By the time a renewal conversation turns difficult, the warning signs were usually visible weeks or months earlier, sitting in usage data, support tickets and engagement patterns that nobody was watching closely enough. This session looks at how customer success teams are building early warning systems that catch churn risk while there is still time to act, covering which signals are genuinely predictive, how to avoid alert fatigue from tracking too many metrics, and what a realistic intervention playbook looks like once a risk score fires. What this session will cover: - Which behavioral signals most reliably predict churn 30 to 90 days out - Why simpler health scores often outperform complex ones - How to build an intervention playbook CSMs will actually use - What proactive retention economics look like compared to reactive save attempts