Content uploaded by Dimitri Petrik
Author content
All content in this area was uploaded by Dimitri Petrik on Sep 30, 2021
Content may be subject to copyright.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
1
Completed Research Paper
Dimitri Petrik
University of Stuttgart
Keplerstr. 17, 70174 Stuttgart Germany
dimitri.petrik@bwi.uni-stuttgart.de
Kleophas Model
University of Stuttgart
Keplerstr. 17, 70174 Stuttgart Germany
kleophas.model@bwi.uni-stuttgart.de
Lion Alexander Drebinger
University of Stuttgart
Keplerstr. 17, 70174 Stuttgart Germany
lion.drebinger@bwi.uni-stuttgart.de
Georg Herzwurm
University of Stuttgart
Keplerstr. 17, 70174 Stuttgart Germany
georg.herzwurm@bwi.uni-stuttgart.de
Abstract
To create customer value, IoT platform providers require complementors to utilize
boundary resources (BR) to create platform-based innovations. Although BRs are an
important determinant in the success formula of digital platforms, the exploration of the
satisfaction potential of BR used by IoT platforms is still in its infancy. In this paper, we
present the results of an empirical satisfaction study among complementors from the
Azure IoT ecosystem. Linking Kano’s theory of the attractive quality with the expectation-
confirmation model, we survey 83 complementors on their satisfaction with the nine
identified technical BR, designed and maintained by Microsoft. Our study offers insights
on the satisfactory potential of technical BR for IoT ecosystem participants. The results
explore the satisfaction potential of the technical BR, and thus support IoT platform
providers in the optimization of the development of technical BR, and enable further
research to derive implications for satisfaction-maximizing provisioning of BR for IoT
platforms.
Keywords: Internet of Things, IoT Platform Ecosystems, Boundary Resources,
Complementor Satisfaction, Kano Survey, Theory of Attractive Quality, Microsoft Azure
Introduction
The digitization of physical assets and processes already plays a central role in many companies for years,
driving the emergence of digital platform ecosystems (Porter and Heppelmann 2014; Beverungen et al.
2020; Parker et al. 2017). In the context of digitization, the Internet of Things (IoT) enables companies to
add value and solve more complex customer problems by combining physical and digital components.
Digitized products are subject to changing architectural rules. They foster the loose coupling of modules in
multi-layered architectures and enable more independent innovations for each module (Baldwin and
Woodard 2009). Digitized products also capture field data and make them available across firm borders to
create added value. These specific features of IoT architecture require an appropriate digital infrastructure,
which can be realized through IoT platforms. IoT platforms offer standardized functionalities for the
integration and management of field data and reduce the complexity of IoT applications’ implementation
(Yoo et al. 2010; Wortmann and Flüchter 2015; Sandberg et al. 2020).
If IoT platforms have open interfaces, they promote the emergence of IoT ecosystems, as complementary
players can develop IoT applications utilizing the accessible platform functionalities (Pauli et al. 2021).
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
2
Enterprise IT companies such as Microsoft, IBM, or SAP utilize the generative potential of third-party
innovation provided by complementary partners. Due to indirect network effects, the external innovation
driven by complementors increases the overall value of the platform ecosystem. Hence, the attraction of
third-parties to join the ecosystem and innovate is vital for platform companies. To foster the complementor
attraction, platform providers often organize third-parties in partner programs and support them through
various platform sponsoring activities (Eisenmann et al. 2009; Ceccagnoli et al. 2012; Yoo et al. 2012;
Engert et al. 2020; Marheine and Pauli 2020; Pauli et al. 2021). In order to maximize the generative effects,
platform companies are required to create an adequate system of rules and incentives for the ecosystem
participants. Thereby they become ecosystem orchestrators and guide the complementors towards the
desired direction (Boudreau and Lakhini 2009; Hein et al. 2019; Beverungen et al. 2020). Platform
boundary resources (BR) build a scientifically recognized concept to foster complementary innovation in
platform ecosystems. IoT platform providers can use BR as instruments to manage the balance between
generativity and restricting measures of the ecosystems’ innovation based on the platform (Ghazawneh and
Henfridsson 2013; Karhu et al. 2018; Hein et al. 2019). Although the information systems (IS) research
offers important findings on the various potentials of BRs (Eaton et al. 2015; Karhu et al. 2020; Hein et al.
2019), little is known about the BR configuration to increase the complementor attraction through
satisfaction. Complementors utilize BRs and evaluate their quality and usefulness during the innovation
process (Petrik and Herzwurm 2020a). It is already known that the satisfaction of users (i.e.,
complementors) results in the continued use of the system (i.e., IoT platform). As BRs represent the points
of contact between complementors and platforms, their quality acts as an antecedent of platform
satisfaction. Therefore, their quality affects the complementor attraction and the use of the IoT platform
(DeLone and McLean 2003; de Ruyter et al. 1997; Bhattacherjee 2001).
