Churn Prediction & Health Scores
Understand member health scores, the factors behind churn prediction, and the at-risk dashboard widget.
Last updated: 05/21/2026
Churn Prediction & Health Scores
Keeping members is more cost-effective than acquiring new ones. Pulse AI's churn prediction system assigns every active member a health score from 0 to 100, giving you an early warning when someone is at risk of canceling. This guide explains how health scores work and how to use them.
Member Health Scores (0-100)
The health score represents how "healthy" a member's relationship with your gym is. A high score means the member is engaged, attending regularly, and paying on time. A low score signals trouble — the member may be drifting away.
| Range | Label | Interpretation |
|---|---|---|
| 80-100 | Thriving | Highly engaged, low cancellation risk. |
| 60-79 | Stable | Consistent activity, healthy relationship. |
| 40-59 | Declining | Engagement is dropping. Worth a proactive check-in. |
| 20-39 | At Risk | Significant warning signs. Intervention recommended. |
| 0-19 | Critical | Very likely to cancel without immediate action. |
You can view any member's health score on their profile page, or filter the Members list by health/at-risk status to see members ranked by score.
Factors Behind Health Scores
Pulse AI analyzes multiple behavioral and transactional data points to calculate each score:
Attendance Frequency
This is the strongest factor. Pulse AI tracks:
- Weekly visit count compared to the member's historical average.
- Attendance trend — Is the member visiting more or less frequently than they were a month ago?
- Class booking vs. attendance — Members who book but do not show up are a warning sign.
- Gap detection — Extended gaps between visits (e.g., no check-in for 10+ days) significantly lower the score.
Payment History
Financial signals matter:
- On-time payments — Consistent, successful payments indicate commitment.
- Failed payments — One failure may be a card issue; repeated failures suggest the member may be passively trying to leave.
- Downgrade requests — Switching to a cheaper plan can be an early indicator.
- Freeze requests — Membership freezes sometimes lead to cancellation after the freeze ends.
Engagement
Beyond attendance, Pulse AI looks at broader engagement:
- Email open rate — Members who open gym communications are more engaged.
- Portal activity — Logging into the member portal, viewing the schedule, or updating profile information.
- Class variety — Members who try different class types tend to stick around longer than those who attend only one.
- Social signals — Referring friends or participating in challenges (if tracked) boosts the score.
Tenure and Contract Status
- Membership length — Members in their first 90 days are statistically more likely to churn, so Pulse AI weighs early-stage engagement more heavily.
- Contract type — Month-to-month members have a lower baseline score than those on annual contracts, reflecting the easier cancellation path.
At-Risk Threshold Configuration
You control what score level triggers an at-risk alert. Churn-prediction thresholds are configured from the Member Health area:
- Set the At-Risk Threshold (default is 40). Any member whose score drops below this value is flagged as at risk.
- Set the Critical Threshold (default is 20) for members needing urgent intervention.
- Click Save.
Adjusting these thresholds depends on your gym's typical churn patterns. If you find you are getting too many false alerts, lower the threshold. If members are canceling before being flagged, raise it.
Dashboard Health Score Widget
The main dashboard includes a Member Health widget showing:
- Health score distribution — A breakdown of how many members fall into each score range (Thriving, Stable, Declining, At Risk, Critical).
- Trend arrow — Whether the overall health of your membership base is improving or declining compared to last week.
- At-risk count — The number of members currently below your at-risk threshold, with a link to view the full list.
- Recent score changes — Members with the biggest score drops in the past 7 days.
Click any section of the widget to drill into the detailed member list.
Using Health Scores Proactively
Do not wait for members to cancel. Use health scores to intervene early:
- Weekly review — Every Monday, check the At-Risk list and assign re-engagement tasks to your team.
- Automated triggers — Combine health scores with follow-up automations to send a friendly check-in when a score drops below a threshold (see At-Risk Member Alerts & Re-engagement).
- Personal training offers — Members with declining scores often respond well to a complimentary personal training session.
- Feedback requests — Reach out to ask what you could do better. Sometimes a simple "we noticed you haven't been in lately" is enough.