Product
Retain
AI Churn Prediction
Use PostHog data to build machine learning model predicting churn risk for proactive intervention.
CaptureProductEmail
$
npx gtm-skills add product/retain/churn-prediction-modelOutcome
Predict top 20% at risk
Leading Indicators
- Prediction accuracy
- Churn rate
- Intervention success
Instructions
-
Define success metric and pass threshold (Predict top 20% at risk).
-
In PostHog, create events for key actions; ensure app tracks all interactions.
-
Instrument critical path with clear PostHog event names.
-
Build minimum viable version with simple UI.
-
Launch to 10-20 users via PostHog feature flag.
-
Track progression in PostHog; create funnels for drop-off analysis.
-
After 7 days, compute metrics vs. threshold.
-
Review PostHog data for patterns and friction.
-
If passed, document and proceed to Baseline; if not, iterate.
-
Record observations for optimization.
Recommendations
Time
5 hours over 1 week
Play-specific cost
Free
Tools
PostHogProduct Analytics