ArticlePDF Available

Digitization and Phase Transitions in Platform Organizing Logics: Evidence from the Process Automation Industry


Abstract and Figures

This paper draws on complex adaptive systems (CAS) theory to explore the transformation of an analog automation product platform as it was infused with extensive and deepening digital capacities over a 40-year period. Our case demonstrates how the deepening digitization of components and functions drives complexity by connecting the platform to multiple social and technical settings and producing new interactions and information exchanges. The increased connectivity and dynamism invited unexpected and significant architectural and organizational shifts that moved the platform toward an ecosystem-centered organizing logic. CAS theory and its notion of constrained generating procedures (CGPs) are used to analyze how new connections and interactions produced a multilevel and nonlinear change in the platform organization. We offer two main contributions. First, we provide a novel empirical analysis of how product platform digitization leads to phase transitions and show the mediating role of three mechanisms in this process treated as CGPs: interaction rules, design control, and stimuli-response variety. Second, we demonstrate the multilevel and recursive nature of digitally driven growth in physical product platforms.
Content may be subject to copyright.
Johan Sandberg and Jonny Holmström
Swedish Center for Digital Innovation, Department of Informatics, Umeå University,
87 Umeå, SWEDEN {} {}
Kalle Lyytinen
Weatherhead School of Management, Case Western Reserve University,
Cleveland, OH 44106 U.S.A. {}
This paper draws on complex adaptive systems (CAS) theory to explore the transformation of an analog auto-
mation product platform as it was infused with extensive and deepening digital capacities over a 40-year
period. Our case demonstrates how the deepening digitization of components and functions drives complexity
by connecting the platform to multiple social and technical settings and producing new interactions and
information exchanges. The increased connectivity and dynamism invited unexpected and significant archi-
tectural and organizational shifts that moved the platform toward an ecosystem-centered organizing logic. CAS
theory and its notion of constrained generating procedures (CGPs) are used to analyze how new connections
and interactions produced a multilevel and nonlinear change in the platform organization. We offer two main
contributions. First, we provide a novel empirical analysis of how product platform digitization leads to phase
transitions and show the mediating role of three mechanisms in this process treated as CGPs: interaction
rules, design control, and stimuli-response variety. Second, we demonstrate the multilevel and recursive nature
of digitally driven growth in physical product platforms.
Keywords: Complexity, platform evolution, phase transition, platform change, digital transformation, digital
innovation, product platform, platform ecosystem, digital control systems, internet of things
Introduction 1
The rapid pervasion of digital technology into physical prod-
ucts has become a prominent driver of complexity, pushing
incumbent firms toward platforms and ecosystems (Yoo et al.
2012). Past studies suggest that increased digitization spurs
change in a firm’s organizing logic by instilling new pro-
perties into product platforms (Gawer 2014; Lee and Berente
2012; Svahn et al. 2017). “Organizing logic” refers to how
the firm designs, manufactures, and distributes its products
and derivative services; the reasons it offers product functions
or services; and the rationale for related organizational
arrangements (Sambamurthy and Zmud 2000; Yoo et al.
2010). However, how infusion with digital components
generates emergent product platform properties that drive
complexity and transitions in the firm’s organizing logics has
received little attention. To fill this gap, we explore the
dynamic of continued embedding of digital capacities and
how it changes the scale and scope of a product platform’s
functions, resulting in punctuations in the firm’s organizing
1The accepting senior editors for this paper were Bill McKelvey, Hüseyin
Tanriverdi, and Youngjin Yoo.
©2020. The Authors. Published by the Management Information Systems
Research Center at the University of Minnesota. This is an open access
article under the terms of the Creative Commons Attribution CC BY License,
which permits use, distribution, and reproduction in any medium, provided
the original work is properly cited.
DOI: 10.25300/MISQ/2020/14520 MIS Quarterly Vol. 44 No. 1, pp. 129-153/March 2020 129
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
logic. We specifically address the following question: How
is increased digitization implicated in product platform transi-
tions and related organizing logics?
Drawing on complex adaptive systems (CAS) theory, we
hypothesize that digitization induces platform transitions via
an emergent change in its architecture (Kretschmer and
Claussen 2016). This dynamic differs starkly from incre-
mental linear change in traditional analog products, related
ecosystems, and associated organizing logics (Cusumano et
al. 1992). CAS theory provides a robust framework for our
analysis because it addresses endogenous change in product
systems that result from emergent interactions within them
and in their environment. We build on Holland’s (1998) con-
cept of constrained generating procedures (CGPs), that is,
networks of relatively simple rules or mechanisms that inter-
actively generate emergence in complex systems and produce
transitions from one systemic state to another. We parti-
cularly examine how digitization changes CGPs linked to
product platforms and triggers new system-level organizing
logics. For this, we apply Hughes’s (1983) concept of the
reverse salient: ongoing digitization of product platforms’
components and/or relationships that block the product
platform’s growth in scale or scope. Through digitization-
driven resolution of a series of reverse salients, cumulative
digital capacities are introduced into product platforms, and
they act as “strange attractors” that continue to drive platform
reorganization and expansion.
The conceptual scaffolding is applied in an exploratory longi-
tudinal case study on product platform change in a leading
process automation manufacturer (ABB) between 1983 and
2016. We trace cumulative effects of successive waves of
digitization on ABB’s automation platform by charting
changes in its architectural designs and ensuing shifts in the
firm’s organizing logic. During this process, ABB trans-
formed from a solely physical product manufacturer to a
producer of versatile hybrid physical–digital systems and
engaged in an increasingly rich range of interactions in a
growing ecosystem with its associated value logics.
Digitization and the Dynamics
of Product Platforms
Platform Types and Change
Platform research is now voluminous, diverse, and cross-
disciplinary and comes with a wide array of concepts
(McIntyre and Srinivasan 2017). This study addresses transi-
tions in product platforms and their temporal and logical
dependencies. We build on Gawer’s (2009, p. 59) suggestion
that platform types represent “stages of evolution of platform
development.” Although some studies offer valuable insights
into such changes and their drivers (e.g., Lee and Berente
2012; Svahn et al. 2017), they do not account for how cumu-
lative technological and organizational change over extended
periods manifests in platform change or how such evolution
triggers shifts in a firm’s organizing logics (Gawer and
Cusumano 2014). Scholars have generally found that as firms
introduce digital elements, they open up their products or
services and, by doing so, they move from an internal
production-oriented logic to an external supply-chain
innovation logic (Thomas et al. 2014). However, how digital
technologies spur such change has received limited attention.
We focus specifically on conditions under which expansive
digitization transforms product platforms and related
organizing logics. This calls for a platform taxonomy that
distinguishes between platform types and associated
organizing logics (Thomas et al. 2014). We draw on Gawer’s
(2014) classification, which recognizes central features of a
spectrum of platform types from a physical product platform
to a digital multisided platform ecosystem, with various inter-
mediate hybrid configurations (Table 1).
Product platform literature originates from studies of engi-
neering and design of industrial products. Product platforms
comprise families of stable shared assets that enable firms to
produce customized derivatives of products for specific
markets and niches more quickly and cheaply (Meyer and
Lehnerd 1997; Robertson and Ulrich 1998; Wheelwright and
Clark 1992). Such assets include product components, pro-
cesses, market knowledge, people, and their relationships:
“the technical architecture of the product or service—as well
as the structure of the underlying capabilities” (Thomas et al.
2014, p. 203). Platform ecosystems literature, which mainly
originates in economics and strategy studies, defines a plat-
form as “a modular structure that consists of tangible and
intangible components (resources) and facilitates the inter-
action of actors and resources (or resource bundles)” (Lusch
and Nambisan 2015, p. 162). This stream focuses on the
strategic and economic effect of controlling (shared) resources
that enable new value-creating interactions among partici-
pants founded on an evolving technological system (Constan-
tinides et al. 2018; Karhu et al. 2018; Parker and Alstyne
2018). Research in this area has addressed the design of
architectures (technological systems) and the economics of
interactions facilitated by associated governance (e.g., gover-
nance rules and boundary resources). Per Tiwana et al.
(2010), architecture partitions a platform “into a relatively
stable platform and a complementary set of modules that are
encouraged to vary, and the design rules binding on both” (p.
677), and governance refers to “who makes what decisions
about a platform” (p. 679).
130 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Table 1. Two Extremes of a Spectrum of Technological Platform Types
Characteristic Internal Product Platform Platform Ecosystem
Description Platform as the stable center of a
product family leading to derivatives.
Platform as a system architecture that supports a collection
of complementary assets for extended value creation and
extraction across multiple actors.
Key constructs Product family; architecture; economies
of scope, modularity; commonality.
Network externalities; complementarities, innovation;
standards; modularity.
Level of analysis The firm, and (for a semi-open product
platform) supply chain participants.
The platform and its evolving ecosystem.
Agents within the firm or external agents
contracted for a specific purpose.
Unbounded and dynamic set of agents, individuals, or firms
that interact for value creation in various, shifting roles.
Nature of
Fixed and semi-closed (not disclosed
externally), one-to-one mapping.
Dynamic and open or semi-open, many-to-many mapping.
Within the firm, supply chain, or selected
Within the ecosystem constrained by IP, governance, and
build-up of boundary resources.
Integration of key functional elements.
Early design decisions ripple through the
development (i.e., early choices remain
strongly inscribed in later designs).
Bundling of core resources facilitates interactions with and
integration of foreseen and unforeseen agents/resources.
Emergent design with frequent adaptations as combinatorial
options arise and are discovered.
Governance and
The platform owner retains decision
rights on what and how modules do
what they do, and interfaces define
The platform owner defines boundaries and openness of
interfaces, what and how modules do what they do through
varying degrees of output, process and input control.
Agents are affected by rules but can engage in autonomous
decision making and action. Significant bidirectional control.
Scale and scope economies for internal
Platform investments shared across
derivatives. Internal value transfer
Scale and scope economies in the value system.
Complementarities provided by open innovation by
unforeseen actors (open source, crowd).
Multisided markets through shared platform co-create and
extract value.
Meyer and Lehnerd 1997; Robertson
and Ulrich 1998; Simpson 2004;
Wheelwright and Clark 1992.
Bresnahan and Greenstein 1999; Eaton et al. 2015; Lusch
and Nambisan 2015; Parker et al. 2017; Tiwana et al. 2010.
Gawer (2014, p. 1245) regards all platforms as “evolving
organizations or meta-organizations that: (1) federate and
coordinate constitutive agents who can innovate and compete;
(2) create value by generating and harnessing economies of
scope in supply or/and in demand; and (3) entail a techno-
logical architecture that is modular and composed of a core
and a periphery.” Governance and architectural choices
enable and constrain distinct sets of interactions among agents
that shape the variety and intensity of a firm’s interactions
with its environment. By affording varying levels of platform
openness and new types of information exchanges, digitiza-
tion modifies participant agents’ behaviors, interactions, and
scope, thereby transforming (gradually or radically) related
organizing logics. Options created by digitization typically
shape the diversity and volume of participating agents and the
coordination and control of their actions. For (internal)
product platforms, agents are primarily located within the
platform-owning firm, and information exchanges primarily
occur within the firm or its immediate supply chain, with
clearly discernible constraints, that is, interactions are rooted
in a managerial hierarchy and stable contractual relations. In
a platform ecosystem, diverse agents such as competitors,
complementors, and users interact through a platform open to
a wide range of exchanges and interactions (see Karhu et al.
2018; Parker and Alstyne 2018). Such increased variation
also creates new options for configuring value creation and
appropriation (Gawer and Cusumano 2014). Economies of
scale and scope change as product platforms move toward
new ways of economizing transactions. Architectural charac-
teristics such as the overall scope of platform functionality
and openness of its interfaces dictate the range of innovation
and production activities and what other agents are allowed or
incentivized to do (Thomas et al. 2014).
MIS Quarterly Vol. 44 No. 1/March 2020 131
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Digitization and Platform Change
Digitization has been recognized as a potent driver toward
platform ecosystems (Parker et al. 2017; Yoo et al. 2012). It
subjects product architectures to new design rules afforded by
novel features of digital technology including reprogram-
mability, data homogenization, decoupling, and distributed-
ness (Lyytinen et al. 2016). Reprogrammability enables
continued, fluid expansion of a product’s functional scope
through programmable von Neumann architecture as long as
new instructions conform to established rules and meanings
of data (Kallinikos et al. 2013; Yoo et al. 2010). New capa-
bilities can be added after initial product design, and new
combinations of functions can be created through integration
(e.g., by gateways and APIs). Data in digital products are
homogenized in that information can be stored, transmitted,
processed, and displayed on multiple devices and networks
and integrated in infinite ways. Data from various sources
can be processed by “general purpose” resources and recom-
bined in novel ways when appropriate instructions are
invented. Functionality is decoupled from material bearers
due to reprogrammability (between devices and services) and
data homogenization (separation between devices/networks
and content). Finally, data and functions can be distributed
across contexts as “transient assemblies of functions, infor-
mation items, or components spread over information infra-
structures” (Kallinikos et al. 2013, p. 360). These new design
rules result in increasingly fluid product boundaries, increased
heterogeneity of use, and expansion of architectural layers
with multiple design hierarchies, product-agnostic com-
ponents, and general standards promoting wider sharing of
information (Lyytinen et al. 2016; Yoo et al. 2010). Although
layered modular architecture and digital design rules
putatively instigate a shift toward platform ecosystems
(Constantinides et al. 2018), we know little about how
endogenous technological change drives platform transitions
or how sequences of novel phenomena happen when digital
capacities are added to product platforms. We also know
little about the factors that make such platform transitions
likely or necessary.
Platform Change as Change of
Complex Adaptive System
Ultimately, product platforms are complicated technical
systems consisting of numerous product components and their
relationships (Simpson 2004). They are complicated because
interactions between components can and must be determined
and predicted during product design irrespective of the
volume of such components and interactions. A complex
system, such as the production organization surrounding a
product platform, in contrast, is “made up of a large number
of parts that interact in a non-simple way” (Simon 1962, p.
