Since January 2020 Elsevier has created a COVID-19 resource centre with

free information in English and Mandarin on the novel coronavirus COVID-

19. The COVID-19 resource centre is hosted on Elsevier Connect, the
company's public news and information website.

Elsevier hereby grants permission to make all its COVID-19-related

research that is available on the COVID-19 resource centre - including this
research content - immediately available in PubMed Central and other
publicly funded repositories, such as the WHO COVID database with rights
for unrestricted research re-use and analyses in any form or by any means
with acknowledgement of the original source. These permissions are
granted for free by Elsevier for as long as the COVID-19 resource centre
remains active.
 Computers & Industrial Engineering 181 (2023) 109344

Contents lists available at ScienceDirect

Computers & Industrial Engineering

journal homepage: www.elsevier.com/locate/caie

The future of industry 4.0 and supply chain resilience after the COVID-19
pandemic: Empirical evidence from a Delphi study
Alexander Spieske *, Maximilian Gebhardt , Matthias Kopyto , Hendrik Birkel , Evi Hartmann
Chair of Supply Chain Management, Friedrich-Alexander University Erlangen-Nuremberg, Lange Gasse 20, 90403 Nuremberg, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: The COVID-19 pandemic has caused major supply chain disruptions and unveiled the pressing need to improve
Supply chain Resilience supply chain resilience (SCRES). Industry 4.0 (I4.0) is a promising lever; however, its future in supply chain risk
Supply Chain Risk management (SCRM) is highly uncertain and largely unexplored. This paper aims to evaluate I4.0′ s potential to
Industry 4.0
improve SCRES in a post-COVID-19 world. Based on current literature and multiple workshops, 13 future pro­
Digital supply chain
COVID-19 pandemic
jections on potential I4.0 application areas in SCRM were developed. A two-round Delphi study among 64 SCRM
Delphi study experts with digital expertise was conducted to evaluate and discuss the projections regarding their probability of
occurrence until 2030, their impact on SCRES, and their desirability. A fuzzy c-means algorithm was applied to
cluster the projections based on the expert assessments. The expert evaluations led to three clusters on I4.0
application in SCRM: Four projections on generating data, increasing visibility, and building digital capabilities
received considerable approval and are reliable to improve SCRES in 2030. Four projections enabling data
sharing and processing were predominantly supported and demonstrated realization potential for 2030. Finally,
five projections that require major supply network adaptations were deemed unlikely to improve SCRES in 2030.
This paper answers several research calls by presenting empirical evidence on the pathway of I4.0 imple­
mentation in SCRM following the COVID-19 pandemic. Moreover, it evaluates a holistic set of technologies and
indicates prioritization potentials to achieve SCRES improvements.

1. Introduction customer satisfaction up to a damaged brand (Basole, 2014). Despite

these experiences and the recognition that resilience to SCDs reveals the
Over the last few years, companies have faced increasing challenges potential for competitive advantage (Hohenstein et al., 2015), many
from supply chain disruptions (SCDs) (Paul & Chowdhury, 2020). The companies failed to prepare for the next “black swan” (Van Hoek, 2020),
reason is twofold: From an internal perspective, firms across industries which became apparent when the COVID-19 outbreak hit the fragile SC
have outsourced and globalized their operations for competitiveness and environment in 2020. Many firms have faced tremendous challenges
cost optimization reasons, which has led to longer, more complex, and along their SCs, including simultaneous supplier failures, production
more vulnerable supply chains (SCs) (Hohenstein, Feisel, Hartmann, & capacity restrictions, and demand uncertainties (Ivanov, 2020). Com­
Giunipero, 2015). In this context, SC members are not just inter­ pany leaders have therefore expressed their intentions to increase
connected operationally but also financially (Wu, Zhang, Pan, & Dolgui, resilience in their SCs for mitigating future SCDs (van Hoek, 2020). In
2021), which further increases vulnerability. From an external this context, supply chain resilience (SCRES) refers to an SC’s “ability to
perspective, disruptive events, including natural disasters and political be prepared for unexpected risk events, responding and recovering
interventions, have increasingly affected operations (Lechler, Canza­ quickly to potential disruptions to return to its original situation or grow
niello, Roßmann, von der Gracht, & Hartmann, 2019). Practical conse­ by moving to a new, more desirable state” (Hohenstein et al., 2015,
quences from SCDs include higher costs, sales losses, and declining p. 108).

Abbreviations: AI, Artificial intelligence; AM, Additive manufacturing; BC, Blockchain; BDA, Big data analytics; CV, Convergence rate; D, Desirability; EP, Ex­
pected probability; FCM algorithm, Fuzzy c-means algorithm; I, Impact on SCRES; I4.0, Industry 4.0; IoT, Internet of things; IQR, Interquartile range; RQ, Research
question; SC, Supply chain; SCD, Supply chain disruption; SCM, Supply chain management; SCRES, Supply chain resilience; SCRM, Supply chain risk management.
* Corresponding author.
E-mail addresses: alexander.spieske@fau.de (A. Spieske), maximilian.gebhardt@fau.de (M. Gebhardt), matthias.kopyto@fau.de (M. Kopyto), hendrik.birkel@fau.
de (H. Birkel), evi.hartmann@fau.de (E. Hartmann).

https://doi.org/10.1016/j.cie.2023.109344

Available online 26 May 2023

0360-8352/© 2023 Elsevier Ltd. All rights reserved.
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Dasaklis, Pappis, and Rachaniotis (2012) already claimed that in­ 2. Literature review and projection development
formation management plays a crucial role during an epidemic. How­
ever, scholars report that traditional information systems are too weak The literature widely discusses different levers to enhance SCRES,
to support SCRES holistically (Pettit, Croxton, & Fiksel, 2019). With with most scholars referring to agility (including visibility, velocity, and
innovative Industry 4.0 (I4.0) technologies evolving, a new opportunity flexibility), collaboration, SCRM culture, and SC (re-)engineering
to strengthen SCRES through generating, exchanging, and processing (Christopher & Peck, 2004; Hohenstein et al., 2015). In this context,
information has emerged (Ali & Gölgeci, 2019; Ivanov, Dolgui, & digital innovations have the potential to improve SCRES levers (Ivanov
Sokolov, 2019). Although the literature consents that successful supply et al., 2019). Since legacy information technology widely lacks the
chain risk management (SCRM) will rely on digital technologies in the possibility of increasing SCRES (Pettit et al., 2019), many scholars see
future (Ivanov et al., 2019; van Hoek, 2020), the road map following the great potential in I4.0 (Ali & Gölgeci, 2019). With the COVID-19
COVID-19 pandemic remains unclear (Ivanov, 2020). pandemic profoundly compromising SCs, the debate on using I4.0 so­
Theoretical concepts to assess I4.0′ s maturity in supply chain man­ lutions to enhance SCRES has gained further momentum (Chowdhury,
agement (SCM) have been developed but not yet applied (Frederico, Sarkar, Paul, & Moktadir, 2020).
Garza-Reyes, Anosike, & Kumar, 2019). De Oliveira and Handfield I4.0 is based on the idea of connected and autonomously interacting
(2019) and Bragazzi (2020) thus reported that I4.0 is still in its infancy, machines, products, and processes within and across company bound­
its implementation level is mostly unclear, and several challenges for its aries (Ivanov et al., 2019). It offers the opportunity to capture, transfer,
successful realization persist. Companies remain uncertain about how to and analyze information within reasonable timelines (Stank et al.,
launch digitalization initiatives; they hesitate to replace their complex 2019). Hofmann et al. (2019, p. 945) stated that I4.0 “will change how
legacy systems and still doubt the economic benefits of I4.0, preventing supply chains are designed and operated,” referring particularly to
application at scale (Ghadge, Er Kara, Moradlou, & Goswami, 2020; better possibilities for coordinating material, information, and financial
Nguyen, Chen, & Du, 2020; Stank, Esper, Goldsby, Zinn, & Autry, 2019). flows. In this context, scholars claim that I4.0 directly supports tradi­
Missing legal frameworks further complicate quick realization initia­ tional resilience levers and thus enhances SCRES (e.g., van Hoek (2020),
tives (Nguyen et al., 2020). A lack of resources and capabilities, as well Belhadi et al. (2021), Modgil, Gupta, Stekelorum, and Laguir (2021)).
as data privacy, security, and accuracy concerns, also hinder imple­ Following a systematic literature review, Spieske and Birkel (2021)
mentation (Ghadge et al., 2020; Handfield, Jeong, & Choi, 2019; recently presented comprehensive evidence on the positive relationship
Horváth & Szabó, 2019). Moreover, I4.0′ s different technologies and between I4.0 and SCRES levers. To give some literature examples, Stank
application areas require identifying and prioritizing the most promising et al. (2019) described that accelerated and more transparent informa­
ones (Frederico et al., 2019; Ivanov et al., 2019). tion flows lead to more visibility and velocity in the SC. Paul and
Nevertheless, scholars have reported occasional SCRES improve­ Chowdhury (2020) noted that sharing accurate and up-to-date infor­
ments from pioneering digitalization projects during the COVID-19 mation with SC partners also improves collaboration. The SC (re-)en­
pandemic. For instance, Javaid et al. (2020) claimed that healthcare gineering lever can be supported through automatically generated SC
SCs have benefited from digital technologies, including effective product maps (Wichmann, Brintrup, Baker, Woodall, & McFarlane, 2020).
allocation and timely delivery. In this context, Bragazzi (2020) noted Nevertheless, certain boundaries for I4.0 to support SCRES exist. For
that additive manufacturing (AM) has helped to create additional pro­ instance, a risk management culture relies predominantly on manage­
duction capacities to overcome medical supplies scarcity. In an auto­ ment engagement, internal collaboration, and employee training
motive context, Belhadi et al. (2021) reported that big data analytics (Chunsheng, Wong, Yang, Shang, & Lirn, 2020) rather than technolog­
(BDA) and the Internet of Things (IoT) have been critical enablers for ical advancement. Establishing organizational routines and procedures
choosing reliable suppliers and determining manufacturing capacities to manage SCDs effectively is another activity driven by traditional
and safety stock levels during the pandemic. organizational learning mechanisms (Scholten, Sharkey Scott, & Fynes,
Scholars expect the exceptional circumstances during the COVID-19 2019). Cross-company relationship management and trust are other
pandemic and the positive I4.0 experiences to be catalysts for SCRM SCRES areas that I4.0 can support (Dubey, Gunasekaran, Bryde, Dwi­
digitalization (Craighead, Ketchen, & Darby, 2020), which implies vedi, & Papadopoulos, 2020; Ivanov et al., 2019) but still strongly rely
further SCRES enhancement (Belhadi et al., 2021). However, the dis­ on personal interaction (Ralston & Blackhurst, 2020).
cussed challenges of implementing I4.0 in SCRM at scale persist. When To discuss I4.0′ s various technologies and application areas with
and how companies will follow the scientific call and broadly leverage topic experts, we developed 13 projections for 2030. We built on a re­
I4.0 solutions to enhance SCRES thus remains uncertain. Already before view of current studies at the intersection of I4.0, SCRES, and the human
the COVID-19 pandemic, scholars stressed these pressing questions factor in SCRM and contextualized them with the challenges of the
(Hofmann, Sternberg, Chen, Pflaum, & Prockl, 2019). Overall, the cur­ COVID-19 pandemic.
rent literature lacks comprehensive and empirical foresight for the post- The COVID-19 pandemic has unveiled major visibility insufficiencies
COVID-19 maturity of I4.0 technologies in SCRM, which indicates a in today’s global and complex SCs (Z. Xu, Elomri, Kerbache, & El Omri,
significant research gap at the intersection of I4.0 and SCRES. Using 2020). When internal, supplier, or transportation malfunctions have
expert knowledge through a Delphi study is recommended to assess caused a disrupted flow of goods, companies have struggled to quickly
future scenarios with high uncertainty (Rowe & Wright, 1999; Winkler, determine points of failure (Craighead et al., 2020). Moreover, exter­
Kuklinski, & Moser, 2015). We therefore leveraged the Delphi method nalities such as infection patterns and governmental interventions have
and asked 64 SCM digitalization experts to assess 13 projections, which developed dynamically and affected SCs over a long period (Craighead
enabled us to answer the following research question (RQ): et al., 2020). Remaining informed of restrictions while steering global
RQ: Given the experiences during the COVID-19 pandemic, how will operations has therefore been a major challenge. Due to these experi­
I4.0 technologies impact SCRES until 2030? ences, companies have reported focusing on improving visibility (van
This paper is organized as follows: First, we review relevant litera­ Hoek, 2020). In this context, scholars consider SC digitalization a central
ture and introduce the projections that the Delphi panel evaluated. lever (Stank et al., 2019). With the IoT and artificial intelligence (AI),
Second, we describe our research methodology. Third, we present our two promising I4.0 solutions for enhancing visibility are evolving. First,
quantitative and qualitative results and discuss them based on our RQ. the IoT refers to a network of machines and products equipped with
Finally, we derive conclusions and implications for theory, practice, and digital technology to interact with one another, humans, or other digital
policy-making, highlight limitations, and propose future research systems autonomously (Queiroz, Pereira, Telles, & Machado, 2021).
directions. Monitoring devices implemented in such networks bear great potential
to collect real-time data and improve objectivity and accuracy (Birkel &

