Analysis of the impact of Customer Value Management (CVM) on increasing Cellular Packet Telkomsel (Study case: PT Telkomsel)

Abstrak

Purpose: This study aims to analyze the impact of the Customer Value Management (CVM) program supported by machine learning on increasing customer purchases of Telkomsel cellular data packages, as well as identifying the key behavioral factors influencing purchasing decision. Methodology/approach: This research employs a quantitative explanatory approach using big data analytics. The dataset consists of customer transaction records over a three-month period (January–March 2023), involving 5.7 million customer data points. A supervised machine learning classification model was developed using the CatBoost Gradient Boosting Decision Tree (GBDT) algorithm to predict customer purchasing propensity, supported by Focus Group Discussions (FGD) with subject-matter experts. Results/findings: The CatBoost model achieved an accuracy of 86% in predicting potential lapsers. The test-and-learn campaign based on CVM personalization resulted in a 6.55% increase in take-up rate and generated a revenue uplift of IDR 141.6 million. The most significant factors influencing purchases were monthly data package revenue, frequency of data usage within specific price ranges, and total monthly data revenue. Conclusion: The findings confirm that CVM implementation supported by machine learning effectively enhances personalized marketing, improves customer targeting, and increases purchasing performance at PT Telkomsel. Limitations: This study is limited to a single company, a three-month observation period, and the use of one machine learning algorithm. Contribution: This study contributes empirical evidence on the effectiveness of integrating CVM and CatBoost-based machine learning in large-scale telecom marketing to optimize customer value and revenue growth.