Churn Analytics Prediction & Classification Using Machine Learning (Case Study : Banking Industry)

Ashardin, Muh. Kasim (2024) Churn Analytics Prediction & Classification Using Machine Learning (Case Study : Banking Industry). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032221039-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (4MB) | Request a copy

Abstract

Industri perbankan merupakan salah satu sektor penting dalam perekonomian nasional. Fungsi utama perbankan Indonesia adalah sebagai penghimpun dan penyalur dana masyarakat. Perusahaan perbankan berlomba-lomba berinovasi dan meningkatkan layanan produk yang ditawarkan ke nasabah. Pada praktek dilapangan, perusahaan berusaha mengakuisisi nasabah bank lain. Fenomena nasabah mengurangi bahkan berhenti menggunakan layanan perbankan suatu bank, lalu nasabah lebih masive bahkan berpindah menggunakan layanan perbankan lainnya dikenal dengan fenomena churn. Fenomena ini tentu saja sangat merugikan perusahaan, sehingga perusahaan perlu membuat manajemen nasabah churn. Tujuan dari penelitian ini adalah membuat model prediksi nasabah churn menggunakan metode machine learning, membuat tampilan dashboard yang efektif untuk keperluan manajerial dan mengambil keputusan strategi manajemen untuk mengurangi nasabah churn. Penelitian ini menggunakan metode binary logistic regression, random forest dan XGBoost. Model terbaik untuk memodelkan nasabah churn adalah model XGBoost dengan nilai akurasi sebesar 97,49%, nilai presisi sebesar 96,72%, spesifisitas sebesar 97,12% dan sensitivitas sebesar 97,91% pada data ujinya. Probabilitas nasabah mengalami churn telah dikelompokkan kedalam tiga kelompok yakni above the line, hard line dan soft line. Dengan melakukan pengelompokkan ini dapat menyelamatkan kerugian perusahaan sebesar Rp1,04 Milyar dari 4.700 nasabah active yang diprediksi akan churn. Dibuat strategi manajemen churn untuk mengurangi nasabah churn, antara lain strategi management dashboard, strategi pengelolaan Relationship Manager dan strategi produk. Strategi management dashboard ditujukan untuk pembaharuan data source (automasi data), manajemen tools visualisasi dashboard dan evaluasi feedback. Strategi pengelolaan Relationship Manager (RM) ditujukan untuk pengelolaan RM Simpanan dan Pinjaman agar dapat mengoptimalkan maintenance nasabah kelolaannya. Strategi produk ditujukan untuk diversifikasi produk agar sesuai dengan kebutuhan nasabah.
=================================================================================================================================
The banking industry is one of the crucial sectors in the national economy. The primary function of the Indonesian banking sector is to gather and channel public funds. Banking companies compete to innovate and enhance the range of products and services offered to customers. In practice, companies strive to acquire customers from other banks. The phenomenon of customers reducing or even ceasing to use the banking services of a particular bank, and opting for services from another bank, is known as the churn phenomenon. This phenomenon is undoubtedly detrimental to companies, prompting the need for customer churn management. The objective of this research is to create a predictive model for customer churn using machine learning methods, develop an effective dashboard for managerial purposes, and make strategic management decisions to reduce customer churn. The research utilizes binary logistic regression, random forest, and XGBoost methods. The best model for modeling customer churn is the XGBoost model, with an accuracy of 97.49%, precision of 96.72%, specificity of 97.12%, and sensitivity of 97.91% on the test data. The probability of customers experiencing churn has been categorized into three groups: above the line, hard line, and soft line. By making this categorization, it is possible to save the company losses amounting to Rp1.04 billion from the predicted churn of 4,700 active customers. A customer churn management strategy is developed to reduce customer churn, including management dashboard strategy, Relationship Manager management strategy, and product strategy. The management dashboard strategy aims to update data sources (data automation), manage visualization dashboard tools, and evaluate feedback. The Relationship Manager (RM) management strategy is focused on managing Savings and Loan RMs to optimize customer maintenance. The product strategy aims to diversify products to align with customer needs.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Churn, Logistic Regression, Random Forest, Visualization, XGBoost
Subjects:	Q Science > QA Mathematics Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) T Technology > T Technology (General) T Technology > T Technology (General) > T385 Visualization--Technique T Technology > T Technology (General) > T58.6 Management information systems
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Muh. Kasim Ashardin
Date Deposited:	02 Feb 2024 08:19
Last Modified:	02 Feb 2024 08:19
URI:	http://repository.its.ac.id/id/eprint/105986

Actions (login required)

View Item