2
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Hartmann, 2020). Application areas include item location, inventory o P.5 (2030): Digital supply chain twins have been widely established to
tracking, and metrics determination (e.g., temperature and pressure improve accurate disruption scenario simulation and alternative SCRM
measurement) (Er Kara, Oktay Fırat, & Ghadge, 2020; Kaur & Prakash measure assessment [digital SC twins].
Singh, 2021). This technology is especially powerful since it enables a
shift from a partial and sequential to a complete and instant information Another I4.0 technology to address visibility constraints in SCs while
supply (van Hoek, 2020). Visibility gains through the IoT can enhance increasing collaboration is blockchain (BC) (P. Xu, Lee, Barth, & Richey,
process risk and overall risk knowledge, as well as the resulting miti­ 2021), an open and cryptographic peer-to-peer network that stores
gation strategies (Birkel & Hartmann, 2020). Second, in an SCRES transactional information on decentralized but identical digital ledgers
context, “AI” is the umbrella term for digital solutions, which can (Li et al., 2020). Since all parties involved in a transaction have timely
autonomously and successfully decide on the next action in a partially access to equal data, SC visibility regarding assets’ provenance, location,
unknown SC environment (Baryannis, Validi, Dani, & Antoniou, 2019). or ownership status (Ivanov & Dolgui, 2021), as well as information
These solutions include all techniques that support a system’s contin­ authenticity and security, can be improved (Li et al., 2020; P. Xu et al.,
uous learning and adaptive decision-making capabilities (Baryannis 2021). Container shipping is among the first sectors to benefit from the
et al., 2019; Queiroz et al., 2021). AI relies on past behavior or under­ BC visibility and security improvements (Nguyen et al., 2020). These
lying data (Stank et al., 2019). A prominent area for applying AI in advancements translate into enhanced communication, trust, and
SCRES is screening external real-time information, including news feeds therefore collaboration across company borders (Dubey et al., 2020;
and social media, for quicker risk identification and assessment (Chae, Kumar, Raut, Narwane, & Narkhede, 2020) and improves all steps (i.e.,
2015; Handfield, Sun, & Rothenberg, 2020). In this context, AI’s risk identification, assessment, and management) of the SCRM process
application is not limited to structured data, as recent progress in natural (Kopyto, Lechler, von der Gracht, & Hartmann, 2020). However, most
language processing techniques reveals (Wichmann et al., 2020). Po­ BC solutions in SCM have not reached maturity, and fundamental
tential outcomes include but are not limited to visual risk maps and challenges remain, including a lack of scalability, incompatibility with
country risk scores (Handfield et al., 2020). Given the visibility chal­ legacy systems, and immature regulatory frameworks (Dubey et al.,
lenges during the COVID-19 pandemic and both technologies’ poten­ 2020; P. Xu et al., 2021). We therefore introduce our sixth projection to
tials, we introduce our first two projections: evaluate BC’s future in building SCRES:

o P.1 (2030): Internet of Things implementation has been heavily extended o P.6 (2030): Blockchain adoption has been substantially increased to
to improve data availability and accuracy for SCM decision-making [IoT improve trusted and reliable information exchange [BC information
data]. exchange].
o P.2 (2030): Artificial intelligence applications have been widely imple­
mented to leverage external real-time information sources for risk iden­ Apart from visibility, the COVID-19 pandemic has also revealed ve­
tification and assessment [AI external data]. locity and flexibility insufficiencies in SCRM. Companies have been
overstrained due to the high number of incidents that require attention
Even when data has been available, processing and gaining action­ and decision-making in their SCs. For instance, infection prevention
able insights from it has been a major challenge during the COVID-19 measures that restrict the number of personnel per area have led to
pandemic (Nikolopoulos, Punia, Schäfers, Tsinopoulos, & Vasilakis, capacity constraints for blue- and white-collar workers (Z. Xu et al.,
2021). Apart from enhancing data quantity and quality, improved data 2020). Digital technologies also offer solutions to these challenges.
processing and sharing are also required to build visibility. SC data Baryannis et al. (2019) claimed that an increasing share of SCRM
availability in organizations has increased exponentially in recent years decision-making can be transferred to autonomous software, allowing
(Vieira, Dias, Santos, Pereira, & Oliveira, 2019). Evolving data and humans to focus on more complex SCRM problems (Ralston
process mining solutions can help to leverage the large, steadily & Blackhurst, 2020). Moreover, cyber-physical systems are becoming a
growing, and primarily unstructured SC data treasure that many com­ credible option to reduce dependency on humans in manufacturing
panies already possess more effectively for SCRM (Choi, Wallace, & processes. These systems integrate physical infrastructure with AI to
Wang, 2018; Er Kara et al., 2020). Visibility, and thus SCRES, can be self-manage operations, allowing companies to facilitate or even replace
enhanced by quickly analyzing this data (Dubey et al., 2021; Oliveira physical human labor with autonomous robots and vehicles (Queiroz
& Handfield, 2019). In this context, van Hoek (2020) recommends et al., 2021; Stank et al., 2019). Cost improvements in robotic devel­
sharing data and important findings with SC partners. Cross-company opment and manufacturing, as well as enhanced operational perfor­
data and communication platforms can enable such data exchange mance, reveal a bright future for this technology (Stank et al., 2019).
(Chen, Dui, & Zhang, 2020). In a final expansion stage that includes all Finally, cloud solutions have emerged in recent years, enabling data
SC members actively contributing data, these platforms can lead to real- access and exchange between decentralized locations (Oliveira
time end-to-end SC visibility, significantly increasing SCRES (Ivanov & & Handfield, 2019). These applications form the foundation for white-
Dolgui, 2021). Furthermore, SC partners’ additional information can be collar workers to use company resources remotely and ensure business
used to implement digital SC twins, representing the real-time SC continuity in times of SCDs. From this discussion, we derived three more
network state and enabling the simulation of alternative disruption projections:
scenarios and SCRM measures (Hosseini, Ivanov, & Dolgui, 2019; Ivanov
& Dolgui, 2021). Such stress test simulations can be valuable decision o P.7 (2030): Autonomous software has assumed responsibility for most
support systems for SCRM managers (Ivanov & Dolgui, 2021). Based on decision-making in SCRM to achieve better mitigation results [autono­
this discussion, we formulated another three projections: mous decision-making].
o P.8 (2030): Autonomous robots have replaced a significant share of
o P.3 (2030): Real-time aggregation of available internal and external data physical labor to reduce human resource dependency during supply chain
has been significantly intensified to improve SC visibility within organi­ disruptions [autonomous robots].
zations [data aggregation]. o P.9 (2030): Remote working and digital collaboration have been imple­
o P.4 (2030): Cross-company data platforms have been widely established mented at scale to ensure business continuity during supply chain dis­
to enable real-time end-to-end supply chain visibility [cross-company data ruptions [remote working].
platforms].
In addition, the COVID-19 pandemic has disrupted several com­
panies’ inbound flows, making SC (re-)engineering a prominent topic.

