23rd Cambridge International Manufacturing Symposium

University of Cambridge, 26 – 27 September 2019

Digital Transformation in the Automotive Supply Chain

Gary Graham a*, Laird Burns b, Patrick Hennelly c
a
Leeds University, Leeds, UK
b
University of Alabama in Huntsville, US
c
Royal College of the Arts London, UK
*
[email protected], [email protected], [email protected]

Abstract
In this paper, we begin the process of developing practice-based research problems and questions
concerning supply chain digitalization in the automotive industry. Specifically, our focus is on
understanding organizational conditions facilitating supply chain and production digitalisation and its role
in performance improvement. Therefore, as a first step we organized a research workshop during last
year’s Cambridge Center for International Manufacturing (CIM) Symposium with 30 executives
representing multiple tiers of the automotive supply chain. The purpose was identifying key themes and
patterns amongst these themes, which provides a basis for a practitioner-based, impact led research
agenda. We presented the executives with ten digital scenarios developed by CIM, and recorded responses
their responses to each scenario in a template of the ten scenarios. Using social network analysis (SNA)
we then interrogated the qualitative data to identify: “inter-codal” patterns, linkages and relationships.
Keywords: Automotive Industry; Workshop; Social Network Analysis

1. Introduction
Digitalisation is the use and adoption of external digital technologies (i.e. resources such as machine learning, IoT,
big data and blockchain) by organizations, to improve supply chain and operational performance (Holmstrom and
Partanen, 2014). Whilst there is a lot of emphasis on the strategic adoption of digital technologies the work on
operational feasibility, research on maturity and supply chain implementation is scarce (Seyedghorban et al.,
2019). In order to explore implementation from a practice-based perspective, this paper presents the findings of an
expert workshop of OEM’s (original equipment manufacturers), OES’s (original equipment suppliers, Tier 1), an
automotive trade association, and high technology consultants that we organized in a research workshop at the
22nd Cambridge International Manufacturing Symposium, in 2018.
We had thirty participants who attended the half-day workshop and participated in a follow up interview survey.
Using KNIME, we deployed text analytics to perform qualitative coding analysis on the results. The aim is to
identify the most common occurring and co-occurring terms within a network encompassing the terms that have
been identified from the ten scenarios. KNIME uses six different steps to process texts: reading and parsing
documents, named entity recognition, filtering and manipulation, word counting and keyword extraction,
transformation and visualization. It provides a neutral coding facility free of human bias.

2. Method
Ten digital scenario topics were presented to the participants/interviewees. These are as follows: i. Automated e-
sourcing (seamlessly connected automated replenishment in line with real-time KPI monitoring with predictive
disruption analytics; ii. Digital factory design - digital 3D modelling system for factory layout design, process and
material for simulation; iii. Real time factory scheduling. Advanced factory execution systems with sensor enabled
smart devices, real time data, KPI monitoring, predictive maintenance; iv. Flexible factory automation - advanced
manufacturing plant/machine reconfiguration; scale flexibility, varied levels of human-robot collaboration; v.
Digital production process - application of digital production processes (i.e. additive manufacturing, continuous
processing) with advanced process analytics; vi. E-commerce fulfilment - web based order management
(configuration, pricing etc) and inventory deployment to multiple points of sale, covering last mile and direct
delivery (all tiers through to end users); vii. Extended supply chain (near) real time monitoring - extended end to
end supply chain visualisation “watch towers” for near real-time monitoring and decision making; viii. digital
product quality - for connecting “traceability islands back from customers to suppliers (not cause analytics); vix.
Digital supply network design - design tools to architect supply network configuration - optimization and
visualisation of site location, capacity, inventory etc; and x. Product lifecycle management for next generation
systems that provide accurate, up-to-date product information accessible throughout the value chain and product
lifecycle.
Each topic scenario was considered by the practitioners who were required to fill out a template with their
responses to five different categories of each scenario. After a description of the scenarios, the thirty delegates

1
were split into six groups of five and then given 30 minutes to discuss as a team and respond individually to each
of the following categories: i. key opportunities; ii. key threats and uncertainties; iii. enabling technologies; iv.
enabling skills, know-how and v. key activities to realising the scenario. A facilitator was on hand to assist with
explaining and clarifying the activities for completing each scenario.

