COMING AUGUST 4, 2025
Introduction
As a gym owner, you know it’s far more cost-effective to retain members than to constantly replace them. But spotting who’s about to drop off — before it happens — isn’t easy.
That’s why we built Risk IQ for GymOS: an intelligent attrition detection system powered by years of member behaviour and retention data.
With Risk IQ, you’ll get early warnings on at-risk members — often days or weeks in advance — giving you the insight and time to take action before they leave.
The traditional approach to attrition
When you start out as a personal trainer working from your garage with a couple of dozen clients, it’s easy to keep tabs on everyone. You know who’s showing up, who’s skipping sessions, and who might be drifting away.
Fast-forward a few years: now you’ve got a full team of coaches, group classes, and hundreds of members. The game has changed, but the risk of member drop-off is still very real.
Even in 2025, most gyms still rely on gut feeling or casual chats between coaches to figure out who might be disengaging. Maybe someone notices they haven’t seen a regular in a while. Maybe they check the booking system and spot a gap.
But attendance tells only part of the story. It doesn’t highlight members who are tapering off slowly, cancelling more often, no-showing, or becoming less responsive. And it certainly doesn’t catch the sudden leavers — the long-time member who cancels everything after a single bad experience.
Often, no one notices until it’s too late.
The lead generation and marketing agencies tell us that we need to focus on lead generation to plug the losses, and what do you know? Conveniently, they offer just the service they say we need!
In many premium small group training gyms in the UK, around 20% of members are on the verge of quitting at any given time. Not all will, of course — but if you can spot the early signs, you have a real shot at changing the outcome. If you can bring in leads and stem losses, that can make a fundamental difference to the long-term health and viability of your gym business.
Bringing the science with Risk IQ
At Quoox, through our GymOS gym management platform, we have long provided our clients with tools to help identify member attrition. However, they, too, predominantly only flagged those for whom “the horse had already bolted.” Whilst it is useful to know that a member hasn’t been in for seven days, it would have been a darn sight more helpful to have spotted the warning signs sooner.
With 7+ years of expansive member behaviour data for gyms like yours, we decided that we could do better. Risk IQ is the culmination of our extensive data and behaviour analysis, algorithm development, evaluation, and testing.
The fundamentals
In the world of premium small-group training, we have the advantage of a “pattern-based product.” Typically, we offer a given number of sessions per week, whether it be 2, 3, 4, or whatever.
Yes, we may then bolster these with the inimitable “unlimited large group” sessions, but even then, we are creatures of habit and typically opt to do the same number of sessions each week. This is key to the power and efficiency of Risk IQ.
By having consistency in our product, this, combined with habitual human traits, enables us to track behaviour. More importantly, it enables us to use that tracked behaviour to spot and analyse anomalies.
The execution
Risk IQ focuses on those with weekly recurring memberships. This is the typical model employed by GymOS clients.
We analyse and track every member’s behaviour within that 24-hour window every day. We then review that behaviour and how it aligns with the member’s behaviour history.
Using our proprietary Risk IQ algorithms, which we have tuned using years’ worth of member behaviour and cancellation data, we then determine a “risk score” for each member. Of course, in reality, this is much more complex than looking for x < y, and that is where Risk IQ works to tune out “false positives”.
Being behaviour-centric, Risk IQ is vulnerable to events such as member sickness or absences. This is why GymOS enables members and coaches to record absences, sickness, membership pauses and other factors that mitigate a member being attributed a high risk score.
The delivery
GymOS clients are presented daily with a list of members for whom there is a risk of their cancelling their membership. Tools are provided to help filter, focus, and evaluate the alerts raised.
A typical implementation in a structured gym operation would be for each mentor to spend a couple of minutes reviewing alerts for their mentees each morning.
Having reviewed an alert (which can be displayed alongside the alert history for that member), a number of actions are available to the reviewer. These include marking the alert as “not a concern”, a “potential concern” for continued review, or they can take action. Amongst the actions available to them is the ability to trigger a mitigation workflow, which the gym can create and configure as they wish.
Alerts for members can be suppressed for a period of time when the reason for the member’s behaviour change is known or when other factors are involved.
The GymOS member summary screen provides a wealth of visual data and clues to help coaches review and assess the veracity of a Risk IQ alert.
Consequential findings
During the development of Risk IQ, there were a number of standout conclusions.
- Poor member behaviour is not being clamped down on, and is causing a downward cycle. For example, some members are ‘reserving’ multiple sessions and then deciding later which to attend. They then cancel the ones they don’t want. However, this then causes other members to do the same – due to the lack of available slots.This behaviour really should be addressed, as some members have crazy cancellation rates of 40%+
- Too many gyms have a problem with capacity. Having oversold their premium time slots, they have not taken on the right sort of members to fill the less popular times.
- A surprising number of cancellations come from those who were regular attendees and then, after a given session, suddenly cancel and give up. These members often get missed, and it could be something as simple as someone said something that upset them.
- In this, the year of our lord 2025, too many gyms are still all about “the in-house session”. Excepting the standard weekly newsletter, lots of gyms still fail to provide additional resources to keep their members engaged beyond the 2-3 hours they spend on the premises.
The beginning, not the end
Risk IQ is released in GymOS as a “beta” function. This is because it is a live, moving algorithm that the Quoox development and analysis teams will continue to refine over time.
As at the point of release, our statistics show that Risk IQ would have correctly predicted approximately 74% of membership cancellations. Many gyms would have then been able to retain a number of those losses. Naturally, other factors affecting these statistics include unrecorded member holidays, membership downgrades, or simply members moving out of the area.
Risk IQ: Fewer surprises. Fewer cancellations. Greater business stability.