468) and exhibits emergence: “the arising of novel and
coherent structures during the process of self-organization”
(Goldstein 1999, p. 49). In product platforms, emergence
occurs in response to increases in complexity driven by digiti-
zation and its design rules, which organically and recursively
shape how a product platform interacts with its social, market,
and institutional environment. The complexity is manifested
in changes in product platform features, enabled interactions,
and related organizing logic (i.e., ways the product organi-
zation interacts internally and externally and organizes its
operations). Overall, digitization triggers increases in variety
and intensity in platform-mediated interactions, which ulti-
mately result in a transition to a new organizing logic. Hence,
we use CAS theory as an analytical lens to examine how the
introduction of digital capacities and design rules create
opportunities or necessities for transitions in organizing logics
of product organizations.
CAS theory generally seeks to explain how importing energy
and information into open systems generates “dissipative
structures” via feedback loops, by which the system acquires
emergent features (Anderson 1999; Nicolis and Prigogine
1989). In our setting, these new resources are digital capa-
cities and associated new information, which result in new
relationships and interactions. A consequent change in the
product platform is viewed as an emergence of an alternative
social order resulting from various feedback-based informa-
tion and resource exchanges within a product platform and
between the platform and its environment. Such change is
called a phase transition when it involves a transformative
shift in the company’s organizing logic following shifts in the
composition of its product platform.
Interactions among various agents (including design and use
processes, related professions, industrial groups, and firms)
jointly produce platform change and the emergence of asso-
ciated order (Benbya and McKelvey 2006). According to
CAS theory, these emergence-generating interactions are
guided by networks of mechanisms expressed in CGPs, which
operate in hierarchies where a basic CGP serves as a sub-
mechanism in a more complex one (Holland 1998). Changes
in CGPs produce new behavioral patterns within and between
agents that subsequently manifest in transitions in system-
level behaviors, producing an alternative organizing logic.
Thus, the generation, adoption, and diffusion of new knowl-
edge and learning through new interactions enabled by GCPs
trigger changes in product platforms (Benbya and McKelvey
2006). The process is influenced by types and levels of con-
nectivity among agents (Morel and Ramanujam 1999) in four
dimensions: diversity, adaptiveness, connectedness, and
mutual dependencies (Page 2010). Diversity refers to degrees
of variations among entities, adaptiveness to the respon-
132 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
siveness of agents’ schemata to new information from the
environment when diversity grows, connectedness to the
amount and level of couplings that govern information flows
among agents, and mutual dependencies to the level of output
interdependency between agents.
In product platforms, architectural and governance structures
and knowledge bases largely determine these four features.
Digitization creates new components, knowledge, and inter-
actions, many from or between new external agents (Lee and
Berente 2012; Yoo et al. 2010). Overall, digitization
increases variety creation and deviation in product platforms
through positive and negative feedback loops manifested,
enabled by changing CGPs (Anderson 1999). As digitization
increases, it significantly affects each of the features. Repro-
grammability and incompleteness raise diversity and needs for
adaptiveness and new knowledge. Decoupling and data
homogenization permit new relationships between product
components by reducing interdependencies and increasing
connectedness, diversity, and adaptiveness.
Generally, as product organizations add digital capacities and
launch new CGPs, they will oscillate between three regions of
change: order, emergent complexity, and edge of discon-
tinuity (Tanriverdi and Lim 2017). Order is treated as a
closed, static system and is not considered here any further.
In the region of emergent complexity, a change in CGPs
produces new interactions with emergent and novel system-
level phenomena, but the system remains focused as its core
structures remain stable, while new schemata, connections,
and potential for self-organization emerge gradually, driven
by organic interactions and coevolution with amplifying
change (Tanriverdi and Lim 2017). Performance in such
states starts to “dance,” that is, the product design and its
environment can change unpredictably, reflected in shifts in
agents’ schemata and growing diversity (Tanriverdi et al.
2010). Our focus is on the edge of discontinuity, where the
system faces a crisis manifested in a growing number of
failures or inadequacies (Page 2010). These curb possibilities
for functional progression of the product platform and its
adaptation by blocking change in critical technological com-
ponents (e.g., the introduction of a new electric car engine
with a different type of digital control). In technology studies,
triggering conditions for strange attractors (situations that can
shift the whole system) have been called reverse salients,
poorly performing subsystems that lag behind the advancing
product performance frontier (Hughes 1983). For partici-
pating agents, a reverse salient is a “complex situation in
which individuals, groups, material forces, historical influ-
ence, and other factors have idiosyncratic, causal roles, and in
which accidents, as well as trends, play a part” (Hughes 1983,
p. 79). Such situations prevent the fulfillment of a platform’s
evolutionary potential.
A reverse salient triggers a search for increase in diversity of
product components and their relationships (Rosenberg 1969),
which creates new “state spaces” for design. Removal of a
reverse salient pushes a product and related organization at an
edge of discontinuity toward novel explorations and creation
of new CGPs. This phase comes with higher diversity, adap-
tiveness, connectedness, and changes in interdependencies
that destabilize the product organization and its environment.
The system may progress to a tipping point (Lamberson and
Page 2012), where it transitions to a qualitatively different
order involving deep changes in conventions, shared assump-
tions, and product functions and their logic. Emergent change
follows removal of the reverse salient through deviation-
amplifying feedback loops. These CAS constructs provide an
initial scaffolding (Table 2) for examining phase transitions
associated with product platforms as a function of increased
digitization resulting in new organizing logics.
Research Site and Methods
Case Selection and Methods
We conducted a longitudinal case study (Eisenhardt 1989)
covering 35 years of digitization of ABB’s product platform.
This is a platform for automating production processes
involving heavy physical machinery across various industries
(e.g., pulp, mining, and energy). Automation product plat-
forms offer ideal settings for studying long-term effects of
digitization because of their capital-intensive nature, longe-
vity (often 40+ years), and central role in contemporary
manufacturing. In the past four decades, extensive digiti-
zation has radically automated and changed the control of
industrial processes. Automation product platforms are
designed as general-purpose products and include numerous
modules that can be flexibly assembled for site-specific
applications. The systems operate in very demanding settings
with a risk of significant environmental hazards. Accordingly,
they must be robust, reliable, and capable of maintaining
continuous operation.
We focus on the gradual expansion of the automation control
system design because architectural principles facilitated by
digital technologies were progressively embraced in ABB’s
products. Although it was initially designed solely to control
local physical production processes, its evolution added many
digital functions, resulting in a growing product scope that
transformed both the product platform’s nature and the firm’s
organizing logic.
Contemporary distributed control systems consist of sets of
interconnected digital processing units that steer dynamic and
MIS Quarterly Vol. 44 No. 1/March 2020 133
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Table 2. Complexity Concepts
Construct Description
Potential Effects of Digitization
of Product Platforms
Structural change when a system (product organization)
passes from one region of complexity to another
(Benbya and McKelvey 2006). A phase transition is
dissipative because structures and associations in the
system change and new ones emerge (Nicolis and
Prigogine 1989). Here, it involves transitions in the focal
company’s organizing logic associated with shifts in the
composition of its product platform.
Digitization challenges product organizations to
revise strategies, innovate, and restructure
products (Henfridsson and Yoo 2014; Svahn et
al. 2017).
Interaction rules with propensity for distinct types of
emergence. CGPs are networks of mechanisms that
generate variety from inputs, actions, and information
coupled to a set of constraints. The networks are fixed
or transitional (linkages are created and dissolved by
mobile agents) and operate in hierarchies where basic
CGPs serves as mechanisms in a more complex CGP
(Holland 1998).
The malleability of digital technology allows
agents to create and dissolve linkages among
components and agents (Benbya and McKelvey
2006; Eaton et al. 2015).
Connectivity Connectivity shapes the system’s propensity for emer-
gent behaviors through diversity, adaptiveness, con-
nectedness, and mutual dependencies. Diversity refers
to qualitative differences among entities. Adaptiveness
reflects the extent to which actors change their schemas
based on feedback. Connectedness measures coup-
lings governing flows of information and resources
among agents. Mutual dependencies capture inter-
dependencies of outputs between agents.
Digitization increases connectivity without
imposing tight couplings where agents are
affected by emergent interaction patterns in the
internal and external systems, new types of
couplings, and different types of interdepen-
dencies (Lee and Berente 2012; Lusch and
Nambisan 2015).
Any subsystem that lags behind in the development of
the technological system because of insufficient
performance and hampers the whole system’s
Reverse salients emerge during digitization
when a product organization is challenged by
trade-offs between backward compatibility
(limited technical improvements) and demand
for an architectural break to radically improve
performance (Kretschmer and Claussen 2016).
continuous manufacturing processes based on predetermined
events (e.g., changes in flow rate or temperature). They are
organized hierarchically in layers, and many functions are
allocated to programmable logic controllers (PLCs) that are
placed in production environments to steer process steps. At
the lowest (process) level, sensors and actuators record sig-
nals and transmit them to controllers with real-time responses
dictated by preprogrammed logic. These signaling actuators
execute commands that change the state of the process. This
happens in milliseconds because the system must respond to
any process changes swiftly to maintain control. Communi-
cation networks connect controllers for system-level moni-
toring and coordination, including analyzing event histories
and orchestrated responses to multiple interactive changes
based on model predictions. Thus, process data are integrated
hierarchically and the highest-level subsystems provide over-
all process control and diverse kinds of information for human
operators, including typical enterprise applications, such as
tracking stock levels, production and maintenance schedules,
production volumes, cost estimates, and quality parameters.
As shown in Table 3, we collected data in two periods
(2007–2010 and 2011–2017) through interviews, participant
observation, and searches of archival data. Indications of
product platform transitions triggered by digitization
prompted an exploratory case study (Eisenhardt 1989). In a
first phase, we observed a project to manage the installed base
of automation systems at various customer sites; in a second
phase we followed a project to increase certified solutions
from third-party developers. This called for mapping the
product platform’s historical evolution (see Appendix A for
134 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Table 3. Data Collection Summary
Data Source Period 1 Period 2 Total
Interviews 14 9 23
Participant observation 5 occasions (17 hours) 12 occasions (66 hours) 17 occasions (83 hours)
Archival data Project descriptions, technical specifications, annual reports, life-cycle and market analyses,
and product pamphlets (56 documents, approximately 3900 pages).
We adopted a process perspective (Van de Ven 1992) because
our initial familiarization with the data revealed path depen-
dencies and long-term cumulative change as salient themes.
We applied inductive procedures in three steps to identify
patterns of change over time (Van de Ven 1992). Each step
included multiple iterations of data reduction, theory-
informed conceptualizations, and triangulation (Langley
1999). The iterations allowed us to triangulate findings empi-
rically and theoretically and clarify and focus on emerging
themes (Miles and Huberman 1994).
First, we open-coded the data to discover key categories and
their properties that characterized platform change (Charmaz
2006). This yielded more than 600 codable moments where
change, its source, or its outcome were identified as sensi-
tizing concepts to organize for subsequent refinement (Bowen
2006). We identified multiple themes indicating the presence
of mechanisms and adaptation strategies. These were further
refined by assessing their internal homogeneity (coherence
within themes) and external heterogeneity (clear and identi-
fiable distinctions) (Patton 1990).
Second, through temporal bracketing, we structured the
analysis around periods of product continuity and discon-
tinuity to allow theorizing of feedback mechanisms, mutual
shaping, and multidirectional causality (Langley 1999). We
identified four evolutionary phases demarcated by transitions
in the product platform, each of which was operationalized as
a radical change in the product platform’s organization
involving implementing a new architecture or opening an
external resource inflow associated with organizational
change. To this end, we formulated a matrix to display and
compare key characteristics of the phases.
Third, we focused on identified phase transitions, then applied
the four sensitizing dimensions of connectivity (diversity,
adaptiveness, connectedness, and mutual dependency) as
analytical filters across the coded instances to identify drivers
of change in each transition. After filtering and categorizing,
we identified 225 instances of change, then reviewed if and
how alterations in the dimensions were associated with
product platform change. As a result, we generalized such
patterns into three generative mechanisms (CGPs), that is,
causal structures that produced observable emergent features
through macro-micro-level interactions (Hedström and Swed-
berg 1998; Henfridsson and Bygstad 2013). We then
challenged this rendition by evaluating conceivable alter-
natives, their grounding in the data (Appendix B), and their
role in phase transitions (see Table 4 later in this paper).
Finally, we labeled the three mechanisms (CGPs) interaction
rules, design control, and stimuli-response variety and
analyzed their role in transitions.
Digitization and Phase Transitions
in ABB’s Product Platform
In the evolution of ABB’s product platforms (Figure 1),
digital capacities triggered change across multiple social and
technical levels and environments, resulting in three platform
transitions offering generalizable insights informed by com-
plexity theory. In each new platform version, functions
expanded to higher levels in the Purdue Reference Model
(PRM, see Appendix C) as new nodes with novel digital
capabilities were added. The increases in scope and openness
of interfaces changed the platform’s adaptability and capacity
to tame complexity. Evolutions between generations were
prompted by the platform’s inability to cope with tensions
stemming from a reverse salient that could not be addressed
within the current product architecture. Here we review the
four platform phases (Appendix D) and then use CAS theory
to review deviation-amplifying mechanisms involved in the
transitions as the reverse salient was resolved through digiti-
zation. This modified prevailing CGPs and drove the firm
toward a new organizing logic.