3
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Infection prevention measures and actual COVID-19 cases have recur­ (Winkler et al., 2015), particularly in SCM (von der Gracht, 2008).
rently affected supply bases and transportation, leading to capacity Second, the Delphi method is an accurate approach to exploring contexts
constraints (Ivanov, 2020). Border crossing restrictions and personnel with insufficient empirical data, where expert knowledge provides the
shortages have also impacted transportation (Z. Xu et al., 2020), which only reliable source of information (Rowe & Wright, 1999). Third, the
has led to severe uncertainty for several companies, even causing pro­ Delphi approach allows drawing on informed opinions from diverse
duction stoppage (van Hoek, 2020). Nevertheless, firms can build on experts (Hirschinger, Spickermann, Hartmann, von der Gracht, & Dar­
I4.0 solutions to strengthen their supply base. AI and BDA offer possi­ kow, 2015), leading to higher accuracy than individual evaluations such
bilities to automize and improve the efficiency and accuracy of supplier as semi-structured interviews and other forms of group evaluations (von
selection and regular audits (Cavalcante, Frazzon, Forcellini, & Ivanov, der Gracht, 2008). Fourth, the method’s guaranteed anonymity prevents
2019). Based on internal and external data, predictive models can be negative group dynamics such as bandwagon and halo effects (Linstone
used to evaluate purchasing markets and develop hedging strategies & Turoff, 1975; Rowe & Wright, 1999).
(Handfield et al., 2019). In this context, suppliers’ maturity to quickly
report delivery capabilities should also be evaluated since companies 3.2. Delphi study design
can benefit from digitalization only if their entire SC can provide critical
data (Kaur & Prakash Singh, 2021; Queiroz et al., 2021). Furthermore, The present foresight study was conducted using a two-round web-
AM – a technology that creates products by successively printing layer based Delphi format and built on a comprehensive projection develop­
upon layer – offers alternatives for designing an SC more resiliently and ment process. It follows the four steps depicted in Fig. 1 and is described
overcoming certain risks (Durach, Kurpjuweit, & Wagner, 2017; Zhang, in detail in the following sub-sections.
Wu, Tang, Feng, & Dai, 2020). Scholars expect this technology to reduce In the context of this study, the Delphi method’s four main advan­
the number of production steps, suppliers, and transportation links, tages presented in the previous section were all decisive. First, the
resulting in more localization and flexibility enhancements (Dolgui, COVID-19 pandemic continues to cause severe disruptions more than
Ivanov, & Sokolov, 2020; Stank et al., 2019). These AM characteristics one year after it began to affect SCs globally. Triggered by this crisis,
also indicate lead time and velocity advantages, which constitute a scholars expect significant structural changes in SCs and foresee a
major SCRES lever (Kaur & Prakash Singh, 2021; Paul & Chowdhury, considerable role for I4.0 technologies in this transformation (Ivanov
2020), especially during emergency situations (Bragazzi, 2020). Given & Dolgui, 2021; van Hoek, 2020). However, due to this crisis’s novelty
these impressions, we introduce three additional projections: and the I4.0 research context, it is highly uncertain how these structural
changes will be realized and which I4.0 technologies will be prioritized
o P.10 (2030): Supplier selection and evaluation have been based exclu­ to improve resilience (Dolgui & Ivanov, 2020). The Delphi method has
sively on big data analytics and artificial intelligence [supplier selection proven to be a reliable instrument for predictive research in various
and evaluation]. contexts related to disruptive technological change (Lechler et al., 2019;
o P.11 (2030): Suppliers not offering real-time information about short- Roßmann, Canzaniello, von der Gracht, & Hartmann, 2018; Rowe
term delivery capabilities have been replaced [supplier replacement]. & Wright, 1999). Therefore, the approach is well suited to answer our
o P.12 (2030): Additive manufacturing has been established at scale to future-oriented RQ on how I4.0 technologies will impact SCRES after the
increase supply and manufacturing flexibility strongly [AM flexibility]. COVID-19 crisis. Second, due to the COVID-19 pandemic’s recency,
reliable, large-scale quantitative information is scarce. Therefore, expert
Scholars consider an organization’s risk sensitivity and culture knowledge is currently the only reliable source of information for future
decisive for successful SCRM, including the right capabilities to manage predictions. Third, a variety of stakeholders who have been affected by
SCRM-related activities (Grötsch, Blome, & Schleper, 2013). In this the COVID-19 pandemic and are engaged in the transition toward a
context, digital skill sets to implement and operate all the discussed more digitalized SCRM could be included in this study. This setting al­
technological innovations have gained importance in recent years lows drawing from first-hand experience with the COVID-19 crisis and
(Zouari, Ruel, & Viale, 2020). However, many firms still struggle to comprehensive subject matter knowledge to create a holistic future
develop these employee capabilities, making a lack of skills and expe­ scenario on I4.0-enabled SCRM, which the literature currently lacks
rience a primary barrier for SCRM digitalization (Hofmann et al., 2019). (Ivanov, 2020). Fourth, the anonymity benefit ensured participants a
As companies must thus address this talent gap internally and externally, protected environment to be able also to share negative experiences and
we developed our final projection: consequences of the COVID-19 pandemic with other experts without
disclosing sensitive information such as their firm affiliation.
o P.13 (2030): Digital capabilities have been established as a core skill
requirement for SCRM external hires and internal upskilling [digital 3.3. Step 1: Projection development
capabilities].
In our study, participants were asked to evaluate future-oriented
3. Research methodology statements. The systematic development of those projections is crucial
for the value, validity, and reliability of the Delphi study (Warth, von der
3.1. The Delphi technique Gracht, & Darkow, 2013). First, an opening workshop with the research
team and three experienced SCM practitioners from different industries
The Delphi technique’s main goal is to predict future developments was held. All practitioners reported that the experiences of the COVID-
(Linstone & Turoff, 1975; von der Gracht, 2008). It is an empirical 19 crisis triggered a reevaluation of digital support for SCRM in their
method building on expert judgments to drive structured group organizations. The group also agreed that a comprehensive foresight
communication and consensus-building on a specific topic (Linstone study on SCRM through I4.0 would be valuable for updating the current
& Turoff, 1975; von der Gracht, 2012). Expert assessments can be pro­ theoretical thinking and practical guidance on digitalization pathways.
vided on a quantitative (e.g., Likert scale rating) or qualitative (e.g., Second, relevant influencing factors on how I4.0 technologies can affect
comment with reasoning) basis (Scheibe, Skutsch, & Schofer, 1975). Its SCRES were collected to build a basis for projection formulation. Mul­
round-based approach allows interaction among participants, individual tiple sources were leveraged to ensure methodological rigor and
reassessment of estimates, and ongoing reflection of answers (Rowe completeness. The sources included (1) an extensive review of academic
& Wright, 1999; von der Gracht, 2008). and practitioner literature, (2) a creative workshop with four re­
The Delphi technique offers four major advantages. First, it is a well- searchers, and (3) five semi-structured interviews with senior SCM ex­
established approach for studying future scenarios with high uncertainty ecutives. Third, a projection formulation workshop was conducted.

4
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Fig. 1. Structure of the Delphi process applied in the study.

Delphi projections were drafted, repeatedly revised, and reflected (1996) recommendations, the number of projections was limited to 13 to
against their theoretical foundations. Commonly accepted guidelines increase the response rate and reduce the likelihood of sparse comple­
regarding formulation (Linstone & Turoff, 1975; Mitchel, 1996; Rowe tion. Consistent with previous Delphi studies in the SCM field (Darkow,
& Wright, 1999) and word count (Salancik, Wenger, & Helfer, 1971) Foerster, & von der Gracht, 2015; Keller & von der Gracht, 2014), the
were considered to ensure validity and reliability. Following Mitchel’s projections were formulated with a 10-year time horizon until 2030 to

5
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

increase creativity and novel thoughts. Finally, the formulated pro­ between early and late responders across both rounds and all dimensions
jections were pre-tested via cognitive interviews with three academics (see Appendix A). The presence of a non-response bias can thus be
and three SC managers with in-depth methodological or subject-specific rejected.
knowledge on SCRM. These interviews helped to ensure the study’s
robustness and validity by cross-validating the projections and testing 3.6. Step 4: Analysis of results
for completeness and comprehensibility (Warth et al., 2013).
Mean values were determined for the EP, I, and D of each projection
3.4. Step 2: Panelist selection after the second round (Keller & von der Gracht, 2014; Warth et al.,
2013). Two additional indicators were calculated for each projection’s
The selection of expert panelists significantly influences a Delphi EP: The convergence rate (CV) describes the difference in standard de­
study’s reliability (Spickermann, Zimmermann, & von der Gracht, viation for the EP from the first to the second Delphi round and provides
2014). One must ensure both deep subject matter expertise and diverse information on overall assessment changes. A negative CV signifies that
perspectives to achieve the most accurate results (Rowe & Wright, participants reevaluated their judgments after reviewing their fellow
2011). For an unbiased selection process, we followed Spickermann panelists’ quantitative and qualitative arguments and veered toward the
et al. (2014) and defined selection criteria, including the type of orga­ group opinion (Rowe & Wright, 1999), which indicates that the group-
nization, present work position, academic status, and supposed level of based consensus-building process worked as intended. The interquartile
expertise on SCRM and digital technologies. We identified 694 experts, range (IQR) can be used to evaluate a projection’s level of agreement
including both SCRM experts from academia and industry representa­ and is generally accepted as an objective and rigorous indicator for
tives with SCRM responsibilities, and invited them to participate in our measuring consensus in a Delphi study (von der Gracht, 2012). Consis­
study. Overall, 64 experts participated in both rounds. This number is tent with several previous Delphi studies in the SCM field, a threshold of
higher than or comparable to many recent Delphi studies in SCM IQR ≤ 25 was set to define consensus (Keller & von der Gracht, 2014;
research (Durach et al., 2017; Gossler, Wakolbinger, & Burkart, 2020; Lechler et al., 2019; Roßmann et al., 2018; Warth et al., 2013). This
Lechler et al., 2019). The final panel represents 15 countries and comes threshold indicates that at least 50% of all estimations fall within a range
from a diverse background of industry (73%) and academia (27%), as of 25 percentage points on the 0%–100% scale for the EP (von der
well as various professional domains (see the third row of Fig. 1). Het­ Gracht, 2012).
erogeneity of expert backgrounds is recommended in Delphi literature Following various recent Delphi studies in the SCM field (Hirschinger
to prevent various potential biases (Rowe & Wright, 2011; Winkler & et al., 2015; Roßmann et al., 2018), we used a fuzzy c-means (FCM)
Moser, 2016). Overall, we achieved a response rate of 9.2%, which, algorithm to assign each projection to a designated cluster according to
given the seniority of the identified experts and the long processing time, its EP, I, and D. We thus established a systematic scenario structure to
is considered adequate and comparable to similar studies’ response rates discuss quantitative results and contextualize the written statements
(Hirschinger et al., 2015; Lechler et al., 2019). (Tapio et al., 2011), which enables a more focused and objective
interpretation of the experts’ judgments.
3.5. Step 3: Execution of the Delphi study To systematically analyze the participants’ qualitative comments, we
applied a coding procedure based on Corbin and Strauss’s (2015) for
In the first Delphi round, all panelists were asked to assess each of the each projection. The written statements were classified according to
13 projections for the year 2030. Consistent with previous Delphi studies their sentiments toward the projection: supportive (arguing for a high
(Keller & von der Gracht, 2014; Lechler et al., 2019; Roßmann et al., rating), negative (arguing for a low rating), balanced/neutral (providing
2018), the following assessment dimensions were selected: pro and contra arguments or more general statements), or non-
applicable (if the comment was incomprehensible). To reduce investi­
• EP: The expected probability of occurrence on a scale from 0% to gator bias, two researchers with experience in content analysis coded
100% each statement individually. Any divergence was discussed until a
• I: The impact on SC performance in the case of an SCD1, based on a consensus was reached, thus improving inter-rater reliability. The ar­
five-point Likert scale (very low = 1, low = 2, medium = 3, high = 4, guments were incorporated into the discussion of long-term quantitative
and very high = 5) expert judgments.
• D: The desirability of occurrence, based on the same five-point Likert To conclude the analysis, the entire research team and two additional
scale academics, who had not been involved in the study, conducted a full-day
workshop to review the Delphi study results and inform the discussion
To foster qualitative data gathering, we asked panel members to add section of this paper.
written comments to their quantitative assessments for each dimension
(Tapio, Paloniemi, Varho, & Vinnari, 2011). After the first round, 4. Results and discussion
descriptive statistics (i.e., means and interquartile ranges per dimen­
sion) for all predictions were derived, and the qualitative comments 4.1. Quantitative Delphi results
were summarized. These results were shared with all participants as
input for the second round, in which panelists could reevaluate their The left side of Table 1 presents the experts’ average judgments for
assessments and add additional qualitative comments. To explore the each projection, the IQRs, and the CVs, while the right side reveals the
existence of a non-response bias among the panelists, we compared the results of coding the experts’ written statements. We found a wide EP
early (initial 10) and late (last 10) responders in both rounds of the range from 38% for P.10 (supplier selection and evaluation) to 89% for
survey for all assessment dimensions. As a Shapiro-Wilk test on P.9 (remote working). The impact was projected high (I ≥ 3.5) for 10 of
normality (p < 0.05) revealed a non-normal distribution of the sample the 13 projections, with P.3 (data aggregation) and P.4 (cross-company
for each of the dimensions (see Appendix A), we conducted a Wilcoxon- data platforms) rated highest (I ≥ 4). In contrast, the impact of P.6 (BC
Mann-Whitney test to explore the differences between the two respon­ information exchange) was rated lowest (I = 2.9). The respondents also
dent groups. The test revealed no significant differences (p < 0.05) assessed eight projections as highly desirable (D ≥ 3.5). Only P.10
(supplier selection and evaluation) indicates a desirability rating below
3, suggesting that experts consider it a threat.
1
SCRES assessment dimension, in line with SCRES definitions (e.g., Hohen­ We observed convergence in ratings across all projections from
stein et al.(2015), Hosseini et al.(2019). round 1 to round 2, which confirms that the Delphi method worked as

