Maneggio
Vol. 3 No. 5 (2026): OCTOBER - MJ

The Application of People Analytics Based on Predictive Modeling to Reduce Employee Turnover in the Banking Sector

Romario Tuhepary (Universitas Victory Sorong)
Vanessa Victoria Micibaroe (Universitas Victory Sorong)
Clara Febriani Wader (Universitas Victory Sorong)
Julian Haumahu (Universitas Victory Sorong)



Article Info

Publish Date
15 Oct 2026

Abstract

Employee turnover presents a significant strategic and financial challenge in the banking sector, leading to substantial recruitment costs, loss of institutional knowledge, disruption of branch operations, and diminished team productivity. Traditional human resource management (HRM) practices rely primarily on retrospective metrics, exit interviews, and managerial intuition, which are descriptive and ill-suited for anticipating future departure risks. This paper investigates the application of People Analytics and predictive machine learning modeling to identify employee attrition risks and support proactive retention strategies in banking institutions. Drawing on recent literature (2022–2026), theoretical frameworks (Resource-Based View, Human Capital Theory, and Technology-Organization-Environment framework), and empirical benchmarks, this study evaluates multi-source employee data taxonomies, comparative predictive modeling algorithms (Logistic Regression, Decision Trees, Support Vector Machines, Random Forest, and Gradient Boosting), key turnover drivers, organizational benefits, implementation challenges, and ethical/privacy safeguards. The literature reveals that non-linear ensemble models, particularly Gradient Boosting (ROC-AUC 0.906) and Random Forest (ROC-AUC 0.891 / 0.890), consistently outperform parametric baseline models in predicting turnover probability. Empirical feature importances highlight job satisfaction, overtime, career development opportunities, compensation growth, and tenure length as the primary predictors of attrition. Furthermore, this paper analyzes regulatory requirements under the EU Artificial Intelligence Act (Regulation EU 2024/1689), which classifies AI workforce analytics as high-risk, necessitating explainable AI (XAI/SHAP), continuous bias auditing, and human-in-the-loop governance. Finally, a structured six-layer implementation framework is proposed to assist banking organizations in transitioning from reactive reporting to evidence-based talent retention.

Copyrights © 2026






Journal Info

Abbrev

MJ

Publisher

Subject

Decision Sciences, Operations Research & Management

Description

The journal publishes original articles on current issues and trends occurring internationally in financial management, marketing management, human-resource management, behavior organizational, good governance, strategic management, business ethics, entrepreneurship, management accounting, manajemen ...