The most expensive weeks produce the least durable members
January demand is real, contested by everybody, and made up disproportionately of people who will stop attending by spring.
Every operator in a market advertises into the same short window, which raises costs sharply at exactly the moment the reported volume looks best.
The members acquired then are the ones acting on a date rather than a decision, and attendance across that cohort falls away faster than any other intake.
So the cost per signup in January flatters the campaign, and the cost per member still attending in June tells a different story.
That argues for measuring by cohort. Which month's members are still active six months later, and what each of those cost, is a calculation most gyms have never run.
It usually produces a reallocation rather than a withdrawal. Spending less into the peak and more across the quieter months, where competition is thinner and the decisions are deliberate, generally improves the number that matters.