1983–1992: Modular Architecture
and Device Digitization
In 1983,2 ASEA released the Master product platform with a
modular architecture and digital components in production
2In 1988, ASEA merged with Brown Boweri, forming ABB, the world’s
largest supplier of electro-technical equipment, with an approximately 20%
share of the $50 billion global electric power market, and about 180,000
MIS Quarterly Vol. 44 No. 1/March 2020 135
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Figure 1. The Platform Evolution Process
process control. Data homogenization introduced loose coup-
lings between devices and the network by reducing compli-
cated physical interface designs, thereby increasing
connectedness and diversity in the system. Both hardware
and software components were almost exclusively developed
internally, but technological advances in the IT industry were
rapidly increasing the affordability and utility of external
digital solutions. Retaining competitive in-house develop-
ment of all process automation components and solutions
became increasingly challenging. Eventually, tight coupling
of data and services and sole use of proprietary components
formed a reverse salient hindering innovation and more effi-
cient integration of components and solutions across automa-
tion systems. Reliance on market selection mechanisms for
noncore components would help cut costs and reduce techno-
logical uncertainty. Services were becoming an increasingly
important revenue source, so refocusing resources toward
core process automation and complementary information-
based services emerged as a key strategic thrust. Reuse of
functional engineering solutions, enabled by a more loosely
coupled digitized design, would create a versatile integrated
engineering environment:
We wanted to construct a common environment, a
lot of this is the same data that you need to transfer
between these systems: CAD, electricity, instrument
databases, etc. All such data transfer was done
manually, it was hard to keep it updated across
systems. Then we felt, this is really added value that
we should do and something we have the com-
petency for, otherwise someone else is going to do
such system integration. (R&D Manager ABB)
1992–1999: Base Service Digitization
and Integration
In 1992, ABB released a new product platform, Advant OSC.
It addressed the reverse salient by introducing an internal
repository for integration of process data and a loosely coup-
led object-oriented software architecture that expanded
product boundaries upward in the PRM hierarchy. The plat-
form was augmented with new functionalities for plant-wide
automation, manufacturing zone engineering, and data
provisioning at the enterprise level. It increased connected-
ness, diversity, and adaptiveness and decreased interdepen-
dencies between service and device. The shift in the
organizing logic focused on core process automation func-
tionality and leveraged new external agents in the design and
manufacture. The new architecture expanded the scope and
openness of platform control points by using open commu-
nication standards, relying on commercial off-the-shelf
solutions (COTS) such as Unix OS and HP hardware, and
integrated data from third-party suppliers. However, ABB
still controlled and developed basic process control-level
hardware and software internally. The integration of all
process data into a single database enabled new information-
based services. Over time, the increased connectedness
created new mutual dependencies due to tight coupling
between external information systems and the database. This
136 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
gradually grew into a reverse salient, because ABB had to
tweak its product configurations for each interface to adapt to
related novel information flows from connected systems.
This hindered digital service innovation because further inte-
gration would exacerbate needs for adaptation in external
A couple of years after releasing Advant we ran into
trouble. The system integrated data, but we couldn’t
get other actors to change their data format, so we
had to assume responsibility for data storage for a lot
of external systems .… we wanted to develop the
functionality, but in doing so, we assumed respon-
sibility for developing other actors’ components.
We realized pretty soon that it wasn’t sustainable.
(R&D Manager, ABB)
Adaptiveness was increased by opening the system interfaces.
This challenged not only perceptions of the nature of the
architecture but also the organizing logic:
We developed APIs for our system and an adap-
tation module that could hook into other systems’
APIs. We started thinking that this idea of holding
information together was interesting for a lot of
actors, not only engineers .… At first, we thought
we were developing an integration platform for engi-
neering tools, then for connecting all the parts of the
control system, then to integrate different products
and units in ABB, and eventually to integrate the
whole world. (R&D Manager, ABB)
1999–2003: Decoupling Service and Content
In 1999, ABB introduced the Aspect Integrator Platform as a
part of its strategic initiative “Industrial IT.” The initiative
pursued a common architectural standard for all of ABB’s
digital product solutions by leveraging Aspect Integrator’s
integrative capacity and providing certified compatible
automation solutions—developed both in house and exter-
nally—for various areas with high reliability. This was ex-
pected to enable ABB to create CGPs that leveraged Aspect
Integrator’s capacity for diversity generation by incorporating
more diverse external agents, but still aptly constrain these
procedures. Although a relatively vague visionary concept,
Industrial IT rapidly turned into a strategic pivot in ABB
because of contextual conditions and amplifications for diver-
sification offered by technology. The goal was to increase
diversity, connectedness with external agents and between the
different automation solutions, and amplify adaptiveness. In
2000, the automation division’s director declared, “Industrial
IT is the future. Within a year, all products from my division
must be Industrial IT-certified.” Many units responded
enthusiastically, but an R&D manager responsible for the
concept noted that nobody really knew what the certification
meant.3 The certification grew rapidly in ABB: 1100
products were certified in 2002, and a year later 35,000.4
During this period, the organization’s identity changed as
process automation concepts and Industrial IT became central
to the firm’s transformation into a digital company:
We are transforming our business portfolio …
incorporating sophisticated software applications.
We apply our expertise to develop creative ways to
integrate our products and systems with our cus-
tomers’ business processes to enhance their produc-
tivity and efficiency. We refer to this integration as
“Industrial IT.” Our increased commitment to
Industrial IT has been supported by our recent stra-
tegic initiatives and our research and development
efforts. Collaboration with our customers and our
commitment to Industrial IT will be further
enhanced by the realignment of our business opera-
tions. (ABB 2001, p. 5)
ABB also explored options to expand product boundaries
using Aspect Integrator and Industrial IT and expand its
integration offerings from process automation to other busi-
ness processes and industrial settings. The organizing logic
expanded to coordinate distributed agents and diverse
technological components through integrative capacity and
compatibility policies. Identity was transformed toward
information-intensive business enabled by the Industrial IT
architecture and realigning operations along customer seg-
ments. This ambitious strategy was abruptly stopped in 2002
when ABB faced severe financial problems due to an acquisi-
tion spree, economic recession (dot-com bust), and mounting
asbestos claims. The firm had to focus on core businesses and
divest a significant portion of the company. During this
period, the platform was migrated to multiple operating
systems and accrued substantial technical debt. Maintaining
versions for all OSs and ensuring component compatibility
became increasingly challenging, error-prone, and costly (a
reverse salient). ABB had to revert its focus, divest a signi-
ficant portion of the company, and consolidate the archi-
tecture by solely building on Windows OS in future products.
3For example, as the IT manager of the power division (with no software at
any level of the scheme) wanted to participate, level 0 certification was
added, defined as availability of all relevant information about a product (i.e.,
service manuals, maintenance information, user manuals, calibration infor-
mation, etc.) in a digital format.
4Many of these at the lowest level: “Each certified product is assigned to a
product suite and is named according to what it does and how it fits into the
Industrial IT system” (ABB 2003, p. 26).
MIS Quarterly Vol. 44 No. 1/March 2020 137
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
2003-2017: Toward a Platform Ecosystem
In 2003 ABB released a new product platform based on the
Aspect Integrator architecture called 800xA. It changed
ABB’s product strategy and organizing logic by packaging all
incremental innovations and new components inherited from
previous years into a single architecture. This positioned
ABB as a full integrator and service provider for automation
systems in industrial environments with a large number of
complementors. The 800xA increased diversity by expanding
connectedness between product components horizontally by
enabling plant-wide data integration and vertically by creating
enterprise system level couplings, and laterally toward other
IT systems through the increased use of COTS interfaces
(such as SAP/R3) and industry standards. Because of rise in
the use of third-party products and reliance on open standards,
800xA was profoundly affected by changes in the external IT
environment. This called for continued maintenance and
adaptation (interdependencies requiring adaptiveness). Rapid,
continual innovation resulted in growing diversity through
technological solutions, making existing components obsolete
and incompatible. Digitization expanded functionality but
increased environmental variability and reduced reliability
and transparency, which conflicted some top priorities in
production environments: safety, minimizing environmental
risks, and avoiding costly downtime:
You don’t want to install updates, that’s a distur-
bance, and there are always risks involved with
installing updates. We minimize updates and only
make software changes when there are obvious
problems or the version of Windows is so old that
it’s not supported anymore. I can’t recall us up-
dating because we needed new functionality. (IT
Manager MineCo)
Production ultimately depends on mechanical systems that
constrain possible efficiency improvements, so clients’
incentives for increasing software functionalities were limited.
Instead, ABB sought to increase diversity by adding func-
tionalities for data analysis. For example, in 2017 ABB
launched a complimentary augmentation service, ABB
Ability, described as a business channel facilitating access to
the installed base of solutions for suppliers and customers
with certified solutions for analyzing operations.5 The
decision to open access to extracted data, rather than platform
services, is in line with the risk-averse context and experience
gained from previous phases of automation system design:
When designing the Aspect Integrator Platform and
organization around Industrial IT we wanted to
build an ecosystem .… We later realized that
opening up too much towards the actual process was
a big risk towards the market and customers. Even
if you certify and test, you might have unexpected
failures and breakdowns .… In an App Store, even
with just 500 external apps, testing all permutations
becomes challenging. We do regression tests
whenever Microsoft has a security update or we
need to fix a bug. So how can we make sure that
the core is not affected? External apps probably
need to operate in the external environment. (R&D
Manager ABB)
Digitally Induced Mechanisms (CGPs)
and Phase Transitions
Our analysis reveals several recursive patterns of change.
Implementing digital capacities to address reverse salients
created new CGPs that fostered continued increases in the
number and diversity of agents interacting with the platform
and their interconnectivity. Digitally induced change was
mediated by three deviation-amplifying mechanisms—
CGPs—that jointly enabled and triggered transitions:
(1) changes in interaction rules, (2) distribution of design
control, and (3) increases in stimulus-response variety. These
CGPs pushed the system to an edge of discontinuity and to
transition into a new organizing logic three times. The pro-
cess was characterized by emergent change and nonlinear
cumulative effects. Changes originally perceived as inno-
cuous technology amendments later proved to have deep
consequences. Digitization scaled product platform functions
upward toward customer operations and related interactions,
expanded information exchanges, and increased the number
of participating agents.
Changes in Interaction Rules: Introduced information-
based couplings profoundly affected interactions enabled and
constrained by the platform architecture. The couplings
separated data from unique physical features of designed
components and their connections by homogenizing internal
data into information-based connections expressed in binary
form. This enabled the use of general-purpose digital tech-
nologies to store, transmit, process and display process data.
The decoupling generated munificent combinatorial options
not possible in earlier systems. Information-based couplings
exponentially increased the numbers of connected elements
and system variety (connectedness and diversity). We refer
to this mechanism as a change in interaction rules, which
results from the increase in the number of (loosely) connected
elements after implementing digitally enabled interactions.
5An “Industrial Internet technology platform and cloud infrastructure. An
open, globally available, digital-industrial ecosystem for customers, partners,
suppliers and developers” (
138 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
The changes in interaction rules need to be viewed in the light
of ABB’s persistent challenge to increase the integrative
capacity while retaining control to maintain security in risk-
avoidant production processes. To ensure high reliability,
ABB needed to carefully consider the positioning of control
points that would allow isolating the production process and
verifying compatibility.
Our customers are terrified about opening up here
[the production process], they even keep their own
IT departments away from it although there are a lot
of servers, clients, etc. Security, reliability, consis-
tency, and real-time properties are extremely impor-
tant. If we compromise on that, then it is goodbye,
our customers are not interested anymore. (R&D
Manager ABB)
Due to this priority, adaptations in interaction rules increasing
diversity and connectedness were implemented gradually.
First, information-based couplings replaced one-to-one
mappings in the physical transport layer with many-to-many
couplings by standardized physical data transfer protocols and
interfaces. This vastly expanded combinatorial options to
trace and monitor production processes and allowed pre-
fabricated hardware modules to provide greater diversity and
connectedness in system control. The ability to combine data
from multiple sources increased the number of connected
elements and enabled a single controller to integrate data from
many sources and expand system control functions. By the
early 1990s, a new reverse salient surfaced when mutual
dependencies from the tight data-service coupling, internal
component development, and proprietary interfaces limited
combinatorial options to reuse logical level engineering
solutions (libraries, configurations, etc.) and integrate data.
ABB restructured its design rules and launched a new
architecture to ease integration and control of system func-
tionalities and abstracted them to software services. The
loose coupling of services and devices increased the number
of connected elements and created new connections to
external components. Thus, object-oriented architecture, open
standards, and data integration from external components
moved ABB from an internally focused organizing logic
toward leveraging external components in an increasing
orientation to system maintenance and engineering services.
Before that, we designed and manufactured all com-
ponents by ourselves, operating system, database,
graphics hardware, and software. Everything, all by
ourselves. We realized that it wasn’t sustainable, we
needed to change. So, in 1992 we introduced the
new architecture, based on Unix and hardware from
HP for workstations. It was a radical change at that
level of the process automation hierarchy. (R&D
Manager ABB)
Over time, the tight couplings to content in external compo-
nents exposed significant mutual dependencies and hindered
further integration because the platform had to ensure full
compatibility with all external systems and their changes. To
address this reverse salient, ABB’s next architecture intro-
duced novel couplings with low levels of mutual depen-
dencies. Physical objects were digitally represented by
“aspect objects” linking to relevant information (drawings,
temperature, control panels, etc.). These couplings made the
product platform’s boundaries fluid and challenged the firm’s
identity. The Aspect Integrator expanded integrative capacity
toward a greater variety of data sources that were no longer
restricted to process automation (increased diversity, con-
nectedness, and adaptiveness). The number of connected
elements rose quickly (already 35,000 certified products in
2003). The new options to connect devices and combine data
led ABB to question the platform boundaries and probe
possibilities of expanding into new industries using the
Industrial IT concept. These integrative capabilities were
seen as foundational for a deeper transformation in
organizing logic and identity:
Through its engagement in Industrial IT, ABB aims
to do for industry what Microsoft did for the office
environment, combining productivity tools in pack-
ages, all the way from process automation to
business automation. (ABB executive committee
member, cited in ABB Review 2001, p. 22)
However, negative feedback from the external system (the
burst of the dot-com bubble) and the rise of new internal
constraints (the financial crisis following the acquisitions and
asbestos crisis) dampened the system-expanding effects of
CGPs. But the transition to many-to-many relations in
components remained the foundation for generating new
growth in diversity and connectedness. With the 800xA, the
number of supported protocols for not only production envi-
ronments but also plant-level systems and enterprise envi-
ronments grew considerably (e.g., IEC61850, WirelessHart),
increasing diversity, connectedness, and need for adaptive-
ness. These extensions not only removed some interde-
pendencies but added some new ones.
