Big Data Analytics for Supply Chain Management: A Literature Review and Research Agenda

Abstract

The main objective of this study is to provide a literature review of big data analytics for supply chain management. A review of articles related to the topics was done within SCOPUS, the largest abstract and citation database of peer-reviewed literature. Our search found 17 articles. The distribution of articles per year of publication, subject area, and affiliation, as well as a summary of each paper are presented. We conclude by highlighting future research directions where the deployment of big data analytics is likely to transform supply chain management practices.

Appendix A. Summary of Articles

Study	Context	Key findings
[29]	ERP selection and implementation.	• ERP users are more mature than non-ERP users in three key indicators: strategic sourcing, category management, and supplier relationship management. • SAP ERP users are more mature than non-ERP users in strategic sourcing, category management, and supplier relationship management
[35]	Exploration of opportunities for research where SCM intersects with data science, predictive analytics, and big data.	• BDA has significant implications for operations and SCM, and presents an opportunity and a challenge to our research and teaching approach. • It is easy to see how data science and predictive analytics apply to SCM but sometimes more difficult to see the direct connection of big data to SCM. • Call for more research on BDA including for Journal of Business Logistics.
[36]	Discussion about big data, predictive analytics, and theory development in the era of a maker movement SC.	• Predictive analytics can be part of the theory building process, even when a given study does not produce or test a specific theory. • Points to the need for strong predictive analytics applications and theory because the disintermediation of the traditional supply chain channels means that consumer behavior has become an integral part of both production and demand. • Call for more research on BDA including for Journal of Business Logistics.
[26]	Discussion of the impact of emerging concepts and technologies (e.g., 3D printing, BDA) on future of SCM.	• All these new concepts and technologies are changing the SCM world. • However, very few firms are proactively managing renewal well. The authors argued that “Paradoxically, past successes often stand in the way, undermining rejuvenation” (p. 21). • Call for more research on these concepts and technologies including for Journal of Business Logistics.
[25]	Exploration of SC game changers.	• Exploration of five emerging “game changers” that represent potential supply chain design inflection points: (1) big data and predictive analytics, (2) additive manufacturing, (3) autonomous vehicles, (4) materials science, and (5) borderless supply chains. • Consideration of four forces that impede transformation to higher levels of value co-creation: (1) supply chain security, (2) failed change management, (3) lack of trust as a governance mechanism, and (4) poor understanding of the “luxury” nature of corporate social responsibility initiatives. • Conclusions: how well managers address sociostructural and sociotechnical issues will determine firm survivability and success (p. 157).
[31]	Exploration of the potential of big data with the latest statistical and machine-learning techniques via the discussion of the Hazy project.	• The high-profile success of many recent BDA-driven systems, also called trained systems, has generated great interest in bringing such technological capabilities to a wider variety of domains. A key challenge in converting this potential to reality is making these trained systems easier to build and maintain.
[33]	Examination of the potential for BDA application in the agricultural sector.	• Integration of data and analysis across business and government entities will be needed for successful implementation (p. 1). • The eventual impact of BDA within the agricultural sector will likely require both organizational and technological innovation (p. 1).
[27]	Exploration of BDA analysis in the context of SCM, followed by a proposal for the use of agent-based competitive simulation as a tool to develop complex decision-making strategies and to stress test them under a variety of market conditions. The authors also propose an extensive set of key performance indicators and apply them to analyze market dynamics.	• When automating business processes, designers should be concerned with business agility and particularly with how the automated process will respond to situations where the standard assumptions of the market may be violated (p. 283). • The use of KPIs may facilitate the process by providing characteristics to measure across the automated SC, and realistic simulation techniques provide rich data sets with which to accurately measure behavior in different situations (p. 283).
[39]	A case study of sensor data collection and analysis in smart city with a focus on smart food supply chain.	• One of the important application areas of the IoT in cities is the food industry. • IoT systems help to monitor, analyze, and manage the food industry in cities. • The proposed smart sensor data collection strategy for IoT has the ability to improve the efficiency and accuracy of provenance (e.g., tracing contamination source and back-tracking potentially infected food in the markets) and minimize the size of the data set at the same time.
