PURPOSE
Segment auto loan customers by early prepayment risk during Federal Reserve rate cycle to support targeted retention and portfolio management strategies.
OBJECTIVE
Segment 30,000+ auto loans using two-stage clustering to identify customer profiles with elevated early payoff risk (>3x baseline) and prioritize retention strategies by product type and risk severity.
PROCESS
Apply two-stage K-means clustering with categorical cross-tabulation to identify optimal customer segmentation approaches for prepayment risk, generating actionable profiles with quantified early payoff rates for portfolio management decisions.
OUTPUT
Validated 3x increase in early prepayment velocity during rate hiking cycle, with one-month payoffs representing critical intervention point (2.76x acceleration). Generated 10 customer clusters expanding to 40 product-specific segments, identifying five high-priority profiles (1.9x-2.8x baseline rates) with actionable retention strategies and quantified portfolio impact.
.png)
.png)