3. Results and findings

Table 1 summarises the key findings of the identification exercise to discover the key opportunities, threats and
enabling skills and technologies and key actions to realise the scenario.

Table 1. Digitalisation opportunities, actions, threats and enabling skills and technologies.
Opportunities • Autonomous predictive • Transparency • Help to understand
maintenance • Map Mobility complexity
• Integrate web-based • Ahead of the • Real-time location
ordering with consumer competition • Scenario planning for
preferences • Reduce end-to- supply chain disruptions
• How to manage capacity end lead time • Predictive analytics
profiles? • Joining systems (storage, inventory)
• Peak pricing models • Security – reduces • Fail prediction and
• No more showrooms/ new cost, requires maintenance
retail modes training • Managing capacity (how
• New • Customer much? How to leverage?)
ownership/subscription satisfaction, • Improved sustainability
models custom detail • Self-organised hyper
• Shared capacity • Real-time efficiency
• How to differentiate in a information • Flexibility in customisation
commoditized world quality • Improved forecasting
• Optimum part production • Control • Dynamic response to
– through life-cycle • Knowing which change (highly focused
• Improved flexibility in sites are affected systems)
face of disruptions and impacts of • Reduced inventory levels
disruptions • Sunk costs leverage
resource efficiency
Threats • Shared capacity • Intrusive • Not having right
• Privacy government competences/capabilities
• System level capacity • Cost • Tariffs
planning • Tier level coding • Brexit/Trade challenges
• Job shifts • Security • Political uncertainty in
• Import/export regimes • Increased other countries
• Software trust uncertainty in • Environmental
• More engineers with high regulatory conditions/disruptions
kill sets required environment
• Tracking • Demand
• Climate change uncertainty
Enabling • Cloud based services • Coms forecasting • Legacy knowledge
Technologies • Real-time simulation data • Simulation/modelling
• Algorithms • Blockchain • R&D, Supply
• Continuous scanning • Digital • AI/Blockchain – parts
• Real-time computing connection authentication
• Connectivity • Tracking ability • RFID
• Cloud computing • Equipment and • Connectivity between
fast different systems
communications
Enabling Skills • Convenience v. cost • Empowering • How to handle dark
• Communication skills – • Technology data/data lakes
negotiator knowledge • Change management skills
• System data analytics • Need for
engineering

2
 • Comprehensive cross- • Big numbers not • Technology for tracking,
functional business specialised transportation and
knowledge • Not automated traceability
• Curious/Risk Taker/Self
Assured
Key actions to • Software competitions • Better quality • VDVP
realise the • Small and independent- supplier • Simulation
scenario led networks • Government • e2e visibility
• Technology push demand monitoring • Knowing what tech is
may come later • Control data useful
• Car sensors connected to sharing • Setting common and
infrastructure sensors • Auditing minimum standards on
• Policy and regulatory • Data protection digital quality
challenges- how will • Institutions and • IBM's
connected cars. regulations • Embedding into University
• Models primarily • Legal education
designed for urban areas • Compliance • Need to deeply integrate
not so good in rural areas. • Collaboration e2e players
• Identify areas of capacity • Skills • Automated vehicles
need • Policy push to • Government regulations
• System-level capacity connect different • People preference
planning stakeholders • Transport times
• SR data trusts • Mapping supply • Delays, parameters or
chain against product and vehicle
products movement

Not shown in the table, but the most popular scenario in terms of depth and breadth of response was that of digital
supply network design. Whilst the least popular by far was the digital product lifecycle management.

3.1 Text analysis

Adopting the text analysis procedures of Roberts (2000) and Mayring (2004) we explored inter-codal patterns
across the ten scenarios and five categories. Pokorny et al. (2016) explains that: “… depicting codes applied to a
text as a network enables the representation and communication of the complex interrelations between codes,
beyond code co-occurrence matrices. Although both represent relations between pairs of codes, the network graph
provides the relation between a given pair of codes in context, and in relation to all the other codes in the network
(i.e., in the workshop)” (p. 1). The networks are also created using predefined criteria (i.e., chronology) to indicate
the relations between codes, rather than on the researcher’s interpretation of code relations within a given type of
coding (Attride-Stirling, 2001; Carley and Palmquist, 1992). This means that measures derived from the relations
between codes are data-driven and not researcher-driven, allowing others the ability to reproduce the graphs from
the same code location data. This offers researchers a way to increase the transparency of the evidence used to
support their research conclusions.

