Big Data Case Study Rp

Information & Management 57 (2020) 103121
Contents lists available at ScienceDirect
Information & Management

journal homepage: www.elsevier.com/locate/im
Actualizing big data analytics affordances: A revelatory case study T

a,⁎ a a b
Christian Dremel , Matthias M. Herterich , Jochen Wulf , Jan vom Brocke
a
Institute of Information Management, University of St. Gallen, Switzerland
b
Institute of Information Systems, University of Liechtenstein, Switzerland
ARTICLE INFO ABSTRACT
Keywords: Drawing on a revelatory case study, we identify four big data analytics (BDA) actualization mechanisms: (1)
Big data analytics enhancing, (2) constructing, (3) coordinating, and (4) integrating, which manifest in actions on three socio-
Affordance theory technical system levels, i.e., the structure, actor, and technology levels. We investigate the actualization of four
Socio-technical approach BDA affordances at an automotive manufacturing company, i.e., establishing customer-centric marketing, pro-
Organizational transformation
visioning vehicle-data-driven services, data-driven vehicle developing, and optimizing production processes.
Organizational benefits
Affordance actualization
This study introduces a theoretical perspective to BDA research that explains how organizational actions con-
tribute to actualizing BDA affordances. We further provide practical implications that can help guide practi-
tioners in BDA adoption.
1. Introduction learning in the realization of BDA value [16,17], it is largely unknown

which organizational actions contribute to realizing BDA value. Orga-
In the digital age, data are leveraged increasingly to derive insights nizational learning characterizes an organization’s concerted efforts to
and make decisions. As a result, the ability to collect, store, transform, enhance existing capabilities and develop novel ones [18,19]. Initial
and build causal or predictive models from data (data analytics) be- scholarly evidence suggests learning broadly applies to an organiza-
comes a competitive factor in many contexts, as it drives how value is tion’s socio-technical system (STS) on the actor level (e.g., employee-
created [1–4]. While, in the past, data were available mainly in small level competencies) [20], the structure level (e.g., organizational
quantities and bound by experimental contexts, technological ad- structures) [21], and the technology level (e.g., technological infra-
vancements allow the generation of digital traces in high volumes, structure) [22]. Further, initial evidence suggests that realizing BDA
velocities, and varieties. Against this background, the capability in value requires the alignment of a firm’s BDA-related socio-technical
terms of managing and making sense of big data has grown rapidly in entities (BDA alignment), including the alignment of strategic plans,
importance for both society and the economy over the last several years experts, and tools [23–25]; an integrating theory which explains the
[5–8]. Beyond enabling data-driven decisions in marketing and cus- constituents and outcomes of BDA alignment, however, is still missing.
tomer research, big data analytics (BDA) has transformed increasingly In summary, in spite of a growing body of practitioner reports (e.g.,
the value propositions of product and service businesses by augmenting [9,26,27]) and academic literature (e.g., [28,29]) on selected socio-
the functionality of physical products [9–11]) and affording in- technical aspects of BDA, scholars struggle in theorizing the value
dividualized services [12]. realization of BDA [28,30–32]. Thus, the existing body of knowledge on
The “value” of BDA, however, is highly dependent on the socio- BDA is still in a nascent stage and lacking empirical evidence and
technical context and rooted in the strategic goals of an organization theoretical foundations with which to illuminate how an organizations’
[13,14]. We consider BDA to be a socio-technical phenomenon which actions lead to the realization of BDA value [6].
rests on the interplay of: (1) technology “to gather, analyze, link, and In Information Systems (IS) research, the concept of affordances has
compare large data sets”, (2) analysis “to identify patterns in order to gained momentum in terms of exploring the value of digital technology
make economic, social, technical, and legal claims from large datasets”, [33–36]. Recent works have introduced the concept of affordance ac-
and (3) a data-driven culture, which represents “the belief that large tualization as a means by which to explore the process of realizing the
data sets offer a higher form of intelligence and knowledge that can value of digital technology and analyze affordances on an organiza-
generate insights that were previously impossible” ([15], p. 663). tional level [35,36]. Accordingly, to address the lack of a theory on the
While extant research indicates the importance of organizational realization of BDA value, we raise the following research question:
⁎
Corresponding author.
E-mail address: [email protected] (C. Dremel).
https://doi.org/10.1016/j.im.2018.10.007
Received 15 October 2017; Received in revised form 25 October 2018; Accepted 27 October 2018
Available online 14 November 2018
0378-7206/ © 2018 Elsevier B.V. All rights reserved.
C. Dremel, et al. Information & Management 57 (2020) 103121
RQ. “How do actions which modify an organization’s socio-technical action which are not required to be performed or actualized [35,41].
system contribute to actualizing BDA affordances?” Depending on the relationship between the actor and environment, an
affordance is perceived “relative to the posture and behavior of the an-
We present the results of a revelatory, in-depth case study on the
imal [or the actor] being considered” ([39], pp. 127–128). In the sub-
actualization of BDA affordances at a leading global automotive man-
sequent development of affordance theory, the relational character of
ufacturing company. The revelatory case study involved the collection
“affordances” was further underlined (cf. [40,42]). Affordances stem,
and analysis of data from 31 key informant interviews triangulated with
accordingly, from the properties of the environment and actor as well as
rich internal case material. Our qualitative data analysis demonstrates
from their relationship [42] and can be seen as an emergent property of
the organizational actions required for the process of the actualization
an actor-environment system [41]. Affordances may change depending
of four affordances: (1) customer-centric marketing, (2) provisioning
on the context (i.e., constraints of and the relationship between the actor
vehicle-data-driven services, (3) data-driven vehicle developing, and
and environment), though the material properties do not [33,39,42]. IS
(4) optimizing production processes. Our results indicate that the ac-
scholars use affordance theory to elaborate on and investigate the con-
tualization of BDA affordances demands four mechanisms of orche-
sequences of IT artifact use in organizations ([43–45]) and the related
strated organizational actions (i.e., enhancing, coordinating, constructing,
organizational changes [46–48].
and integrating). Our discovery of these mechanisms at the structure, the
Affordances may not only be analyzed on an individual level, but
actor, and the technology levels of a company’s STS extends, in parti-
also on an organizational level. Strong et al. [35], for example, study
cular, prior research on BDA alignment by distinguishing organiza-
the implementation of an electronic health records system in a multi-
tional actions of intellectual, social, and operational BDA alignment.
site medical group and investigate the affordances which relate to
Further, this discovery extends the current knowledge on how a com-
group-level goals. Accordingly, Strong et al. [35] p. 74 refer to an af-
pany develops BDA capability by eliciting two modes of organizational
fordance as “an organizational affordance to the extent that the po-
learning, incremental and radical learning, on the individual level as
tential actions enabled are associated with achieving organizational-
well as the organizational level. These findings motivate and inform
level immediate concrete outcomes in support of organizational level
further research which explores different modes of BDA alignment and
goals.” In our research, we regard affordances as organizational-level
organizational learning in the realization of BDA value.
action possibilities stemming from the material properties of IS artifacts
The remainder of the article is structured as follows. In the back-
and the socio-technical characteristics of an organization as well as
ground section, we introduce key concepts and relate them to the extant
from their recursive interrelationships [35,44,48].
body of knowledge. In Section 3, we present our research approach,
From an affordance-theoretic perspective, BDA technologies and
comprising the illustration of the case context, data collection, and data
tools provide the means to process, store, and collect a vast volume of
analysis. Section 4 illustrates our results by presenting our identified
data characterized by variety, variability, and velocity [49,114]. These
BDA affordances briefly and detailing their actualization. In Section 5,
technologies enable a flexible handling of incomplete, inconsistent,
we integrate our research findings theoretically against the background
ambiguous, heterogeneous, and agnostic data [49]. On this technolo-
of our uncovered actualization mechanisms. Section 6 discusses the
gical basis, affordances (i.e., BDA value potential) on the task level
implications for theory, implications for practice, and limitations. Sec-
emerge, such as the improvement in decision-making processes, pro-
tion 7 concludes our research study.
ducts, and services [50–52]. The actualization of BDA affordances, in
turn, results in BDA value.
2. Background Prior literature on BDA value distinguishes the following value ca-
tegories ([32], p. 239): “creating transparency” (e.g. [53]), “enabling
In the following section, we discuss the theoretical framework used experimentation to discover needs, expose variability, and improve
to analyze the realization of BDA value. In this framework, we integrate performance” (e.g. [54,55]), “segmenting populations to customize
affordance theory with STS theory. Affordance theory allows us to actions” (e.g. [26,56]), “replacing/supporting human decision making
conceptualize the action potential which results from big data tech- with automated algorithms” (e.g., [51,52,57]), and “innovating new
nologies on an organizational level. STS theory, in turn, enables us to business models, products, and services” (e.g. [7,49,58],). Only one
frame value realization as the process of a recursive shaping of social prior study, to the best of our knowledge, uses an affordance-theoretic
and technical entities [33,37,38]. In the next subsection, we introduce perspective in order to study BDA value [12]. Lehrer et al.’s analysis
our affordance theoretical lens and discuss prior literature on BDA focuses on how the technological features of BDA stimulate business to
value. In the subsequent subsection, we explain our STS-theoretic per- consumer (B2C) service innovation and does not focus on the processes
spective on value realization and describe prior literature related to the through which affordances are actualized.
implementation of BDA. Mapping the field of BDA value through extensive, systematic lit-
erature reviews, Mikalef, et al. [30] and Wamba et al. [32] highlight the
2.1. An affordance-theoretic perspective on big data analytics value lack of empirical research and call for further research on the phe-
nomenon of interest in order “to understand the mechanisms and pro-
To study the perceived value potential of BDA, we use affordance cesses through which big data can add business value to companies”
theory as our theoretic lens. In IS research, affordance theory is em- ([30], p. 1). Moreover, the current body of knowledge is dominated by
ployed broadly to conceptualize the action possibilities which a tech- either conceptual or technologically focused studies [14,32].
nical artifact – in our case, big data technologies and tools – affords an
actor through its materiality (i.e., constituent materials of a technical 2.2. A socio-technical perspective on big data analytics value realization
artifact) [33,34].
Gibson [39] first coined the term “affordance” as part of his research While the concept of affordance is well-established in IS research at
in perceptual psychology to explain the behavior of actors (e.g., animals an individual level only, the study of technology-induced organiza-
or people) in regard to their environment. According to Gibson [39], an tional change requires theoretical extensions to the affordance concept
actor interacts with a physical object due to its value in a specific-use which allow a description of the interdependencies between aggregated
context (e.g., a bed to lay down upon) rather than its qualities (e.g., a bed technologies and larger social collectives [59]. For this reason, several
is horizontal, constructed out of wood, and knee-high). “What we per- researchers articulate the need for further theory which supports the
ceive when we look at objects are their affordances, not their qualities” study of affordance actualization at the organizational level (e.g.,
([39], p. 134). Hence, any affordance needs to be perceived before it can [36,60]). For example, Volkoff and Strong [36], p.242 acknowledge the
be actualized by an actor [39,40]. Affordances are prerequisites to an need to build “mid-range theories of IT implementation and IT-enabled
2
Fig. 1. Affordance Actualization Model (adapted from [33,66,67]).
organizational change that focus on specific technologies and specific system changes, the social system adjusts “to damp out the impact of
organizational goals” and call for research that “seeks to identify me- the innovation” ([68], p. 25). As a consequence, an STS is subjected
chanisms for organizational change.” The actualization of affordances continuously to incremental and punctuated changes to one or more of
in organizations is influenced by several social and technical factors, its socio-technical entities [64]. Framing a STS as a “complex web of
such as employee expertise, organizational processes and procedures, mutual causality” ([69], p. 13) allows us to conceptualize the process of
controls, boundary-spanning approaches, and social capabilities [48]. actualization as modifications to socio-technical entities which result in
Thus, studying the actualization of affordances at the organizational realized action potentials.
level requires a broad recognition of the socio-technical context of an Fig. 1 summarizes our theoretical model. Affordances are action
organization which may “stimulate […] actualization in a variety of potentials at an STS’s task level. Affordance actualization consists of a
ways” ([60], p. 87). recursive shaping of socio-technical entities, which is reflected in the
Against this background, Strong et al. [35] introduce a model for the organizational activities related to modifying entities at the structure,
actualization of organizational affordances which includes organiza- actor, and technology levels. Structure-, actor-, and technology-level
tional goals, the organizational context, and goal-directed actions. An actions then lead to the actualized affordance at the task level.
extension of this actualization model considers actions, intended and Interpreting BDA as a socio-technical phenomenon and studying the
unintended outcomes, as well as adjusted actions [61]. While these actualization of BDA affordances through an STS lens allows us to detail
works introduce key concepts with which to study actualization, they the socio-technical antecedents of value realization. The current body
consider actualization activities as a unidimensional construct; they of knowledge on BDA, according to Mikalef et al. [70], focuses largely
lack a granular perspective in terms of which adjustments of the or- on the technical aspects and characteristics of big data and pays little
ganization (which we conceptualize as a STS) influence affordance attention to “the organizational changes they entail and how they
realization and thus form part of the actualization process. should be leveraged strategically” ([30], p. 1). Though several authors
In our theoretical model, we complement the affordance perspective (e.g., [16,70,30]) adopt a dynamic capability perspective to provide
with STS theory to distinguish actualization activities analytically by first conceptual works toward the development of a “BDA capability,”
capturing which modifications of the STS affect the realization of af- there is no theory on how organizations realize value from BDA [6].
fordances. STS theory is a well-established perspective in IS research More precisely, how perceived affordances are actualized through or-
which, still today, is employed frequently to study technology-induced ganizational actions is unknown. Recent studies on BDA suggest that
organizational change (e.g., [62–65]). STS theory distinguishes the actualizing BDA affordances involves actions which affect an organi-
social system and its socio-technical entities actors and structures as well zation’s socio-technical entities at the structure, actor, and technology
as the technical systems and its socio-technical entities tasks and tech- levels.
nologies [66,67]. Actors are, among other things, characterized by Since BDA poses novel structural challenges, such as the allocation
capabilities and a shared culture, structures are characterized by project of analytical knowledge and capabilities and new functional inter-
organizations and institutional arrangements, technologies by tools and relationships between business units and the IT unit, an organization
technological platforms, and tasks by the processes which are required needs to perform structure-level actions; examples include the estab-
to fulfill work or the delivery of services [64,66,67]. lishment of new organizational structures and work methods in order to
STS are inherently dynamic; they evolve through a recursive support and grow cross-departmental collaboration [49,71–73].
shaping of social constructs and technical infrastructure [37,38]. The Against this backdrop, the implementation of a dedicated BDA unit is
organizational adoption of novel technologies entails changes in an proposed in order to implement the organizational frame and acquire
organization’s social system (e.g., new roles, modified organizational and develop the capabilities and knowledge with which to actualize
structures, and changes in cultural norms). Such modifications, in turn, BDA affordances ([13,50,73,74]) especially when trying to monetize
enable the realization of technology affordances [33]. As a technical big data [52]. In addition, several authors find that the use of agile
3
methods for data science projects, such as Scrum, Crisp, and Kanban, consists of STS modifications and how this process links to realizing
contributes to the success of BDA projects (e.g., [75,76]). business value are largely unexplored [6,30,49,50,84]. The realization
Actor-level actions address the development of novel employee cap- of BDA value is by no means assured, and the complex nature of or-
abilities (e.g., business acumen, and analytical and technical cap- ganizational designs challenges the effective use of BDA [30,85,86].
abilities) and the creation of a commonly shared mind-set concerning Therefore, there are calls for further research on: (1) “enterprise-
the role of data as a critical success factor [15,77]. The culture of the wide BDA […] that maximizes the potential for competitive advantage
company, which is characterized by the beliefs and practices of the in different types of industries and for different organizational cultures
individual employees, should reflect that the use and analysis of data and governance archetypes” ([87], pp. XX–XXI) and (2) on “the pro-
are critical for improving firm performance [15,77]. Prior research cesses and structures necessary to orchestrate […] resources into a firm-
shows the significant effects of a data-driven culture on competitive wide capability”([30], p. 23). In particular, empirical studies on BDA
advantage [77] as well as on a company’s ability to innovate products are scarce. Therefore, Günther et al. [6], p. 192 calls for empirical
[58]. The development of advanced employee skills in data science studies to develop “theories on big data value realization.”
methodologies is an additional critical success factor, particularly for Thus, with our research, we pursue the objectives: (1) to theorize on
companies which digitize their business models [4,78]. the socio-technical modifications which constitute the actualization of
Technology-level actions address the implementation of BDA tech- BDA affordances and (2) to explore the interrelationships between ac-
nology and tools, such as Hadoop clusters or in-memory databases [22]. tualization activities and BDA affordances.
Organizations are confronted with an unprecedented volume of data,
and key benefits arise from large-scale integration and processing 3. Research design and method
[79,80]. An important source of value lies in the exploitation of pre-
viously unused data sources, such as text or log data, which has, in By drawing on the aforementioned theoretical foundations, this
some cases, been considered as noise or garbage in the past [80,81]. As work aims to investigate how affordances related to the BDA phe-
a consequence, the processing of big data necessitates the use of new nomenon are actualized in the context of an automotive manufacturing
data base structures, such as HBase, to leverage new data file systems, company. We chose an inductive qualitative research design due to: (1)
such as the Hadoop Distributed File System and Google File System, as the novelty of the topic and (2) the lack of prior research on BDA and
well as the dynamic allocation of computational resources through resulting affordances [88–91]. Specifically, we conducted an in-depth
solutions such as YARN [22]. To this end, new computational ap- case study of one of the largest automotive corporations worldwide
proaches to the analysis of data (e.g., machine learning and neural (AutomotiveGroup). We chose the automotive industry because harnes-
networks) as well as the process of data acquisition (e.g., connection of sing BDA is pivotal to this industry [10,92,93]. Acknowledging the
new data sources through data loggers) are a prerequisite [81]. In paucity of in-depth empirical work on the actualization of BDA affor-
summary, technology-level actions enable the realization of BDA value dances [6], we elaborate on the actualization of BDA affordances at
through better insights from different, diverse, and new data sources AutomotiveGroup in a revelatory single case study [91]. In line with Yin
(e.g., social media, wearables, and radio-frequency identification [91], we opted for a revelatory case study design for two reasons. First,
(RFID)) and computational approaches which uncover patterns, corre- the phenomenon of interest has been inaccessible to previous in-
lations, or other previously unknown information pertaining to this vestigators due to its novelty and due to investigators’ limited accessi-
data [58]. bility to large organizations. Second, the insights can potentially help to
Summarizing the above aspects, socio-technical entities, among better explain the research phenomenon of a company’s actualization of
others, are related to data-oriented culture, practices, technologies, and BDA affordances.
analytical processes (i.e., accessing, examining, aggregating, and ana- Table 1 provides an overview of our overall research approach. In
lyzing evidence) which need to be adapted in order to actualize BDA what follows, we go into details regarding the case context, data col-
affordances [82,83]. The design of the actualization process which lection, and data analysis.
Table 1
Research Approach with Illustrations (structure based on [61]).
Methodological Consideration Illustration
Data Collection
Selection of Case The revelatory case was selected based on the ability to study the The case organization in the automotive industry represents one of
phenomenon of interest in a real-world-setting with high strategic the largest players. Harnessing BDA and driving the corresponding
relevance. change is reflected in the organization’s vision.
Choice of interviewees Snowball sampling [94] was used to obtain an adequate set of We stopped adding additional interviewees when saturation of a
interviewees. specific perspective was reached.
Conduct of the interviews A semi-structured interview guideline was developed based on the Interviews were conducted in two rounds. An explorative, open first
recommendations of Schultze and Avital [95]. round allowed us to gather rich information on the overall context
and goals of the organization. A more focused second round, targeting
mainly affordance actualization, was conducted to obtain additional
insights on the actualizations of specific affordances.
Inclusion of multiple data Our findings are rooted in a rich set of data sources, taking into account Mainly internal documents such as strategic working papers,
sources contextual information and allowing us to capture the complexity of the presentations, and memos were used for triangulation purposes.
phenomenon of interest from various perspectives. Besides, publicly available documents were considered.
Validity and/or reliability To limit bias, we considered various perspectives on the phenomenon of In the interviews, additional questions were asked when a good
considerations interest by recruiting a broad set of interviewees. The integrity of our understanding of the points being made by the interviewees was
data was ensured by working exclusively with recorded and transcribed needed.
interviews.
Data Analysis
Data triangulation 182 additional documents were reviewed to triangulate findings When condensing insights from the data, we made sure that we
derived from interviews. looked for converging evidence between multiple sources.
Validity and/or reliability Throughout the process of analyzing the data, we were aware that We were careful with regard to the statements of interviewees. For
considerations interview statements could be subject to the personal biases of instance, interviewees in business functions tended to underestimate
interviewees related to their roles and responsibilities. the required technological implementation efforts.
4
3.1. Case context rounds. The objective of the first interview round was to explore BDA
affordances broadly and gain an understanding of the affordances and
BDA can be regarded as the key underlying technology which fuels their socio-technical contexts. Thus, the first round of interviews had a
trends in the automotive industry, such as connectivity, autonomous rather explorative character and included representatives of Auto-
driving, and car sharing [10,92,93]. In order to be innovative in terms motiveGroup’s seven suborganizations. Interview partners were selected
of products and business models, the automotive industry has to master in such a manner as to gather contextual information and cover a broad
the most dramatic shift in years and draw on BDA to face the challenges variety of roles and responsibilities with regards to BDA. The second
of digitization [3,96]. Accordingly, organizations in the automotive interview round, which focused on deriving insights into affordance
sector rely heavily on BDA to achieve customer and service orientation actualization, zeroed in on AutomotiveGroup’s suborganization Pre-
[93,97]. In this regard, BDA provides the foundation for developing miumCar for two main reasons. First, in the initial interview round, we
new vehicle-data-driven services and business models [9,93,96,98]. As learned that PremiumCar acts as the innovation leader in the Auto-
a consequence, the effective actualization of BDA affordances has a motiveGroup. Even though other suborganizations perceived similar
strategic impact on the competitive edge of automotive organizations action potentials with regard to BDA, they did not start the actualiza-
[10,92]. One interviewee described how BDA relates to the growing tion of the perceived affordances. PremiumCar was the only sub-
importance of digitization in the automotive industry as follows: organization which tackled broadly the actions required to realize the
value related to BDA. Second, as the actualization activities for different
“All topics which we take care of through analytics are enablers of better
affordances are complex and interrelated, our focus on PremiumCar
functions for the customer or result in product optimization and, hope-
allowed us to study actualization activities with sufficient depth. Prior
fully, a big cost reduction to aid our product development process […]
to conducting the interviews, an interview guideline was developed
Moreover, look to other industries that successfully changed their busi-
following the work of Schultze and Avital [95] (see Appendix A). Based
ness models and how they work today: Pretty much every successful
on the respective knowledge of the interviewees and the interview
business model which somehow tries to be [a] digital [business model]
context, additional questions were asked. In particular, we conducted
uses data as its energy source. That is what we are trying to achieve.”
31 interviews within AutomotiveGroup. Table A1 in Appendix A pro-
Interviewee A2
vides an overview on the two interview rounds which highlights the
In consequence, the automotive industry being an industry which is interviewees’ roles. Interviews were conducted by two senior re-
affected heavily by data analytics and digital technology [92,97,99], it searchers and lasted between 25 and 108 min. They were transcribed
serves as an adequate context in which to study the actualization of based on audio recordings, resulting in 410 pages of text. In addition to
BDA affordances (i.e., value realization of BDA). the interviews, internal documents and publicly available data (182
In the course of a research initiative which was kicked off in 2015 documents in total) were used to corroborate findings.
with AutomotiveGroup we obtained in-depth insights into the actuali-
zation of BDA affordances. AutomotiveGroup represents one of the lar- 3.3. Data analysis
gest multinational automotive manufacturing companies in terms of
global sales and market share and comprises seven suborganizations To analyze the gathered case data, we follow the well-established
which target distinct customer segments ranging from premium sports recommendations of Corbin and Strauss [100]. Specifically, we pursued
cars to affordable compact cars: LuxCar, PremiumCar, MediumCar, a step-wise coding which consisted of open, axial, and selective coding
MobilityServiceCo, TrucksCo, CarCo, and CommodityCar. LuxCar is a car in order to elaborate on the actualization of affordances and their re-
manufacturer specializing in producing high-performance sports cars, quired organizational actions and actualized outcomes.
SUVs, and sedans for the luxury segment. LuxCar has approximately In the open-coding stage, codes emerged through case write-ups and
25,000 employees world-wide, making an annual net income of ap- summaries which were used to condense the transcripts and obtain an
proximately USD 2.3 billion. PremiumCar, targeting customers in the initial overview of all case data [91]. Codes were initially developed
premium segment, is known for its technological advancement and inductively due to the novelty of the topic. In the axial coding stage, we
produces approximately 2 million cars a year while making an annual condensed the data based on the dimensions of the affordance and STS
net income of approximately USD 4.6 billion and employing around theories (see Fig. 1). Along the dimensions of technology and task, we
90,000 people world-wide. Moreover, PremiumCar serves as the digital identified relevant aspects of the technical system. Our analysis re-
innovation leader within the AutomotiveGroup. MediumCar, which tar- garding the social system was structured along the actor characteristics
gets the mass market, earns around USD 2.0 billion per year while and structural aspects of the organization. We aggregated emerging
producing approximately 6 million cars a year with approximately codes to identify reoccurring themes. Selective coding allowed us to
600,000 employees world-wide. CommodityCar employs around 14,000 finally sharpen our focus on the relations between the identified con-
employees and has an annual output of approximately 400,000 cars per cepts. Specifically, we identified organizational actions which related to
year for a USD 0.1 billion net income. CarCo produces approximately 1 affordance-specific socio-technical modifications on the structure,
million cars a year with its 30,000 employees world-wide while accu- actor, and technology levels. During coding, we continuously corro-
mulating a profit of USD 0.12 billion per year. TrucksCo manufactures borated the detailed insights derived from analyzing the interviews by
high-class trucks and buses for commercial purposes with the help of its constantly comparing and triangulating these insights with the results
50,000 employees, achieving a profit of USD 0.6 billion per year. Lastly, obtained from analyzing the internal (e.g., internal presentations) and
MobilityService acts as one of the biggest mobility and financial service external case material (e.g., public statements).
providers in Europe with its 11,000 employees to make a USD 1.8 In order to analyze the data and manage the collected data in a
million profit every year. systematic way, we used ATLAS.ti as our computer-assisted qualitative
data analysis software. Over the three coding stages, a total of 396
3.2. Data collection codes were generated.
In order to obtain in-depth qualitative data, exploratory interviews 4. Results

served as primary source and were conducted with managers at
AutomotiveGroup. Specifically, as the focus of our investigation lies in In the following section, we present briefly the four BDA affordances
the actualization of BDA affordances, we selected managers and senior which we identified in the initial interview round. Thereafter, we detail
executives responsible for BDA within AutomotiveGroup for the inter- the actions required for the actualization of these four affordances at
views (see Table A1 in Appendix A). Interviews were conducted in two PremiumCar.
5
4.1. Affordances of big data analytics

“We [try] to define [the functions of] our brake based on large-scale
customer usage data.” Interviewee G1
4.1.1. Affordance 1: establishing customer-centric marketing
This affordance is characterized by the organizational goal of im- The R&D department at PremiumCar leverages BDA to ensure the
proving marketing effectiveness through personalized customer inter- safety of new car models based on on-road test data. In doing so,
actions, which is enabled by analyzing traces of digital and nondigital PremiumCar identifies anomalies such as misconfigurations in the
customer interactions. CarCo, for example, uses A/B testing to improve electric braking system.
the design of the online car configurator and predict customer pre-
ferences early on in the configuration process. CommodityCar tracks its “In research and development, we analyze an immense amount of data to
customers’ behaviors across different online platforms and, based on improve our cars and safeguard their functioning.” Interviewee B1
this data, calculates customer profiles which support the dealer’s pre-
Additionally, PremiumCar increases the coverage of tested car
paration of personalized sales offers:
functions constantly. This is particularly important for testing (semi-)
“We join data either from touch points like [our] car configurator […] to automatic driving features, which require a great number of test sce-
generate customer journeys, and […] we combine this with social media narios.
data […] to optimize the targeting of the customers.” Interviewee E1
PremiumCar tracks the behavior of its customers on its digital touch
4.1.4. Affordance 4: optimizing production processes
points in real time, particularly on its different homepages and the car
This affordance aims at leveraging production and sensor data for
configurator. This enables the assessment and measurement of sales
the optimization of production processes. In particular, the focus lies on
funnel efficiency in real time. Moreover, PremiumCar uses socio-de-
leveraging a smart factory approach which involves creating a highly
mographic and behavioral customer data (e.g., purchase history) for
accurate virtual representation (digital twin) of the entire production
lead management, which has already resulted in an increase in the lead
facilities and using real-time data to minimize production costs.
conversion rate by 10 percent in Spain and France.
MediumCar, for instance, undertakes several steps to avoid production
down times, such as a near real-time identification of failures in
4.1.2. Affordance 2: provisioning vehicle-data-driven services screwing processes and the predictive maintenance of production fa-
This affordance is characterized by the organizational goal of im- cilities:
proving customer offerings through vehicle-data-driven services which
“Big data analytics helped [MediumCar] to minimize blackouts in pro-
complement the vehicle manufacturers’ core products. TrucksCo, for
duction. That is to predict when a machine might break down in ad-
instance, produces commercial trucks with which their customers
vance.” Interviewee L1
transport goods. Using predictive maintenance, TrucksCo tries to reduce
truck breakdowns and outage times: CarCo tries to optimize its rate of production through aligning the
production and sales processes:
“In particular, we focus on a predictive maintenance service […] we use
the onboard connectivity to transfer the data, and we provide remote and “Based on car configurations […] we can make predictions of sales
predictive maintenance services to the customers.” Interviewee C1 opportunities […] this enables sales and product development […] [to
analyze] which car components and configurations will actually be sold
MediumCar works on integrating car location data with real-time
more, so [they] can better plan supply and production.” Interviewee E1
traffic information, such as traffic light data. Based on this information,
MediumCar would like to enhance its navigation services. PremiumCar PremiumCar conducts automatic car disposition planning to improve
works on fleet management services which use car location information yard management and uses energy analytics to reduce its facilities’
to optimize route scheduling, support energy-efficient driving, and energy consumption. Through a proactive analysis of errors during the
identify security-related incidents, such as theft. production process, PremiumCar also tries to reduce waste.
“If you realize an error of the press shop at the final step of painting or
4.1.3. Affordance 3: data-driven vehicle developing the final montage, a lot of value creation already has taken place.”
This affordance targets analyzing the data acquired through on-road Interviewee M2
tests and vehicle usage data in order to improve vehicle safety and
functionality. TrucksCo, for example, tries to establish a feedback loop Table 2 provides an overview of the four affordances, associated
from the actual truck usage to the requirement definitions for future organizational goals, and examples from the interviewed suborganiza-
truck systems and functionality improvements: tions.
Table 2
Affordances, Organizational Goals, and Examples.
Affordance Organizational Goal Examples
Affordance 1: Establishing Improvement of marketing effectiveness • Improve design of car configurator through A/B testing
Customer-Centric Marketing through personalized customer interactions • Personalize sales offers through customer profiling based on click data
• Increase lead conversion rate through analysis of socio-demographic data and purchase history
Affordance 2: Provisioning Improvement of customer offerings • Reduce outage time of trucks through the use of predictive maintenance
Vehicle-Data-Driven Services • Improve navigation services with traffic light data and high-resolution traffic information
• Provide fleet management and security services based on car location information
Affordance 3: Data-Driven Improvement of car safety and functionality • Improve brake system functionality based on usage data
Vehicle Developing • Identify anomalies during on-road tests
• Increase the coverage of on-road tests
Affordance 4: Optimizing Decrease in production costs • Near real-time identification of failures in screwing processes
Production Processes • Predictive maintenance of production facilities
• Demand-adaptive production of cars
• Reducing production costs through improved yard, energy, and error management
6
4.2. Actualization of affordances self-actualization. Second, PremiumCar’s company headquarters are

