The Shape of Retention
Two of your members came into your gym the same number of times this month. On every report you can pull, they look identical — same visit count, same days since last seen, same row in your dashboard. One of them will still be here in two years. The other is already leaving. They just haven’t told you, and they may not know it themselves yet.
The difference between them isn’t how often they show up. It’s the shape of how they show up.
We studied nearly six years of month-by-month attendance at a single facility — not a snapshot of where each member stood, but the full trajectory of how their engagement moved over time. The finding ran against the way the entire industry reads retention. Engagement isn’t a line. It doesn’t climb steadily when a member is thriving or decline steadily when they’re slipping. It moves in waves — busy stretches and quiet stretches, a personal rhythm of peaks and dips. 87.6% of the members we modeled — everyone with a year or more of continuous history — follow this pattern.
And inside that wave are two properties that separate the members who stay from the members who leave. Neither one is a visit count. Neither one is visible to a system built to count attendance. We call them Baseline Habit Strength and Drift Rate — and once you can see them, our two members stop looking identical.
I built a way to see six representative shapes from patterns we found in the data. Give it a watch and then let’s talk about what you saw.
The wave has a center, and a direction
Every member’s engagement wave is organized around a gravitational center. That’s Baseline Habit Strength. A member swings above it in a strong month and below it in a slow one, but they keep coming back to it. A high baseline means the habit has structural support during a temporary dip. A low baseline means even a good month won’t carry someone very far.
Then there’s the magnitude of the waves themselves — what’s happening to the size of the swings over time. We call that the Drift Rate, and it has three states:
- Expanding — the swings get bigger. Habits are building, and the trajectory points toward long-term retention.
- Compressing — the swings get smaller. Not a crisis yet, but you can see the energy draining out of the pattern. Watch for the pattern to accelerate.
- Collapsing — the wave is flattening fast (warning: a flat line isn’t calm). We’ll come back to that, because it’s the most important finding in the issue.
Baseline tells you where a member lives. Drift Rate tells you which way they’re going. Together they are the shape — and two members with the exact same attendance this month can sit on completely different baselines, drifting in opposite directions. That’s the answer to the question we started with.
What the shapes actually look like
The video walks all six. Let’s dig into three:
The ‘Rookie Warrior’ looks like a problem and isn’t. Violent week-to-week swings, a schedule that looks chaotic on any report. A visit-count system flags every bad week. But the baseline is high and the drift is expanding — that volatility is a member building a habit, not losing one. The shape tells you to leave them alone. They’re fine.
The ‘Social Veteran’ looks fine and isn’t. High baseline, long tenure, the kind of member you never worry about. Nothing in a visit count will flag them — they’re still showing up. But their wave is compressing. The energy is quietly draining, month over month, underneath an attendance number that still looks healthy. This is the member who cancels and blindsides you. Not because the warning wasn’t there, but because your reports were never built to read it.
The ‘Trough False Alarm’ looks like they’re slipping but they aren’t — A genuinely healthy member, strong baseline, expanding wave, simply caught at the bottom of a normal cycle the week you happened to look. A visit-count alert sees the dip and screams at risk. The shape sees a member who is about to bounce, and tells you to save your outreach for someone who actually needs it. Misread this member and you pay twice: wasted effort on someone who was always going to be fine, and the slow erosion of trust that comes from crying wolf.
Imagine the Social Veteran and the Trough False Alarm side by side, because that’s the point of this entire issue. The system that misreads the Trough False Alarm as dying is the same system that misses the Social Veteran who actually is. One snapshot, two opposite errors, in opposite directions, on the same day. Seeing the shapes corrects both.
A flat line isn’t calm
The instinct, looking at all this movement, is to want it to settle down — to read volatility as risk and steadiness as health. The data says the reverse. The members who stayed weren’t the ones with the smoothest attendance. They were the ones whose waves held their size or grew — swings and all. The peaks and troughs are the signs of life. A member who has stopped moving has stopped engaging.
So the reflex to relax when a member goes quiet and consistent is backwards. Swings that hold their size around a strong baseline are health. Flatness — even flatness at a respectable level — is the beginning of the end.
This changes what “at risk” even means. Risk was never about low attendance in any given period. Operators can see risk as a collapsing or compressing drift of baseline habit strength — a shape that’s assembled from snapshots over time. It’s why seeing the shape early is the whole game: by the time a flattening member finally shows up in your visit counts as “gone quiet,” the drift has been collapsing for months. The shape was the early warning the count can’t give you.
The shape is the early warning the count was never able to give you.
Read the shape, not the number
Your members’ visit counts are real and useful even if hard to unpack. But the shift in thinking is simple — stop reading a member as a number, and start reading them as a shape. The question moves from how many times did this member come in to where’s their baseline, and which way are they drifting?
Three questions to hold up against whatever process you run today:
- When a member has a slow month, can I tell a normal trough from a real collapse — or do I treat every dip the same?
- Which of my “fine” members — good attendance, long tenure — might be quietly compressing right now, one disruption away from a cancellation that blindsides me?
- Am I focusing on members who are genuinely at risk to leave, or on members who are just at the bottom of a healthy wave and would have bounced on their own?
Your reports give you a number, and that number tells you how many times a member showed up. It can’t tell you the shape of their engagement — and the shape is where cancellations happen, long before the number ever moves.
P.S. — There’s a property of the wave we didn’t get into here: not just its shape, but where a member sits in their cycle on the day you reach them. Two members with the exact same healthy shape can need completely different things depending on whether they’re cresting or troughing the week you make contact. Getting a read on that — the timing, not just the trajectory — is the next layer of the work. More soon.