3.2 Network graphs

The creation of network graphs also allows researchers an opportunity to view their data in a novel way that
provides insights and drives further analysis. In Figure 1 below, the network diagram presents the key co-occurring
relationships between the different nodes. We have shaded in red the highest density of node relationships. The
network statistics are summarised in Table 2. This enables thematic clusters to be identified and provided an
important basis for complex and practice driven research insights (i.e. phenomena-driven research and applied
problem identification, with respect to digitalisation (Gallien et al., 2015)). Enabling (skills and technologies) had
the most connections to the other terms (n = 68). The highest density of connections it had were linked to “quality”,
“after-market”, “control”, “government”, “testing”, “real-time”, “management” and “products”. These themes
were the most-strongest connected across the ten scenarios. This preliminary identification of emerging nodal
relationships provides an important precursor for more detailed critical analysis of relationship patterns as we
progress the research project to extended case study development (Pokorny et al., 2016).

3
23rd Cambridge International Manufacturing Symposium
University of Cambridge, 26 – 27 September 2019
Density of edges

Figure 1. Network of co-occurring relationships across the ten scenarios.

4
23rd Cambridge International Manufacturing Symposium
University of Cambridge, 26 – 27 September 2019

Table 2. Network statistics.

4. Conclusion
With 68 connections across the network, the theme of enabling (skills and technology) is central to the discussion
by our automotive executives on digitalisation and opportunities. As well an enabling digitalisation it was
associated with emerging opportunities across the ten scenarios (with respect to “quality”, “inventory”, “control”
and (preventing) “stock-piling”). Though there were also a lot of threats identified concerning digitalization with
respect to “risks”, “disruption”, “privacy”, “costs” and “intrusiveness”. By far the most interest and data was
provided by the executives for the “digital supply network design” scenario. The least popular was “product life-
cycle management”. Therefore, with respect to practice-based problem and topic identification to guide our future
research investigation, it would appear that the scenario topic of investigation which has the most interest from
practitioners, would be that of supply network design. Further, that we would be looking at problems, phenomna
and insight-based research exploring supply network design realisation from the perspectives of enabling (skills,
technology) and its close assocation with issues of quality, technology, products, regulations, analytics, technical,
management and software. There would be a need to balance this with an analysis of the threats, risks and
disruptions of automotive organizations implementing digital supply network design.

Acknowledgments
We would like to acknowledge the work of Dr Jagjit Singh Srai for his invaluable support with the promotion and
scenario facilitation.

References
Attride-Stirling, J., 2001. Thematic networks: An analytic tool for qualitative research. Qualitative Research, Vol.
1, pp. 385–405.
Carley, K.M., Palmquist, M., 1992. Extracting, representing, and analyzing mental models. Social Forces, Vol.
70, pp. 601–636.

5
Gallien, J., Graves, S.C., Scheller-Wolf, A., 2015. OM Forum—Practice-based research in operations
management: What it is, why do it, related challenges, and how to overcome them. Manufacturing and Service
Operations Management, Vol. 18, No. 1, pp. 5–14.
Holmstrom, J., Partanen, J., 2014. Digital manufacturing-driven transformations of service supply chains for
complex products. Supply Chain Management - An International Journal, Vol. 19, No. 4, pp. 421–430.
Mayring, P., 2004. Qualitative content analysis. A Companion to Qualitative Research, Vol. 1, pp. 159–176.
Pokorny, J.J., Norman, A., Zanesco, A.P., Bauer-Wu, S., Sahdra, B.K., Saron, C.D., 2016. Network analysis for
the visualization and analysis of qualitative data. Psychological Methods, Vol. 23, No. 1, pp 169–183.
Roberts, C.W., 2000. A conceptual framework for quantitative text analysis. Quality and Quantity, Vol. 34, No.
3, pp. 259–274.
Schepis, D., 2011. Social network analysis from a qualitative perspective. Conference: Australian and New
Zealand Marketing Academy Conference, Perth. November 2011.
Seyedghorban, Z., Tahernejad, H., Meriton, R, Graham G., 2019. The architecture of supply chain digitalisation
research. Production Planning and Control, In Press.