In summary, digitally enabled interactions, facilitated by
malleability, data homogenization, and decoupling of content
from service and physical devices, increased connectedness
and diversity and reshaped mutual dependencies.
Information-based couplings allowed platform owners, users,
and component developers to expand searches for combina-
torial options among new elements, facilitating the discovery
of novel forms of value creation. Such new rules accelerated
platform change by promoting faster and broader spread of
digitization. Information-based couplings loosely coupled
previously tightly coupled components and identified and
MIS Quarterly Vol. 44 No. 1/March 2020 139
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
removed inefficiencies expressed in a reverse salient. The
combinatorial options resulted in scope extensions and
increased integrative capacity, which was advantageous for
innovation and efficiency. However, changes in connected-
ness had to be balanced with security concerns in the risk-
averse context. When information-induced perturbations
exceeded the system’s coping capacity and possibility for
control, operational risks rose in ways that could cascade
proposed solutions into chaos.
Distribution of Design Control: Digitization led to a gradual
release and distribution of design control in each new version
of ABB’s product platform. This recurrently triggered emer-
gent changes that challenged the prevailing organizing logic.
Distribution of control altered diversity, called for adaptive-
ness, and created new interdependencies. The hybrid cyber-
physical nature of automation systems generated tensions
related to temporal differences in the growth of diversity and
adaptiveness among digital and physical product components.
Digital component developers were more distributed and had
little or no connection to specific use contexts of process
control with unique physical components. They could inno-
vate more freely without significant centralized control, and
their innovations could be accessed through digital channels
at arm’s length without significant physical interactions.
We refer to this mechanism as distribution of design control,
defined as the decentralization of design rights to hetero-
geneous agents interacting with the platform. The distribution
is enabled by the digital design rules and related structural
composition of digital objects. This enabled fluidity (repro-
grammability) in services and data processing and created
dynamic relations between digital objects (because of the
separation of content and services). In digital components
functions, interfaces, and use rules could be readily edited,
reconfigured, and recombined by agents with some decision
rights, if sufficient openness to design components was
granted. These properties enabled continuous reconfiguration
of decision rights over components that granted external
agents higher degrees of freedom for recombining
components and system assemblies. The relative speed and
ease of recombination increased product variability (i.e.,
dynamics and variety of elements and their couplings within
and across products in use), heightening diversity and con-
nectedness and calling for increased adaptiveness within the
The Master product platform increased diversity and need for
adaptiveness at production sites through information-based
coupling, but it did not diversify decision rights regarding the
design and manufacturing of platform components much.
Because development was kept in-house, diversity remained
low in terms of output and innovating agents. However, the
loose information coupling in the Advant platform increased
diversity and connectedness by reconfiguring decision rights
on the supply side. As new types of process data were inte-
grated into a single database, ABB had to adapt to changes in
the external systems providing data. This increased diversity
called Advant to migrate across multiple OSs (HP-UX
8.x;9.x;10.x, Digital Unix, Open VMS, Windows NT) and
progressively embrace general communication standards,
continuing to distribute decision rights and increasing
The transition to the Aspect Integrator platform was tightly
linked to “the transformation of the ABB group into a
knowledge- and service-based company” (ABB 2000, p. 2).
The change sought to balance the need to generate greater
product variability and versatility while keeping control over
the overall solution. Through the Industrial IT certification
process, ABB sought to instill control points to ensure
reliability and compatibility while leveraging component
developers’ innovations and generating revenues from
platform access. Although the strategy was never fully
implemented, the new couplings offered by open, generic
APIs to other systems in the Aspect Integrator architecture
enabled ABB to cope with increased diversity and inter-
dependencies arising from the Advant architecture. This
opening up significantly increased variability in terms of
diversity of third-party components and adaptiveness
requirements toward OS, hardware, and other software
changes. Also, acquired automation systems (e.g., Alfa Laval
and Elsag Bailey) had to be incorporated around the millen-
nium, and ABB’s economic crisis necessitated cost
reductions. The firm decided to reduce internal diversity and
interdependencies by limiting the 800xA platform to run on
Microsoft Windows OS. However, this move restricted
ABB’s control over decisions on product variability and
compatibility, raising additional concerns:
Since we’ve moved away from making all products
and systems, both ABB and the customers have
ended up in a situation where we don’t control the
development of IT components. Hardware suppliers
have their own life cycles, and Microsoft releases
their updates and patches frequently .… We need to
follow their development and customers need to
follow both their [IT component suppliers] and ours.
It’s much more dynamic, in a way everything floats.
(ABB Project manager)
For example, one platform user made a strategic decision to
avoid all but essential updates after a substantial shutdown
caused by the incompatibility between a new system version
and firmware on a circuit board in an old machine. This
increased diversity was considered an operational risk
because automation system configurations had become highly
dynamic and required constant updates (adaptiveness).
140 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Throughout the evolutionary process, distribution of decision
rights extended the platform’s functionality but also obliged
ABB to repeatedly reconfigure the platform and operations to
balance diversity and adaptiveness requirements. This be-
came a significant driver in shifting organizing logic. Using
digital design rules increased search spaces, expanded
functional variability, and gave the platform provider access
to new pools of agents and more extensive resource base.
However, the mechanism reduced ABB’s control over techno-
logical trajectories and rate of change as reprogrammability
increased calls for adaptiveness. It also changed structural
complexity by stimulating local diversity in products, which
created new dependencies in local contexts (e.g., firmware in
production environments).
The digitization of ABB’s product platforms created hybrid
cyber-physical systems with tensions arising from the differ-
ence in the evolutionary pace between digital and physical
components. ABB had to carefully consider what adaptation
strategies to deploy as a response to such tensions, that is,
whether new digital capacities will remove the tension or
generate new ones. This inertia has restrained ABB from
establishing a full two-sided market and can be interpreted as
an effect of system constraints embodied in CGPs that set
requirements for change in system functions and their rela-
tions and conditions their use.
Increase in Stimulus-Response Variety: Digital compo-
nents in ABB’s platform vastly increased functional variety.
Variety increase was founded on the ease of programming
new functions and enhanced capacities for transmitting,
processing, storing, and representing digital process data.
Functional variety spurred environmental interactions, which
became new sources of value creation and service provi-
sioning. Each adaptation relieved specific concerns, but over
time exposed the product platform to growing stimuli variety
(in terms of sources and types). This was produced by
increasingly open product boundaries and emergent inter-
action patterns. When increases in stimulus variety could not
be efficiently tamed, more adaptations added new digital
capacities with connectedness to new control levels that
increased diversity. The increase in stimulus-response variety
mechanisms refers to how digital components enrich and
expand the scope of environmental interactions by increasing
product functionality and the variety of stimuli related to it.
The stimulus-response variety mechanism played a significant
role in scaling platform digitization. It expanded the set of
agents interacting with the product, and such additions took
place after each phase transition, except perhaps the last one
(which intended to reduce stimuli variety). Expanded plat-
form functionality increased connectedness toward new
environments and generated distinct levels of diversity, adap-
tiveness, and mutual dependencies with emergent outcomes.
These effects were mostly unforeseen and strengthened
across phases, pushing the platform into new phase transi-
tions. Between phases, such effects became increasingly
important when introduced digital components enabled new
types of interactions at higher organizational levels. For
example, during the first phase, the proprietary design and in-
house component development limited stimuli variety. The
configuration impeded ABB’s ability to increase functional
variety, which became a fundamental driver to transition to
Advant architecture. The effects of stimuli variety were more
profound for the Aspect Integrator, where technologically
distinct potential for connectivity dimensions became an
important driver of firm-wide radical change. Advant’s new
digital components enabled integration across previously
separate engineering systems, but these interactions created
additional dependencies. ABB had to invent new ways of
increasing its adaptiveness toward updates in external
vendors’ systems. The response introduced another platform
layer with new digital components and APIs, boosting the
scope and number of information-based couplings and further
increasing stimulus-response variety:
Realizing that everybody needs an integration plat-
form, we started thinking about business models for
our platform. We can sell it to Siemens [main com-
petitor]. We need to be open and offer this to
everybody, not just automation. We had discussions
with arms makers, construction companies. We
formed alliances with Microsoft, Intel, Accenture.
In addition to using it for internal components, we
wanted to certify external components with business
models based on licensing. Then the IT bubble
burst and ABB had a financial crisis so it was
back to basics. (ABB R&D manager)
Similarly, introducing open APIs and communication stand-
ards and building on Microsoft technology in the 800xA
reduced system fragmentation, decreased the number of plat-
form versions, and enabled new types of component-level
innovation. These changes increased the rate of adaptation to
the external environment and triggered an inflow of inno-
vation. But they also increased maintenance requirements by
pushing customers toward dynamic software maintenance
and frequent upgrades, which raised operational risk. Thus,
the transition from the Aspect Integrator to the 800xA did not
necessarily expose the product platform to new environments,
but it changed the nature of interactions due to new types of
dependencies. Most resources had to be devoted to co-
evolving the platform with changes in the digital environment
to maintain compatibility with the physical production:
MIS Quarterly Vol. 44 No. 1/March 2020 141
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
The main part of development costs is devoted to
maintaining the architecture … these systems are so
complex. I’m not sure about the exact number but I
know that a couple of years ago we had ten million
lines of code in the platform. (ABB Engineer)
Increases in stimuli-response variety generate self-reinforcing
cycles of change that alter the complexity in the system’s
agents and elements and influence the platform’s ability to
tame it. As reflected in the case, connectedness toward new
environments altered diversity, adaptiveness, and mutual
dependencies. This resulted from continued scaling of digital
components in the platform, as digitization provided flexible
means to extend the platform’s functional scope. In each
cycle, further digitization increased variety in system func-
tions to respond effectively to reverse salient and related
tensions. Over time, the increased functional variety created
new spaces for novel environmental interactions, leading to
further amplification of the variety of stimuli and calling for
further digitization.
The evolutionary pattern in ABB’s product platform suggests
that substantial architectural change that expands platform
scope will trigger transitions in the organizing logic when the
new type of stimuli call for significant shifts in the agent’s
schemata, connections, and mutual dependencies. However,
when negative firm-level and environmental feedback is
present, agents will selectively engage in architectural change
to lower the stimuli variety by curbing environmental inter-
actions (e.g., reducing numbers of supported OS and targeted
markets). Thus, stimuli-response variety configurations can
and should be orchestrated by platform providers. However,
outcomes of digitally induced platform change are not easily
foreseen because they are mediated by emergent and often
unintended change in CGPs, which generate emergent
system-level outcomes. Table 4 summarizes how the intro-
duction of digital capacities produced these effects and
triggered phase transitions over time.
Implications and Conclusions
Distinct platform types have been viewed as evolutionary
stages (Gawer 2009, 2014), and previous studies suggest that
digitization triggers such evolutionary transitions toward
ecosystem-based organizing logics (Svahn et al. 2017; Yoo et
al. 2010). However, we know little about how digitization
drives such platform transitions and what mechanisms make
them more likely or necessary. At ABB, waves of digitization
were followed by significant organizational outcomes as they
changed the boundaries of the platform scope, scale, and its
sources of value creation and extraction (Bharadwaj et al.
2013; Nambisan et al. 2017). These were enabled by the
traits and design rules of the digital artifacts (Kallinikos et al.
2013; Lyytinen et al. 2016). The resultant blurring of bound-
aries through the infusion of digital capacities pushed the
incumbent firm to gradually transition to platform ecosystems
(Thomas et al. 2014; Yoo et al. 2010). Our analysis reveals
three mechanisms modeled as CGPs that triggered such
Research Implications
Prior studies have shown that digitization expands product
scale and scope and change the organizing logic of its hosting
firm (Lee and Berente 2012; Svahn et al. 2017). However,
there is a paucity of studies of mechanisms that undergird
deep cumulative changes in the firm’s organizing logic. To
address this gap, we built on CAS theory and synthesized the
role of interactions during emergence (CGPs) (Holland 1998;
Page 2010), digitization (Kallinikos et al. 2013; Yoo et al.
2010), and platform evolution (Constantinides et al. 2018; de
Reuver et al. 2017; Gawer 2014) into a conceptual
scaffolding that allowed us to trace the cumulative effects of
waves of digitization on ABB’s automation platform and
ensuing shifts in the firm’s organizing logic. We identified
three endogenous CGPs and related networks of
mechanisms—interaction rules, design control, and stimuli-
response variety— through which digitization triggered
product platform transitions and shaped new organizing
logics. These mechanisms offer an analytical apparatus for
future studies of how digitization unfolds in platform
transitions. Whereas the initial typology by Gawer (2014)
identifies essential distinctions among technological
platforms, the variations in platform types we identified and
their multifinality as complex adaptive systems suggest that
a more fine-grained classification of (digital) platform types
is essential for advancing understanding of digitization’s
impact on product platforms.
Outcomes of the three mechanisms are not deterministic.
They have highly interactive composite effects, although they
were not examined in detail in this study (Henfridsson and
Bygstad 2013). The mechanisms also show that macro-level
change affects interactions at the micro-level, which, in turn,
produces emergent macro-level orders (Hedström and Swed-
berg 1998). Thus, a new organizing logic emerges from
myriad interactions enabled and enforced by the three digital
(CGPs) mechanisms and other (nondigital) mechanisms
(CGPs) (Holland 1998). These mechanisms have fractal pro-
perties in that they operate at multiple levels simultaneously
with large variations in their scope and intensity. As a
product platform accrues digital capacities, process outcomes
cannot be understood solely by analyzing static
configurations of its social and technical elements
(Tanriverdi et al. 2010;
142 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Table 4. Characteristics of Phase Transitions
Construct Transition 1 Transition 2 Transition 3
Proprietary components and inter-
faces, and tight data service
coupling hinder logical engineering
solutions’ reuse, performance of
general-purpose digital components
and integration.
