Nobody Is Born Consistent
Consistency is what a routine produces, and the best evidence on who stops going is thinner than the numbers make it look
By Dildem Smolcic · Certified Reformer Pilates instructor
Key takeaways
- Sperandei and colleagues followed 5,240 members of one Rio de Janeiro gym and found 63 per cent stopped before their third month, with fewer than 4 per cent lasting a continuous year, though the study measured cancelled payments rather than attendance and the authors say attendance frequency was not recorded, so it describes membership survival rather than gym-going (Journal of Science and Medicine in Sport, 2016).
- Phillippa Lally's 2010 study of 96 people put the median time to automaticity at 66 days, with a range of 18 to 254 days, and found that missing a single day did not measurably damage the habit curve.
- If-then plans have a moderate effect on whether people reach a goal, taken across every kind of goal anybody has tested. Narrowed to physical activity the effect is small, the kind of difference that shows up across thousands of people and would be invisible in any one of them, and that smaller figure is the one an exercise claim has to use.
- Self-efficacy, a person's belief they can do the specific session in the specific situation, was the most consistently supported predictor of attendance in a 2026 systematic review of adult group fitness classes, ahead of social support.
- Alan Marlatt's abstinence violation effect names the variable that decides what a missed week costs: whether the person attributes the lapse to the situation or to themselves.
Consistency is an outcome, and it arrives late. It is what a routine produces once that routine has run long enough that turning up no longer takes a decision, and it goes when the routine goes. Nobody is issued it at birth. The people who look as though they were are further along the same process, running a structure they built at a point nobody was watching.
The industry sells the problem the other way round, as a shortage of a personal quality that some people have and the rest need to go and find, which is the same mistake the self-control literature spent a decade correcting, and it is why a missed fortnight gets read as a finding about the person rather than as a fortnight.
I teach Reformer Pilates, and the January intake is the most predictable thing in my working year. The people who arrive in the first fortnight almost never arrive calm. Something happened. A photograph, a comment that stayed, a birthday, a break-up, an event with a date on it. They book five sessions in a week, cut sugar the same Monday, set an alarm an hour earlier than the one they already ignore, and start counting things they have never counted before. The intensity is real, and in years of watching it I have never found it predictive of anything.
By the third or fourth week something ordinary happens, a child is ill or a deadline lands or a holiday interrupts the pattern or somebody sleeps badly for six nights running, and the week gets missed.
What follows is the part I have watched often enough to trust. The missed week gets read as a verdict on the person rather than as a week, and the booking stops. Almost nobody leaves in a temper. They rebook once, move it, then stop replying, and the membership carries on being paid for months after the last visit.
What I can see from the front of a studio is who stops coming. I cannot see why. The accounts people give when they come back, sometimes a year later, are reconstructions made afterwards, and I have no way of separating those from what actually happened. So the rest of this is other people's data, and part of the job is being honest about how thin some of it is. The one thing it agrees on is that [exercise adherence](https://thewhyweproject.com/articles/what-is-exercise-adherence) has never tracked discipline very closely.
Cancellations
The best available figure on gym dropout comes from a single gym in central Rio de Janeiro. Sandro Sperandei and his colleagues tracked 5,240 members of one medium-sized fitness centre over ten years and reported that 63 per cent stopped before their third month, and that fewer than 4 per cent lasted a continuous year. It was published in the Journal of Science and Medicine in Sport in 2016 and it is the closest thing the field has to a real answer.
Dropout there means stopping payment. The study counted cancellations, and the authors state that attendance frequency was not recorded, so what the curve describes is membership survival. An article about why people quit should say that plainly, because the thing being counted is a cancelled direct debit, and most people stop going some time before they stop paying. The behavioural cliff is therefore earlier and steeper than 63 per cent by three months, and nothing in that dataset can tell you by how much.
The other limit is scale. One gym, one city, between 550 and 700 active members at a time, mean age 31, 58.8 per cent of them women, and 74.7 per cent inactive before they joined. Corroboration exists and it does not agree on the size of the effect. A 2019 study in BMJ Open followed new members across 25 fitness clubs in Oslo and found roughly 37 per cent still exercising regularly at twelve months, against Sperandei's fewer than 4 per cent still paying at a continuous year. The direction is the same in both. The distance between them is a lesson in how much the measure decides the answer.
The six-week figure, the one that gave this piece its working title, has no parent at all. It traces to no study, and to no industry report either.