STAYNEX: An Explainable, AI-Powered Customer Churn Intelligence and Retention Decision-Support System for E-Commerce

Customer churn erodes revenue in e-commerce, yet most retention teams learn about a lost customer only after the last purchase. This paper presents STAYNEX, a web-based churn intelligence system that predicts the churn probability of each customer, assigns a risk tier and a business segment, explains the behavioural drivers behind every score, recommends retention actions, and lets analysts test what-if scenarios before spending marketing budget. The system combines a cleaning and feature engineering pipeline built on recency, frequency and monetary indicators plus composite engagement and return-cancellation measures with a Random Forest classifier, a lightweight explainability engine that relies only on native tree importances, a rule-based recommender, a what-if simulator, a revenue-at-risk estimator and a rule-based analyst, all served through a Flask application with interactive dashboards. On a synthetic dataset of 5,000 customers with a churn rate of 27.0 percent, the model reaches a ROC-AUC of 0.76 and an accuracy of 0.76 on a held-out test set. A logistic regression baseline reaches 0.78 AUC, which shows that the synthetic data are largely linear and that the contribution of this work is the integrated decision-support pipeline rather than a new classifier. A case study shows that a combined intervention lowers the predicted churn probability of a critical-risk customer from 0.87 to 0.58. We discuss limitations and outline validation on real transaction data.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-13
DOI
https://doi.org/10.5281/zenodo.24184914
Primary Topic
Customer churn and segmentation
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

STAYNEX: An Explainable, AI-Powered Customer Churn Intelligence and Retention Decision-Support System for E-Commerce

Karan Rautela, Vardaan Gaur, Hardik Kaushik
1 links
Zenodo (CERN European Organization for Nuclear Research)
Customer churn and segmentation
article

STAYNEX: An Explainable, AI-Powered Customer Churn Intelligence and Retention Decision-Support System for E-Commerce

Karan Rautela, Vardaan Gaur, Hardik Kaushik
article en

Abstract

Customer churn erodes revenue in e-commerce, yet most retention teams learn about a lost customer only after the last purchase. This paper presents STAYNEX, a web-based churn intelligence system that predicts the churn probability of each customer, assigns a risk tier and a business segment, explains the behavioural drivers behind every score, recommends retention actions, and lets analysts test what-if scenarios before spending marketing budget. The system combines a cleaning and feature engineering pipeline built on recency, frequency and monetary indicators plus composite engagement and return-cancellation measures with a Random Forest classifier, a lightweight explainability engine that relies only on native tree importances, a rule-based recommender, a what-if simulator, a revenue-at-risk estimator and a rule-based analyst, all served through a Flask application with interactive dashboards. On a synthetic dataset of 5,000 customers with a churn rate of 27.0 percent, the model reaches a ROC-AUC of 0.76 and an accuracy of 0.76 on a held-out test set. A logistic regression baseline reaches 0.78 AUC, which shows that the synthetic data are largely linear and that the contribution of this work is the integrated decision-support pipeline rather than a new classifier. A case study shows that a combined intervention lowers the predicted churn probability of a critical-risk customer from 0.87 to 0.58. We discuss limitations and outline validation on real transaction data.

Zenodo (CERN European Organization for Nuclear Research)
Customer churn and segmentation
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon