Data-related risks for the use of machine learning in retail customer demand forecasting

Purpose: The use of machine learning in customer demand forecasting is reliant on quality data sources. Data should be governed and managed appropriately to ensure that customer demand forecasting is accurate. Most retailers, however, do not understand the technology and are unable to identify all t...

Full description

Saved in:
Bibliographic Details
Main Authors: Lee-Ann Pietersen, Riaan J. Rudman
Format: Article
Language:English
Published: AOSIS 2025-05-01
Series:South African Journal of Business Management
Subjects:
Online Access:https://sajbm.org/index.php/sajbm/article/view/4766
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Purpose: The use of machine learning in customer demand forecasting is reliant on quality data sources. Data should be governed and managed appropriately to ensure that customer demand forecasting is accurate. Most retailers, however, do not understand the technology and are unable to identify all the risks. The purpose of this study is to identify significant data-related risks which arise from the use of machine learning for customer demand forecasting. Design/methodology/approach: A structured literature review was conducted to obtain an understanding of machine learning used for customer demand forecasting and data governance mechanisms required to appropriately manage data assets. Using this understanding, the data governance principles and objectives of the Data Management Body of Knowledge developed by The Global Data Management Community (DAMA DMBOK) and Control Objectives for Information and Related Technologies 2019 (COBIT-2019) governance frameworks were used to identify the data-related risks in a comprehensive manner. Findings/results: Several significant data-related risks arising from the implementation of machine learning for retail customer demand forecasting were identified. These risks link to each stage and component of the machine learning system development life cycle. Practical implications: The risks can be used by internal and external auditors, as well as those charged with governance and other management functions within an organisation, to identify and evaluate risks arising from the use of machine learning within their organisation. Originality/value: While previous studies identify risks on an ad hoc basis, this study used the COBIT-2019 and DAMA DMBOKv2 governance frameworks as the foundation for the identification of risks to ensure completeness and rigour of the risks identified.
ISSN:2078-5585
2078-5976