located in a mid-sized city which is less attractive to young adults. In
In the following section, we describe essential organizational ac- order to introduce a new means of attracting data scientists,
tions and explain how they contribute to actualizing the four BDA af- PremiumCar founded an innovative start-up-like analytics subsidiary
fordances at PremiumCar. We provide an organizational chart for in a big city nearby.
PremiumCar and a complete list of all affordance-related organizational
“The data scientists in the [analytics subsidiary] have substantial in-
actions which we discovered in Fig. B1, Appendix B and Table C1,
dependencies because they can focus on and design the analytics methods
Appendix C.
and approaches without having to handle the day-to-day business. Being
located in a more influential city, the analytics subsidiary is far more
4.2.1. Actualization of affordance 1: establishing customer-centric
attractive to the rare talents we try to recruit for our analytics projects.”
marketing
Interviewee E2
4.2.1.1. Structure-level actions. Within the marketing unit, analytics
efforts had been scattered. The associated lack of a managed The subsidiary can design its own business processes (such as em-
development of analytical capabilities and the lack of knowledge ployee recruitment) without having to comply with the bureaucratic
exchange between employees in analytics projects had resulted in the processes of PremiumCar. The introduction of the analytics subsidiary
marketing unit’s inability to support BDA projects. In order to channel helped PremiumCar to attract data scientists and increase its capability
the establishment of analytical capabilities within the marketing unit of establishing BDA-driven customer intelligence.
and centralize marketing activities, the chief of marketing ordered the In the IT unit, analytical capabilities had been scattered because the
installation of a marketing analytics subunit. IT unit is organized into organizationally and physically separate sub-
units following the departmental structures of PremiumCar’s business
“We decided on the management level that we want to move towards
units. As a consequence, the IT unit had limited capability to design and
doing things independently in order to develop BDA knowledge and skill
operate big data infrastructures. To bring together all of the BDA ex-
in house. In particular, with the topic data, we will have a competitive
perts, PremiumCar’s IT unit established a big data competence center
edge and will face a problem if our external providers know more than
which cuts across the individual IT subunits.
we do, as they might sell the same [services] to our direct competitors.”
Interviewee I2 “The implemented competence network mainly targets employees in the
IT subunits. In our company, we faced initially a serious issue with ex-
Starting off as a small group, this subunit extended its headcount to
perts on analytics in the IT unit. The big data competence center was a
approximately 20 people on two management levels. The marketing
first step to bring together all existing and future experts to share their
analytics subunit is now responsible for marketing-related data analy-
expertise across the IT unit and, consequently, to be a partner for units
tics topics as well as a dedicated BDA strategy which enables customer-
like the [marketing analytics] subunit.” Interviewee A2
centric marketing. It has initiated several BDA projects, including sales
forecasting and the analytical design of car configuration packages In particular, it organizes synchronization meetings in which IT
based on online configuration data, as well as a data quality program. unit’s current technological solutions are discussed. Further, it operates
Further, the marketing analytics subunit took over the project lead for a a document space for knowledge exchange between the big data ex-
marketing analytics service initiative at the marketing unit. This in- perts. This led to a substantial increase in the analytical capabilities of
itiative involves consolidating the requirements of the different stake- the IT unit and the ability to act as a respected partner of the marketing
holders (i.e., subunits of the marketing unit, car dealers, and car im- analytics subunit in BDA projects.
porters) and managing service implementation. The installation of the
marketing analytics subunit led to a bundling of know-how and a 4.2.1.3. Technology-level actions. When the marketing analytics subunit
managed development of analytical capabilities within the marketing was founded, it was confronted with greatly disintegrated legacy
unit. business intelligence (BI) infrastructures. As a consequence, the
The marketing unit’s BDA projects are not carried out by the mar- marketing analytics subunit was unable to implement BDA scenarios
keting analytics subunit exclusively, but also involve the IT analytics which required a sophisticated analytics infrastructure. Consequently,
subunit and PremiumCar’s independent analytics subsidiary, which in order to store and analyze customer-related data centrally, a
bundles data science knowhow. However, not having the adequate centralized marketing BDA platform was set up:
governance instruments in place meant that PremiumCar had risked a
“We implemented, in collaboration with the IT unit, Teradata as the data
misalignment of the three organizational entities. As a consequence, a
warehouse with the SPSS modeler as an analytics solution and Tableau
joint steering committee was established which is responsible for BDA
as visualization software to unify the technological stack used in mar-
project portfolio management and prioritization:
keting.” Interviewee C2
“We meet up on a weekly basis in a steering committee to align all our
This BDA platform provided the capabilities required to conduct
efforts from [the analytics subsidiary], [the IT analytics subunit] and
advanced BDA projects based on customer data.
[the marketing analytics subunit]. With four managers, two for our unit,
Not only the infrastructure, but also the customer data had been
we have to align our strategic goals and discuss the main challenges we
scattered across the marketing unit, which had resulted in the inability
are facing, may it be a business issue or a technological one.” Interviewee
to aggregate rich customer information. As a consequence, the mar-
C2
keting analytics unit initiated the integration of customer data in the
This steering committee includes the head of and a product manager marketing BDA platform.
from the marketing analytics subunit, the head of and a product man-
“The biggest challenge was the fact that no one paid attention to what
ager from the IT analytics subunit, and the head of the analytics sub-
happens with the data and what has emerged over the last five, six, seven
sidiary. Thus, it provides a joint platform for coordinating BDA projects
years of operational systems. There existed very few individual examples
which support customer-centric marketing at PremiumCar.
with both a clean logic and a clean data model and an appropriate link
with other systems.” Interviewee A1
4.2.1.2. Actor-level actions. PremiumCar had considerable difficulties in
attracting new recruits with much-needed data science competencies Customer data from approximately 230 data sources, such as online
due to two reasons. First, the hierarchical structure of a large customer platforms and car dealer systems, were copied little by little
corporation is unattractive for young potential recruits searching for into the centralized storage. Integration further involved the matching
7
Table 3
Actions of Establishing a Customer-Centric Marketing.
Structure Level Actor Level Technology Level
- Constructing a marketing analytics subunit - Recruiting data scientists, big data architects, and - Implementing a Teradata- and Tableau-based
- Aligning marketing analytics subunit, the IT analytics visualization experts in a start-up-like analytics technological infrastructure
subunit, and the analytics subsidiary through a steering subsidiary - Integrating customer and dealer data from dealer
committee - Newly bringing together big data experts in the IT systems and customer portals
unit through a big data competence center - Delivering dynamic analytics services through a
Tableau solution instead of offline dashboards
of schemas and customer identifiers. This action supported the access to The first project addressed the equipping of vehicles with data
and full exploitation of different customer data sources. loggers; the second project targeted building an integration architecture
The analytics experts in the marketing unit had traditionally com- to enable the analysis of vehicle-sensor data across different models.
piled analysis results manually into static reports owing to the lack of a The marketing analytics subunit’s funding of these two projects was a
reporting and visualization platform. This static reporting and lack of key structure-level action because providing access to vehicle data re-
automation had resulted in heavy manual workloads and constant issues presented a crucial actualization step toward vehicle-data-driven ser-
with outdated customer data. In order to enable dynamic reporting and vices.
to increase the level of automation, the marketing unit introduced a
Tableau-based infrastructure which supports dynamic dashboards. Users 4.2.2.2. Actor-level actions. PremiumCar had used sequential processes
can now adjust data views dynamically and conduct analyses themselves. initially to develop vehicle-data-driven services, which had led to
insufficient coordination between product managers and data
“We built now a Tableau server through which we deliver our analytics
scientists. In order to improve the interworking of product managers
service […] Previously, data snapshots were used and sent as dashboards
from the R&D unit and the data scientists from the analytics subsidiary,
via mail, for instance, as PDFs or sometimes as offline Tableau dash-
PremiumCar introduced the scrum process. The iterative nature of
boards. Now, we integrate these dashboards step-by-step.” Interviewee
scrum helped to streamline the explorative design of vehicle-data-
C2
driven services and service implementation.
The introduction of the Tableau-based infrastructure decreased
“We have to develop a flexibility and agility in regard to our releases.
significantly the data scientists’ repetitive and manual labor and im-
Instead of two major releases, we need – as it is characteristic for agile
proved the supply of customer information to PremiumCar’s marketers.
approaches – high-frequency micro releases […] Scrum is one possibility
Table 3 summarizes the key actions used to establish a customer-
to achieve this and to get our product management and the developing
centric marketing approach at the structure, actor, and technology le-
team together.” Interviewee N2
vels.
All involved data scientists update each other and the respective
4.2.2. Actualization of affordance 2: provisioning vehicle-data-driven product manager in a brief daily scrum call. Projects consist of weekly
services sprints. Each sprint represents a one-week development cycle within
4.2.2.1. Structure-level actions. The development of vehicle-data-driven which data scientists explore specific user stories. These actor-level
services had required the collaboration of the R&D unit, which manages actions improved the collaboration between product managers and data
the vehicle-embedded systems and vehicle data, and the marketing scientists and reduced significantly the risk of misguided efforts in
unit, which has customer expertise and manages the development of terms of implementing vehicle-data-driven services.
customer services. However, these two units traditionally had limited It had been clear to PremiumCar’s R&D unit from early on that the
interactions: development of digital vehicle-data-driven services, comparable to en-
gineering vehicle technology, would determine eventually
“At first, it proved difficult to get hold of vehicle data, as our vehicles
PremiumCar’s competitive market advantage. However, the R&D unit
were not able to send data and, even more importantly, we were not
had lacked data science experts with the data-processing skills which
collaborating until then with the R&D unit.” Interviewee E2
are indispensable for vehicle-data-driven services. For this reason,
Consequently, the marketing analytics subunit and the R&D unit PremiumCar had relied heavily on external analytics consultants. In
joined forces by installing a joint task force. This task force was re- order to establish in-house competence in developing BDA applications
sponsible for defining use cases for vehicle-data-driven services and in the R&D unit, PremiumCar introduced co-coding for projects with
initiating first pilots. Establishing this collaboration was an essential external analytics consultants.
prerequisite for the subsequent actions taken to collect and store cen-
“We use co-coding to avoid being dependent on the analytics knowledge
trally vehicle data as well as to implement vehicle-data-driven services.
of external partners […] that way we make sure that we develop the
Early initiatives to implement vehicle-data-driven services had
required analytics know-howwithin our company.” Interviewee N2
faced two major obstacles. First, there had been very limited techno-
logical capability with which to collect vehicle data because the Through this action, the analytics competencies of the R&D unit’s
tracking of vehicle sensors had been limited to a few maintenance-re- employees improved substantially. Today, the R&D unit develops ve-
lated car incidents. Second, the embedded systems in different vehicles hicle-data-driven services largely independently of company-external
had used inconsistent data models and identifiers, which had prohibited support.
an integrated use of sensor data from different vehicle models. As a
consequence, the marketing analytics subunit used its strategic budget 4.2.2.3. Technology-level actions. New cross-unit collaborations at
to finance two projects in order to overcome these obstacles: PremiumCar, such as between the R&D unit and the marketing unit,
resulted in BDA scenarios which required the integration of large and
“To finance our first projects with the R&D unit, we were in the lucky
diverse data sources and high processing capabilities. Since the
position that we had been granted a strategy budget from our top man-
individual units had historically been satisfied with dedicated-purpose
agement to heavily push forward our joint analytics efforts.” Interviewee
BI infrastructures, no infrastructure could meet these new
F2
requirements. In order to address these demands, the IT unit
8
Table 4
Actions of Provisioning Vehicle-Data-Driven Services.
- New collaborating of the marketing analytics - Improving the interworking of product managers and - Constructing a cross-unit Hadoop platform
subunit and R&D unit software developers with scrum - Improving the exchange of data between vehicles and the
- Central funding of marketing-R&D analytics - Co-coding with external partners analytics platform with car data loggers
projects
implemented a central BDA platform which is open to all of the R&D analytics subunit was installed. This subunit is responsible for
PremiumCar’s units. It is designed as an on-premises platform which the analysis of sensor signals, bus system signals, and overall product
supports large-scale analyses and reporting. For large-scale analyses, it test data before the actual start of production. In addition, the subunit
covers the entire Hadoop ecosystem, including cluster storage, Spark, took over the task of conceptualizing a technological infrastructure for
and the support of diverse analytical programming languages such as vehicle sensor data analyses. Establishing this R&D analytics subunit
SQL, R, and Python. For visualization purposes, the platform includes was a key structure-level action because the combination of analytical
the Tableau and SAP BusinessObjects server. capabilities and vehicle electronics expertise is indispensable for data-
driven vehicle developing.
“The platform receives data from our cars, production data (this means
The R&D unit is organized in a decentralized fashion into several
we try to connect our robots from the assembly lines), diagnosis data
vehicle-component-specific R&D subunits with historically developed
from our dealerships […] we try to consolidate our data streams across
work structures. Owing to an initial skepticism within some of these
the company with the platform.” Interviewee S2
vehicle-component-specific R&D subunits, the R&D analytics subunit
This platform represents a fundamental enabler for the development struggled initially to gain access to information about all relevant ve-
of vehicle-data-driven services and supports, among others, the ex- hicle-data sources. In order to ensure the R&D analytics subunit’s access
ploration of remote maintenance, fleet management, and advanced to relevant vehicle data sources, the R&D unit introduced a standar-
navigation scenarios. dized data approval process. Chief data stewards now serve as re-
Traditionally, the logging functionality of the vehicle-embedded presentatives for their R&D subunits and have the power to authorize
systems had been limited to a fault memory which logged few main- access to the data sources for which the respective subunit is re-
tenance-relevant events. This data had been transferred to PremiumCar’s sponsible. The approval process also includes stakeholders from the
systems very sporadically when dealers read out this fault memory legal unit (with particular expertise related to privacy issues) and from
during maintenance services. Vehicle-data-driven services, such as the IT unit (with particular expertise related to technological topics).
predictive maintenance, however, rely on detailed and up-to-date
“Every department has to nominate a chief data steward. Supported by a
sensor data, which were not logged at that time. In order to capture a
purchased data management software tool, we will implement a quick
broad array of sensor data and transfer it to the central BDA platform on
process to ensure the access to data.” Interviewee A1
a regular basis, PremiumCar decided to mount data loggers in new ve-
hicles. The process to gain access to data sources requires that the project
leader fills in a data request online. The stakeholders involved in the
“Together with R&D, we have to lay the technological foundation for
approval process (e.g., system and data owners) have the right to object
accessing vehicle data. […] At first, we absolutely had to build in data
to a data access request. If, within 1 week, no objection is made, the
loggers within our vehicles to allow the over-the-air-transfer of data.”
data access request is approved automatically. This data approval
Interviewee E2
process facilitates significantly data-driven vehicle developing through
The data loggers are connected to the vehicles’ data transmission avoiding project setbacks because of a lack of access to data.
system and thus allow an instant over-the-air transfer of sensor data.
Having access to this rich vehicle data at near real-time in the central 4.2.3.2. Actor-level actions. The R&D unit had been faced with a
BDA platform improved substantially PremiumCar’s ability to explore shortage of employees who combined vehicle engineering knowledge
and design innovative vehicle-data-driven services. with profound analytical capabilities, both of which are required for
Table 4 summarizes the key actions taken to provision vehicle-data- vehicle-data-driven engineering.
driven services at the structure, actor, and technology levels.
“In particular, when we look at cars, the domain-specific knowledge is
key. You are just not able to analyze any data of the car without having a
4.2.3. Actualization of affordance 3: data-driven vehicle developing
proper understanding of the car itself […] our cars are technically very
4.2.3.1. Structure-level actions. Owing to a lack of capabilities related to
complex.” Interviewee O2
the analysis of large volumes of data and the use of a scalable
infrastructure, PremiumCar’s R&D unit had been unable to cope with In order to equip domain experts in R&D with analytical capabilities,
the sheer amount of data a car produces during an on-road test. During the R&D unit began to take part in PremiumCar’s company-wide BDA
such tests, a single vehicle produces more than one gigabyte of data per education program. This program consists of five modules which are
second. Collaborations with external analytics providers had proved separated into three levels (i.e., basic, advanced, and expert). The first
difficult because the sensor data which the in-car systems produce are module, digital business, focuses on an overview of methods and ap-
highly specific. proaches (e.g., CRISP-DM, a modeling approach for data mining, and
design thinking, an interdisciplinary innovation methodology) as well as
“We at R&D collaborated initially with a lot of external partners…
portfolio management. The second module, data architecture, covers
but we had to realize that it is really difficult to find a partner who
relational and nonrelational database systems. At the expert level, the
could deliver components which work properly and take the do-
foundations of SAP HANA and Hadoop on Hortonworks are addressed.
main-specific internal car systems into account.” Interviewee R2
The third module, data analytics, comprises methods with which to
As a consequence, the R&D unit had only been able to analyze very structure and cleanse data as well as foundations in R and Python.
specific errors and safeguard limited and clearly defined car function- Machine learning and artificial intelligence, which comprise the fourth
alities. In order to help expand the coverage and depth of on-road tests, module, cover supervised and unsupervised learning methods as well as
9
Table 5
Actions of Data-Driven Vehicle Developing.
- Constructing an R&D analytics subunit for car - Improving employee BDA competencies through a - Constructing an R&D Hadoop platform for on-road
usage data central education program test data
- Using a standard data approval process - Recruiting platform architects for the IT unit
- Establishing a closer exchange between security
experts from the IT in-car and IT security subunits
neural networks. The last module, data strategy, illustrates strategic data. […] Though we have the technological infrastructure and processes
trends in analytics and data governance. This education program im- to collect and store the data, we had to extend our technological land-
proved significantly the ability of engineers in the R&D unit to partici- scape to allow analyzing these high volumes of data.” Interviewee R2
pate in data-driven vehicle developing.
In order to enhance the existing data storage system with analytics
Initially, the conceptualization of a dedicated analytics platform had
components, the R&D analytics subunit initiated the implementation of
been done by the R&D analytics subunit, which had, however, turned to
a Hadoop-based analytics platform for vehicle test data analysis. This
the IT unit for help with scalability issues it could not resolve. But the IT
platform includes Spark and a Hadoop distribution of MapR as analytics
unit had initially been unable to resolve these issues owing to a lack of
engines. It supports interactive programming through Jupyter and
expertise in building and maintaining BDA platforms. The IT unit re-
Python. This platform provided the R&D analytics subunit, for the first
acted consequently to the R&D unit’s new technology demands and
time, with the technological means by which to exploit fully the col-
recruited BDA platform architects.
lected test data for data-driven vehicle developing.
“For us, it is crucial that we have the manpower to evaluate the tech- Table 5 summarizes the key actions taken for data-driven vehicle
nological solutions available on our own and to choose the best suitable developing at the structure, actor, and technology levels.
technological stack for our use cases […] in this regard, [the IT unit]
took a huge step forward and hired platform architects to bring our en-
4.2.4. Actualization of affordance 4: optimizing production processes
deavor forward.” Interviewee R2
4.2.4.1. Structure-level actions. With regard to managing the production
After recruiting platform architects, the IT unit and R&D analytics facilities’ IT systems, the production unit had historically collaborated
subunit could collaboratively provide the needed employee-level skill with the manufacturers of the production machines and not with the IT
set for data-driven vehicle developing independently of external sup- unit. In its efforts to implement a smart factory approach, the
port. production unit required a competent partner who supported the
In addition to using sensor data from on-road tests, PremiumCar implementation of a secured digital twin of the production facilities.
further had started to inform vehicle development with customer-gen- Because the IT unit had by then developed considerable knowledge in
erated vehicle usage data which was transferred to PremiumCar’s cen- managing big data infrastructures, the production unit decided to
tral systems over-the-air. Historically, sensor data from PremiumCar’s collaborate with the IT unit.
sold vehicles had only been transferred through secured access methods
“We started to collaborate heavily with the IT unit as they built an
to central systems in the rare event of car maintenance being done at
analytics platform for [PremiumCar]. The IT unit supports us with the
PremiumCar’s repair shops. For this reason, employees in the IT in-car
toolset on which we can work and also takes care of all the licensing
subunit very rarely had contact with security experts in the IT security
issues you face when you use enterprise solutions.” Interviewee O2
subunit. The extension of the vehicles’ functionality to transfer usage
data over-the-air required a further intensification of the collaboration In the course of this novel collaboration, the chief of production and
between the car and IT security experts. Therefore, regular meetings the chief information officer hosted a hackathon. The goal was to
between both subunits’ security experts were set up in order to align connect to skilled and creative data scientists. The hackathon led to 18
their current data-related security issues. use cases, of which 12 were followed up on within the production unit.
Overall, the IT unit’s analytical capabilities contributed substantially to
“Initially, our vehicles and our IT landscape had to be seen as disjoint
implement the technological infrastructure and generation of in-
systems. They only interacted in rare cases, such as in the case of
novative use cases for a smart factory approach.
maintenance work […] accordingly, both the vehicles and the IT land-
A plethora of digitization efforts had been scattered across
scape were separately secured […] Today, however, the IT security ex-
PremiumCar’s production unit and had been insufficiently aligned to
perts and in-car security experts work closely together, as we are
improve PremiumCar’s production processes in a sustainable manner. In
otherwise not able to look at all aspects of security throughout the chain
order to improve the structural support for its digitization efforts, the
from the car to our IT landscape and vice versa.” Interviewee N2
production unit implemented a board of digital officers. Each subunit
The active exchange between the car and IT security experts sup- within the production unit nominates a digital officer who is re-
ports data-driven vehicle developing through ensuring end-to-end se- sponsible for his/her subunit’s digitization efforts and consolidating
curity measures. these efforts at the production unit level.
“Besides coordinating our digitization efforts, which are mostly analytics
4.2.3.3. Technology-level actions. An immense amount of data is
topics, they try to avoid redundancies within the production unit. They
produced during an on-road test which requires the use of hard
have to find answers to:“Which data do we need from which production
drives within the test vehicle for storage. These hard drives are
site? What should our system landscape look like over a five- or ten-year
collected and the respective data is stored within an existing data
time horizon?” Interviewee O2
storage system in the R&D unit. However, the R&D analytics subunit
had lacked an adequate analytics infrastructure with which to process The committee not only prioritizes digital topics, which are mostly
and analyze this high volume of data with advanced analytics methods. analytics topics, and grants budgets, but also instructs subunits within
the production unit to take over tasks. This structure-level action im-
“During on-road tests, our test cars produce more than a gigabyte of
proved significantly the management of analytics efforts within the
10
Table 6
Actions of Optimizing Production Processes.
- New collaborating between the production and IT units - Establishing a social community for BDA in the - Digitizing interfaces to production
- Improving structural digitization support through a board of digital production unit facilities
officers in the production unit - Offering a data-awareness program for - Implementing a data lake for the
management-level employees production unit
production unit, leading to an avoidance of redundancies and ensuring facilities allowed, for the first time, the unified and centralized access to
strategic prioritization when optimizing production processes. previously untouched data points and thus was a precondition for the
data-driven optimization of production processes.
4.2.4.2. Actor-level actions. Because analytics experts at the individual While it had been possible for some production subunits to manage
production subunits had traditionally focused on optimizing their local production facilities digitally, the global distribution of
specific production steps, there was a lack of knowledge exchange PremiumCar’s production facilities around the world required a central
and collaboration among the analytics experts at the production unit. In solution in order to collect and store production data from across the
order to ensure employee-level collaboration and bring together globe. Owing to the inability to integrate production data on a global
analytics experts, the production unit launched a social community. scale, PremiumCar had not been able to leverage synergies through a
This social community uses an enterprise social software in order to global optimization of, among others, yard management, energy man-
communicate and exchange information and further hosts information agement, and error management. In order to integrate the data from the
exchange meetings on a regular basis. world-wide production facilities (e.g., data from KUGA robots, press
shop, or other production facilities) into the company-wide BDA plat-
“Our analytics efforts are supported by regular workshops within our
form, the production unit initiated the implementation of a data lake.
production unit which are organized by a colleague from another sub-
unit. He actively pushes our social community in order to ensure colla- “Our production unit is dominated by the continuous improvement pro-
boration and a common understanding in regard to our analytics efforts cess mindset. Thus, we try to support as well as we can the process of