6
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Table 1
Quantitative results from the expert panel and sentiment analysis of qualitative comments per projection.
Quantitative results Qualitative comments’ trend

No. Projection EP IQR CV I D Supportive Negative Balanced/ Neutral n/a Sum

[0–100%] [0–100] [-100%-100%] [1–5] [1–5]
P.1 IoT data 73% 20.3 − 26% 3.8 4.1 64 9 23 3 99
P.2 AI external data 60% 40 − 10% 3.5 3.7 38 21 38 2 99
P.3 Data aggregation 78% 34.3 − 14% 4.5 4.6 55 3 20 1 79
P.4 Cross-company data platforms 64% 30 –22% 4.4 4.3 36 6 34 2 78
P.5 Digital SC twins 57% 31.8 − 11% 3.8 3.9 32 9 26 – 67
P.6 BC information exchange 48% 30.8 − 12% 2.9 3.1 22 36 31 – 89
P.7 Autonomous decision-making 53% 34.3 − 17% 3.5 3.3 12 5 51 – 68
P.8 Autonomous robots 57% 40 − 13% 3.5 3.2 13 12 44 2 71
P.9 Remote working 89% 17.5 − 8% 3.8 4.2 37 7 15 – 59
P.10 Supplier selection and evaluation 38% 34.8 − 11% 3.1 2.9 5 22 44 1 72
P.11 Supplier replacement 47% 40 − 11% 3.4 3.3 12 12 45 – 69
P.12 AM flexibility 50% 37 − 10% 3.5 3.5 16 14 36 – 66
P.13 Digital capabilities 80% 22 − 7% 3.8 4.1 38 1 6 – 45
Sum 380 157 413 11 961

Note: EP: Expected probability; I: Impact; D: Desirability; CV: Convergence (i.e., decrease in standard deviation).

intended (Rowe & Wright, 1999). Overall, standard deviation decreased very little industry experience with most of their careers spent in
by 13.2% from round 1 to round 2 and fell between 7% and 26% across research institutions. Hence, the little variation in results from both sub-
all projections, revealing a high convergence level to group consensus groups further strengthens the Delphi results’ credibility and validity as
(Roßmann et al., 2018; Rowe & Wright, 1999). Applying the previously the experts attained comparable estimates regardless of their back­
defined threshold of IQR ≤ 25, we conclude that consensus was reached grounds (Kopyto et al., 2020; Roßmann et al., 2018).
for three projections (P.1, P.9, and P.13). Leveraging the FCM approach regarding EP, I, and D, we determined
Since further insights can be derived from reviewing the differing three projection clusters with four to five projections each (see Fig. 2).
assessments of subgroups, we classified our panel into industry (n = 47) Cluster 1 comprises four projections with high ratings across di­
and academia (n = 17) subgroups. As a Shapiro-Wilk test confirmed a mensions (EP > 70%, I > 3.5, D > 3.5). Two projections with an EP over
non-normal distribution for each of the three dimensions in both rounds 80% directly affect conditions and requirements for employees. Panel­
(p < 0.05), we also conducted a Wilcoxon-Mann-Whitney test to assess ists believe that companies will make digital capabilities a requirement
differences between responses for each of the projections in the sub­ for SCRM personnel (P.13) and expect remote working to become a core
groups (see Appendix A). The results demonstrate that EP and D as­ lever to ensure business continuity (P.9). P.9 and P.13 also yielded the
sessments for P.10 (supplier selection and evaluation) and D for P.11 lowest conversion rates among all projections (-7% and − 8%, respec­
(supplier replacement) differed significantly (p < 0.05). No other sig­ tively) and an IQR ≤ 25, which indicates that participants are particu­
nificant deviations were found among the subgroups. This is an inter­ larly confident in their foresight and have reached consensus.
esting result as the academia panelists are rather distinct, having no or Cluster 2 comprises four projections evaluated with EPs between

Fig. 2. Classification of projections based on FCM method.

7
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

57% and 64%, suggesting slightly positive tendencies regarding their P.1 indicates a relatively low expected occurrence. This assessment can
realization until 2030. All four projections have thus been controver­ be explained by participants cautioning against the IoT’s high required
sially rated and discussed (IQR > 31). In terms of impact and desir­ investments, which constitutes a significant implementation barrier.
ability, three projections show average ratings between 3 and 4, while Participants also stressed the IoT’s potential incompatibility with other
P.4 indicates ratings above 4. technologies in a company’s digital infrastructure as IoT implementa­
Cluster 3 comprises the last five projections. With comparably low tion requires considerable adaptation to companies’ hardware and leg­
occurrence (EP < 55%), impact (I ≤ 3.5), and desirability (D ≤ 3.5) acy systems.
scores, these statements were considered little relevant until 2030. This Data aggregation (P.3). Delphi panelists expect that aggregating
cluster also includes the least probable and desirable projection (P.10: existing external and internal data in real-time for improved SC visibility
EP = 38%, D = 2.9) and the least impactful one (P.6: I = 2.9). However, and risk identification will be established in 2030. The panel further
it is essential to note that all statements were discussed controversially, considers this projection the most impactful and most desirable of all,
and some experts still see high probability and impact potential for all which indicates a significant influence on future SCRM. Explaining the
statements, as IQRs > 25 indicate. high likelihood of occurrence, experts consider this projection a
continuation of an ongoing trend. Many panelists have stressed that data
4.2. Qualitative results aggregation would be an essential baseline for technology-enabled de­
cision support systems and enhanced visibility to detect disruptions such
The experts provided 961 written statements (see right side of as those caused by COVID-19 earlier. It should be an initial priority
Table 1), which equals an average of 15 comments per participant. before more advanced solutions are implemented. Panelists who were
Moreover, 75% of all participants submitted at least one comment, more critical of P.3 indicated that company silos and fragmented data
which suggests a high level of interaction. In the following, we present systems often constrain holistic data aggregation. These organizational
the experts’ qualitative reasoning for their quantitative evaluations barriers must be overcome before effective collection and processing
along the three clusters. does become possible.

4.2.1. Cluster 1: Focus topics 4.2.2. Cluster 2: Promising candidates

Digital capabilities (P.13). Panel members expect that ensuring the Cross-company data platforms (P.4). Participants highlighted that
digital capabilities of the workforce will be crucial in 2030. They COVID-19 forcefully revealed how often SC visibility still ends at the
stressed P.13′ s desirability, highlighting that digital-savvy employees boundaries of a company’s operations. Panelists particularly stressed
will be more capable of data-driven decision-making, leading to an that data platforms will play a key role in addressing future disruptions
enhanced SCRM culture, including earlier and more precise risk iden­ by reducing information silos, offering more SC collaboration opportu­
tification and mitigation. One panel member stated that “none of the nities, and accelerating decision-making in crises, emphasizing such
concepts mentioned in this study can be implemented and maintained platforms’ scalability. Furthermore, the participants expressed only
without digital skill sets.” Some panelists even argued that implemented moderate confidence that P.4 will be realized in 2030, thus differing
I4.0 applications could be harmful to SCRES if employees administered from their trust in the other high-impact, high-desirability projections of
them without digital experience. Some panelists believe that this pro­ Cluster 1. They stressed significant implementation barriers for cross-
jection will become a reality even before 2030, and the trend will company data platforms, including SC partners’ reluctance to share in­
accelerate due to the COVID-19 crisis. However, a continuous lack of formation, high initial investment costs, and incompatible legacy sys­
digital talent within companies and on the labor market has recurrently tems. They highlighted the technology’s security to protect company-
been cited by the experts as a significant barrier to realizing P.13. specific data and intellectual property as a prerequisite for platform-
Remote working (P.9). During the COVID-19 pandemic, remote based cross-company collaboration. Panel members also cited power
working has become integral across SCs to ensure business continuity. dependence as an obstacle to realizing this projection.
Papanikolaou and Schmidt (2020) found that organizations with higher AI external data (P.2). Participants generally agreed on AI’s impor­
remote working capability have revealed significantly greater robust­ tance for screening external real-time information sources to enhance
ness against COVID-19-related disruptions. Panelists suppose that the visibility lever by identifying and assessing risks for a company’s SC
remote working capabilities will improve faster coordination for pan­ (Handfield et al., 2020). In this context, frequent reference to social
demics. Specifically, it should accelerate an organization’s reaction media data was made. While its large volume makes it difficult for
speed for white-collar workers and enable more flexible contingency humans to monitor, it is expected to provide quick insights in the case of
planning. Some experts have also mentioned that more decentralized disruptions. As the panel emphasized, AI could recognize pandemic-
and digital interactions will allow for significant cost savings, which can related risks and others with similar global effects earlier. AI will also
safeguard critical cash flows in an SCD. Although remote working ca­ prevent human bias in risk assessment, which is essential in high-
pabilities are currently mostly considered a reactive measure, our results impact/low-probability disruptions. However, as the average EP of
suggest that having such options in place will be an SCRM capability in 60% and an IQR of 40 indicate, participants were undecided on whether
itself and safeguard business continuity during future disruptions. this technology will be widely established in 2030. Critical panelists
Nevertheless, experts deem P.9′ s impact contingent on the SCD; while stressed AI’s low maturity, which makes it vulnerable to false positives.
remote working represents an effective SCRES lever in a pandemic such Moreover, some experts warned that companies might get carried away
as the COVID-19 crisis, it might not prove as decisive in a more decen­ by inflated expectations and forestall implementing more fundamental
tralized disruption, such as a financial shock. digital solutions that should be prioritized initially. Some panelists
IoT data (P.1). COVID-19 has exposed the critical need for more therefore argued that companies should spend their limited resources on
visibility throughout SCs, and panelists expect that the IoT will be highly other digitalization projects or using AI on internal data first.
leveraged in 2030 to improve SCRES through higher data availability Digital SC twins (P.5). Most experts consider digital SC twins a
and accuracy for SCRM decision-making. Participants see the technol­ powerful instrument for SCRM. As one participant stated, “current
ogy at the center of SC executives’ agendas even now. Moreover, pan­ decision-making in risk management includes many assumptions.”
elists connect the anticipated increase in IoT adoption to the expected Establishing digital twins will address and reduce this lack of accuracy in
trend toward more regionalization and increased safety stocks induced simulations, thus improving SCRES through three main levers, as the
by COVID-19. They argue that managing higher and more globally panelists noted: First, precise simulations will help to establish more
dispersed stock levels will require even more visibility regarding prod­ resilient SC designs proactively before a disruption occurs. Second, these
ucts’ statuses and locations. Compared to other projections in Cluster 1, advanced simulation capabilities will enable more proactive risk