[38]	Review of the current research on OI that uses streaming data and proposes an approach to design intelligent operational dashboards for SCM systems.	• With the BDA advantage, live streaming data can be processed to build intelligent dashboards providing insights for management teams (p. 9).
[34]	Discussion of the third industrial revolution.	• The Third Industrial Revolution (TIR) is based on the confluence of three major technological enablers: big data analytics, adaptive services and digital manufacturing (p. 257). • These three major technological enablers underpin the integration or mass customization of services and/or goods. • The TIR potential: - is about the integration of services and goods into “servgoods”; - is about the integration of demand and supply chains; - requires more big data analytics, adaptive services, digital manufacturing, mass customization and other white-collar professionals; - minimizes the need for outsourcing and offshoring; and - can subsume mass production within its broader mass customization framework. • As for concerns, TIR: - makes uneducated or undereducated men and women jobless; - aggravates cybersecurity, privacy and confidentiality problems; - aggravates the economic and social divide between the rich and poor within a country; and - aggravates the economic and social divide between the have and have-not countries (p. 293).
[37]	An ontology-driven approach for distributed information processing in SC environments.	• Using supply chain event ontology based on the ABC and SCOR models, the study shows that automatic access rules on local ontology and the use of ontology mapping could facilitate the realization of heterogeneous data integration, and thus foster facilitate distributed decision-making.
[24]	Analysis of the extended analytics ecosystem.	• The extended analytics ecosystem includes individuals and groups who use analytics functions, and involves several key roles and elements including: executive sponsors (e.g., chief marketing officer (CMO), chief financial officer (CFO), or chief operating officer (COO)), data owners, subject-matter experts, business users, external analytics ecosystem places the organization within the wider data supply chain (e.g., incorporate data from the SC network to multiply the value generated or provide focal firm data and analytics to other firms in the SC network), customers, external data providers, external data consumers, cloud analytics platform and business analytics as services, and big data analytics vendors and consultants. • Analytics is a game changer that will revolutionize how individuals, businesses, and society can use technology. However, the full value of analytics can be realized only when applied to integrated data from multiple sources and when insights are immediate and actionable. • Analytics should be explored gradually to understand what value can be gained from it, but this exploration should be done in a way that enables it to grow so more value can be obtained. • “Viewing big data analytics as an ecosystem provides the understanding of how to chart the way to start small while enabling growth to achieve advanced levels of maturity and value. By observing the success or failure in building a big data analytics capability in small and large organizations, several recommendations can be adopted” (p. 5).
[32]	Work breakdown structure method based on information SC.	• Using the guiding ideology of the MapReduce programming model, historical data, maps and reduce operations, the authors argue that it is possible to trace the source of an SC’s unstable link.
[30]	Selection strategies related to the problem of partnership choice of SC in the context of 3D printing and BDA based on analytic hierarchy process and fuzzy synthetic evaluation.	• Analytic hierarchy process and fuzzy synthetic evaluation may reduce the influence of subjective factors on partner choice, enhance the accuracy and reliability of partner choice and strengthen the competitiveness of SC enterprises.
[23]	Introduction to a general concept to model and analyze logistical state data, in order to find irregularities and their causes and dependences within SCs.	• To perfectly manage an efficient and effective supply chain with a continuous and undisturbed flow of goods, it is possible to use data mining methods on logistical state data to filter irregularities and their causes.
[41]	Introduction to problems and benefits of data quality for data science, predictive analytics, and big data in SCM.	• The growing importance of data to SC managers should lead to an amplified awareness and sensitivity to their need for high-quality data products. • The results of decisions based on poor quality data could be costly. - SC managers should begin to view the quality of the data products they depend upon for decisions in much the same way they view the quality of the products their SC delivers. - Managers who appreciate the value of data products that are accurate, consistent, complete, and timely should consider the potential for using control methods to improve the quality of data products, much as these methods improved the quality of manufactured products (p. 78).