at the production unit.” Interviewee P2 connecting data from all of our production facilities to our technological
infrastructure and making the data analyzable. […] We are producing
Through this social community, the production unit improved the
across the globe, which makes it even more challenging.” Interviewee B2
sharing of best practices among analytical-savvy engineers in terms of
optimizing production processes. Based on these data, 300,000 screwing processes, for instance, were
The smart factory efforts at the production unit had been struggling analyzed to identify proactively errors in early car production phases,
with strong opposition from middle-level managers. Many managers thus avoiding cost-intensive repairs in later production phases. The data
feared losing decision-making competencies and a devaluation of their lake at the production unit was a key technology-level action because
domain-specific skills owing to increased production automation. For providing accessibility to the global production data represented a
these reasons, they had been reluctant to change their established work crucial actualization step toward optimizing production processes.
practices. In order to educate these managers and receive their buy-in, Table 6 summarizes the key actions taken for optimizing production
the production unit hosted a series of data awareness workshops. processes at the structure, actor, and technology levels.
“We offer data awareness workshops on the management level. But you
5. Discussion
always have to consider that you will not reach all employees through the
management or top-down approach. Thus, you have to complement it
We conducted a revelatory case study at AutomotiveGroup, one of the
with bottom-up approaches.” Interviewee O2
largest automotive corporations worldwide, in order to study the ac-
These workshops were complemented by events in which employees tualization of BDA affordances, which has been identified as lacking a
of the decentralized units presented their initial success with applying research base [6,30,32]. We have identified four BDA affordances,
BDA to achieve smart factory approaches. Through achieving commit- namely: 1) establishing customer-centric marketing, 2) provisioning of
ment on the management level, additional projects and, particularly, vehicle-data-driven services, 3) data-driven vehicle developing, and 4)
more capacities could be acquired when optimizing production pro- optimizing production processes, and we gathered detailed information
cesses with smart factory approaches. on affordance actualization at the structure, actor, and technology levels.
Analyzing the actions taken, we see evidences for four underlying
4.2.4.3. Technology-level actions. In order to implement a full digital mechanisms used by PremiumCar to advance affordance actualization:
representation of its production facilities, PremiumCar’s production unit enhancing, constructing, coordinating, and integrating.
had to gain access to the data of all of the production machines. Access
to the data from many of PremiumCar’s production facilities, however, - Enhancing refers to the further development of a company’s socio-
had not been feasible initially as, in the majority of the cases, no digital technical entities.
interfaces existed. In order to establish digital connections, the - Constructing refers to the implementation of novel socio-technical
production unit decided to build dedicated interfaces and leverage entities.
company-internal platforms owing to the operational criticality of these - Coordinating refers to improving the interworking of socio-technical
interfaces. entities.
- Integrating refers to establishing novel interfaces between socio-
“If we are not able to connect to our machines using protocols such as
technical entities.
OPCUA, we enable our machines to connect to our LAN- or WLAN-
infrastructure and try to access the data using different approaches […]
We refer to these mechanisms as “actualization mechanisms,” and
for instance, screen scraping.” Interviewee M2
we have detailed specific actions on the structure, actor, and technology
Establishing these interfaces enabled the production machines to be levels which refer to each mechanism as we found them in the data in
read out on a regular basis. The digitization of interfaces to production Table 7.
11
Table 7
Actualization Mechanisms and Exemplary Organizational Actions.
Actualization Structure Level Actor Level Technology Level
Mechanism
Enhancing Further development of structural entities: Further development of employee competencies: Further development of information
- Using a standard data approval process - Improving employee BDA competencies technologies:
- Improving structural digitization support through a central education program - Constructing an R&D Hadoop platform for
through a board of digital officers in the - Co-coding with external partners on-road test data
production unit - Offering a data-awareness program for - Delivering dynamic analytics services
management-level employees through a Tableau solution instead of
offline dashboards
Constructing Installing novel structural entities: Hiring employees with novel competencies: Implementing novel information
- Constructing a marketing analytics subunit - Recruiting data scientists, big data architects, technologies:
- Constructing an R&D analytics subunit for car and visualization experts in a start-up-like - Implementing a Teradata- and Tableau-
usage data analytics subsidiary based technological infrastructure
- Recruiting platform architects for the IT unit - Constructing a cross-unit Hadoop platform
Coordinating Improving the interworking of structural entities: Improving the interworking of employees: Improving the interfacing of information
- Aligning marketing analytics subunit, the IT - Improving the interworking of product technologies:
analytics subunit, and the analytics subsidiary managers and software developers with - Improving the exchange of data between
through a steering committee scrum vehicles and the analytics platform with car
- Central funding of marketing-R&D analytics - Establishing a closer exchange between data loggers
projects security experts from the IT in-car and IT - Implementing a data lake for the
security subunits production unit
Integrating Installing novel collaborations between structural Generating novel collaborations between Establishing novel interfaces between
entities: employees: information technologies:
- New collaborating of the marketing analytics - Newly bringing together big data experts in - Integrating customer and dealer data from
subunit and R&D unit the IT unit through a big data competence dealer systems and customer portals
- New collaborating between the production and center - Digitizing interfaces to production facilities
IT units - Establishing a social community for BDA in
the production unit
The first mechanism is enhancing. PremiumCar, for instance, for- While these results assert the importance of organizational learning for
malized the data approval process within the R&D unit and introduced BDA capability, there has been, to the best of our knowledge, no prior
the role of digital officers in the production unit; both actions represent research on how firms develop BDA capability through organizational
structural-level enhancements. At the actor level, PremiumCar launched learning. The organizational learning literature distinguishes two polar
an education program and implemented alternative mechanisms to learning mechanisms [101–103]. Incremental learning refers to the im-
further develop employee capabilities. At the technology level, provement of an established capability through the modification of the
PremiumCar improved its IT systems, among others, by enhancing a individual building blocks of this capability, for example, by optimizing
storage system for on-road test data with analytical system components an operational routine. Radical learning refers to the reframing of a si-
and improving reporting technologies with dynamic Tableau dash- tuation and acquisition of novel capabilities, for example, by hiring
boards. Through these actions, PremiumCar developed further the employees with competencies in a novel and sought area. Furthermore,
company’s socio-technical entities in order to actualize BDA affor- learning in organizations takes place at the individual level (i.e., novel
dances, which we refer to as the actualization mechanism enhancing. employee knowledge) as well as at the organizational level (i.e., novel
The second mechanism is constructing. PremiumCar, for instance, organizational structures and technological capabilities) [18,104].
constructed analytics subunits in the marketing and R&D units, which In PremiumCar’s actualization of BDA affordances, incremental
are actions taken on the structure level. At the actor level, PremiumCar learning at the employee level manifests in actor-specific instantiations
recruited data scientists to its analytics subsidiary and increased sub- of the enhancing mechanism. An example of incremental learning at the
stantially the headcount of big data platform architects in the IT unit. At individual level is the employee education program because it signals
the technology level, PremiumCar introduced novel technologies, most an enhancement of employee competencies. Incremental learning at the
notably a central BDA platform based on Amazon Web Services and organizational level manifests in structure- or technology-specific in-
Hadoop, as well as a Hadoop-based analytics platform in the R&D unit. stantiations of the enhancing mechanism. The introduction of a stan-
Through these actions, PremiumCar implemented novel socio-technical dardized data approval process in PremiumCar’s production unit, for
entities in order to actualize BDA affordances, which we refer to as the example, is evidence of incremental learning at the organizational level
actualization mechanism constructing. because it represents a structure-specific enhancement of the (pre-
Results provided by enhancing and constructing provide insights for viously informal) data approval practices.
the ongoing discussion on organizational learning. Specifically, we in- Radical learning at the employee level manifests in actor-specific
terpret instances of the enhancing mechanism as evidence of incre- instantiations of the constructing mechanism. An example of radical
mental learning. Similarly, we interpret instances of the constructing learning at the individual level is the recruitment of data scientists with
mechanism as evidence of radical learning. This, we argue in the fol- competency profiles, which PremiumCar’s employees did not match, for
lowing, extends the current knowledge of how firms develop BDA the start-up-like analytics subsidiary; this action signals the construc-
capability. tion of novel employee competencies. Radical learning at the organi-
The sparse research on BDA capability development accentuates the zational level manifests in structure- or technology-specific instantia-
fact that a firm’s intensity of organizational learning represents a de- tions of the constructing mechanism. The installation of a novel
terminant of BDA capability [16,17]. In their empirical validation of the marketing analytics subunit, for example, represents radical learning at
BDA capability construct, Gupta and George [16] consider the intensity the organizational level because it entailed the construction of a new
of organizational learning as a formative factor for BDA capability, structure-specific entity.
measured with several items which describe the perceived ability of a To summarize our discussion of the first two mechanisms enhancing
firm to search for, acquire, assimilate, and apply relevant knowledge. and constructing, our case study demonstrates that PremiumCar conducts
12
incremental and radical learning on the individual and organizational alignment. The installation of a committee which steers jointly the
levels to actualize BDA affordances. By detailing how PremiumCar im- marketing and IT units’ analytics activities, for example, led to a shared
proved incrementally socio-technical entities through enhancing and understanding of the BDA objectives and priorities which facilitated the
how PremiumCar radically introduced novel socio-technical entities actualization of customer-centric marketing. Social BDA alignment de-
through constructing, we provide a link to the extant literature on or- notes a state in which BDA employees and employees from other busi-
ganizational learning [101–103] and contribute to an understanding of ness functions understand and are committed to the business and BDA
how firms develop BDA capability. mission and objectives (cf. [107]). Wamba et al. [25] consider social BDA
The third mechanism is coordinating. PremiumCar, for instance, im- alignment as a subconstruct in their operationalization of BDA capability.
proved the structure-level alignment of the marketing and IT units with They measure the general level to which BDA experts and line people
a steering committee which brings together key stakeholders for the exchange information and coordinate joint work and refer to it as BDA
marketing analytics service initiative. At the actor level, PremiumCar coordination. Our case study’s actor-level actions which we allocate to
introduced scrum in order to align the work of product managers and the mechanisms of coordinating and integrating introduce a fine-grained
software developers. At the technology level, PremiumCar, among other perspective on how firms increase social alignment in their efforts to
actions, improved the exchange of sensor data between vehicles and a actualize BDA affordances. The introduction of a social community
central analytics platform with the use of car data loggers. Through platform, for example, fueled the personal exchanges between analytics
these actions, PremiumCar intended to improve the interworking of experts in the production unit and, in this way, supported the actuali-
socio-technical entities in order to actualize BDA affordances, which we zation of optimizing production processes. Operational BDA alignment
refer to as the actualization mechanism coordinating. refers to the alignment of business infrastructure and processes with BDA
The fourth mechanism is integrating. PremiumCar, for instance, in- infrastructure and processes (cf. [108]). Ghasemaghaei et al. [24] ad-
troduced the collaboration between the R&D unit and the marketing dress operational BDA alignment and empirically show that the fit be-
analytics subunit in order to support the development of vehicle-data- tween BDA tools and data, between BDA tools and people, as well as
driven services on a structure level. At the actor level, a newly created between BDA tools and tasks, moderates the impact of data analytics use
big data competence center, for example, brought together big data on firm agility. Our case study’s technology-level actions which we al-
experts who were distributed across the IT unit. At the technology level, locate to the integrating and coordinating mechanisms represent further
customer data owned by the car dealers were newly integrated with actions in support of operational BDA alignment beyond a focus on BDA
customer data collected through PremiumCar’s web platforms. Through tools and thus broaden the understanding of how firms implement op-
all of these actions, PremiumCar intended to establish novel interfaces erational BDA alignment. The integration of customer data from dealer
between socio-technical entities in order to actualize BDA affordances, systems and PremiumCar’s online platforms in Affordance 1, for example,
which we refer to as the actualization mechanism integrating. led to an increased alignment of customer analytics systems to Pre-
Insights related to coordinating and integrating contribute to the ex- miumCar’s customer-facing business processes and thus facilitated the
isting body of knowledge by providing a detailed understanding of how actualization of customer-centric marketing.
firms align BDA-related socio-technical entities and thus extend the To summarize the insights achieved through our work for the on-
currently limited body of knowledge on BDA alignment. going BDA alignment discussion [23–25], our study leads to the rich
Many prior studies take a resource-based perspective on the con- discovery of actions which represent integrating and coordinating me-
stituents of BDA capability and discuss tangible, intangible, and human chanisms on the structure, actor, and technology levels. This discovery
resources (e.g., [16,17,30]). The actualization mechanisms enhancing extends the currently limited knowledge on how firms achieve BDA
and constructing are congruent with this resource-based perspective on alignment, particularly in the intellectual, social, and operational di-
BDA capability for the following three reasons. First, enhancing and mensions of alignment.
constructing socio-technical entities at the technology level addresses In sum, we conclude that companies actualize BDA affordances
the formation of tangible resources, such as data, infrastructure, and through enhancing, constructing, coordinating, and integrating socio-
software. Second, enhancing and constructing structure-level entities technical entities. The case study also highlights specific actions
addresses the formation of intangible resources, such as business ana- PremiumCar has taken regarding each mechanism on the structure,
lytics governance and prioritization of business analytics investments. actor, and technology levels so that, overall, a balanced portfolio of
Finally, enhancing and constructing actor-level entities addresses the actions has been created. These mechanisms and the identified actions
formation of human resources with a focus on, for example, technical characterize conjunctively the “recursive shaping” of socio-technical
knowledge and analytical skills. However, the results of our research entities [33,37,38].
show that actualization is not merely a result of providing individual
socio-technical entities through enhancing and constructing. Actualiza- 6. Implications
tion further requires the coordination and integration of these entities.
With this finding, we address the open research question as to how BDA 6.1. Implications for theory
resources must be aligned in order to reach sustainable performance
effects [25]. Prior studies on BDA have primarily focused on either the potential
Our findings support the results of prior researchers who have value of BDA (e.g., [6,109]) or on the organizational resources and
highlighted the importance of aligning BDA-related socio-technical en- capabilities for BDA (e.g., [16,30]). The link between effective appli-
tities [17,105,106] and enrich the present understanding of how align- cations of BDA and associated organizational preconditions, however, is
ment is implemented in different alignment dimensions, as follows. Prior still omitted in existing research [6]. This link, however, is essential in
IT alignment research distinguishes several alignment dimensions, in- creating knowledge on how to generate business value through BDA.
cluding intellectual, social, and operational alignment [107,108]. In- Our work contributes to the current body of knowledge on the value
tellectual BDA alignment deals with how business strategy is supported by realization of BDA in several ways. We identify affordances which drive
BDA strategy (cf. [108]). Based on a general assessment of how BDA automotive manufacturing companies to adopt BDA. Thus, we respond
plans align with the companies’ objectives and support the companies’ to the call for reserach of Abbasi et al. [87] to study concrete and in-
direction, Akter et al. [23] empirically show that intellectual BDA dustry-specific scenarios on how enterprise-wide BDA maximizes the
alignment moderates the impact of a BDA capability on firm perfor- potential for competitive advantage and shed light on how BDA can
mance. Our case study’s structure-level actions, which we allocate to the provide and deliver value in the automotive industry.
mechanisms of coordinating and integrating, detail the understanding of Further, we adopt a theoretical perspective on BDA actualization by
how firms actualize BDA affordances through achieving intellectual BDA connecting affordance theory with the STS theory. Inspired by Volkoff
13
and Strong [47], we study affordance actualization by gaining detailed each mechanism and each socio-technical level. In this manner, our
insights into the adaptation of organizational structure, actors, and BDA study can provide initial support for a company in developing and
technology. In doing so, we identify the organizational actions required transforming its STS to realize value from BDA.
to actualize and thus profit from BDA. Our findings contribute to the
general IS affordance literature by supporting a better understanding of 6.3. Limitations
how configurations of the STS relate to the actualization of affordances,
as called for by Zammuto et al. [48]. Due to the interpretive nature of our revelatory case study, we
Moreover, we respond to Günther et al. [6], Mikalef et al. [30], and cannot claim that we have explored exhaustively the actualization of
Wamba et al. [32] call for empirical studies on the realization of BDA BDA affordances. The statements in our interviews could have been
value. We contribute an actions-oriented perspective to the currently interpreted differently by other people. In particular, the fact that we
sparse theory on the organizational-level formation of BDA capability. were in intense contact with our case organization between October
Our theorization on the BDA affordance actualization represents a type 2016 and October 2017 bears the risk of inconsistent interpretations of
II mid-range theory (“theory for explaining”) [110]. With our re- early interviews and interviews conducted at later stages. To take this
velatory case study, we do not aim at generating testable predictions, threat into account, we were careful when analyzing early data and
but, instead, to provide a theory which is “new and interesting […] to aimed at triangulating findings between multiple data sources.
explain something that was poorly or imperfectly understood before- Moreover, throughout the process of analyzing the data, we tried to
hand” ([110], p. 625). create clear chains of evidence, considered multiple viewpoints, and
We examine the process of building BDA capability, as called for by involved multiple researchers in data coding and analysis.
Mikalef et al. [70], and discover four general mechanisms behind how a Further, as we used a single case study approach to gain rich, in-
company actualizes BDA affordances: enhancing, constructing, co- depth insight, the generalizability of our findings may be limited.
ordinating, and integrating. Our discovery of these mechanisms on the However, we base our empirical findings on well-accepted IS theories
structure, actor, and technology levels of a company’s STS extends and propose four general actualization mechanisms (i.e., enhancing,
particularly prior research on BDA alignment by relating the mechan- constructing, coordinating, and integrating) which emerged from our case
isms to extant IT alignment theory [107,108] and by distinguishing the results to support further empirical research.
organizational actions of intellectual, social, and operational BDA Our case study can be considered as a revelatory case as our case
alignment. Further, this discovery extends the current knowledge on organization serves as a prime example for the actualization of BDA
how a company develops BDA capability by relating the mechanisms to affordances when similar evidences or emergent theories on the ac-
theories on organizational learning [101–103] and by eliciting two tualization of BDA still need to be developed. Before generalizing our
modes of organizational learning, incremental and radical learning, on results to a further extent, however, additional research is required to
the individual level as well as on the organizational level. prove that the results of this study also hold true in different contexts
and industries.
6.2. Implications for practice
7. Conclusion
Despite the increasing application of BDA, it is not well understood
how firms actualize BDA affordances successfully. In this research, we The process of realizing value from BDA has long been a Pandora’s
uncover structure-, actor-, and technology-level actions resulting in the box for both scholars and practitioners. Insights from our revelatory case
actualization of four affordances at PremiumCar. Our revelatory case study of PremiumCar shed light into the actualization of four BDA affor-
study provides actionable inspirations for organizational actions which dances which address customer-centric marketing, data-driven services
contribute to the realization of business value through BDA. and development approaches, and optimized production processes. With
The findings suggest that firms should leverage jointly four me- regard to our introductory research question of how socio-technical ac-
chanisms to actualize BDA affordances: enhancing, constructing, co- tions contribute to actualizing, our results indicate that the enhancing,
ordinating, and integrating. As an example of enhancing, a company must constructing, coordinating, and integrating of socio-technical entities
complement its analytics systems with state-of-the-art BDA technolo- conjunctively form the mechanisms which lead to an affordance being
gies to enable scalable analytics, as the enhancement of PremiumCar’s actualized. Our socio-technical perspective, which frames actualization as
storage system shows (see Affordance 3). As an example of constructing, the process of a recursive shaping of social and technical entities
companies must implement novel structural entities to pool BDA ex- [33,37,38], enabled us to discover evidence of incremental and radical
pertise and to create the organizational frame for pursuing BDA in- learning on both the organizational and individual levels. Further, it al-
itiatives, as PremiumCar’s marketing analytics subunit shows (see lowed us to distinguish three modes of BDA alignment (i.e., intellectual,
Affordance 1). Firms, however, should not only focus on constructing social, and organizational alignment) which PremiumCar leverages.
novel socio-technical entities or on enhancing existing socio-technical We foresee three avenues for further research based on the results of
entities on the structure, actor, and technology levels. Rather, they this study. First, future research should analyze in other contexts how
should also align these entities by coordinating and integrating them. the four proposed mechanisms have a bearing on the actualization of
As an example of coordinating, companies must improve the inter- BDA affordances. Supplementary insights into how these mechanisms’
working of BDA stakeholders, as the introduction of a scrum process to effectiveness is contingent on factors such as industry, level of digita-
coordinate product managers and software developers shows (see lization, or size would further increase our knowledge on the actuali-
Affordance 2). As an example of integrating, companies must bring to- zation of BDA affordances. Second, our findings motivate studies on
gether disparate data sources, as the integration of customer data from how incremental and radical organizational learning mechanisms in-
dealer systems and customer portals shows (see Affordance 1). teract and shape the development of BDA capability. Particularly, we
In summary, our findings can help guide practitioners in crafting require further knowledge on the interrelationship of individual- and
their BDA strategies and will be particularly insightful for traditional organizational-level learning and the combined effects on BDA cap-
businesses seeking to profit from BDA, for instance, in form of product ability formation. Third, our findings evidence that organizational ac-
or service innovations. The discussed actualization mechanisms can, tions on intellectual, social, and operational BDA alignment contribute
together with the exemplary actions which we discovered at to actualizing BDA affordances. Further research should study in
PremiumCar and instantiate these mechanisms (see Table 7), be used as greater detail the interrelationships of the different types of BDA
an organizational framework. Companies can map their planned and alignment and BDA value. Particularly, studies on whether and how the
undertaken organizational actions in order to reflect on the coverage of roles of intellectual, social, and operational BDA alignment are
14
contingent on a company’s strategic orientation would enhance greatly issue, namely Patrick Mikalef, Ilias O. Pappas, Paul A. Pavlou, and John
our knowledge on BDA value realization. Krogstie, which helped us improve this article with each revision.
Moreover, we express our gratitude to all interviewees for sharing their
Acknowledgements experiences and knowledge making this research possible. Finally, we
would like to thank the journal manager Sudharson Aiswarya and the
The authors are very grateful for the valuable feedback of the re- production team of Information and Management.
viewers, the editor-in-chief Patrick Chau, and the editors of this special
Appendix A. Interview Protocol
Interview Introduction
The accumulating evidence of the potential benefits provide a legitimate reason to consider big data analytics (BDA) as a sustainable phe-
nomenon rather than a buzzword [[1],50]. BDA opens up new business opportunities, such as using “real-time information from sensors, radio
frequency identification and other identifying devices to understand business environments at a more granular level, to create new products and
services, and to respond to changes in usage patterns” ([1], p. 22). BDA is based on a continuous stream which delivers real-time data from
heterogeneous sources and demands new approaches with which to structure and interpret the gathered data [[1]]. The implementation of BDA
poses challenges regarding the development of appropriate organizational competencies [7,111] because the “expanding sea of data […] is either too
voluminous or too unstructured to be managed and analyzed through traditional means” ([1], p. 22).
Those challenges originate from the vast amount of data which come in both structured and unstructured forms and from various sources, such as
the Web, social media, or the Internet of Things [49,112]. This leads to specific and novel implications for organizations on a procedural (e.g., new
forms of decision-making), organizational (e.g., new employee competencies), and technological (e.g., new platform and tools) level. Accordingly,
the adoption of BDA requires organizational transformations as well as the development of specific analytical (e.g., identification of new patterns)
and technological capabilities (e.g., the effective deployment of big data platforms within the current IT landscape) [4,50,72,113].
In this interview, we aim to get an in-depth understanding of the actualization of BDA affordances. In particular, we focus on required orga-
nizational actions as well as their concrete outcomes.
• Interview Questions
The following questions were asked selectively during interviews. If it proved necessary, follow-up questions were asked to gain a deeper
understanding. The interviews took approximately 45 min. The interviews were recorded if consent was given to enable the anonymized analysis
using qualitative research tools.
• 1. Introduction and General Information

