low-attendance members up to attending once or twice a week was worth £141,260. Those are projections from observed data rather than measured outcomes, so treat them as direction rather than promise. The direction is not subtle.

Scale it down to a real independent gym. At 200 members, the 16-saves-per-1,000 figure works out to roughly three memberships a month that you keep instead of losing. Three memberships a month, compounding, against software that costs less than one of them.

The version you can actually run this week

None of the six signals require anything exotic. They require knowing who came in, when, and how that compares to their own normal.

If your check-ins are digital, you can build the whole thing from attendance data:

• Every Monday, pull everyone with no visit in 8+ days. Message them individually, not as a group blast. Reference something specific.
• Every month, pull everyone under four visits in the last 30 days. This is your at-risk list.
• For anyone in their first 90 days, treat a single missed week as urgent rather than normal. The 27% figure says this window is worth disproportionate effort.
• Get first-timers into one class. Just one. The 26% gap between class-goers and floor-only members is the cheapest retention win available to you.
• Give members a way to see their own progress, in lifts and in food, so that month three has evidence in it rather than just effort.

At Inkuity, this is what the check-in data and member app are built around: QR check-ins that make attendance queryable, analytics that surface who's drifting, WhatsApp messaging so the follow-up reaches somebody rather than sitting in a group chat, and a member portal where people track workouts, log meals by photo, and watch a streak build.

But the tooling is downstream of the idea, and the idea is free: your members tell you they're leaving, in their attendance, weeks before they tell you at the desk. Most gyms simply aren't set up to hear it.

Where to start

Pick one signal. The seven-day gap is the easiest, and it has the best evidence behind it of the six.

Pull the list this week. Send five messages. See what comes back.

If you want the data to be there when you look, setting up a gym on Inkuity takes a few minutes and costs nothing while we're in early access. Print the QR code, add your members, and in a month you'll have something worth reading.