Growing mutual dependencies from
tight coupling to external components,
while requirements grow for further
integration of internal and external
business processes.
Trade-offs in resource consumption
associated with ensuring compatibility
across OSs and increased functionality in
use environments.
Object-oriented architecture, open
standards, shared database, and
general-purpose components (OS,
hardware) in manufacturing zone
(PRM levels 2 “Area supervisory
control” and 3 “Site manufacturing
operations and control”).
Object architecture with APIs,
adaptation module, and component
Standardization on Windows OS and
increased use of COTS and
communication/service standards.
Change in
Decoupling of data and service
gives combinatorial options for
software and data use increasing
connectedness among software
Loose couplings toward external soft-
ware services reduce mutual depen-
dencies and drive connectedness by
allowing data to be combined from
external sources and accessed from
the enterprise zone with little effort.
Support for communication standards
increases connectedness across produc-
tion components and production sites, and
toward business processes at the enter-
prise zone (PRM levels 4 “Site business
planning” and 5 “Enterprise network”).
of design
Couplings to external software
components interacting with PRM
level 3 (“Site manufacturing
operations and control”) increase
interdependencies and
adaptiveness requirements.
New interfaces and support for
standards increase internal and
environmental connectedness, and
diversity and adaptiveness toward
the IT industry solutions and busi-
ness environments.
Distribution of decision rights to the
IT industry environment.
Certification scheme to reconfigure
decision rights.
Higher reliance on COTS and support
for standards increase diversity and
adaptiveness requirements by distrib-
uting decision rights to external actors.
To ensure security and compatibility
among diverse and adaptive connected
components, ABB strengthens input
control through testing and requirements
for updated software.
Change in
Increased connectedness and
mutual dependencies toward
external software components
expose the product platform to
stimuli from the IT industry.
Functionality to integrate data
sources Increase connectedness to
business environments and elimin-
ates the anchoring in process auto-
mation, increasing both internal
(e.g., the Industrial IT certification
program) and external (e.g., across
technologies and industries) stimuli
(increase interdependencies, diver-
sity, and requirements for adap-
Platform functionality for wide
range of OSs and acquired control
Decreased scope of use environments
lowers environmental stimuli. Simul-
taneous increased reliance on COTS,
general-purpose standards, and con-
nectedness through the internet signi-
ficantly increase stimuli (diversity and
adaptiveness requirement) from the IT
Change in
From internal product-centric con-
trol over design and manufacturing
to coping with dependences in
physical production environment
into specializing in process auto-
mation functionality and integrating
external competencies and
Product platform governance
structures to coordinate distributed
set of agents and diverse techno-
logical components through
integrative capacity and compati-
bility verification.
Identity challenged by a transfor-
mation of business portfolio toward
high-return business enabled by
Industrial IT architecture, and
realigning operations along
customer segments.
Product platform with high degree of
internal design and manufacturing for
production process components while
leveraging communication standards to
integrate data across the enterprise and
innovative capabilities within the
The product platform viewed as
dedicated to process automation.
Organizational restructuring from
customer segments into process
automation business.
MIS Quarterly Vol. 44 No. 1/March 2020 143
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Yoo et al. 2012; Yoo et al. 2010). Order emerges performa-
tively from micro-level phenomena during platform design
and use. For example, at ABB initially narrow digitization
effort eventually led to unexpected and unintended changes
that transitioned the product platform toward a platform
ecosystem. Across and within each phase, the change was
subject to “increasing returns,” that is, capabilities previously
assimilated within the system could be leveraged later and
were necessary to enter the next stage (Kurzweil 2004).
Digitization eventually changed from handling direct informa-
tion about the physical process to dedicated sets of informa-
tion functions and services to coordinating production
processes and full process systems. Ultimately, this generated
new kinds of relationships and interactions among external
partners in areas where ABB had insufficient knowledge and
skills or could now leverage economies of scale or scope. Ini-
tial digitization in the device and network layers enabled to
progress to loosely coupled services and content, causing
CGPs to generate substantial platform complexity. Conse-
quently, in later stages digitization covered substantial parts
of ABB’s and its customers’ operations (with more significant
and broader effects) and involved multiple subsystems simul-
taneously. This sequencing provides insights into how digiti-
zation processes are likely to unfold in product platforms and
also suggests the need for future research on their path
Practical Implications
Our study has a several practical implications. First, it under-
scores that digitization expands the scope of product plat-
forms in terms of both functions and participating agents. The
consequent increases in diversity, adaptiveness, connected-
ness, and dependencies drives complexity and generates
effects that are difficult (or impossible) to predict. Seemingly
innocuous digitization can have strategically significant
effects, which calls for strategic heedfulness to digital tech-
nology design and its role in orchestrating internal and
external relations and interactions. Leveraging integrative
capacity of digital technology calls for continuously tuning
interactions and dynamically reconfiguring technical and
organizational resources. Second, digitization can give archi-
tectural leverage in terms of production, innovation, and
transactions (Thomas et al. 2014). Our research suggests that
such processes can be stimulated in sequences—at ABB we
observed changes in production, then innovation, and finally,
to some degree, transactions. Managers should consider the
kind of leverage they seek through digitization and review
sequences and path dependencies for achieving it while recog-
nizing that unpredictable changes may be unleashed. For
high-reliability settings, such analysis should consider
security and avoid pursuing (for example) transaction lever-
age through multisided markets, if reliability in the local
context cannot be guaranteed. Third, our study shows that the
idea of a reverse salient can serve as a valuable point of
departure for an organizational transition. Platform providers
should consider not only immediate resolution but also how
responses can be used as strategic pivots. Fourth, while
digital capacities enable firms to expand target markets
radically, such growth needs to be aligned with organizational
capacity and current platform use. If these cannot be
reasonably aligned, providers should consider exploring
expansion options separately. Finally, pace and patterns of
change strongly differ in physical and digital architecture.
Our findings suggest that managing these differences through,
for example, appropriate control points is essential for
successfully digitizing product platforms.
Three limitations of this study should be noted. First, we
focused on an industry that is highly dependent on very
reliable machinery with long life cycles. The differences in
dynamics between the digital and physical architectures (and
hence some of the identified adaptive tensions) are pro-
nounced in the focal context. It would be beneficial to
examine the digitization of other types of platforms. Second,
we examined platform evolution in a world-leading firm in a
capital-intensive high-tech engineering industry. Processes
and consequences of physical environments’ digitization may
vary substantially in other contexts. Future studies should
investigate the dynamics of product platforms with different
resource bases. Third, industrial control systems are used in
numerous sectors, and innovation adoption patterns might be
different across use contexts. For example, most platform
users we studied operate in industries that have been highly
profitable in recent decades. Platforms in more cost-
dependent industries might be subject to other dynamics.
To conclude, our study provides novel empirically grounded
insights on the role of digitization in product platform transi-
tions by empirically demonstrating the multilevel and recur-
sive nature of digitally driven growth in product platforms.
We hope that this work will stimulate future research to
identify additional complex patterns and mechanisms to
advance richer theorizing of the impact of digitization on
platform transitions.
The authors wish to thank seminar participants at Case Western
Reserve University, University of Georgia, and University of Texas
at Austin for their valuable feedback. We are also grateful to the
senior editors and the review team for their constructive feedback,
thoughtful guidance, and support. This research is supported by the
Marianne and Marcus Wallenberg Foundation.
144 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
ABB. 2000. ABB Group Annual Report 1999, ABB Ltd., Zurich
(r e t rie v ed fr om h ttps ://l i bra r y .e. a bb.c o m /pu b l ic/
Ozp a5p7i N0iCxWTHLAesM5o4YDZ8q7xz7XtGc/oHkC0
ABB. 2001. 2000 Annual Report of ABB as filed with the United
States Securities and Exchagne Commission, form 20-F
(retrieved from
ABB 2003. 2002 Annual Report of ABB as filed with the United
States Securities and Exchagne Commission, form 20-F
(retrieved from
ABB Review. 2001. “Industrial IT and Collaborative Commerce:
Accelerating Next-Generation eMarketplaces,” ABB Review,
eT echnology Se ction, pp. 20-22 (retrie ved fr om
Anderson, P. 1999. “Perspective: Complexity Theory and Organi-
zation Science,” Organization Science (10:3), pp. 216-232.
Benbya, H., and McKelvey, B. 2006. “Toward a Complexity
Theory of Information Systems Development,” Information
Technology & People (19:1), pp. 12-34.
Bharadwaj, A., El Sawy, O. A., Pavlou, P. A., and Venkatraman, N.
2013. “Digital Business Strategy: Toward a Next Generation of
Insights,” MIS Quarterly (37:2), pp. 471-482.
Bowen, G. A. 2006. “Grounded Theory and Sensitizing Concepts,”
International Journal of Qualitative Methods (5:3), pp. 12-23.
Bresnahan, T. F., and Greenstein, S. 1999. “Technological Com-
petition and the Structure of the Computer Industry,” Journal of
Industrial Economics (47:1), pp. 1-40.
Charmaz, K. 2006. Constructing Grounded Theory: A Practical
Guide through Qualitative Analysis, Thousand Oaks, CA: SAGE
Constantinides, P., Henfridsson, O., and Parker, G. G. 2018.
“Introduction—Platforms and Infrastructures in the Digital Age,”
Information Systems Research (29:2), pp. 381-400.
Cusumano, M. A., Mylonadis, Y., and Rosenbloom, R. S. 1992.
“Strategic Maneuvering and Mass-Market Dynamics: The
Triumph of VHS over Beta,” Business History Review (66:1), pp.
de Reuver, M., Sørensen, C., and Basole, R. C. 2017. “The Digital
Platform: A Research Agenda,Journal of Information Tech-
nology (33:2), pp. 1-12.
Eaton, B., Elaluf-Calderwood, S., Sørensen, C., and Yoo, Y. 2015.
“Distributed Tuning of Boundary Resources: The Case of
Apple’s iOS Service System,” MIS Quarterly (39:1), pp.
Eisenhardt, K. 1989. “Building Theories from Case Study
Research,” Academy of Management Review (14:4), pp. 532-550.
Gawer, A. 2009. Platforms, Markets and Innovation, Cheltenham,
UK: Edward Elgar Publishing.
Gawer, A. 2014. “Bridging Differing Perspectives on Techno-
logical Platforms: Toward an Integrative Framework,” Research
Policy (43:7), pp. 1239-1249.
Gawer, A., and Cusumano, M. A. 2014. “Industry Platforms and
Ecosystem Innovation,” Journal of Product Innovation Manage-
ment (31:3), pp. 417-433.
Goldstein, J. 1999. “Emergence as a Construct: History and
Issues,” Emergence (1:1), pp. 49-72.
Hedström, P., and Swedberg, R. 1998. “Social Mechanisms: An
Introductory Essay,” in Social Mechanisms: An Analytical
Approach to Social Theory, P. Hedström and R. Swedberg (eds.),
Cambridge, UK: Cambridge University Press, pp. 1-31.
Henfridsson, O., and Bygstad, B. 2013. “The Generative
Mechanisms of Digital Infrastructure Evolution,” MIS Quarterly
(37:3), pp. 907-931.
Henfridsson, O., and Yoo, Y. 2014. “The Liminality of Trajectory
Shifts in Institutional Entrepreneurship,” Organization Science
(25:3), pp. 932-950.
Holland, J. H. 1998. Emergence: From Chaos to Order, Reading,
MA: Perseus.
Hughes, T. P. 1983. Networks of Power: Electrification in Western
Society, 1880–1930, Baltimore: Johns Hopkins University Press.
Kallinikos, J., Aaltonen, A., and Marton, A. 2013. “The Ambi-
valent Ontology of Digital Artifacts,” MIS Quarterly (37:2), pp.
Karhu, K., Gustafsson, R., and Lyytinen, K. 2018. “Exploiting and
Defending Open Digital Platforms with Boundary Resources:
Android’s Five Platform Forks,” Information Systems Research
(29:2), pp. 479-497.
Kretschmer, T., and Claussen, J. 2016. “Generational Transitions
in Platform Markets—The Role of Backward Compatibility,”
Strategy Science (1:2), pp. 90-104.
Kurzweil, R. 2004. “The Law of Accelerating Returns,” in Alan
Turing: Life and Legacy of a Great Thinker, C. Teuscher (ed.),
Berlin: Springer, pp. 381-416.
Lamberson, P., and Page, S. E. 2012. “Tipping Points,” Quarterly
Journal of Political Science (7:2), pp. 175-208.
Langley, A. 1999. “Strategies for Theorizing from Process Data,”
Academy of Management Review (24:4), pp. 691-710.
Lee, J., and Berente, N. 2012. “Digital Innovation and the Division
of Innovative Labor: Digital Controls in the Automotive
Industry,” Organization Science (23:5), pp. 1428-1447.
Lusch, R. F., and Nambisan, S. 2015. “Service Innovation: A
Service-Dominant Logic Perspective,” MIS Quarterly (39:1), pp.
Lyytinen, K., Yoo, Y., and Boland Jr., R. J. 2016. “Digital Product
Innovation Within Four Classes of Innovation Networks,”
Information Systems Journal (26:1), pp. 47-75.
McIntyre, D. P., and Srinivasan, A. 2017. “Networks, Platforms,
and Strategy: Emerging Views and Next Steps,” Strategic
Management Journal (38:1), pp. 141-160.
Meyer, M. H., and Lehnerd, A. P. 1997. The Power of Product
Platforms, New York: Free Press.