• 1.1. How long have you been working in your current role and what are your main tasks?
• 2. Main Objectives
• 2.2. Which potential benefits (i.e., affordances) led to the adoption of BDA for the OEM which you are related to?
• 2.3. What organizational changes have been/are undertaken to achieve these potential benefits (i.e., affordances)?
• 2.4. What challenges did the OEM face to profit from BDA?
• 2.5. What mid- or long-term goals have been achieved by the OEM with BDA? Which individual, departmental, or company-wide actions were/
are necessary to achieve them?
• 3. Detailed Questions
• 3.1. Social System
• 3.1.1. Please explain how digitalization and, in particular, BDA are anchored in the organizational structure.
• 3.1.2. Why was this design chosen (advantages/disadvantages)?
• 3.1.3. What roles have been created because of BDA and why?
• 3.1.4. What were the success factors for establishing BDA within the company?
• 3.1.5. Which governance mechanisms were helpful for anchoring BDA (e.g., committees, new processes, structures, etc.) (especially regarding the
analysis of car data)?
• 3.1.6. What is the role of the IT department in the successful use of BDA?
• 3.1.7. Do collaborations with external service providers take place regarding BDA initiatives and projects?
• 3.1.8. Which capabilities are critical to the successful profiting from BDA?
• 3.1.9. How are these capabilities built up or brought into the company?
• 3.1.10. Are there any classical organizational capabilities that should be developed or modified in particular? Or do certain capabilities gain in
importance?
• 3.1.11. To profit from (big) data analytics, should it be anchored in the culture of the company (i.e., a data-driven culture)?
• 3.1.12. Has such a data-driven culture been implemented in the past or is it recognizable? Which characteristics are recognizable?
• 3.1.13. Is the access to the data always possible, or are there any resistant forces recognizable in the company? How is data access guaranteed?
• 3.2. Technical System
• 3.2.1. What technologies are used in this context and why was this technology chosen?
• 3.2.2. What data sources have been/are used? Are traditional data sources (such as ERP, CRM) used? What are the most valuable data sources and
why?
• 3.2.3. How is technological access to relevant data sources ensured (e.g., 5 G for car data streams)?
• 3.2.4. To what extent do the findings generated through applying BDA technologies affect organizational decision-making processes?
15
• 3.2.5. How are the findings passed on to internal or external customers (e.g., provisioned through analytics services or one-off projects/in-
itiatives)?
• 3.3. What else
• 3.3.1. Against the background of these questions would you like to add something?
Interview Participants
Table A1
Overview of Interviewees.
Interviewee Role Years in Role Interview Details Suborganization
Round 1. Exploring Affordances

