Predictive Churn
Using historical behaviour and statistical or machine-learning models to estimate which players are likely to stop playing soon.
Definition
Predictive churn models learn patterns that precede a player going dormant — falling session frequency, shrinking deposits, longer gaps, lost streaks — and output a churn-probability score for each active player. Unlike churn rate, which reports what already happened, this looks forward so CRM teams can intervene before a valuable player leaves, typically with targeted messages or offers. The value lies in acting on high-probability, high-value players rather than spending on everyone. As with all retention modelling, using churn scores to keep at-risk players engaged must be balanced against safer-gambling duties, since some players lapse precisely because they are stepping back from harm.
Worked example
A model might flag a player whose weekly sessions fell from five to one and whose deposits halved as 80% likely to churn within 30 days, triggering a retention journey — subject to a check that the pull-back is not a positive safer-gambling signal.
Why it matters
For a learner, it shows the shift from reporting the past to predicting the future. For a professional, it is a high-value CRM capability that raises a direct ethical question about retaining players who may be self-protecting.