8
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

identification, assessment, and mitigation, which – according to several experiences during the COVID-19 pandemic revealed.
panelists – has been lacking during the COVID-19 crisis. Third, from a Supplier selection and evaluation (P.10). Participating experts were
reactive perspective, firefighting will require fewer resources and be rather open to general automation in SCRM decision-making (P.7) but
quicker. Nevertheless, although SC data availability has significantly rejected the specific option of basing supplier selection and evaluation
increased within organizations (Vieira et al., 2019), panelists are skep­ exclusively on quantitative analyses. Although several panelists are
tical that digital twins will be established in 2030 since their full capa­ convinced that AI’s and BDA’s relevance will increase and improve the
bilities depend on unrestricted cross-organizational SC visibility, end-to- SCRES (re-)engineering lever, they widely stressed the importance of
end data aggregation, and data sharing between all stakeholders. The soft factors in supplier audits and negotiations. In this context, they
experts thus reported that enabling prerequisites such as projections P.1, referred explicitly to cultural proximity, buyers’ experience and gut
P.3, P.4, and P.13 must be achieved first. feelings, relational trust, and personal impressions from site visits.
Autonomous robots (P.8). The projections’ aggregated EP, I, and D Panelists also indicated that overly digitalized selection processes and
responses indicate no clear inclination to replace physical human labor the corresponding loss of interpersonal skills could ultimately harm
with autonomous robots for improved SCRES. Although some panelists collaborative buyer–supplier relationships. Nevertheless, participants
expect the experiences during the COVID-19 crisis to trigger a conscious presented two forms of hybrids in supplier selection. Panelists believe
effort for more production automation, they consider all SC risks too that supplier pre-selection and commodity material procurement will
diverse to be handled autonomously. While they pointed out positive soon be managed exclusively by intelligent software. Remarkably, the
risk-mitigating effects of robots during pandemics and for operations Wilcoxon-Mann-Whitney test revealed significant deviations in occur­
with little variability, the experts also highlighted the SCRES-improving rence and desirability ratings between panelists’ organizational types.
capabilities of humans. For instance, they emphasized humans’ higher Contrasting industry, academia reached consensus (IQR = 10) for a
flexibility and improvisation capacity, which they agreed is more much lower probability rating (EP = 29%) and noted less desirability (D
important during SCDs than under normal operations. Furthermore, = 2.4). This result indicates a significant and fairly manifested discon­
robots must be reprogrammed for every risk-mitigating modification of nect between industry and research on the future of digital procurement
processes, which may cost valuable time during disruptions. Panelists support, possibly caused by the lack of practice exposure of the
generally concluded that full automation is undesirable for SCRM, with academia panelists to the applications in question.
one panelist stating that “robots can be even harmful.” Overall, the ex­ Supplier replacement (P.11). (Re-)engineering the SC by replacing
perts see potential in robots for reducing certain risks but mainly believe suppliers not able or willing to share data on short-term delivery capa­
they will be employed for cost, efficiency, or lack of skilled labor rea­ bility was controversially discussed. Overall, experts expect consider­
sons, not SCRM-motivated ones. able SCRES improvements from ensuring higher supplier visibility. They
revealed that certain SC partners’ extended data sharing during the
4.2.3. Cluster 3: Wider circle COVID-19 pandemic has helped to detect bottlenecks in the SC and
BC information exchange (P.6). Panelists were skeptical about BC’s quickly identify alternative sources in the case of supplier capacity
potential to enhance collaboration, particularly trust and reliability, constraints. However, P.11 has a comparably low expected probability,
between SC entities until 2030. Most experts argued that current poor resulting from different priorities in supplier selection criteria, as the
BC implementation and understanding levels will hinder significant panelists confirmed. They believe that companies will continue priori­
SCRM improvements. Nevertheless, in addition to the speed of adoption, tizing a supplier’s cost and quality characteristics over potential visi­
the technology’s general advantageousness was questioned. Experts also bility benefits. Moreover, they stressed that some suppliers possess
pointed out that less advanced data exchange systems will suffice in certain unique value propositions, which translates into great negotia­
many situations, and only a limited number of truly beneficial use cases tion power. If these firms refuse more transparency – a scenario most
for BC exist. Experts therefore cautioned against overengineering with panelists anticipate – customers will have to accept these suppliers’
BC as a far more advanced solution than required for the problems to be business conditions and renounce visibility ambitions instead of
solved. Incidentally, several experts revealed that a lack of SC collabo­ replacing these players.
ration has not been apparent during the COVID-19 pandemic. Therefore, AM flexibility (P.12). The projection on AM’s potential to improve SC
BC’s possibilities for strengthening this SCRES lever might thus be more flexibility received comparably high impact and desirability scores.
advantageous for SCDs other than a pandemic. Moreover, missing reg­ Experts explained their ratings by highlighting AM’s advantages
ulatory guidelines and required high investments were discussed as regarding commodity and spare part supply, lead times, and local
further implementation barriers. Especially the second challenge was backup capacities. Additionally, different panelists mentioned the
repeatedly stressed to explain BC’s low desirability. We also observed a technology’s potential to reduce supplier numbers and tiers. In this
higher occurrence rating (EP = 52%) and a consensus among academic context, experts claimed that higher AM adoption would have helped to
participants (IQR = 22), which demonstrates more optimism about BC manage the COVID-19 pandemic by reducing supplier dependency. To
within this group. illustrate, many panelists referred to successful ad-hoc AM applications
Autonomous decision-making (P.7). The projection on autonomous during COVID-19, such as the decentralized 3D printing of personal
SCRM decision-making received the highest probability score in this protective equipment for hospitals and other care facilities to decrease
cluster. Nevertheless, many experts highlighted that humans will still supplier dependence and cope with considerable shortages. Although
assume the most important decisions in SCRM in 2030. The unforesee­ sound reasons for advancing the technology further exist, panelists
able spectrum of SCDs and liability regulations that require human indicated that they doubt implementation at scale will occur until 2030.
involvement were discussed as the main reasons. For more critical in­ Three main reasons were discussed: First, supply network adoptions
stances going beyond operational day-to-day and low-impact activities, would be significant, requiring more time. Second, several companies’
panelists expect digital technologies to support rather than replace parts portfolios are too complex to be produced predominantly by AM.
human decision-making. According to most panelists, a high number of Experts rather see high-risk mitigation potential for niche industrial use
time-consuming SCRM activities overstrained employees in 2020; cases, such as spare part production. Third, the required investments
comprehensive support from more digital solutions would have been and operational costs would be too high to scale AM, based on today’s
highly welcomed. A potential use case is thus intelligent software pre­ technological status.
senting next-best-action options, providing the human decision-maker
with a reliable basis for quickly assessing and choosing between 4.3. Discussion
different mitigation options. Such applications can free up employee
capacity for more complex tasks, which can be highly valuable, as recent The presented quantitative and qualitative results reveal that several