A1 Group Head of Data Analytics & Strategy ∼2 years Online Meeting, 35 min AutomotiveGroup, Group Strategy
B1 Head of IT Analytics Subunit ∼2 years Face-to Face, 56 min PremiumCar
C1 Head of Mobility Services ∼1 year Online Meeting, 30 min MediumCar
D1 Head of Navigation Systems ∼2 years Online Meeting, 30 min MediumCar
E1 Head of Digitization Customer Journey ∼2 years Online Meeting, 37 min CarCo
F1 Head of CRM & Data Analytics ∼1 year Online Meeting, 27 min MobilityServiceCo
G1 Head of Development Big Data Analytics ∼1 year Online Meeting, 30 min TrucksCo
H1 Product Owner, Big Data Analytics Platform ∼1 year Online Meeting, 42 min TrucksCo
I1 Head of Market Intelligence (Data Science & Analytics) ∼1 year Online Meeting, 69 min CommodityCar
J1 Consultant, Data Analytics ∼4 years Online Meeting, 66 min LuxCar
K1 Partner Data Privacy Consultancy, Automotive ∼8 years Online Meeting, 45 min AutomotiveGroup, Privacy Consultancy
L1 Consulting Director IoT/Analytics, Automotive Sector ∼6 years Online Meeting, 20 min AutomotiveGroup, IT-Consultancy
Round 2. Actualization of Affordances
A2 Head of Marketing Analytics Subunit ∼5 years Face-to-Face, 60 min PremiumCar
B2 Head of Big Data Competence Center ∼2 years Face-to-Face, 79 min PremiumCar
C2 Product Owner, Marketing Analytics Service Initiative ∼2 years Face-to-Face, 79 min PremiumCar
D2 Product Owner, IT Analytics Subunit ∼2 years Face-to-Face, 35 min PremiumCar
E2 Product Manager, Car Data #1 ∼2 years Face-to-Face, 49 min PremiumCar
F2 Product Manager, Car Data #2 ∼4 years Face-to-Face, 44 min PremiumCar
G2 Product Manager, Marketing Analytics Service Initiative #1 ∼4 years Online Meeting, 44 min PremiumCar
H2 Product Manager, Marketing Analytics Service Initiative #2 ∼2 years Face-to-Face, 44 min PremiumCar
I2 Product Manager, Marketing Analytics Service Initiative #3 ∼2 years Face-to-Face, 63 min PremiumCar
J2 Product Manager, Marketing Analytics Service Initiative #4 ∼2 years Face-to-Face, 42 min PremiumCar
K2 Manager, Privacy for Marketing Analytics Service Initiative ∼1 year Face-to-Face, 25 min PremiumCar
L2 Manager, Digitization Strategy ∼3 years Online Meeting, 108 min PremiumCar
M2 Manager, Digitization of Production Processes ∼2 years Online Meeting, 66 min PremiumCar
N2 Head of R&D Analytics Subunit ∼5 years Online Meeting, 25 min PremiumCar
O2 Business Analyst, Production Analytics Subunit ∼2 years Online Meeting, 47 min PremiumCar
P2 Data Scientist, Production Analytics Subunit ∼2 years Online Meeting, 30 min PremiumCar
R2 Project Lead, R&D Analytics Subunit ∼2 years Online Meeting, 42 min PremiumCar
S2 Data Platform Architect, Analytics Subsidiary ∼2 years Online Meeting, 44 min PremiumCar
T2 Data Scientist, IT Analytics Subunit ∼1 year Online Meeting, 42 min PremiumCar
Appendix B. Organizational Chart of PremiumCar
Fig. B1. Organizational Chart of PremiumCar (Simplified Excerpt).
16
Appendix C. Full List of Organizational Actions
Table C1
Complete List of Organizational Actions for the Actualization of Affordances.
Organizational Actions Empirical Evidence Affordances STS Dimension
(Actualization Mechanism)
Collocating IT analytics and marketing analytics “To further improve on our collaboration with the IT analytics subunit, we are 1 Structure-Level
subunits now collocated in the same building and on the same floor.” Interviewee A2 (Enhancing)
Using a standard data approval process “Every department has to nominate a chief data steward. Supported by a 1 Structure-Level
purchased data management software tool, we will implement a quick process to 2 (Enhancing)
ensure the access to data.” Interviewee A1 3
Constructing a marketing analytics subunit “We decided on the management level that we want to move towards doing things 1 Structure-Level
independently in order to develop BDA knowledge and skill in house. In particular, (Constructing)
with the topic data, we will have a competitive edge and will face a problem if our
external providers know more than we do, as they might sell the same [services] to
our direct competitors.” Interviewee I2
Constructing an IT analytics subunit “Through the creation of the [analytics subunit at IT], we had, for the first time, a 1 Structure-Level
unit which was responsible for our IT tasks for all the analytics efforts we took (Constructing)
over. Beforehand, it was so difficult to find a supporting partner in the IT unit, as
no one really felt responsible.” Interviewee A2
Aligning marketing analytics subunit, the IT analytics “We meet up on a weekly basis in a steering committee to align all our efforts from 1 Structure-Level
subunit, and the analytics subsidiary through a [the analytics subsidiary], [the IT analytics subunit] and [the marketing analytics 2 (Coordinating)
steering committee subunit]. With four managers, two for our unit, we have to align our strategic
goals and discuss the main challenges we are facing, may it be a business issue or a
technological one.” Interviewee C2
Improving employee BDA competencies through a “We try to educate [that way] our employees and our top management […] we 1 Actor-Level (Enhancing]
central education program want to give them an understanding of the world of data at [PremiumCar].” 2
Interviewee L2 3
4
Initiating a top-down cultural change (through “Data analytics is an immense cultural topic because it holds true that data results 1 Actor-Level (Enhancing]
strategy implementation and leadership) in knowledge, knowledge is power […] to share and to provide access to this data
is perceived of as a danger or risk by some people […] Through our top
management who put analytics and digitization on their strategic agenda, we
could slowly get our colleagues at the marketing unit on board.” Interviewee A2
Ramping up employee knowledge and skills through “In case we do not have the skills and knowledge, we try to collaborate initially 1 Actor-Level (Enhancing)
collaborating with external service providers with external providers to learn and to enable ourselves to slowly take care of the
projects on our own.” Interviewee C2
Recruiting analytics-savvy and experienced business “Through the management decision to internalize knowledge and skills we had the 1 Actor-Level (Constructing)
analysts for the marketing analytics subunit possibility to slowly grow our capacities at the [marketing analytics subunit].”
Interviewee A2
Recruiting data scientists, big data architects, and "The data scientists in the [analytics subsidiary] have substantial independencies 1 Actor-Level (Constructing)
visualization experts in a start-up-like analytics because they can focus on and design the analytics methods and approaches 2
subsidiary without having to handle the day-to-day business. Being located in a more
influential city, the analytics subsidiary is far more attractive to the rare talents we
try to recruit for our analytics projects." Interviewee E2
Leveraging scrum as an agile development process to “When I first started in my role, I had frustrated employees due to the number of 1 Actor-Level (Coordinating)
align employees of marketing BDA projects tasks they had to take care of and the lack of transparency of the tasks they
worked on. So, we started to implement scrum as an agile method.” Interviewee B1
Bringing together previously distributed marketing “At the marketing analytics subunit, we consolidate all analytics experts and 1 Actor-Level (Integrating)
analytics experts and consolidating respective projects of the marketing unit.” Interviewee C2
marketing analytics projects
Newly bringing together big data experts in the IT “The implemented competence network mainly targets employees in the IT 1 Actor-Level (Integrating)
unit through a big data competence center subunits. In our company, we faced initially a serious issue with experts on 4
analytics in the IT unit. The big data competence center was a first step to bring
together all existing and future experts to share their expertise across the IT unit
and, consequently, to be a partner for units like the [marketing analytics]
subunit." Interviewee A2
Delivering dynamic analytics services through a “We built now a Tableau server through which we deliver our analytics service 1 Technology-Level
Tableau solution instead of offline dashboards […] Previously, data snapshots were used and sent as dashboards via mail, for (Enhancing)
instance, as PDFs or sometimes as offline Tableau dashboards. Now, we integrate
these dashboards step-by-step.” Interviewee C2
Implementing a Teradata- and Tableau-based “We implemented, in collaboration with the IT unit, Teradata as the data 1 Technology-Level
technological infrastructure warehouse with the SPSS modeler as an analytics solution and Tableau as (Constructing)
visualization software to unify the technological stack used in marketing.”
Interviewee C2
Integrating customer and dealer data from dealer “The biggest challenge was the fact that no one paid attention to what happens 1 Technology-Level
systems and customer portals with the data and what has emerged over the last five, six, seven years of (Integrating)
operational systems. There existed very few individual examples with both a clean
logic and a clean data model and an appropriate link with other systems.”
Interviewee A1
Leveraging an existing R&D committee to achieve “[It] was crucial to take all established stake holders by the hand. You have to 2 Structure-Level
visibility clarify that the process before was not the worst one and it made sense, but that (Enhancing)
(continued on next page)
17
Table C1 (continued)