Miles, M. B., and Huberman, A. M. 1994. Qualitative Data
Analysis: An Expanded Sourcebook, Thousand Oaks, CA:
SAGE Publications.
Morel, B., and Ramanujam, R. 1999. “Through the Looking Glass
of Complexity: The Dynamics of Organizations as Adaptive and
Evolving Systems,” Organization Science (10:3), pp. 278-293.
Nambisan, S., Lyytinen, K., Majchrzak, A., and Song, M. 2017.
“Digital Innovation Management: Reinventing Innovation
MIS Quarterly Vol. 44 No. 1/March 2020 145
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Management Research in a Digital World,” MIS Quarterly
(41:1), pp. 223-238.
Nicolis, G., and Prigogine, I. 1989. Exploring Complexity, New
York: Freeman.
Page, S. E. 2010. Diversity and Complexity, Princeton, NJ:
Princeton University Press.
Parker, G., and Van Alstyne, M. 2018. “Innovation, Openness, and
Platform Control,” Management Science (64:7), pp. 3015-3032.
Parker, G., Van Alstyne, M., and Jiang, X. 2017. “Platform Eco-
systems: How Developers Invert the Firm,” MIS Quarterly
(41:1), pp. 255-254.
Patton, M. Q. 1990. Qualitative Evaluation and Research Methods,
Newbury Park, CA: SAGE Publications.
Robertson, D., and Ulrich, K. 1998. “Planning for Product
Platforms,” Sloan Management Review (39:4), pp. 19-31.
Rosenberg, N. 1969. “The Direction of Technological Change:
Inducement Mechanisms and Focusing Devices,” Economic
Development and Cultural Change (18:1), pp. 1-24.
Sambamurthy, V., and Zmud, R. 2000. “Research Commentary:
The Organizing Logic for an Enterprise’s IT Activities in the
Digital Era—A Prognosis of Practice and a Call for Research,”
Information Systems Research (11:2), pp. 105-114.
Simon, H. A. 1962. “The Architecture of Complexity,” Pro-
ceedings of the American Philosophical Society (106:6), pp.
Simpson, T. W. 2004. “Product Platform Design and Customi-
zation: Status and Promise,” Artificial Intelligence for
Engineering Design, Analysis and Manufacturing (18:1), pp.
Svahn, F., Mathiassen, L., and Lindgren, R. 2017. “Embracing
Digital Innovation in Incumbent Firms: How Volvo Cars
Managed Competing Concerns,” MIS Quarterly (41:1), pp.
Tanriverdi, H., and Lim, S.-Y. 2017. “How to Survive and Thrive
in Complex, Hypercompetitive, and Disruptive Ecosystems? The
Roles of IS-Enabled Capabilities,” in Proceedings of the 38th
International Conference on Information Systems, Seoul.
Tanriverdi, H., Rai, A., and Venkatraman, N. 2010. “Research
Commentary—Reframing the Dominant Quests of Information
Systems Strategy Research for Complex Adaptive Business
Systems,” Information Systems Research (21:4), pp. 822-834.
Thomas, L. D., Autio, E., and Gann, D. M. 2014. “Architectural
Leverage: Putting Platforms in Context,” Academy of Manage-
ment Perspectives (28:2), pp. 198-219.
Tiwana, A., Konsynski, B., and Bush, A. 2010. “Platform
Evolution: Coevolution of Platform Architecture, Governance
and Environmental Dynamics,” Information Systems Research
(21:4), pp. 675-687.
Van de Ven, A. H. 1992. “Suggestions for Studying Strategy
Process: A Research Note,” Strategic Management Journal
(13:S1), pp. 169-188.
Wheelwright, S. C., and Clark, K. B. 1992. Revolutionizing
Product Development: Quantum Leaps in Speed, Efficiency, and
Quality, New York: Simon and Schuster.
Yoo, Y., Boland, R. J., Lyytinen, K., and Majchrzak, A. 2012.
“Organizing for Innovation in the Digitized World,” Organi-
zation Science (23:5), pp. 1398-1408.
Yoo, Y., Henfridsson, O., and Lyytinen, K. 2010. “Research
Commentary: The New Organizing Logic of Digital Innovation:
An Agenda for Information Systems Research,” Information
Systems Research (21:4), pp. 724-735.
About the Authors
Johan Sandberg is an associate professor at the Swedish Center for
Digital Innovation at the Department of Informatics, Umeå
University. His research interests relate to the transformative effects
of digital innovation on organizing. His research has been pub-
lished in journals such as Journal of the Association for Information
Systems, European Journal of Innovation Management, and
Business & Information Systems Engineering.
Jonny Holmström is a professor at the Swedish Center for Digital
Innovation at the Department of Informatics, Umeå University. He
is the director and co-founder of Swedish Center for Digital Inno-
vation and writes, consults, and speaks on topics such as digital
innovation, digital transformation, and digital entrepreneurship. His
work has appeared in journals such as Communications of the AIS,
Design Issues, European Journal of Information Systems, Informa-
tion and Organization, Information Systems Journal, Information
Technology and People, Journal of the AIS, Journal of Information
Technology, Journal of Strategic Information Systems, Research
Policy, and The Information Society.
Kalle Lyytinen (Ph.D., Computer Science, University of Jyväskylä;
Dr. h.c. Umeå University, Copenhagen Business School, Lap-
peenranta University of Technology) is Distinguished University
Professor and Iris S. Wolstein Professor of Management Design at
Case Western Reserve University, and a distinguished visiting
professor at Aalto University, Finland. He is among the top five IS
scholars in terms of his h-index (85); he is a LEO Award recipient
(2013), AIS Fellow (2004), and the former chair of IFIP WG 8.2
“Information Systems and Organizations.” He has published more
than 400 refereed articles and edited or written over 30 books or
special issues. He conducts primarily research on digital innovation
and transformation, design work, requirements in large systems, and
the emergence of digital infrastructures.
146 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Appendix A
Data Collection
We collected data in two periods (2007–2010 and 2011–2017) through interviews, participant observation, and inspection of both internal and
publicly available archival data. Indications of the transformations triggered by digitization emerged during our study that prompted an
exploratory case study (Eisenhardt 1989; Yin 2003) of ABB’s control system evolution. The first data collection period involved field
observations at five sites (three ABB offices and two mining plants) and 14 interviews (10 with ABB staff with positions including sales
manager, mining automation unit manager, project manager, and process engineer, and four with informants from a large mining company:
a site IT manager, site maintenance manager, and two technicians. We collected extensive documentation, such as product specifications,
market analyses, project reports, and descriptions of installations at user sites.
During the second data collection period, we followed a project aimed at identifying ways of increasing certified solutions from third-party
developers to offer new functions for ABB’s 800xA process automation product. This period included nine interviews with third-party
developers and representatives of process industries (pulp and mining companies), participation in meetings and workshops, and collection of
documents (both ABB internal reports and public documents). We attended several project meetings and workshops on lessons from historical
design decisions and the concept’s development, including three days of training on development and integration. Finally, we gathered
substantial amounts of both internal and external archival data such as ABB’s annual reports, technical specifications for each product release,
related market analyses, and product pamphlets.
In each interview, we asked semi-structured open-ended questions (Spradley 1979) about informants’ roles and responsibilities, the product
platform’s history and status, and the informants’ interactions with it. We recorded and transcribed each interview, which lasted on average
an hour (with significant variance upward), compiled researchers’ notes of meetings and workshops, and verified transcripts’ correctness with
the interviewees.
Eisenhardt, K. 1989. “Building Theories from Case Study Research,” Academy of Management Review,(14:4), pp. 532-550.
Spradley, J. P. 1979. The Ethnographic Interview, New York: Holt, Rinehart and Winston.
Yin, R. K. 2003. Case Study Research: Design and Methods (3rd ed.), Thousand Oaks: SAGE Publications.
MIS Quarterly Vol. 44 No. 1/March 2020 147
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Appendix B
Coding Extracts
Table B1. Phase Transition 1
Construct Dimension Illustrative Examples from the Data
Digital Design Rules
“It provided a low-level interface in C and C++ and ran on HPUX. Since then, OMF has been used
in Advant OCS and MES as the communication middleware. Several thousand Advant OCS have
been sold in recent years. In 1993 and 1994 the focus was on performance and stability
improvements. In 1995 the first application frameworks (C++, Python) were developed and a port
to OpenVMS was made for a customer. 1996 was the year when the Smalltalk framework was
finished and a port to Digital Unix was introduced.
Data homogenization
“OMF makes process and production data available to the majority of computer programmers
and users, even those not necessarily involved in the industrial control field. For instance, it is
easy to develop applications in Microsoft Word, Excel, and Access to access process information.”
“One major problem OMF had to solve was the support for heterogeneous environments
(platforms, languages, communication protocols, field busses etc.). At this time it was recognized
that the commitment to a full object-oriented approach promised the biggest advantages with
regard to modularity, extendibility, portability and interoperability.”
Distributedness “We wanted to create a common engineering environment where information was shared, so we
created a large database in Oracle that the different applications could interact with.”
Connectedness “Our control system builder can work toward the data base, CAD for electricity, piping and
instrument diagrams. Then you can start re-using libraries etc.”
“Products such as programmable control systems use sophisticated sensor technology to
monitor, regulate and optimize a wide variety of technical and environmental control
processes in factories, shipyards, theatres and even private homes.”
“The system is now developed further using components based on new, standard technologies.
During this development, further new components become available on the market. ABB faced
this issue more than once.”
Mutual dependencies
“A typical example of such an incompatible change, is a change in the communication protocol
between OMF clients and servers. All versions of OMF must be able to talk to each other to make
the system flexible and open”
Interaction rules
“Advant OCS can be configured in a multitude of ways, depending on the size and complexity
of the process. The initial investment can consist of stand-alone process controllers and,
optionally, local operator stations for control and supervision of separate machines and process
sections. Subsequently, several process controllers can be interconnected and, together with
central operator and information management stations, build up a control network. Several control
networks can be interconnected to give a complete plant network which can share centrally
located operator, information and engineering workplaces.”
Design control
“A couple of years after we released Advant we ran into trouble. The system integrated data, but
we couldn't get other actors to change their data format, which meant that we had to assume
responsibility for data storage for a lot of external systems .… we wanted to develop the
functionality, but in doing so we assumed responsibility for developing other actors’
components. We realized pretty soon that it wasn't sustainable.”
“But even for a big company like ABB, with more than 220,000 employees, the time for proprietary
solutions expires. Indeed only functionality, which is not applicable as COTS (components off
the shelf) will be developed and maintained by ABB in the future.”
Note: The highlighted text (bold italics) exemplifies passages from the data that served as basis for the coding.
148 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Table B2. Phase Transition 2
Construct Dimension Illustrative Examples from the Data
Digital Design Rules
“All functionality was then realized as aspect systems. We built aspect systems, graphical
builders, graphical presentations, history modules, trend modules, reporting modules, control
builders, all functionality in the control system was integrated through this platform.”
Data homogenization
“We intend to achieve this common architecture by using software to link our various technical
platforms together and by improving compatibility in succeeding generations of each of our
products and systems.
Decoupling “So we developed APIs for our system and an adaptation module that could hook into APIs in
other systems."
“In an inventory in 2000, we went through all the customers’ systems and created our own
database, which was great for a year or so. Then the customers started replacing parts, or
maybe some consultant or some other company did. Then the information we created was
basically useless. This kind of information is great, but updating procedures must be more or less
fully automatic. Every little thing has some kind of intelligence, everything with a plug has
“We also intend to expand our Industrial IT offering. This means, in addition to creating a single
Industrial IT architecture as discussed above, we will increasingly emphasize products and
systems that can link our business processes with those of our customers.
“At about this time we started thinking that this idea of holding information together was interesting
for a lot of actors, not only engineers. It´s interesting for operators, maintenance staff, production
managers etc. ... and then we started to think about the next generation of control systems. The
basic architectural philosophy emerged during these years; at first we thought we were developing
an integration platform for engineering tools, then for connecting all the different parts of the
control system, then to integrate different products and units in ABB, and eventually to integrate
the whole world.
”We had discussions with large construction companies, weapons manufacturers. We formed
alliances with Microsoft, Intel and Accenture. In addition to using it for internal components, we
wanted to certify external components with business models based on licensing.”
“All functionality was then realized as aspect systems. We built aspect systems, graphical
builders, graphical presentations, history modules, trend modules, reporting modules, control
builders, all functionality in the control system was integrated through this platform.
Mutual dependencies “…developing a common architecture across the range of our products and systems so that they
can be easily combined with each other and with our customers’ systems.”
Interaction rules
“So we developed APIs for our system and an adaptation module that could hook into APIs in
other systems. Then we didn’t have to care, applications could store data wherever they wanted,
we just made sure we could access the data they produced through the APIs they published.”
Design control
“Then the idea was, OK, we have this platform for our control system. That is an architecture and
integration platform aimed at internal use. Why don’t we also use it externally? Systems,
products, components, the idea was that you should be able to certify everything. We realized
that this was an extremely good concept and that we could start certifying external products too.
We could start certifying the PCs that we build our system on, network components, routers,
gateways, switches, stuff like that. We can certify software that is to be connected.”
“It turned into a competition in ABB about who had most certified products. This is where it got out
of hand. Then the division manager became the CEO of ABB. Everybody who wanted to be
something had to talk about Industrial IT and AIP, whether they knew what it was or not … in
one of the annual reports, ABB was a huge company with about 200,000 employees, but the only
product mentioned was the Aspect Integrator Platform.”
Note: The highlighted text (bold italics) exemplifies passages from the data that served as basis for the coding.