9
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

I4.0 technologies will be widely established in 2030 and considerably projections according to each one’s probability of occurrence, impact,
benefit resilience levers such as agility, collaboration, SC (re-)engi­ and desirability. Our analysis includes quantitative and qualitative ap­
neering, and SCRM culture. This overall finding validates various proaches and employs the FCM algorithm to ensure transparency in the
scholars’ conceptual arguments that future SCRES will increasingly experts’ evaluations. Important implications for theory, practice, and
depend on I4.0 after COVID-19 (e.g., Ivanov et al., 2019; van Hoek, policy-making were derived from the results.
2020) and indicates that the positive relationship between I4.0 and
SCRES confirmed for previous literature (Spieske & Birkel, 2021) will 5.1. Implications for theory
only strengthen in the future.
By interpreting the results for individual clusters and projections, we First, our paper is one of the primary studies that integrate multiple
can deduce more specific confirmations and contradictions in relation to I4.0 technologies in SCRES-related foresight research. We followed the
existing literature. Various “focus topics” and “promising candidates” calls of Zouari et al. (2020) and Ivanov et al. (2019), who demanded
enable data collection, exchange, and enhanced transparency. The pri­ empirical and more holistic investigations into different I4.0 solutions in
oritization of such technologies is consistent with Zouari et al. (2020), SCRM. Whereas most studies at the intersection of I4.0 and SCRES are
who found that digital connectivity and systems that foster information conceptual or have a narrow empirical focus (e.g., focusing on specific
sharing strongly influence SCRES and act as an enabler for more case studies), ours offers extensive quantitative and qualitative evidence
advanced digital tools. The results further empirically confirm argu­ from a broad range of experts. Second, our findings reveal that all
ments by Ivanov and Dolgui (2021) and Horváth and Szabó (2019), who investigated I4.0 technologies improve SCRES and are expected to be
claim that such solutions are the base requirement for end-to-end visi­ increasingly established until 2030. Thus, we can empirically confirm
bility and further SC collaboration to increase SCRES. While experts the claims that SCRM practices’ future success will increasingly depend
concur with the literature that AI will build on this foundation and offer on I4.0 (Ivanov et al., 2019; Ivanov & Dolgui, 2021). This development
far more precise and faster SC disruption prediction capabilities than is relevant for any kind of future SCD, including but not limited to
humans (Baryannis et al., 2019), wide-reaching maturity lies beyond a pandemics. At the same time, we contribute a nuanced perspective on
10-year horizon. The high rating and supportive qualitative reasoning I4.0 for SCRM by showing that the human factor will remain crucial in
for the remote working (P.9) and digital capabilities (P.13) propositions SCRM in 2030. While SCRES benefits of technologies such as BDA and AI
are consistent with observations of Zouari et al. (2020), who demon­ are clearly acknowledged, fully-automated risk management processes
strated that an organization’s digital maturity, including its employees’ or decision-making are not be expected and human judgement will
digital competencies and systems connectivity, has a more substantial remain critical. Connecting behavioral perspectives such as stakeholder
positive effect on SCRES than the adoption of specific digital tools. The or principal-agent theory with the I4.0-SCRM literature might therefore
fact that experts made clear that they do not expect fully automated be advantageous to accompany the transition. Third, while significant
decision-making and supplier selection (P.7 and P.10), as sometimes implementation progress can be expected, our analysis also reveals that
advocated in literature (e.g., Cavalcante et al., 2019), reinforces the the full range of I4.0 technologies will not yet be mature in 2030. We
notion that human capabilities and assessments will continue to play an found that visibility-enabling levers, including IoT applications and real-
essential role in future SCRM. This supports previous research claiming time data exchange, and SCRM culture-related levers, such as digital
that mostly routine activities will be automated (e.g., Ralston workforce capabilities, represent base requirements for digital SCRM.
& Blackhurst, 2020) and contradicts scholars’ claims that several These technologies will have initial priority as they build the necessary
relationship-based decision-making criteria such as personal experience, baseline for more sophisticated digital SCRES-improving approaches.
trust, and gut feeling will be less critical (Ocampo, Abad, Cabusas, Furthermore, our analysis suggests that the future road map for SCRES-
Padon, & Sevilla, 2018). Furthermore, experts’ evaluations of the BC enabling I4.0 technologies continues much beyond the 2030 horizon. On
(P.6) and the AM (P.12) projections refute some aspects attributed to a different note, the comparatively low ratings and qualitative data for
these technologies by previous works. While several participants the “wider circle” cluster indicate that applicable use cases for the
referred to BC’s visibility and velocity benefits as a more promising way included I4.0 technologies have not been developed or their benefit is
of enhancing SCRES, they contradicted previous research (e.g., Min, not sufficiently convincing. Particularly BC and AM revealed a more
2019; P. Xu et al., 2021) by denying significant collaboration im­ limited application scope than claimed in previous literature. Finally, we
provements. Previous foresight research on AM claimed a considerable provide empirical foresight for the post-COVID-19 path and the maturity
impact of the technology on a broad scope of processes in various in­ of I4.0 technologies in a SCRES context, thus offering the research
dustries within the next five to ten years (Durach et al., 2017). The re­ community insights into how the intersection of SCRM and I4.0 must be
sults of this Delphi study rather point to a longer time to maturity conceptualized after COVID-19.
(beyond 2030) and a more limited scope of beneficial use cases. In
conjunction with the evaluation of other projections in Cluster 3, all 5.2. Implications for practice
referring to the (re-)engineering lever, we can infer that far-reaching
digitalization initiatives demanding considerable supply network ad­ From a managerial perspective, our study results offer both tactical
aptations are unlikely to occur after the COVID-19 pandemic. and strategic guidance. First – and most importantly – our study con­
firms the benefits of I4.0 application in SCRM and encourages practi­
5. Conclusion tioners toward accelerated SC digitalization. Therefore, managers
should assume a holistic perspective and consider I4.0 technologies’
The COVID-19 crisis has put SCRM to the test in a wide array of potential to improve resilience when deciding on implementing digital
industries. Moreover, since “the supply chain will be [only] as good as solutions in their SCs. Second, we identified that several SCRES-
the digital technology behind it” (Ivanov et al., 2019, p. 838), scholars improving I4.0 solutions face similar implementation barriers. Com­
have increasingly called for I4.0 solutions to mitigate SCDs (Chowdhury mon challenges include high investment costs, a lack of digital capa­
et al., 2020). Given the experiences during the COVID-19 pandemic, we bilities, a lack of compatibility with legacy or fragmented data systems,
analyzed the momentum of digitalization and the future role of I4.0 in and organizational silos within and between companies. Managers are
strengthening SCRES. thus advised to design their companies’ digitalization processes holis­
We built on existing I4.0 and SCRM literature and developed 13 tically and enable various I4.0 technologies by addressing common
projections on how I4.0 technologies will impact SCRES for 2030. barriers together. Third, with this study’s results, practitioners can pri­
Conducting a Delphi study, we surveyed 64 SCRM experts from industry oritize implementing I4.0 technologies in a more nuanced fashion. By
and academia, who assessed and discussed the future-oriented clarifying the future road map of I4.0 support for SCRM, we enable

10
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

managers to compare the expected development to their companies’ qualitatively identifies the projections’ main implementation barriers,
I4.0 maturity and inform their decision-making frameworks for future we have not investigated their interrelations, levels of influence, or ways
digitalization initiatives. Finally, managers can deduct from our study to overcome them. Future research is particularly important for pro­
that not all impactful and desirable changes through I4.0 technologies jections of the “promising candidates” and “wider circle” clusters as
must involve the structural reconfiguration of internal processes and panelists expect a considerable impact on SC performance and major
applications. For instance, practitioners can considerably increase their implementation challenges. Third, our study design fostered individual
digital SCRM activities’ effectiveness by focusing on common enabling assessments of the investigated I4.0-related SCRM solutions. As our
prerequisites, such as IoT implementation for data availability and ac­ experts’ comments revealed, further understanding of their in­
curacy, data visibility and internal and external information sharing, terdependencies could increase SCRM performance. Fourth, the Delphi
establishing cross-company data platforms, and fostering digital capa­ panel is mainly comprised of experts from European countries and the
bilities. Our results suggest that these comparatively mature technology USA. Similar studies with respondents from other geographies, partic­
levers can improve SCRES as well as or even better than more advanced ularly Asia given its critical contribution to global SCs, might yield
applications such as AI, BC, and digital twins. varying results due to alternate economic and regulatory environments.
Fifth, the study compared evaluations of industry and academia sub-
5.3. Implications for policy groups with no significant deviations besides projections on digital
enablement of supplier assessment and management. While the little
Our study confirms a significant contribution of I4.0 technologies to variations indicate alignment between the two groups, scholars should
SCRES by 2030. It must be in public authorities’ interest to support aim for higher exposure to practice on the dissent topics and initiate
measures that can mitigate the negative consequences of disruptive stronger academia-industry exchange. Finally, we consciously limited
events such as the COVID-19 pandemic. Several policy implications can the number of projections in our sample to ensure completion of the
thus be drawn from our results. First, we demonstrated that visibility- survey, even though other unexplored topics exist and should be
enhancing digital solutions can build an essential foundation for investigated.
improving SCRES. Policymakers should therefore incentivize or even This research did not receive any specific grant from funding
enforce more information sharing among companies while ensuring agencies in the public, commercial, or not-for-profit sectors.
trust in data protection. Such efforts could include establishing
comprehensive and harmonized data security standards, common but CRediT authorship contribution statement
anonymized industry databases, or minimum information sharing re­
quirements. Second, policymakers should consider subsidy programs for Alexander Spieske: Conceptualization, Data curation, Investiga­
I4.0 solutions for SCRM. As Delphi participants recurrently discussed, tion, Methodology, Validation, Visualization, Project administration,
high investments are among the most severe barriers to implementing Writing – original draft. Maximilian Gebhardt: Conceptualization,
I4.0 at scale. Third, policymakers should aspire to create the appropriate Data curation, Investigation, Methodology, Validation, Project admin­
legal frameworks for these transformative technologies, enabling com­ istration, Writing – original draft. Matthias Kopyto: Conceptualization,
panies to capture the benefits of I4.0 for SCRM while ensuring the Data curation, Investigation, Methodology, Validation, Writing – orig­
required legal certainty. Finally, governments should support the rapid inal draft. Hendrik Birkel: Conceptualization, Investigation, Method­
buildup of digital capabilities as it represents a core requirement for the ology, Supervision, Validation, Writing – original draft, Writing – review
success of I4.0 for better SCRM. Such measures could include more & editing. Evi Hartmann: Conceptualization, Methodology, Supervi­
holistic financial support for research and education programs at uni­ sion, Validation, Writing – review & editing.
versities or comprehensive re-skilling of the labor market.
Declaration of Competing Interest
5.4. Limitations and future research
The authors declare that they have no known competing financial
As with any scientific paper, this study’s limitations provide oppor­ interests or personal relationships that could have appeared to influence
tunities for further research. First, it empirically confirms several use the work reported in this paper.
cases of I4.0 technologies for SCRES improvement but lacks insights into
their precise implementation. We encourage researchers to investigate Data availability
implementation practices of I4.0 in SCRM empirically, for instance,
using an in-depth single-case study approach. Second, while our study No data was used for the research described in the article.

Appendix

Appendix A. . Results of statistical tests.

1. Shapiro-Wilk test on normality

Dimension p-value

EP 0.0000000000002405*
I <0.00000000000000022*
D <0.00000000000000022*

Note: *p < 0.05 allows to reject the null-hypothesis on a normal distribution of the expert assessments.
2. Wilcoxon-Mann-Whitney test on non-response bias
Dimension p-value

EP 0.238
I 0.647
D 0.104

11
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Note: *p < 0.05 allows to reject the null-hypothesis that two samples belong to the same population (here: comparison of initial 10 and last 10 re­
sponders’ (=approximation for non-responders) assessments).
3. Wilcoxon-Mann-Whitney test on subgroup differences
No. Projection Dimension p-value

P.1 IoT data EP 0.669

P.1 IoT data I 0.928
P.1 IoT data D 0.987
P.2 AI external data EP 0.885
P.2 AI external data I 0.661
P.2 AI external data D 0.511
P.3 Data aggregation EP 0.371
P.3 Data aggregation I 0.366
P.3 Data aggregation D 0.520
P.4 Cross-company data platforms EP 0.855
P.4 Cross-company data platforms I 0.212
P.4 Cross-company data platforms D 0.402
P.5 Digital SC twins EP 0.958
P.5 Digital SC twins I 0.697
P.5 Digital SC twins D 0.705
P.6 BC information exchange EP 0.293
P.6 BC information exchange I 0.425
P.6 BC information exchange D 0.705
P.7 Autonomous decision-making EP 0.168
P.7 Autonomous decision-making I 0.411
P.7 Autonomous decision-making D 0.753
P.8 Autonomous robots EP 0.891
P.8 Autonomous robots I 0.182
P.8 Autonomous robots D 0.645
P.9 Remote working EP 0.523
P.9 Remote working I 0.060
P.9 Remote working D 0.419
P.10 Supplier selection and evaluation EP 0.029*
P.10 Supplier selection and evaluation I 0.785
P.10 Supplier selection and evaluation D 0.039*
P.11 Supplier replacement EP 0.849
P.11 Supplier replacement I 0.576
P.11 Supplier replacement D 0.024*
P.12 AM flexibility EP 0.311
P.12 AM flexibility I 0.472
P.12 AM flexibility D 1.000
P.13 Digital capabilities EP 0.969
P.13 Digital capabilities I 0.876
P.13 Digital capabilities D 0.942

Note: *p < 0.05 allows to reject the null-hypothesis that two samples belong to the same population (here: comparison of academic and industry
experts’ assessments).

