Retention Intelligence uses AI to estimate which members are likely to leave your club, and what's contributing to that risk. This article explains where the predictions come from and how to use them.
What the model looks at
The model learns from patterns in your club's historical data, such as playing activity, who members play with, membership seniority, age and gender. It compares each member against members who have left in the past and estimates how similar their pattern is.
Predictions are updated monthly. The "Last updated" date in the top right shows when the current data is from.
Churn risk
Each member is placed in one of four risk bands: Low, Limited, Elevated or High. The band reflects how likely the model thinks it is that the member will leave, not a certainty. Some members marked High will stay, and some marked Low will leave.
Drivers
Drivers show the factors the model weighted most for a member. They're a starting point for understanding the risk, not a confirmed reason. A member with "Age" as a driver isn't leaving because of their age. It means members with a similar profile have left more often.
How accurate is it?
Accuracy depends on how complete and recent your club's data is. The more history the model has, the better it gets. Predictions work best for spotting groups and trends, and are less reliable for any single member.
Using predictions well
Use Retention Intelligence to decide where to focus your attention, for example who to reach out to first. Combine it with what you and your staff know about your members. Avoid using a prediction as the only basis for a decision about an individual member.