you can improve it through analytics […] So, we took part actively in the R&D
committee to push our topic forward." Interviewee E2
Implementing a centralized committee landscape for “We install in every [suborganization] two committees […]. The first [(the 1 Structure-Level
data analytics (i.e., analytics board and office) analytics board)] consists of the top management of each suborganization. The 2 (Constructing)
second committee, the analytics office, comprises managers of the middle 3
management, who are nominated by their business department heads. These 4
managers represent the business department’s interests and make decisions in the
name of the respective business departments.” Interviewee A1
Central funding of marketing-R&D analytics projects “To finance our first projects with the R&D unit, we were in the lucky position that 2 Structure-Level
we had been granted a strategy budget from our top management to heavily push (Coordinating)
forward our joint analytics efforts.” Interviewee F2
Supporting joint analytics efforts through top “We received support from both the top managers from the R&D and the 2 Structure-Level
management in the R&D and marketing units marketing unit. That is something you absolutely need for new topics in our (Coordinating)
company.” Interviewee E2
New collaborating of the marketing analytics subunit “At first, it proved difficult to get hold of vehicle data, as our vehicles were not able 2 Structure-Level
and R&D unit to send data and, even more importantly, we were not collaborating until then (Integrating)
with the R&D unit.” Interviewee E2
Co-coding with external partners “We use co-coding to avoid being dependent on the analytics knowledge of 2 Actor-Level (Enhancing)
external partners […] that way we make sure that we develop the required
analytics know-how within our company.” Interviewee N2
Increasing the employee headcount for car-related “For us, it helped a lot that our top management not only ensured our financial 2 Actor-Level (Constructing)
analytics topics in the marketing unit budget, but also allowed us to recruit new employees for our topics.” Interviewee
F2
Improving the interworking of product managers and “We have to develop a flexibility and agility in regard to our releases. Instead of 2 Actor-Level (Coordinating)
software developers with scrum two major releases, we need – as is characteristic for agile approaches – high- 3
frequency micro releases […] Scrum is one possibility to achieve this and to get
our product management and the developing team together.” Interviewee N2
Constructing a cross-unit Hadoop platform “The platform receives data from our cars, production data (this means we try to 2 Technology-Level
connect our robots from the assembly lines), diagnosis data from our dealerships 4 (Constructing)
[…] we try to consolidate our data streams across the company with the
platform.” Interviewee S2
Improving the exchange of data between vehicles and “Together with R&D, we have to lay the technological foundation for accessing 2 Technology-Level
the analytics platform with car data loggers vehicle data. […] At first, we absolutely had to build in data loggers within our 3 (Coordinating)
vehicles to allow the over-the-air-transfer of data.” Interviewee E2
Constructing an R&D analytics subunit for car usage “We at R&D collaborated initially with a lot of external partners…but we had to 2 Structure-Level
data realize that it is really difficult to find a partner who could deliver components 3 (Constructing)
which work properly and take the domain-specific internal car systems into
account.” Interviewee R2
Recruiting platform architects for the IT unit “For us, it is crucial that we have the manpower to evaluate the technological 3 Actor-Level (Constructing)
solutions available on our own and to choose the best suitable technological stack
for our use cases […] in this regard, [the IT unit] took a huge step forward and
hired platform architects that bring our endeavor forward.” Interviewee R2
Establishing a closer exchange between security “Initially, our vehicles and our IT landscape had to be seen as disjoint systems. 3 Actor-Level
experts from the IT in-car and IT security They only interacted in rare cases, such as in the case of maintenance work […] (Coordinating)
subunits accordingly, both the vehicles and the IT landscape were separately secured […]
Today, however, the IT security experts and in-car security experts work closely
together, as we are otherwise not able to look at all aspects of security throughout
the chain from the car to our IT landscape and vice versa.” Interviewee N2
Constructing an R&D Hadoop platform for on-road “During on-road tests, our test cars produce more than a gigabyte of data. […] 3 Technology-Level
test data Though we have the technological infrastructure and processes to collect and store (Constructing)
the data, we had to extend our technological landscape to allow analyzing these
high volumes of data.” Interviewee R2
Improving structural digitization support through a “Besides coordinating our digitization efforts, which are mostly analytics topics, 4 Structure-Level
board of digital officers in the production unit they try to avoid redundancies within the production unit. They have to find (Enhancing)
answers to: Which data do we need from which production site? What should our
system landscape look like over a five- or ten-year time horizon?” Interviewee O2
Constructing a production analytics subunit “We as a team are responsible for finding ways to improve our production 4 Structure-Level
processes through analytical methods […] Thus, we first try to access the data (Constructing)
available to us and, in collaboration with the other [production subunits], identify
use cases.” Interviewee O2
Top management support for the analytics efforts in “Once our top management understood the value of analytics, they supported us 4 Structure-Level
the production unit as best as they could. For instance, through additional budgets.” Interviewee O2 (Coordinating)
New collaborating between the production and IT “We started to collaborate heavily with the IT unit as they built an analytics 4 Structure-Level
units platform for [PremiumCar]. The IT unit supports us with the toolset on which we (Integrating)
can work and also takes care of all the licensing issues you face when you use
enterprise solutions” Interviewee O2
New collaborating of production unit with legal “When we optimize and digitize our production processes, the work our employees 4 Structure-Level
offices to assert privacy aspects are doing in our production assemblies becomes transparent […] of course, our (Integrating)
work council is concerned with the privacy of our employees, so we closely discuss
this aspect with our colleagues from legal.” Interviewee M2
Offering a data-awareness program for management- “We offer data awareness workshops on the management level. But you always 4 Actor-Level (Enhancing)
level employees have to consider that you will not reach all employees through the management or
top-down approach. Thus, you have to complement it with bottom-up
approaches.” Interviewee O2
(continued on next page)
18
Table C1 (continued)

