At-Risk Member Detection: The Attendance Signals That Predict Cancellation 6 Weeks Out

A drop from 3+ visits per week to 1 visit per week, sustained over two consecutive weeks, predicts member cancellation within 6 weeks with 71% accuracy. Studios that identify this pattern at week 2 and intervene personally retain 50% of those members. Studios that identify it at the cancellation request retain almost none. The entire advantage is in the timing — and the timing requires automation.

Why At-Risk Detection Beats Win-Back Campaigns

Win-back campaigns are expensive and low-return — they target members who have already decided to leave. At-risk detection targets members who are leaving in slow motion and haven't made a final decision yet.

The decision psychology is different at each stage. An at-risk member — someone whose attendance has dropped but who hasn't cancelled — is still in the relationship. They may have lost motivation, hit a life disruption, or simply drifted. A personal outreach at this stage can re-engage them before the drift becomes a decision.

A churned member has already emotionally exited the studio. The friction of returning is high, and a "we miss you" email rarely overcomes it. Win-back rates for churned members average 8–15%. At-risk recovery rates with personal outreach average 40–55%.

The resources required are the same — staff time and a message. The return is 3–5× higher when the timing is right.

What Are the Primary and Secondary At-Risk Signals?

Not all attendance changes are equal. The signals vary in predictive power and urgency.

Primary signal — sustained visit drop:

• Member attended 3+ times/week for 4+ weeks
• Drops to 1 visit/week
• Pattern sustained for 2+ consecutive weeks
• Action: immediate outreach within 48 hours of identifying the pattern

Secondary signals — worth monitoring but lower urgency:

• Missing 2+ consecutive bookings without a cancellation notification
• Shift from advance booking (5+ days ahead) to same-day-only booking
• Absence from a preferred instructor's class that the member regularly attended
• Membership renewal date approaching with no renewal action taken
• Failed payment that hasn't been resolved in 5+ days

Contextual signals — require human judgment:

• A member who mentions a life event (new job, new baby, moving) at check-in
• A reduction in the range of class types booked (was trying 4 formats, now only attends 1)
• Reduced class attendance paired with a reduction in retail purchases

The primary signal should trigger automation. Secondary signals should be visible on staff dashboards. Contextual signals require front-desk training to recognize and act on.

How Do You Configure At-Risk Flagging?

The automation rule has two components: the trigger condition and the output.

Trigger condition: Rolling 2-week visit count drops 50% or more below rolling 4-week average. A member who averaged 3.5 visits/week over the previous four weeks and logged 1 visit in each of the last two weeks meets the trigger. The rolling average accounts for members with naturally variable attendance (frequent travelers, parents with variable schedules) without excluding them from monitoring.

Output: Flag on a staff dashboard, not an automated email to the member. The flag should show:

• Member name and photo
• Previous visit frequency (rolling 4-week average)
• Current visit frequency (rolling 2-week average)
• Days since last visit
• Membership type and next billing date
• Last class attended and instructor

With this information visible, a staff member can make a personal, informed outreach — "Hey Marcus, we haven't seen you in reformer since the 12th — are you okay? Your Thursday 7am spot is there whenever you're ready" — not a generic automation response.

What Should the Outreach Message Say?

Personal, specific, and low-pressure. Three elements:

Acknowledgment. Reference the specific absence without being surveillance-y about it. "We missed you in Tuesday's class" is natural. "Our system shows you haven't attended in 11 days" is not.

Check-in. Ask an open question that signals genuine care, not a sales agenda. "Is everything okay?" or "Anything we can help with?" opens a conversation without assuming the member is leaving.

Easy re-entry. Offer a frictionless next step. "Your regular Thursday slot is open this week if you want to jump back in." Don't offer discounts or incentives in the first message — that signals you know they're leaving and cheapens the relationship. Save the offer for a follow-up if the initial check-in doesn't convert.

How Do You Measure Intervention Effectiveness?

Two metrics: intervention rate and recovery rate.

Intervention rate: (At-risk members who received personal outreach within 7 days of flagging) / (total at-risk members flagged). Target: 80%+. If this is below 60%, the process is failing — staff are seeing flags but not acting. Investigate the workflow: are flags visible? Is there a clear owner for outreach? Is the outreach expected behavior or treated as optional?

Recovery rate: (At-risk members who returned to pre-drop attendance within 30 days of intervention) / (at-risk members who received intervention). Target: 40%+. If this is below 25%, the outreach content or timing needs adjustment. Review the specific messages being sent and identify whether the issue is message quality, timing (too late in the at-risk window), or channel.

Track both monthly. Both metrics are leading indicators of churn rate improvements that will show up in the headline churn number 4–8 weeks later.

For the churn calculation framework to contextualize these numbers, see the studio churn rate guide. For the win-back sequence to apply after members do cancel, see the studio win-back sequence guide. For the full retention system, see the studio client retention playbook.