References Chen, L., Dui, H., & Zhang, C. (2020). A resilience measure for supply chain systems
considering the interruption with the cyber-physical systems. Reliability Engineering
& System Safety, 199, Article 106869. https://doi.org/10.1016/j.ress.2020.106869
Ali, I., & Gölgeci, I. (2019). Where is supply chain resilience research heading? A
Choi, T.-M., Wallace, S. W., & Wang, Y. (2018). Big data analytics in operations
systematic and co-occurrence analysis. International Journal of Physical Distribution &
management. Production and Operations Management, 27(10), 1868–1883. https://
Logistics Management, 49(8), 793–815. https://doi.org/10.1108/IJPDLM-02-2019-
doi.org/10.1111/poms.12838
0038
Chowdhury, M. T., Sarkar, A., Paul, S. K., & Moktadir, M. A. (2020). A case study on
Baryannis, G., Validi, S., Dani, S., & Antoniou, G. (2019). Supply chain risk management
strategies to deal with the impacts of COVID-19 pandemic in the food and beverage
and artificial intelligence: State of the art and future research directions. International
industry. Operations Management Research, 1–13. https://doi.org/10.1007/s12063-
Journal of Production Research, 57(7), 2179–2202. https://doi.org/10.1080/
020-00166-9
00207543.2018.1530476
Christopher, M., & Peck, H. (2004). Building the resilient supply chain. International
Basole, R. C. (2014). Supply network structure, visibility, and risk diffusion: A
Journal of Logistics Management, 15(2), 1–13. https://doi.org/10.1108/
computational approach. Decision Sciences, 45(4), 753–789.
09574090410700275
Belhadi, A., Kamble, S., Jabbour, C. J. C., Gunasekaran, A., Ndubisi, N. O., &
Chunsheng, L., Wong, C. W., Yang, C.-C., Shang, K.-C., & Lirn, T. (2020). Value of supply
Venkatesh, M. (2021). Manufacturing and service supply chain resilience to the
chain resilience: Roles of culture, flexibility, and integration. International Journal of
COVID-19 outbreak: Lessons learned from the automobile and airline industries.
Physical Distribution & Logistics Management, 50(1), 80–100. https://doi.org/
Technological Forecasting and Social Change, 163, Article 120447. https://doi.org/
10.1108/IJPDLM-02-2019-0041
10.1016/j.techfore.2020.120447
Corbin, J. M., & Strauss, A. L. (2015). Basics of qualitative research: Techniques and
Birkel, H. S., & Hartmann, E. (2020). Internet of Things – the future of managing supply
procedures for developing grounded theory (4th ed.). Thousand Oaks, CA: Sage
chain risks. Supply Chain Management: An International Journal, 25(5), 535–548.
Publications.
https://doi.org/10.1108/SCM-09-2019-0356
Craighead, C. W., Ketchen, D. J., & Darby, J. L. (2020). Pandemics and supply chain
Bragazzi, N. L. (2020). Digital technologies-enabled smart manufacturing and Industry
management research: Toward a theoretical toolbox. Decision Sciences, 51(4),
4.0 in the post-COVID-19 era: Lessons learnt from a pandemic. International Journal
838–866. https://doi.org/10.1111/deci.12468
of Environmental Research and Public Health, 17(13), 4785. https://doi.org/10.3390/
Darkow, I.-L., Foerster, B., & von der Gracht, H. A. (2015). Sustainability in food service
ijerph17134785
supply chains: Future expectations from European industry experts toward the
Cavalcante, M. I., Frazzon, E. M., Forcellini, F. A., & Ivanov, D. (2019). A supervised
environmental perspective. Supply Chain Management: An International Journal, 20
machine learning approach to data-driven simulation of resilient supplier selection
(2), 163–178. https://doi.org/10.1108/SCM-03-2014-0087
in digital manufacturing. International Journal of Information Management, 49, 86–97.
Dasaklis, T. K., Pappis, C. P., & Rachaniotis, N. P. (2012). Epidemics control and logistics
https://doi.org/10.1016/j.ijinfomgt.2019.03.004
operations: A review. International Journal of Production Economics, 139(2), 393–410.
Chae, B. (2015). Insights from hashtag #supplychain and twitter analytics: Considering
https://doi.org/10.1016/j.ijpe.2012.05.023
twitter and twitter data for supply chain practice and research. International Journal
of Production Economics, 165, 247–259. https://doi.org/10.1016/j.ijpe.2014.12.037

