Customer Segmentation for Targeted Churn Intervention in the Telecom Industry
(2024) DABN01 20241Department of Economics
Department of Statistics
- Abstract
- The telecom industry falls among one of the most progressive markets that operates at the level of advanced technologies and has the fastest rate of growth all over the world, including from the perspective of 5G technology, the integration of digital products, and the constant increase in demand for broadband internet. In such a competitive environment, customer attrition has
been known to be a major problem, which results in a loss of revenues apart from the increased costs of acquiring new customers. Most of the conventional churn prediction models proved to be unsuitable for addressing the complex issue of different customer expectations, and therefore poor retention strategies were developed.
This research focuses on integrating... (More) - The telecom industry falls among one of the most progressive markets that operates at the level of advanced technologies and has the fastest rate of growth all over the world, including from the perspective of 5G technology, the integration of digital products, and the constant increase in demand for broadband internet. In such a competitive environment, customer attrition has
been known to be a major problem, which results in a loss of revenues apart from the increased costs of acquiring new customers. Most of the conventional churn prediction models proved to be unsuitable for addressing the complex issue of different customer expectations, and therefore poor retention strategies were developed.
This research focuses on integrating churn prediction models, including Logistic Regression,Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine, K nearest neighbors, Neural Network, and XG boost, with segmentation techniques including K-means clustering, Hierarchical clustering, DBSCAN clustering, and Gaussian mixture modeling. These
models are then compared, and the effect of clustering on CHURN analysis is then measured using a telco dataset. Thus, ROC-AUC, accuracy, F1 score as well as the measures for clustering efficiency, including the Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Index are applied.
Assuming that intervention opportunities are aligned with identified segments, approaches to dealing with key factors affecting churn across each segment are suggested. The findings of this research may be useful for telecom companies to formulate better retention models, minimize costs, and increase relevant metrics.
Keywords: churn prediction models, clustering algorithms, churn intervention strategies, customer churn. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9185851
- author
- Aliyeva, Gunel LU
- supervisor
-
- Michal Kos LU
- organization
- course
- DABN01 20241
- year
- 2024
- type
- H1 - Master's Degree (One Year)
- subject
- language
- English
- id
- 9185851
- date added to LUP
- 2025-09-17 08:30:51
- date last changed
- 2025-09-17 08:30:51
@misc{9185851,
abstract = {{The telecom industry falls among one of the most progressive markets that operates at the level of advanced technologies and has the fastest rate of growth all over the world, including from the perspective of 5G technology, the integration of digital products, and the constant increase in demand for broadband internet. In such a competitive environment, customer attrition has
been known to be a major problem, which results in a loss of revenues apart from the increased costs of acquiring new customers. Most of the conventional churn prediction models proved to be unsuitable for addressing the complex issue of different customer expectations, and therefore poor retention strategies were developed.
This research focuses on integrating churn prediction models, including Logistic Regression,Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine, K nearest neighbors, Neural Network, and XG boost, with segmentation techniques including K-means clustering, Hierarchical clustering, DBSCAN clustering, and Gaussian mixture modeling. These
models are then compared, and the effect of clustering on CHURN analysis is then measured using a telco dataset. Thus, ROC-AUC, accuracy, F1 score as well as the measures for clustering efficiency, including the Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Index are applied.
Assuming that intervention opportunities are aligned with identified segments, approaches to dealing with key factors affecting churn across each segment are suggested. The findings of this research may be useful for telecom companies to formulate better retention models, minimize costs, and increase relevant metrics.
Keywords: churn prediction models, clustering algorithms, churn intervention strategies, customer churn.}},
author = {{Aliyeva, Gunel}},
language = {{eng}},
note = {{Student Paper}},
title = {{Customer Segmentation for Targeted Churn Intervention in the Telecom Industry}},
year = {{2024}},
}