Customer Churn Prediction for Consumer-Facing Delivery Apps
Abstract
Customer disengagement in non-contractual on-demand services must be inferred from activity rather than explicit cancellation, motivating repeated risk assessment. This study proposes a rolling-window prediction and decision-support framework using 401,164 bookings from 101,765 customers of a doorstep car-wash platform. Matched 30/30-, 15/15-, and 7/7-day observation/evaluation windows generated customer states with 54 behavioral and operational features. Logistic Regression, Random Forest, XGBoost, LSTM, GRU, and Temporal ResNet were tuned and evaluated using customer-grouped partitions, with recall prioritized. LSTM achieved the highest recall at 30, 15, and 7 days (96.0%, 96.4%, and 90.9%), while the highest ROC-AUC values were 0.944, 0.951, and 0.954. Retrospective 30-day analysis identified engagement-depth-specific behavioral patterns, including service failure (65.8%), promotion dependence (52.7%), inactivity (62.0%), and subscription loss with price burden (50.2%), which informed candidate retention actions. External Instacart replication achieved 94% recall and 0.9777 ROC-AUC with 30-day XGBoost, providing preliminary evidence of cross-dataset applicability.
Keywords: Customer Churn, Rolling Window, Behavioral Modeling, Decision Support, Customer Retention, Machine Learning
DOI: 10.54941/ahfe1008254
Cite this paper
More from this volume
- AI-Enhanced Live-Streamed English Language Learning: Designing and Evaluating an Intelligent Platform for Chinese Learners
- "They Will Move": Confirmation Bias, Expectation Frames, and Premature Monitoring Drop Off in COLREG Encounters
- Automation transparency communication does not improve supervisory control performance, but shapes operator trust
- The influence of neuroticism on startle and surprise response: a flight simulator study
- EEG-Based Classification of Reward, Penalty, and Movement Events Across Four Video Games of Varying Complexity
- The Effects of Task Similarity and Time Pressure on Task Performance Following Interruptions in Complex Human-Computer Interaction Systems
- Trusting the Chain, Not the Black Box: A Provenance-Based Architecture for Human-Machine Teaming in Military Autonomy
- Design of a Simplified Humanoid Robot Face: Cultural Differences and Emotional Synchronization
- From Visual Observation to Auditory Interaction: A Human-Centered Tangible Music System Based on Physical Objects
- The Impact of Scarcity-Based Dark Patterns on User Performance, Perceived Darkness, and Workload in Subscription Services
- A Behavioral Observation Tool for Diagnosing Human–AI Interaction Profiles in Engineering Design Education
- The Mediated Moon: Semantic Mediation Between Gesture and Generative Image


AHFE Open Access