12
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Dolgui, A., & Ivanov, D. (2020). Exploring supply chain structural dynamics: New international expert panel. Technological Forecasting and Social Change, 161, Article
disruptive technologies and disruption risks. International Journal of Production 120330. https://doi.org/10.1016/j.techfore.2020.120330
Economics, 229, Article 107886. https://doi.org/10.1016/j.ijpe.2020.107886 Kumar, S., Raut, R. D., Narwane, V. S., & Narkhede, B. E. (2020). Applications of Industry
Dolgui, A., Ivanov, D., & Sokolov, B. (2020). Reconfigurable supply chain: The X- 4.0 to overcome the COVID-19 operational challenges. Diabetes & Metabolic
network. International Journal of Production Research, 58(13), 4138–4163. https:// Syndrome, 14(5), 1283–1289. https://doi.org/10.1016/j.dsx.2020.07.010
doi.org/10.1080/00207543.2020.1774679 Lechler, S., Canzaniello, A., Roßmann, B., von der Gracht, H. A., & Hartmann, E. (2019).
Dubey, R., Gunasekaran, A., Bryde, D. J., Dwivedi, Y. K., & Papadopoulos, T. (2020). Real-time data processing in supply chain management: Revealing the uncertainty
Blockchain technology for enhancing swift-trust, collaboration and resilience within dilemma. International Journal of Physical Distribution & Logistics Management, 49(10),
a humanitarian supply chain setting. International Journal of Production Research, 58 1003–1019. https://doi.org/10.1108/IJPDLM-12-2017-0398
(11), 3381–3398. https://doi.org/10.1080/00207543.2020.1722860 Li, Z., Guo, H., Barenji, A. V., Wang, W. M., Guan, Y., & Huang, G. Q. (2020).
Dubey, R., Gunasekaran, A., Childe, S. J., Fosso Wamba, S., Roubaud, D., & Foropon, C. A sustainable production capability evaluation mechanism based on blockchain,
(2021). Empirical investigation of data analytics capability and organizational LSTM, analytic hierarchy process for supply chain network. International Journal of
flexibility as complements to supply chain resilience. International Journal of Production Research, 58(24), 7399–7419. https://doi.org/10.1080/
Production Research, 59(1), 110–128. https://doi.org/10.1080/ 00207543.2020.1740342
00207543.2019.1582820 Linstone, H. A., & Turoff, M. (Eds.). (1975). The Delphi method: techniques and applications.
Durach, C. F., Kurpjuweit, S., & Wagner, S. M. (2017). The impact of additive Boston, MA: Addison-Wesley Publishing.
manufacturing on supply chains. International Journal of Physical Distribution & Min, H. (2019). Blockchain technology for enhancing supply chain resilience. Business
Logistics Management, 47(10), 954–971. https://doi.org/10.1108/IJPDLM-11-2016- Horizons, 62(1), 35–45. https://doi.org/10.1016/j.bushor.2018.08.012
0332 Mitchel, V. (1996). Assessing the reliability and validity of questionnaires: An empirical
Er Kara, M., Oktay Fırat, S.Ü., & Ghadge, A. (2020). A data mining-based framework for example. Journal of Applied Management Studies, 5, 199–208.
supply chain risk management. Computers & Industrial Engineering, 139, Article Modgil, S., Gupta, S., Stekelorum, R., & Laguir, I. (2021). AI technologies and their
105570. https://doi.org/10.1016/j.cie.2018.12.017 impact on supply chain resilience during COVID-19. International Journal of Physical
Frederico, G. F., Garza-Reyes, J. A., Anosike, A., & Kumar, V. (2019). Supply Chain 4.0: Distribution & Logistics Management, ahead-of-print.. https://doi.org/10.1108/
Concepts, maturity and research agenda. Supply Chain Management: An International IJPDLM-12-2020-0434
Journal, 25(2), 262–282. https://doi.org/10.1108/SCM-09-2018-0339 Nguyen, S., Chen, P.-S.-L., & Du, Y. (2020). Risk identification and modeling for
Ghadge, A., Er Kara, M., Moradlou, H., & Goswami, M. (2020). The impact of Industry blockchain-enabled container shipping. International Journal of Physical Distribution
4.0 implementation on supply chains. Journal of Manufacturing Technology & Logistics Management, 51(2), 126–148. https://doi.org/10.1108/IJPDLM-01-2020-
Management, 31(4), 669–686. https://doi.org/10.1108/JMTM-10-2019-0368 0036
Gossler, T., Wakolbinger, T., & Burkart, C. (2020). Outsourcing in humanitarian logistics Nikolopoulos, K., Punia, S., Schäfers, A., Tsinopoulos, C., & Vasilakis, C. (2021).
– status quo and future directions. International Journal of Physical Distribution & Forecasting and planning during a pandemic: Covid-19 growth rates, supply chain
Logistics Management, 50(4), 403–438. https://doi.org/10.1108/IJPDLM-12-2018- disruptions, and governmental decisions. European Journal of Operational Research,
0400 290(1), 99–115. https://doi.org/10.1016/j.ejor.2020.08.001
Grötsch, V. M., Blome, C., & Schleper, M. C. (2013). Antecedents of proactive supply Ocampo, L. A., Abad, G. K. M., Cabusas, K. G. L., Padon, M. L. A., & Sevilla, N. C. (2018).
chain risk management – a contingency theory perspective. International Journal of Recent approaches to supplier selection: A review of literature within 2006–2016.
Production Research, 51(10), 2842–2867. https://doi.org/10.1080/ International Journal of Integrated Supply Management, 12(1/2), 22–68. https://doi.
00207543.2012.746796 org/10.1504/IJISM.2018.095683
Handfield, R., Jeong, S., & Choi, T. (2019). Emerging procurement technology: Data de Oliveira, M. P. V., & Handfield, R. (2019). Analytical foundations for development of
analytics and cognitive analytics. International Journal of Physical Distribution & real-time supply chain capabilities. International Journal of Production Research, 57
Logistics Management, 49(10), 972–1002. https://doi.org/10.1108/IJPDLM-11-2017- (5), 1571–1589. https://doi.org/10.1080/00207543.2018.1493240
0348 Papanikolaou, D., & Schmidt, L. D. (2020). Working remotely and the supply-side impact of
Handfield, R., Sun, H., & Rothenberg, L. (2020). Assessing supply chain risk for apparel COVID-19. National Bureau of Economic Research Working Paper Series, No. 27330.
production in low cost countries using newsfeed analysis. Supply Chain Management: https://doi.org/10.3386/w27330
An International Journal, 25(6), 1–19. https://doi.org/10.1108/SCM-11-2019-0423 Paul, S. K., & Chowdhury, P. (2020). A production recovery plan in manufacturing
Hirschinger, M., Spickermann, A., Hartmann, E., von der Gracht, H. A., & Darkow, I.-L. supply chains for a high-demand item during COVID-19. International Journal of
(2015). The future of logistics in emerging markets-fuzzy clustering scenarios Physical Distribution & Logistics Management, 51(2), 104–125. https://doi.org/
grounded in institutional and factor-market rivalry theory. Journal of Supply Chain 10.1108/IJPDLM-04-2020-0127
Management, 51(4), 73–93. https://doi.org/10.1111/jscm.12074 Pettit, T. J., Croxton, K. L., & Fiksel, J. (2019). The evolution of resilience in supply chain
Hofmann, E., Sternberg, H., Chen, H., Pflaum, A., & Prockl, G. (2019). Supply chain management: A retrospective on ensuring supply chain resilience. Journal of Business
management and Industry 4.0: Conducting research in the digital age. International Logistics, 40(1), 56–65. https://doi.org/10.1111/jbl.12202
Journal of Physical Distribution & Logistics Management, 49(10), 945–955. https://doi. Queiroz, M. M., Pereira, S. C. F., Telles, R., & Machado, M. C. (2021). Industry 4.0 and
org/10.1108/IJPDLM-11-2019-399 digital supply chain capabilities. Benchmarking: An International Journal, 28(5),
Hohenstein, N.-O., Feisel, E., Hartmann, E., & Giunipero, L. (2015). Research on the 1761–1782. https://doi.org/10.1108/BIJ-12-2018-0435
phenomenon of supply chain resilience. International Journal of Physical Distribution Ralston, P., & Blackhurst, J. (2020). Industry 4.0 and resilience in the supply chain: A
& Logistics Management, 45(1/2), 90–117. https://doi.org/10.1108/IJPDLM-05- driver of capability enhancement or capability loss? International Journal of
2013-0128 Production Research, 58(16), 5006–5019. https://doi.org/10.1080/
Horváth, D., & Szabó, R. Z. (2019). Driving forces and barriers of Industry 4.0: Do 00207543.2020.1736724
multinational and small and medium-sized companies have equal opportunities? Roßmann, B., Canzaniello, A., von der Gracht, H. A., & Hartmann, E. (2018). The future
Technological Forecasting and Social Change, 146, 119–132. https://doi.org/10.1016/ and social impact of Big Data Analytics in Supply Chain Management: Results from a
j.techfore.2019.05.021 Delphi study. Technological Forecasting and Social Change, 130, 135–149. https://doi.
Hosseini, S., Ivanov, D., & Dolgui, A. (2019). Review of quantitative methods for supply org/10.1016/j.techfore.2017.10.005
chain resilience analysis. Transportation Research Part E: Logistics and Transportation Rowe, G., & Wright, G. (1999). The Delphi technique as a forecasting tool: Issues and
Review, 125, 285–307. analysis. International Journal of Forecasting, 15(4), 353–375. https://doi.org/
Ivanov, D. (2020). Predicting the impacts of epidemic outbreaks on global supply chains: 10.1016/S0169-2070(99)00018-7
A simulation-based analysis on the coronavirus outbreak (COVID-19/SARS-CoV-2) Rowe, G., & Wright, G. (2011). The Delphi technique: Past, present, and future prospects
case. Transportation Research Part E: Logistics and Transportation Review, 136, Article - introduction to the special issue. Technological Forecasting and Social Change, 78(9),
101922. https://doi.org/10.1016/j.tre.2020.101922 1487–1490. https://doi.org/10.1016/j.techfore.2011.09.002
Ivanov, D., & Dolgui, A. (2021). A digital supply chain twin for managing the disruption Salancik, J. R., Wenger, W., & Helfer, E. (1971). The construction of Delphi event
risks and resilience in the era of Industry 4.0. Production Planning & Control, 32(9), statements. Technological Forecasting and Social Change, 3, 65–73. https://doi.org/
775–788. https://doi.org/10.1080/09537287.2020.1768450 10.1016/S0040-1625(71)80004-5
Ivanov, D., Dolgui, A., & Sokolov, B. (2019). The impact of digital technology and Scheibe, M., Skutsch, M., & Schofer, J. (1975). Experiments in Delphi methodology. In
Industry 4.0 on the ripple effect and supply chain risk analytics. International Journal H. A. Linstone, & M. Turoff (Eds.), The Delphi method: techniques and applications (pp.
of Production Research, 57(3), 829–846. https://doi.org/10.1080/ 262–287). Boston, MA: Addison-Wesley Publishing.
00207543.2018.1488086 Scholten, K., Sharkey Scott, P., & Fynes, B. (2019). Building routines for non-routine
Javaid, M., Haleem, A., Singh, R. P., Haq, M. I. U., Raina, A., & Suman, R. (2020). events: Supply chain resilience learning mechanisms and their antecedents. Supply
Industry 5.0: Potential applications in COVID-19. Journal of Industrial Integration and Chain Management: An International Journal, 24(3), 430–442. https://doi.org/
Management, 5(4), 507–530. https://doi.org/10.1142/S2424862220500220 10.1108/SCM-05-2018-0186
Kaur, H., & Prakash Singh, S. (2021). Multi-stage hybrid model for supplier selection and Spickermann, A., Zimmermann, M., & von der Gracht, H. A. (2014). Surface- and deep-
order allocation considering disruption risks and disruptive technologies. level diversity in panel selection - exploring diversity effects on response behaviour
International Journal of Production Economics, 231, Article 107830. https://doi.org/ in foresight. Technological Forecasting and Social Change, 85, 105–120. https://doi.
10.1016/j.ijpe.2020.107830 org/10.1016/j.techfore.2013.04.009
Keller, J., & von der Gracht, H. A. (2014). The influence of information and Spieske, A., & Birkel, H. (2021). Improving supply chain resilience through industry 4.0:
communication technology (ICT) on future foresight processes - results from a A systematic literature review under the impressions of the COVID-19 pandemic.
Delphi survey. Technological Forecasting and Social Change, 85, 81–92. https://doi. Computers & Industrial Engineering, 158, Article 107452. https://doi.org/10.1016/j.
org/10.1016/j.techfore.2013.07.010 cie.2021.107452
Kopyto, M., Lechler, S., von der Gracht, H. A., & Hartmann, E. (2020). Potentials of Stank, T., Esper, T., Goldsby, T. J., Zinn, W., & Autry, C. (2019). Toward a digitally
blockchain technology in supply chain management: Long-term judgments of an dominant paradigm for twenty-first century supply chain scholarship. International

13
A. Spieske et al. Computers & Industrial Engineering 181 (2023) 109344

Journal of Physical Distribution & Logistics Management, 49(10), 956–971. https://doi. Journal of Production Research, 58(17), 5320–5336. https://doi.org/10.1080/
org/10.1108/IJPDLM-03-2019-0076 00207543.2020.1720925
Tapio, P., Paloniemi, R., Varho, V., & Vinnari, M. (2011). The unholy marriage? Winkler, J., Kuklinski, C.-P.-J.-W., & Moser, R. (2015). Decision making in emerging
Integrating qualitative and quantitative information in Delphi processes. markets: The Delphi approach’s contribution to coping with uncertainty and
Technological Forecasting and Social Change, 78(9), 1616–1628. https://doi.org/ equivocality. Journal of Business Research, 68(5), 1118–1126. https://doi.org/
10.1016/j.techfore.2011.03.016 10.1016/j.jbusres.2014.11.001
Van Hoek, R. (2020). Research opportunities for a more resilient post-COVID-19 supply Winkler, J., & Moser, R. (2016). Biases in future-oriented Delphi studies: A cognitive
chain – closing the gap between research findings and industry practice. International perspective. Technological Forecasting and Social Change, 105, 63–76. https://doi.org/
Journal of Operations & Production Management, 40(4), 341–355. https://doi.org/ 10.1016/j.techfore.2016.01.021
10.1108/IJOPM-03-2020-0165 Wu, D., Zhang, C. [Chenghao], Pan, Y., & Dolgui, A. (2021). The supply chain effects on
Vieira, A. A. C., Dias, L. M. S., Santos, M. Y., Pereira, G. A. B., & Oliveira, J. A. (2019). order strategy of cross-shareholdings. International Journal of Production Research, 59
Simulation of an automotive supply chain using big data. Computers & Industrial (22), 6848–6863. https://doi.org/10.1080/00207543.2020.1828639.
Engineering, 137, 1–14. https://doi.org/10.1016/j.cie.2019.106033 Xu, P., Lee, J., Barth, J. R., & Richey, R. G. (2021). Blockchain as supply chain
Von der Gracht, H. A. (2008). The Delphi technique for futures research. In H. A. von der technology: Considering transparency and security. International Journal of Physical
Gracht (Ed.), The Future of Logistics (pp. 21–68). Wiesbaden: Gabler. Distribution & Logistics Management, 51(3), 305–324. https://doi.org/10.1108/
Von der Gracht, H. A. (2012). Consensus measurement in Delphi studies. Technological IJPDLM-08-2019-0234
Forecasting and Social Change, 79(8), 1525–1536. https://doi.org/10.1016/j. Xu, Z., Elomri, A., Kerbache, L., & El Omri, A. (2020). Impacts of COVID-19 on global
techfore.2012.04.013 supply chains: Facts and perspectives. IEEE Engineering Management Review, 48(3),
Warth, J., von der Gracht, H. A., & Darkow, I.-L. (2013). A dissent-based approach for 153–166. https://doi.org/10.1109/EMR.2020.3018420
multi-stakeholder scenario development — the future of electric drive vehicles. Zhang, F., Wu, X., Tang, C. S., Feng, T., & Dai, Y. (2020). Evolution of operations
Technological Forecasting and Social Change, 80(4), 566–583. https://doi.org/ management research: From managing flows to building capabilities. Production and
10.1016/j.techfore.2012.04.005 Operations Management, 29(10), 2219–2229.
Wichmann, P., Brintrup, A., Baker, S., Woodall, P., & McFarlane, D. (2020). Extracting Zouari, D., Ruel, S., & Viale, L. (2020). Does digitalising the supply chain contribute to its
supply chain maps from news articles using deep neural networks. International resilience? International Journal of Physical Distribution & Logistics Management, 51(2),
149–180. https://doi.org/10.1108/IJPDLM-01-2020-0038

14