MIS Quarterly Vol. 44 No. 1/March 2020 149
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Table B3. Phase Transition 3
Construct Dimension Illustrative Examples from the Data
Digital Design Rules
“800xA Performance Improvements (since 5.1)
Reduced Engineering New CDE editor, new Batch editor, FD
Improvements, Electrical Integration, Foundation Fieldbus …
More Opportunity Multisystem Integration for larger systems, Asset Optimization, Alarm Mgt, Control
Room(EOWs), History, SmartClient, new interfaces
More Communications Interfaces
Reduced server computers, Virtualized Servers and clients
More sanctioned combinations
Fewer Connectivity Servers CS Server footprint
IEC 61850 throughput FF Subnets per CS
Fewer Controllers Required ×2-3 AC 800M performance
IAC for automation and safety
More effective deployment of IEC61850
Improved availability Vmware vmotion high availability, sans storage
Redundant Ethernet IP
Larger system support ×2 no of network areas
X2 tags in system
x2 no of clients, MSI supported features
Fewer cabinets needed, Digital High density S800 I/O, new power supplies,”
“ABB has developed a component that bridges this language divide. The Fieldbus Plug (FBP) is a
compact accessory that takes information from equipment in the field and converts it to any industry-
standard protocol. It is, in effect, a translator or interpreter.
“The OPC connectivity that's part of the System 800xA’s integration platform enables connection of third-
party DCS controllers and PLCs. Once connected, the data becomes part of the system in the same way
as other integrated ABB hardware and software components.”
“A key element of this division’s product offering is its System 800xA process automation platform. This
product extends the capability of traditional process control systems, introducing advanced functions such
as batch management, asset optimization and field device integration which plug in’ to a common user
environment. The same user interface may also be used to manage components of existing multiple ABB
control systems that have been installed in the market over approximately the past 25 years.”
At the core of System 800xA is its integration platform, which enables ABB to provide a powerful evolution
path for its large installed base of control systems to System 800xA Operations. There are significant
benefits in evolving to the latest hardware and software versions if you own one of ABB’s traditional control
systems. However, these benefits extend even further when integrating other equipment such as 3rd
party controllers and PLCs.”
150 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Seamless integration to various ABB and third-party systems:
ABB Control Systems: Infi90, Melody, MOD 300, DCI, Freelance, DCI, Satt-line
Third-party controllers: OPC, Honeywell TDC3000, PLCs
BatchManagement, AssetOptimization, Device Integration, Information Management
EnterpriseConnect, CMMSIntegration, VideoOne, Matrikon AlarmInsights, OPC History Access for
Trends, Documentation.”
“As a pioneering technology leader in digital solutions, with an installed base of more than 70 million
connected devices and 70,000 control systems, ABB is uniquely positioned to support its customers’
digital transformation.”
“A consultant that I really respect told us this. Ten years ago, I knew about the whole system. Today, I
understand the controller, the rest I don’t understand. Everything is so specialized so you only really
understand your little part. This means that trouble shooting involves ten-twelve persons directly, just
coordinating that is demanding.”
“A yearly subscription to the Automation Sentinel program provides the following deliverables:
Licenses for new versions of system software
Software maintenance updates
Extended support for System 800xA software versions, up to seven years
Technical phone support to assist in system problem troubleshooting
On-line website access for downloads to assist in system maintenance
Software updates
Firmware updates
User manuals
Software release notes
Product technical bulletins
Software security management:
Microsoft security patch validation status reports
Third-party virus scanner qualification
Personal computer hardware qualifications for compatible replacement PC models for new and existing
software versions
Device library management updates for System 800xA for PROFIBUS, FOUNDATION Fieldbus and
Auto notification by e-mail
Technical updates
Product release information”
“Previously, control systems were developed in-house by different suppliers. More or less everything was
developed in-house, something on all circuit boards, they built and assembled it themselves. Now
Windows is used as a foundation, servers … and we know what is happening in the world of PCs,
everything spins faster and faster. New operating systems and new versions. And we’re seeing this
development in control systems as well. You're forced to update more frequently, even control systems.”
“The challenge for manufacturers is that different fieldbus devices often speak entirely different languages,
depending on the standard communications protocol they use. One might use a protocol called DeviceNet;
another Profibus; a third might use the AS-interface protocol. Most devices cannot be interchanged without
modification. ABB has developed a component that bridges this language divide. The Fieldbus Plug
(FBP) is a compact accessory that takes information from equipment in the field and converts it to any
industry-standard protocol. It is, in effect, a translator or interpreter. ABB believes the FBP will have
significant implications for its low-voltage products business, currently worth some $2 billion a year.”
Design control
“Automation Sentinel is the system lifecycle management program that extends support for, and the value
of, existing ABB control systems, protecting our customers' system investments. The program provides the
best overall ROI for past, present and future automation control system software expenditures. The
Automation Sentinel Program assists system owners in actively managing their lifecycle system costs
and investments. With this program, system owners can decide when to update to newer versions of
system software based on their system lifecycle plan and business objectives. In addition, customers
receive consistent support through the complete lifecycle of their system.”
“‘There is a constant flow of upgrades from ABB’s side, not only from them actually but also from
Microsoft, a lot of security-related updates.’ Most updates are done to avoid problems, but there are of
course also improvements in functionality. Between versions four and five for example, there are functional
improvements. But there are also minor ones and service packs, they are often corrections when they
discover problems.”
Note: The highlighted text (bold italics) exemplifies passages from the data that served as basis for the coding.
MIS Quarterly Vol. 44 No. 1/March 2020 151
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Appendix C
Table C1. A Functional Model of Process Production Systems: The Purdue Reference Model as Applied
in ISA-99 (adapted from CISCO 2013)
Zone Layer Description
Level 5: Enterprise
Centralized IT systems and functions, Enterprise Resource Management, business-to-business and
business-to-customer services
Level 4: Site
Business Planning
Extension of the enterprise network, basic business administration performed through standard IT
services. Access to internet, e-mail and enterprise applications. Non-critical plant systems such as
manufacturing execution systems and plant reporting such as inventories.
Demilitarized Zone: Provides a buffer zone where services and data can be shared between Manufacturing and Enterprise zones
Level 3: Site
Operations and
The highest level of the distributed control system, manages plant-wide automation functions. Reporting
data such as cycle times and predictive maintenance, detailed production scheduling, asset and material
management, control room workstations, patch launch server, file server. Applications primarily based
on standard computing equipment and operating systems.
Level 2: Area
Applications and functions associated with supervision and operation of each area such as operator
interfaces, alarms, and control room workstations. Communicates with controllers in level 1 and shares
data with levels 3 and/or 4 and 5 via the DMZ.
Level 1: Basic
Process Control
Controllers that steer automation of the process based on input from level 0.
Level 0: Process Input and output units such as sensors and actuators that measure and perform the functions of the
manufacturing system. Solutions that can operate without alteration for extended periods.
CISCO. 2013. Converged Plantwide Ethernet (CPwE) Design and Implementation Guide (
verticals/CPwE/CPwE_chapter2.html#wp1002608; accessed February 6, 2013).
152 MIS Quarterly Vol. 44 No. 1/March 2020
Sandberg, Holmström, & Lyytinen/Digitization & Phase Transitions
Appendix D
Table D1. Summary of Platform Phases
Characteristic Phase 1: Master Phase 2: Advant Phase 3: AIP Phase 4: 800xA
New functionality for
reducing complicated
interactions among
physical devices.
New functionality for
reducing complicated
interactions among specific
information management
components through tight
New functionality for
taming complexity asso-
ciated with connectivity
among non-specified
information management
components through loose
New functionality for
interactions with foreseen
and unforeseen agents
and assets in the
Firm (mainly division level)
due to internal design and
Firm (mainly division level)
and strategically selected
suppliers (HP, Unix,
collection of information
system providers).
Firm level and ecosystem
with certified external
Unbounded and dynamic
ecosystem of agents with
large variations in
contextual anchoring.
Firm capabilities. Firm, suppliers, Unix
developers, and selected
information management
system providers.
Potentially unlimited pool
of external capabilities.
Potentially unlimited pool
of external capabilities.
Nature of
Fixed and closed, one-
to-one mapping.
Focused on physical
Adapted to internalize
data from external
systems, one-to-one
Focused on engineering
(site configuration).
API’s and adaptation
module to exchange
data, many-to-many
Core of the firm’s
business strategy.
Based on general
purpose IT standards,
Focused on connec-
tivity within and across
sites, and toward
general purpose IT
Architecture Device modularization
and loose coupling in
physical environment
between devices and
Vertical increase in
connectedness at PRM
levels 0-2 (Process,
Basic Process Control,
Area Supervisory
Services and devices
loosely coupled to
increase adaptiveness
toward general purpose
IT components, tight
coupling between
services and content.
Increases in connected-
ness horizontally at PRM
levels 0-2 (Process,
Basic Process Control,
Area Supervisory
Control), and vertically
toward level 3 (Site
Operations and Control).
Loose coupling
between services and
content, and among
Vertical and horizontal
increases in connected-
ness, PRM levels 4
(Site Business
Planning) and 5
(Enterprise Network)
Tight coupling to
Windows OS. Open
and disclosed inter-
faces, shared
Horizontal expansion in
connectedness at user
sites. Vertical expan-
sion toward (e.g.,
remote clients). Lateral
expansion toward the IT
Governance Design decision rights
and control retained
Design decision rights
and control retained for
hardware in PRM level 1
(Basic Process Control)
but distributed to sup-
pliers for level 2 (Area
Supervisory Control).
For software integration
high levels of distribution
due to tight couplings
that ABB has to
Attempts to implement
formal input control
measures through
certification program
and decrease
dependencies on
components with
distributed decision
Significant decision
rights distributed to
Microsoft. Input control
through substantial
testing and certification
of software versions.
MIS Quarterly Vol. 44 No. 1/March 2020 153
154 MIS Quarterly Vol. 44 No. 1/March 2020
... An increasing number of established companies across industries have introduced innovation platforms to create flourishing ecosystems of applications (Khanagha et al., 2020;Reynolds & Yetton, 2015;Sandberg et al., 2020;Sebastian et al., 2017;Svahn et al., 2017). Innovation platforms provide an extensible software core on top of which third parties -also referred to as complementors -create applications (Cusumano et al., 2019;Tiwana et al., 2010). ...
... Building on these findings on platform governance, IS scholars have studied established companies that have implemented innovation platforms or evolved corporate IT platforms into more open innovation platforms (Gawer, 2014;Gawer & Cusumano, 2002;Sandberg et al., 2020;Svahn et al., 2017). They contend that such a strategy increases the complexity of interactions as a hierarchy-based organisational logic evolves into an ecosystem-centred organisational logic (Gawer, 2014;Sandberg et al., 2020). ...
... Building on these findings on platform governance, IS scholars have studied established companies that have implemented innovation platforms or evolved corporate IT platforms into more open innovation platforms (Gawer, 2014;Gawer & Cusumano, 2002;Sandberg et al., 2020;Svahn et al., 2017). They contend that such a strategy increases the complexity of interactions as a hierarchy-based organisational logic evolves into an ecosystem-centred organisational logic (Gawer, 2014;Sandberg et al., 2020). Consequently, the focus of IT governance shifts from hierarchies to "ecosystem governance" (Gawer, 2014(Gawer, , p. 1244 or "innovation governance" (Svahn et al., 2017, p. 249), thereby generating competition between established companies and external complementors (Gawer, 2014;Svahn et al., 2017;see also Förderer et al., 2018). ...
... Today, firms are increasingly integrating digital components (e.g., software) into physical product environments (e.g., cars) to tap into new avenues for value creation (e.g., self-driving cars) Hukal and Henfridsson 2018). To leverage the integrative capacity of digital technologies, firms continuously tune and dynamically reconfigure technical and organizational resources towards certain goals (Sandberg et al. 2020). This is a challenging task, however, since existing resources together with accumulated knowledge bases are catered towards existing products that are characterized by stable revenue prospects or market boundaries (March 1991;O'Reilly and Tushman 2013). ...
... In other words, theorizing how or whether organizing digital innovation leads to digital transformation is a worthwhile research inquiry (Gregory et al. 2019). Similarly, from an operational viewpoint, prior studies have generally suggested that as firms integrate digital components into physical products, they transition from an internal production-oriented logic to an external supply-chain innovation organizing logic (Thomas et al. 2014;Sandberg et al. 2020). Yet, how such changes in one layer, (e.g., technology) causes ripple effects in other layers (e.g., strategy or internal firm structures), has received limited attention (Sandberg et al. 2020;Gregory et al. 2019). ...
... Similarly, from an operational viewpoint, prior studies have generally suggested that as firms integrate digital components into physical products, they transition from an internal production-oriented logic to an external supply-chain innovation organizing logic (Thomas et al. 2014;Sandberg et al. 2020). Yet, how such changes in one layer, (e.g., technology) causes ripple effects in other layers (e.g., strategy or internal firm structures), has received limited attention (Sandberg et al. 2020;Gregory et al. 2019). Understanding the interconnection of these layers is important for identifying the drivers and trajectories of digital transformation and better managing its emergent and contingent nature (Staykova 2018;Warner and Wäger 2019). ...
Conference Paper
Full-text available
The development of digital product platforms is a prevailing trend in many industries. As firms incorporate digital technologies into established product categories, they need to manage tensions on multiple organizational layers including strategy, technology and structure. Recent findings suggest that tensions are most likely to be managed by creating resource and coordination flexibility. This paper reports a longitudinal case study of developing a digital product platform. By drawing on organizational ambidexterity theory, we identify four mechanisms-re-scripting, centralizing, redirecting and decoupling-through which the firm creates resource and coordination flexibility. The resulting resource and coordination flexibility in turn lead to the transformation of the firm's strategy, technology and internal structure. The contribution of this study is in adopting an internal perspective and a bottom-up approach which help to theorize the evolution of digitized products into digital product platforms in an emergent way.
... In contrast, to repeatedly studied ecosystems (Saarikko, 2016;Hannah and Eisenhardt, 2018), the cocreated outcome in the AM ecosystem remains physical, although the input is partially digital. Whilst ecosystems based on software, services or digital products seems to be a well-researched field, ecosystems with physical outcome need a distinct look in some terms (Autio and Thomas, 2016;Nischak and Hanelt, 2019;Sandberg et al., 2020). Physical products differ from software on the basis of various properties. ...