Recruiting data scientists and data analysts for the “We slowly started to recruit our own personal for analytics topics; in particular, 4 Actor-Level (Constructing)
production analytics subunit data scientist and analysts like myself […] our production processes are quite
specific and require at some point more than analytical understanding.”
Interviewee P2
Establishing a social community for BDA in the “Our analytics efforts are supported by regular workshops within our production 4 Actor-Level (Integrating)
production unit unit which are organized by a colleague from another subunit. He actively pushes
our social community in order to ensure collaboration and a common
understanding in regard to our analytics efforts at the production unit.”
Interviewee P2
Connecting production data to the central analytics “Even though we implement a dedicated data lake for production, our main goal is 4 Technology-Level
platform to use our central analytics platform.” Interviewee O2 (Integrating)
Digitizing interfaces to production facilities “If we are not able to connect to our machines using protocols such as OPCUA, we 4 Technology-Level
enable our machines to connect to our LAN- or WLAN-infrastructure and try to (Integrating)
access the data using different approaches […] for instance, screen scraping.”
Interviewee M2
Implementing a data lake for the production unit “Our production unit is dominated by the continuous improvement process 4 Technology-Level
mindset. Thus, we try to support as well as we can the process of connecting data (Coordinating)
from all of our production facilities to our technological infrastructure and making
the data analyzable. […] We are producing across the globe, which makes it even
more challenging.” Interviewee B2
References [20] S. Debortoli, O. Müller, J.vom Brocke, Comparing business intelligence and big
data skills, Bus. Inf. Syst. Eng. 6 (5) (2014) 289–300.
[21] G. Slinger, R. Morrison, Will organization design Be affected by big data? J. Organ.
[1] T.H. Davenport, P. Barth, R. Bean, How ‘Big data’ is different, MIT Sloan Manage. Des. 3 (3) (2014) 17–26.
Rev. 54 (1) (2012) 43–46. [22] I.A.T. Hashem, I. Yaqoob, N.B. Anuar, S. Mokhtar, A. Gani, S. Ullah Khan, The rise
[2] T.H. Davenport, S. Kudyba, Designing and developing analytics-based data pro- of ‘big data’ on cloud computing, Inf. Syst. 47 (2015) 98–115.
ducts, MIT Sloan Manage. Rev. 58 (1) (2016) 83–89. [23] S. Akter, S.F. Wamba, A. Gunasekaran, R. Dubey, S.J. Childe, How to improve firm
[3] C. Dremel, M. Herterich, J. Wulf, J.-C. Waizmann, W. Brenner, How AUDI AG performance using big data analytics capability and business strategy alignment?
established big data analytics in its digital transformation, MIS Q. Exec. 16 (2) Int. J. Prod. Econ. 182 (2016) 113–131.
(2017) 81–100. [24] M. Ghasemaghaei, K. Hassanein, O. Turel, Increasing firm agility through the use
[4] A. McAfee, E. Brynjolfsson, “Big data. The management revolution.” The man- of data analytics: the role of fit, Decis. Support Syst. 101 (2017) 95–105.
agement revolution, Harvard Bus Rev. 90 (10) (2012) 61–67. [25] S.F. Wamba, A. Gunasekaran, S. Akter, S.J. Ren, R. Dubey, S.J. Childe, Big data
[5] R.D. Galliers, S. Newell, G. Shanks, H. Topi, Datification and its human, organi- analytics and firm performance: effects of dynamic capabilities, J. Bus. Res. 70
zational and societal effects: the strategic opportunities and challenges of algo- (2017) 356–365.
rithmic decision-making, J. Strateg. Inf. Syst. 26 (3) (2017) 185–190. [26] J. Bughin, M. Chui, J. Manyika, Clouds, big data, and smart assets: ten tech-en-
[6] W.A. Günther, M.H. Rezazade Mehrizi, M. Huysman, F. Feldberg, Debating big abled business trends to watch, McKinsey Q. 56 (1) (2010) 75–86.
data: a literature review on realizing value from big data, J. Strateg. Inf. Syst. 26 [27] J. Manyika, M. Chui, B. Brown, J. Bughin, R. Dobbs, C. Roxburgh, Big Data: the
(3) (2017) 191–209. Next Frontier for Innovation, Competition, and Productivity, McKinsey Global
[7] C. Loebbecke, A. Picot, Reflections on societal and business model transformation Insitute, 2011.
arising from digitization and big data analytics: a research agenda, J. Strateg. Inf. [28] B. Baesens, R. Bapna, J.R. Marsden, J. Vanthienen, J.L. Zhao, Transformational
Syst. 24 (3) (2015) 149–157. issues of big data and analytics in networked business, Mis Q. 40 (4) (2016)
[8] J. Wulf, T. Mettler, W. Brenner, Using a digital services capability model to assess 807–818.
readiness for the digital consumer, MIS Q. Exec. 16 (3) (2017). [29] H. Chen, R.H.L. Chiang, V.C. Storey, Business intelligence and analytics: from big
[9] N. Henke, J. Bughin, M. Chui, J. Manyika, T. Saleh, B. Wiseman, G. Sethupathy, data to big impact, Mis Q. 36 (4) (2012) 1165–1188.
The Age of Analytics: Competing in a Data-driven World, (2016) http://www. [30] P. Mikalef, I.O. Pappas, J. Krogstie, M. Giannakos, Big data analytics capabilities: a
mckinsey.com/business-functions/mckinsey-analytics/our-insights/the-age-of- systematic literature review and research agenda, Inf. Syst. E-business Manag.
analytics-competing-in-a-data-driven-world. (2017) 1–32.
[10] A. Luckow, K. Kennedy, F. Manhardt, E. Djerekarov, B. Vorster, A. Apon, [31] P. Tambe, Big data investment, skills, and firm value, Manage. Sci. 60 (6) (2014)
Automotive big data: applications, workloads and infrastructures, 2015 IEEE 1452–1469.
International Conference on Big Data (2015) 1201–1210. [32] S.F. Wamba, S. Akter, A. Edwards, G. Chopin, D. Gnanzou, How ‘big data’ can
[11] D. Opresnik, M. Taisch, The value of Big Data in servitization, Int. J. Prod. Econ. make big impact: findings from a systematic review and a longitudinal case study,
165 (2015) 174–184. Int. J. Prod. Econ. 165 (2015) 234–246.
[12] C. Lehrer, A. Wieneke, J. Vom Brocke, R. Jung, S. Seidel, How big data analytics [33] P.M. Leonardi, Materiality, sociomateriality, and socio-technical systems: what Do
enables service innovation, forthcoming, J. Strateg. Inf. Syst. (2018). these terms mean? How are they different? Do We need them? in: P.M. Leonardi,
[13] R. Sharma, S. Mithas, A. Kankanhalli, Transforming decision-making processes: a B.A. Nardi, J. Kallinikos (Eds.), Materiality and Organizing, Oxford University
research agenda for understanding the impact of business analytics on organisa- Press, 2012, pp. 24–48.
tions, Eur. J. Inf. Syst. 23 (4) (2014) 433–441. [34] D. Robey, B. Raymond, C. Anderson, Theorizing information technology as a
[14] G. Troilo, L.M. De Luca, P. Guenzi, Linking data-rich environments with service material artifact in information systems research, in: P.M. Leonardi, B.A. Nardi,
innovation in incumbent firms: a conceptual framework and research propositions, J. Kallinikos (Eds.), Materiality and Organizing, Oxford University Press, 2012, pp.
J. Prod. Innov. Manage. 34 (5) (2017) 617–639. 216–236.
[15] D. Boyd, K. Crawford, Critical questions for big data, Inf. Commun. Soc. 15 (5) [35] D.M. Strong, S.A. Johnson, B. Tulu, J. Trudel, O. Volkoff, L.R. Pelletier, I. Bar-On,
(2012) 662–679. et al., A theory of Organization-EHR affordance actualization, J. Assoc. Inf. Syst.
[16] M. Gupta, J.F. George, Toward the development of a big data analytics capability, 15 (2) (2014) 53–85.
Inf. Manage. Big Data Commerce 53 (8) (2016) 1049–1064. [36] O. Volkoff, D.M. Strong, Affordance theory and how to use it in IS research,
[17] P. Mikalef, M. Boura, G. Lekakos, J. Krogstie, Complementarities Between Routledge Handbooks Online, (2017).
Information Governance and Big Data Analytics Capabilities on Innovation, [37] W.J. Orlikowski, Using technology and constituting structures: a practice Lens for
Proceedings of the 26th European Conference on Information Systems (2018). studying technology in organizations, Organ. Sci. 11 (4) (2000) 404–428.
[18] M.M. Crossan, H.W. Lane, R.E. White, An organizational learning framework: from [38] W.J. Orlikowski, S.V. Scott, 10 sociomateriality: challenging the separation of
intuition to institution, Acad. Manag. Rev. 24 (3) (1999) 522–537. technology, work and organization, Acad. Manag. Ann. 2 (1) (2008) 433–474.
[19] J.G. March, Exploration and exploitation in organizational learning, Organ. Sci. 2 [39] J.J. Gibson, The Ecological Approach to Visual Perception, Routledge, Boston,
(1) (1991) 71–87.
19
1986 Revised. [75] J. Saltz, I. Shamshurin, K. Crowston, Comparing data science project management
[40] A. Chemero, An outline of a theory of affordances, Ecol. Psychol. 15 (2) (2003) methodologies via a controlled experiment, Proceedings of the 50th Hawaii
181–195. International Conference on System Sciences (2017).
[41] T.A. Stoffregen, Affordances as properties of the animal-environment system, Ecol. [76] J. Saltz, I. Shamshurin, C. Connors, Predicting data science sociotechnical ex-
Psychol. 15 (2) (2003) 115–134. ecution challenges by categorizing data science projects, J. Assoc. Inf. Sci.
[42] I. Hutchby, Technologies, texts and affordances, Sociology 35 (2) (2001) 441–456. Technol. 68 (12) (2017) 2720–2728.
[43] A. Majchrzak, M.L. Markus, Technology affordances and constraints in manage- [77] G. Cao, Y. Duan, A path model linking business analytics, data-driven culture, and
ment information systems (MIS), in: E. Kessler (Ed.), Encyclopedia of Management competitive advantage, Proceedings of the 22nd European Conference on
Theory, Sage Publications, 2012Forthcoming. Information Systems (2014).
[44] M.L. Markus, M. Silver, A foundation for the study of IT effects: a new look at [78] M. Ghasemaghaei, S. Ebrahimi, K. Hassanein, Generating valuable insights
DeSanctis and Poole’s concepts of structural features and spirit, J. Assoc. Inf. Syst. through data analytics: a moderating effects model, Proceedings of the 36th
9 (10) (2008). International Conference on Information Systems (2016).
[45] T. Mettler, J. Wulf, Physiolytics at the workplace: affordances and constraints of [79] P. Gölzer, P. Cato, M. Amberg, data processing requirements of industry 4.0 - use
wearables use from an employee’s perspective, Inf. Syst. J. (2018). cases for Big data applications, in Proceedings of the 23rd European Conference on
[46] P.M. Leonardi, When flexible routines meet flexible technologies: affordance, Information Systems (2015).
constraint, and the imbrication of human and material agencies, Mis Q. 35 (1) [80] A. Marton, M. Avital, T.B. Jensen, Reframing open Big data, Proceedings of the
(2011) 147–167. 21st European Conference on Information Systems (2013).
[47] O. Volkoff, D.M. Strong, Critical realism and affordances: theorizing IT-Associated [81] Y. Yoo, It is not about size, J. Inf. Technol. 30 (1) (2015) 63–65.
organizational change processes, Mis Q. 37 (3) (2013) 819–834. [82] C. Dremel, S. Overhage, S. Schlauderer, J. Wulf, Towards capability model for big
[48] R.F. Zammuto, T.L. Griffith, A. Majchrzak, D.J. Dougherty, S. Faraj, Information data analytics, Proceedings Der 13. Internationalen Tagung Wirtschaftsinformatik
technology and the changing fabric of organization, Organ. Sci. 18 (5) (2007) (WI 2017) (2017).
749–762. [83] C. Holsapple, A. Lee-Post, R. Pakath, A unified foundation for business analytics,
[49] I.D. Constantiou, J. Kallinikos, New games, new rules: big data and the changing Decis. Support Syst. 64 (Suppl. C) (2014) 130–141.
context of strategy, J. Inf. Technol. 30 (1) (2015) 44–57. [84] G. George, M.R. Haas, A. Pentland, Big data and management, Acad. Manag. J. 57
[50] T. Davenport, Big Data @ Work: Dispelling the Myths, Uncovering the (2) (2014) 321–326.
Opportunities, Harvard Business Review Press, 2014. [85] D.Q. Chen, D.S. Preston, M. Swink, How the use of big data analytics affects value
[51] M.L. Markus, New games, new rules, new scoreboards: the potential consequences creation in supply chain management, J. Manag. Inf. Syst. 32 (4) (2015) 4–39.
of big data, J. Inf. Technol. 30 (1) (2015) 58–59. [86] M. Kowalczyk, P. Buxmann, An ambidextrous perspective on business intelligence
[52] S. Woerner, B.H. Wixom, Big data: extending the business strategy toolbox, J. Inf. and analytics support in decision processes: insights from a multiple case study,
Technol. 30 (1) (2015) 60–62. Decis. Support Syst. 80 (2015) 1–13.
[53] R. Bärenfänger, B. Otto, H. Österle, Business value of in-memory technology - [87] A. Abbasi, S. Sarker, R.H. Chiang, Big data research in information systems: to-
multiple-case study insights, Ind. Manag. Data Syst. 114 (9) (2014) 1396–1414. ward an inclusive research agenda, J. Assoc. Inf. Syst. 17 (2) (2016) 3.
[54] H.-M. Chen, R. Schütz, R. Kazman, F. Matthes, How lufthansa capitalized on big [88] K.M. Eisenhardt, Building theories from case study research, Acad. Manag. Rev. 14
data for business model renovation, MIS Quarterly Executive 16 (1) (2017) 19–34. (4) (1989) 532–550.
[55] K. Tiefenbacher, S. Olbrich, Increasing the level of customer orientation-a Big data [89] M.D. Myers, Qualitative research in information systems, Mis Q. 21 (2) (1997)
case study from insurance industry, Proceedings of the 23rd European Conference 241–242.
on Information Systems (2015). [90] S.X. Sarker, Tanya, Qualitative studies in information systems: a critical review
[56] M. Kowalczyk, P. Buxmann, Big data and information processing in organizational and some guiding principles, Mis Q. 37 (4) (2013) iii–xviii.
decision processes: a multiple case study, Bus. Inf. Syst. Eng. 6 (5) (2014) [91] R.K. Yin, Case Study Research: Design and Methods, 5th edition, Sage
267–278. Publications, Thousand Oaks, CA, 2008.
[57] S. Newell, M. Marabelli, Strategic opportunities (and challenges) of algorithmic [92] Deloitte, Big Data and Analytics in the Automotive Industry. Automotive Analytics
decision-making: A call for action on the long-term societal effects of ‘datification, Thought Piece, (2015) https://www2.deloitte.com/content/dam/Deloitte/uk/
J. Strateg. Inf. Syst. 24 (1) (2015) 3–14. Documents/manufacturing/deloitte-uk-automotive-analytics.pdf.
[58] Y. Duan, G. Cao, Understanding the impact of business analytics on innovation, [93] SAS Institute Inc, The Connected Vehicle. Big Data, Big Opportunities, (2015)
Proceedings of the 23rd European Conference on Information Systems (2015). https://www.sas.com/content/dam/SAS/en_us/doc/whitepaper1/connected-
[59] D. Robey, C. Anderson, B. Raymond, Information technology, materiality, and vehicle-107832.pdf.
organizational change: a professional Odyssey, J. Assoc. Inf. Syst. Atlanta 14 (7) [94] M.D. Myers, M. Newman, The qualitative interview in IS research: examining the
(2013) 379–398. craft, Inf. Organ. 17 (1) (2007) 2–26.
[60] B. Bygstad, B.E. Munkvold, O. Volkoff, Identifying generative mechanisms through [95] U. Schultze, M. Avital, Designing interviews to generate rich data for information
affordances: a framework for critical realist data analysis, J. Inf. Technol. 31 (1) systems research, Inf. Organ. 21 (1) (2011) 1–16.
(2016) 83–96. [96] A. Gissler, O. Clemens, C. Knackfuß, F. Kupferschmidt, Are You Ready for Pole
[61] Y. Tim, S.L. Pan, S. Bahri, A. Fauzi, Digitally enabled affordances for community- Position? Driving Digitization in the Auto Industry, (2016) https://www.
driven environmental movement in rural Malaysia, Inf. Syst. J. (2017). accenture.com/us-en/insight-pole-position-driving-digitization-auto-industry.
[62] B. Bygstad, P.A. Nielsen, B.E. Munkvold, Four integration patterns: a socio-tech- [97] M. Mocker, N.O. Fonstad, How AUDI AG is Driving Toward the Sharing Economy,
nical approach to integration in IS development projects, Inf. Syst. J. 20 (1) (2010) MIS Q. Exec. 16 (4) (2017) 279–293.
53–80. [98] McKinsey & Company, Connected Car, Automotive Value Chain Unbound, (2014)
[63] M. Durkin, G. Mulholland, A. McCartan, A socio-technical perspective on social http://www.mckinsey.com/industries/automotive-and-assembly/our-insights/
media adoption: a case from retail banking, Int. J. Bank Mark. 33 (7) (2015) connected-car-automotive-value-chain-unbound.
944–962. [99] C. Dremel, J. Wulf, M. Annegret, W. Brenner, Understanding the value and or-
[64] K. Lyytinen, M. Newman, Explaining information systems change: a punctuated ganizational implications of Big data analytics? The case of AUDI AG, Proceedings
socio-technical change model, Eur. J. Inf. Syst. 17 (6) (2008) 589–613. of the 37th International Conference on Information Systems (2017).
[65] S. Seidel, J.C. Recker, J. Vom Brocke, Sensemaking and sustainable practicing: [100] J. Corbin, A. Strauss, Grounded Theory Research: Procedures, Canons, and
Functional affordances of information systems in green transformations, Manage. Evaluative Criteria, Qual. Sociol. 13 (1) (1990) 3–21.
Inf. Syst. Q. 37 (4) (2013) 1275–1299. [101] A.C. Edmondson, The local and variegated nature of learning in organizations: a
[66] R.P. Bostrom, J.S. Heinen, MIS problems and failures: a socio-technical perspec- group-level perspective, Organ. Sci. 13 (2) (2002) 128–146.
tive. Part I: the causes, Mis Q. 1 (3) (1977) 17–32. [102] D. Lavie, Capability reconfiguration: an analysis of incumbent responses to tech-
[67] H.J. Leavitt, Applied organizational change in industry: structural, technological nological change, Acad. Manag. Rev. 31 (1) (2006) 153–174.
and humanistic approaches, Handbook of Organizations, Rand McNally & Co., [103] A.S. Miner, S.J. Mezias, Ugly duckling no more: pasts and futures of organizational
Chicago, 1965, pp. 1144–1170. learning research, Organ. Sci. 7 (1) (1996) 88–99.
[68] P.G. Keen, Information systems and organizational change, Commun. ACM 24 (1) [104] D.H. Kim, The Link Between Individual and Organizational Learning, MIT Sloan
(1981) 24–33. Manage. Rev. 35 (1) (1993).
[69] E.L. Trist, The evolution of socio-technical systems: a conceptual framework and [105] P. Mikalef, J. Krogstie, R. Wetering, I. van de Pappas, M. Giannakos, Information
an action research program, Ontario Ministry of Labour, Ontario Quality of governance in the Big data era: aligning organizational capabilities, Proceedings of
Working Life Centre, Toronto, 1981. the 51th Hawaii International Conference on System Sciences (2018).
[70] P. Mikalef, V. Framnes, F. Danielsen, J. Krogstie, D.H. Olsen, Big data analytics [106] P.P. Tallon, R.V. Ramirez, J.E. Short, The information artifact in IT governance:
capability: antecedents and business value, Proceedings of the 21st Pacific Asia toward a theory of information governance, J. Manag. Inf. Syst. 30 (3) (2013)
Conference on Information Systems (2017). 141–178.
[71] D. Kiron, P.K. Prentice, R.B. Ferguson, Innovating with analytics, MIT Sloan [107] Y.E. Chan, B.H. Reich, IT alignment: what have we learned? J. Inf. Technol. 22 (4)
Manage. Rev. 54 (1) (2012). (2007) 297–315.
[72] S. Miranda, I. Kim, D.D. Wang, Whose talk is walked? IT decentralizability, vendor [108] J.E. Gerow, J.B. Thatcher, V. Grover, Six types of IT-business strategic alignment:
versus adopter discourse, and the diffusion of social media versus Big data, an investigation of the constructs and their measurement, Eur. J. Inf. Syst. 24 (5)
Proceedings of the 36th International Conference on Information Systems (2015). (2015) 465–491.
[73] M.E. Porter, J.E. Heppelmann, How smart, connected products are transforming [109] O. Müller, M. Fay, J. vom Brocke, The effect of big data and analytics on firm
companies, Harv. Bus. Rev. 93 (10) (2015) 96–114. performance: an econometric analysis considering industry characteristics, J.
[74] T.H. Davenport, J.G. Harris, Competing on Analytics: the New Science of Winning, Manag. Inf. Syst. 35 (2) (2018) 488–509.
Harvard Business Press, 2007. [110] S. Gregor, The nature of theory in information systems, Manage. Inf. Syst. Q. 30
20
(3) (2006) 611–641. digital business innovation and business potentials arising from physical products aug-
[111] R. Agarwal, V. Dhar, Editorial —big data, data science, and analytics, Inf. Syst. mented with digital technology (e.g., the Internet of Things). In particular, he investigates
Res. 25 (3) (2014) 443–448. the impact of digitized products on value co-creation in smart service systems and
[112] A. Bhimani, Exploring big data’s strategic consequences, J. Inf. Technol. 30 (1) business/service innovation.
(2015) 66–69.
[113] A. Chatfield, C. Reddick, W. Al-Zubaidi, Capability challenges in transforming Jochen Wulf ([email protected]) is lecturer and fellow of the International
government through Open and Big data: tales of Two cities, Proceedings of the Postdoctoral Fellowship program at the University of St.Gallen. Prior to this, he was as-
36th International Conference on Information Systems (2015). sistant professor at the Institute of Information Management at University of St.Gallen
[114] P. Goes, Editor’s comments: big data and is research, Manage. Inf. Syst. Q. 38 (3) (IWI-HSG), Switzerland. His research focuses on socio-technical aspects of applying large-
(2014) iii–viii. scale data processing systems, consumer-centricity and IT service management. Jochen
has authored more than 50 scientific publications. His research has been published in
Christian Dremel ([email protected]) is a research associate at the Institute of journals such as Information Systems Journal (ISJ), MIS Quarterly Executive (MISQE),
Information Management (IWI-HSG) at the University of St.Gallen. He holds an M.Sc. Business & Information Systems Engineering (BISE) and Electronic Markets. He further pre-
from the University of Bamberg. His research focuses on the successful adoption and sented at leading conferences such as International Conference on Information Systems
assimilation of big data analytics. In particular, he investigates the organizational (ICIS) and European Conference on Information Systems (ECIS).
transformations (e.g., organizational structures, governance mechanisms and capabilities)
required to profit from big data analytics through providing analytics-as-a-service. His Jan vom Brocke ([email protected]) is Professor of Information Systems, the Hilti
research has been published in journals such as the MIS Quarterly Executives (MISQE), Endowed Chair of Business Process Management and Director of the Institute of
Electronic Markets, and Journal of Information Technology Teaching Cases, and presented at Information Systems at the University of Liechtenstein. His research focuses on business
conferences such as the International Conference on Information Systems (ICIS). process management and related aspects of digital innovation and transformation. He has
published, among others, in MIS Quarterly, Journal of Management Information Systems,
Matthias M. Herterich ([email protected]) is a research associate at the Journal of Information Technology, European Journal of Information Systems, Information
Institute of Information Management (IWI-HSG) at the University of St.Gallen. Currently Systems Journal, Communications of the ACM, and MIT Sloan Management Review. He has
he is a visiting academic at the Center for Design Research (CDR) at the Mechanical held various editorial roles and leadership positions in Information Systems research and
Engineering department of Stanford University. He holds an M.Sc. from the University of education.
Mannheim, partially conducted at the University of Florida. His research focuses on
21

Big Data Case Study Rp

Uploaded by

Copyright:

Available Formats

Big Data Case Study Rp

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Big Data Case Study Rp

Uploaded by

Copyright:

Available Formats

Information & Management 57 (2020) 103121

Contents lists available at ScienceDirect

Information & Management

Actualizing big data analytics affordances: A revelatory case study T

ARTICLE INFO ABSTRACT

1. Introduction learning in the realization of BDA value [16,17], it is largely unknown

Fig. 1. Affordance Actualization Model (adapted from [33,66,67]).

In order to obtain in-depth qualitative data, exploratory interviews 4. Results

4.1. Affordances of big data analytics

4.2. Actualization of affordances self-actualization. Second, PremiumCar’s company headquarters are

Appendix A. Interview Protocol

• 1. Introduction and General Information

Round 1. Exploring Affordances

Appendix B. Organizational Chart of PremiumCar

Fig. B1. Organizational Chart of PremiumCar (Simplified Excerpt).

Appendix C. Full List of Organizational Actions

Organizational Actions Empirical Evidence Affordances STS Dimension

Organizational Actions Empirical Evidence Affordances STS Dimension

You might also like