... Physical products differ from software on the basis of various properties. Examples of differences are the need for interface standardization or the possibility of generating indirect network effects, as is the case with digital products as reprogramming at the device level is not possible (Jacobides et al., 2018;Sandberg et al., 2020). Against this background, we propose the following research question: Which determinants shape the ecosystem in AM, and how can they contribute to design business models in a generic AM ecosystem? ...
Conference Paper
Full-text available
Despite the increasing managerial awareness for ecosystems to organize complex value propositions, little is known about how different roles can establish their business models (BM) in ecosystems. AM drives innovations in the product design and manufacturing fields predominantly across companies, indicating the eco-systemic organization of value creation without orchestrating and dominant keystone actors yet. This paper explores ecosystem determinants by analyzing the dynamic additive manufacturing (AM) paradigm. We conduct an empirical study with companies from the AM domain to visualize their value activities and define generic roles in the interdependent value creation process, adopting the e³-value methodology. By exploring these ecosystem determinants, our results aid practitioners in positioning their BMs in the AM domain and generate descriptive insights for the orchestrator BM design in a dynamic domain without orchestrating keystones.
... Debates about digital technologies' effects on organizations and on the nature of work have intensified recently. This may largely be due to the rapid emergence and scaling of new digital technologies (Huang et al., 2017), together with new and increased concerns about the consequences associated with their increasing levels of complexity (Sandberg et al., 2020). The new digital technologies developed and implemented in recent years have had major effects, with largescale transformations of occupations, job roles, and entire industries (Acemoglu & Autor, 2011;Mokyr et al., 2015). ...
Full-text available
Strategies and means for selecting and implementing digital technologies that realize firms' goals in digital transformation have been extensively investigated. The recent surge in artificial intelligence (AI) technologies has amplified the need for such investigation, as they are being increasingly used in diverse organizational practices, creating not only new opportunities for digital transformation but also new challenges for managers of digital transformation processes. In this article, I present a framework intended to assist efforts to address one of the first of these challenges: assessment of organizational AI readinessdthat is, an organi-zation's ability to deploy AI technologies to enable digital transformation, in four key dimensions: technologies, activities, boundaries, and goals. I show that this framework can facilitate analysis both of an organization's current sociotechnical AI status and of the prospects for the technology's fuller value-adding, sociotech-nical deployment. The AI readiness framework invites fuller theorizing of the roles that AI candand willdplay in digital transformation.
... In terms of their technical maturity, one can distinguish between the provision of searchable app catalogs, app stores that are capable of installing complements, and integrations of external app stores in the IIoT platform [11,28]. According to extant research, an industrial app store that is integrated into an IIoT platform primarily serves the following four goals [11,22,28,29]: ...
Conference Paper
Full-text available
App stores represent an already established concept for distributing software-intensive solutions that complementary extend the value of digital platforms. In the context of industrial digital platforms in the IIoT domain, the benefits of app stores are not yet fully understood or shared by the majority of platform users. This article contributes to the instantiation of app stores in digital industrial platforms using empirical data to reveal the functionalities expected from an industrial app store by IIoT platform users on the one hand and IIoT platform providers on the other. The increased understanding of the app stores integrated in IIoT platforms provides a design-oriented gain in knowledge for research and practice. Based on an empirical survey, we analyze which functionalities of app stores are demanded by IIoT platform users. This enables us to formulate five propositions for adapting app stores for IIoT to increase their success in this domain.
... (Baiyere et al., 2020, Markus and Loebbecke, 2013, Carlo et al., 2012, designing digital platforms (c.f. (Henningsson et al., 2021, Sandberg et al., 2020, Singh and Hess, 2017, Granados and Gupta, 2013, effects on organisational culture (c.f. Dremel et al., 2017, (Karimi andWalter, 2015) and generating value (c.f. ...
... As a result of this study, human capabilities, the competitiveness of leaders, core talent and subject were ascertained to be the most critical factors. Many previous studies [39,62,63] stress that the success of digital transformation strategies mainly depends on selection and concentration through digital human resource fostering and organizational change management. If new organizations such as AI research institutes and big data management organizations for digital transformation strategies are set up, cooperation and conflicts with existing organizations are necessary. ...
Full-text available
Firms’ digital environment changes and industrial competitions have evolved quickly since the Fourth Industrial Revolution and the COVID-19 pandemic. Many companies are propelling company-wide digital transformation strategies based on artificial intelligence (AI) technology for the digital innovation of organizations and businesses. This study aims to define the factors affecting digital transformation strategies and present a decision-making model required for digital transformation strategies based on the definition. It also reviews previous AI technology and digital transformation strategies and draws influence factors. The research model drew four evaluation areas, such as subject, environment, resource, and mechanism, and 16 evaluation factors through the SERM model. After the factors were reviewed through the Delphi methods, a questionnaire survey was conducted targeting experts with over 10 years of work experience in the digital strategy field. The study results were produced by comparing the data’s importance using an Analytic Hierarchy Process (AHP) on each group. According to the analysis, the subject was the most critical factor, and the CEO (top management) was more vital than the core talent or technical development organization. The importance was shown in the order of resource, mechanism and environment, following subject. It was ascertained that there were differences of importance in industrial competition and market digitalization in the demander and provider groups.
The main trend of digital development of the transport industry is platformization. Digital platforms are able to integrate all transport and logistics processes into a single system, and the expansion of interfaces and created values allows the platform to grow into an ecosystem. However, today there is no clear answer to the question of how free/restricted access to the digital transport platform should be for both service providers (service operators) and consumers. The purpose of the article is to substantiate the architecture of building digital platforms and ecosystems focused on the transport environment, in which any service can be easily included with the full preservation of existing functionality and the potential for its development, without losing its identity and subscriber base. The research is based on the platform ecosystem approach, as well as the stages of the evolution of digital platforms in the B2B industrial context, which have already been described in modern literature. The article presents the prospects for the development of digital platforms with a "transport core" and transformation into an open digital ecosystem. This study expands views on the evolution of platform transport ecosystems and the development of digital transport and related services.
Dem soziologischen Digitalisierungsdiskurs mangelt es an zweierlei: einerseits an einem Dialog zwischen bestehenden theoretischen Ansätzen; andererseits an einer systematischen Bestimmung des Status von Organisation im Digitalisierungsprozess. Beiden Mängeln wird im vorliegenden Beitrag Abhilfe geschaffen, indem auf Basis bestehender soziologischer Digitalisierungsansätze ein Vier-Phasen-Modell des Prozesses der Digitalisierung angefertigt wird. Neben Technisierung, Datafizierung und Vernetzung wird Organisation als Formalisierung dabei eine wesentliche Rolle spielen. Die Regel für die Verknüpfung der vier Phasen liefert der Sinnbegriff aus der Systemtheorie Niklas Luhmanns. Mithilfe eines derart ausgearbeiteten Digitalisierungsansatzes wird es möglich, den Blick stärker als bisher für drei empirische Fragen zu sensibilisieren: 1) Wie verschiebt sich gegenwärtig das praktisch wirksame Bild der Digitalisierung? 2) Welchen Einfluss haben die digitalisierenden Organisationen bei dieser Verschiebung? 3) Inwiefern geht mit dieser Verschiebung grundsätzlich ein verändertes Bild des Organisationalen einher?
Full-text available
Digital platforms can be opened in two ways to promote innovation and value generation. A platform owner can open access for third-party participants by establishing boundary resources, such as APIs and an app store, to allowcomplements to be developed and shared for the platform. Furthermore, to foster cooperation with the complementors, the platform owner can use an open-source license boundary resource to open and share the platform's core resources. However, openness that is too wide renders the platform and its shared resources vulnerable to strategic exploitation. To our knowledge, platform strategies that promote such negative outcomes have remained unexplored in past research. We identify and analyze a prominent form of strategic exploitation called platformforking in which a hostile firm, i.e., a forker, bypasses the host's controlling boundary resources and exploits the platform's shared resources, core and complements, to create a competing platform business. We investigate platform forking on Google's Android platform, a successful open digital platform, by analyzing the fate of five Android forks and related exploitative activities. We observe several strategies that illustrate alternative ways of bundling a platform fork from a set of host, forker, and other resources. We also scrutinize Google's responses, which modified Android's boundary resources to curb exploitation and retain control. In this paper, we make two contributions. First, we present a theorization of the competitive advantage of open digital platforms and specifically expose platform forking as an exploitative and competitive platform strategy. Second, we extend platform governance literature by showing how boundary resources, which are mainly viewed as cooperative governance mechanisms, are also used to combat platform forking and thus sustain a platform's competitive advantage.
Full-text available
In the last few years leading-edge research from information systems, strategic management, and economics have separately informed our understanding of platforms and infrastructures in the digital age. Our motivation for undertaking this special issue rests in the conviction that it is significant to discuss platforms and infrastructures concomitantly, while enabling knowledge from diverse disciplines to cross-pollinate to address critical, pressing policy challenges and informstrategic thinking across both social and business spheres. In this editorial,we reviewkey insights from the literature on digital infrastructures and platforms, present emerging research themes, highlight the contributions developed from each of the six articles in this special issue, and conclude with suggestions for further research.
Full-text available
Suppose that a firm in charge of a business ecosystem is a firm in charge of a microeconomy. To achieve the highest growth rate, how open should that economy be? To encourage third-party developers, how long should their intellectual property interests last? We develop a sequential innovation model that addresses the trade-offs inherent in these two decisions: (i) Closing the platform increases the sponsor's ability to charge for access, while opening the platform increases developer ability to build upon it. (ii) The longer third-party developers retain rights to their innovations, the higher the royalties they and the sponsor earn, but the sooner those developers' rights expire, the sooner their innovations become a public good upon which other developers can build. Our model allows us to characterize the optimal levels of openness and of intellectual property (IP) duration in a platform ecosystem. We use standard Cobb-Douglas production technologies to derive our results. These findings can inform innovation strategy, choice of organizational form, IP noncompete decisions, and regulation policy.
Full-text available
As digital platforms are transforming almost every industry today, they are slowly finding their way into the mainstream information systems (ISs) literature. Digital platforms are a challenging research object because of their distributed nature and intertwinement with institutions, markets and technologies. New research challenges arise as a result of the exponentially growing scale of platform innovation, the increasing complexity of platform architectures and the spread of digital platforms to many different industries. This paper develops a research agenda for digital platforms research in IS. We recommend researchers seek to (1) advance conceptual clarity by providing clear definitions that specify the unit of analysis, degree of digitality and the sociotechnical nature of digital platforms; (2) define the proper scoping of digital platform concepts by studying platforms on different architectural levels and in different industry settings; and (3) advance methodological rigour by employing embedded case studies, longitudinal studies, design research, data-driven modelling and visualisation techniques. Considering current developments in the business domain, we suggest six questions for further research: (1) Are platforms here to stay? (2) How should platforms be designed? (3) How do digital platforms transform industries? (4) How can data-driven approaches inform digital platforms research? (5) How should researchers develop theory for digital platforms? and (6) How do digital platforms affect everyday life?
Full-text available
Past research provides instructive yet incomplete answers as to how incumbent firms can address competing concerns as they embrace digital innovation. In particular, it offers only partial explanations of why different concerns emerge, how they manifest, and how firms can manage them. In response, we present a longitudinal case study of Volvo Cars’connected car initiative. Combining extant literature with insights from the case, we argue that incumbent firms face four competing concerns—capability (existing versus requisite), focus (product versus process), collaboration (internal versus external), and governance (control versus flexibility)—and that these concerns are systemically interrelated. Firms must therefore manage these concerns cohesively by continuously balancing new opportunities and established practices.
Full-text available
Rapid and pervasive digitization of innovation processes and outcomes has upended extant theories on innovation management by calling into question fundamental assumptions about the definitional boundaries for innovation, agency for innovation, and the relationship between innovation processes and outcomes. There is a critical need for novel theorizing on digital innovation management that does not rely on such assumptions and draws on the rich and rapidly emerging research on digital technologies. We offer suggestions for such theorizing in the form of four new theorizing logics, or elements, that are likely to be valuable in constructing more accurate explanations of innovation processes and outcomes in an increasingly digital world. These logics can open new avenues for researchers to contribute to this important area. Our suggestions in this paper, coupled with the six research notes included in the special issue on digital innovation management, seek to offer a broader foundation for reinventing innovation management research in a digital world.
- This paper describes the process of inducting theory using case studies from specifying the research questions to reaching closure. Some features of the process, such as problem definition and construct validation, are similar to hypothesis-testing research. Others, such as within-case analysis and replication logic, are unique to the inductive, case-oriented process. Overall, the process described here is highly iterative and tightly linked to data. This research approach is especially appropriate in new topic areas. The resultant theory is often novel, testable, and empirically valid. Finally, framebreaking insights, the tests of good theory (e.g., parsimony, logical coherence), and convincing grounding in the evidence are the key criteria for evaluating this type of research.
Research summary: A substantial and burgeoning body of research has described the influence of platform‐mediated networks in a wide variety of settings, whereby users and complementors desire compatibility on a common platform. In this review, we outline extant views of these dynamics from the industrial organization (IO) economics, technology management, and strategic management perspectives. Using this review as a foundation, we propose a future research agenda in this domain that focuses the on the relative influence of network effects and platform quality in competitive outcomes, drivers of indirect network effects, the nature and attributes of complementors, and leveraging complementor dynamics for competitive advantage. Managerial summary: In many industries, such as social networks and video games, consumers place greater value on products with a large network of other users and a large variety of complementary products. Such “network effects” offer lucrative opportunities for firms that can leverage these dynamics to create dominant technology platforms. This article reviews current perspectives on network effects and the emergence of platforms, and offers several areas of future consideration for optimal strategies in these settings. Copyright © 2016 John Wiley & Sons, Ltd.