To tackle the question of how to attract complementary actors and trigger network effects in platform-based
IoT ecosystems, we examined BRs from the satisfaction perspective. By applying Kano’s theory of attractive
quality (Kano et al. 1984; Löfgren and Witell 2008), we conducted an empirical satisfaction study among
complementors in the Microsoft Azure IoT ecosystem, evaluating their experience with the utilized BR.
Overall, this paper presents the combination of two satisfaction measurement techniques: the Kano survey
and the evaluation of self-stated satisfaction. IS research already recognized that user satisfaction affects
the success of stakeholders - i.e., ecosystem actors such as platform providers (DeLone and McLean 2003).
Considering the preliminary work of Petrik and Herzwurm (2020a) on the analysis of BRs in industrial IoT
(IIoT), we noticed that almost all technical BRs (TBR) missed the potential to create exceptionally high
satisfaction. Consequently, this motivated us to explore this phenomenon more deeply, using only TBRs as
a research object in the survey (De Reuver et al. 2018). Hence, we explore the impact of TBRs on
complementor satisfaction in platform-based IoT ecosystems to answer the following research question:
How can TBRs affect the satisfaction of complementors in platform-based IoT ecosystems? From the
platform provider perspective, our findings offer a practical value for platform companies to understand
the satisfactory potential of TBRs as an instrument in IoT ecosystem development. Furthermore, our
findings contribute to the research on IoT platform ecosystem design and generate additional knowledge
on the BRs as levers for innovational processes (McIntyre and Srinivasan 2017; Jacobides et al. 2018). Our
findings also contribute to IS theory by paving the way to adopt the IS success model to the IoT platform
domain (DeLone and McLean 2003) since holistic theories with established constructs such as the IS
success model are still lacking in the context of IoT platform ecosystem design (Floetgen et al. 2021).
Background
Platform-based IoT Ecosystems
Prior work distinguishes at least three perspectives on digital platforms, taking into account the seven
recognized ecosystem concepts (Thomas et al. 2014; Kapoor 2018; Suominen et al. 2019). IoT platforms
combine both of Cusumano's value-added platform concepts (Pauli et al. 2021). We define IoT platforms as
a part of a digital infrastructure for processing data from digitized products and providing digital services.
Therefore, the concept of innovation platforms is most applicable to define them (Cusumano et al. 2019).
Innovation platforms provide and combine specific technological building blocks for the execution and
development of dependent peripheral modules. The primary purpose of innovation platforms is to solve
existing problems through complementary modules (Cusumano et al. 2019; Hevner and Malgonde 2019).
In addition to the innovation platform concept, IoT platforms fit the software platform definition by Tiwana
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
3
et al. (2010), defined as an “extensible codebase of a software-based system that provides core functionality
shared by the modules that interoperate with it and the interfaces through which they interoperate” (Tiwana
et al. 2010). Similarly, market-ready IoT platforms usually offer extensible software architectures to solve
the existing customer problems through IoT applications, often combining software and hardware
components (Wortmann and Flüchter 2015; Beverungen et al. 2020). Although an innovation platform
does not always have to be openly accessible to third-parties in the beginning, in order to cope with the
heterogeneity of IoT use cases and application areas (Cusumano et al. 2019; Endres et al. 2019; Leminen et
al. 2019), lots of IoT platform providers on the market offer third-parties access to the platform building
blocks (Petrik and Herzwurm 2020a; Beverungen et al. 2020; Pauli et al. 2021). Platform providers
encourage third-party users of the platform to contribute ‘vertical’ complements for the specific use cases.
Such a platform logic goes in line with the focus on core competencies for each value-creating actor in the
IoT ecosystem. The existing complexity of IoT applications is often depicted in multi-layered technology
stacks (Porter and Heppelmann 2014; Wortmann and Flüchter 2015). This logic allows the IoT platform
provider to combine ‘best-in-class’ complementors for each module of the IoT application. With the help of
the building blocks within the platform core, IoT platforms reduce the development effort for the
complementors and shorten their time-to-market, which leads to an increase in the overall profitability of
the complement creation (Kude et al. 2012; Koch and Kerschbaum 2014; Cusumano et al. 2019). IoT
platforms act as middleware systems, facilitating the interoperability between different networked assets
and provide the scalability needed to process large volumes of data (Guth et al. 2018; Mineraud et al. 2016).
Extending these core benefits of IoT platforms, the general logic of attracting complementors plays a key
role in the success of the platform, as indirect network effects apply for platform-based IoT ecosystems
(McIntyre and Srinivasan 2017; Schüler 2020). Hence, the value of an IoT platform is largely dependent on
the variety and the quality of the complements, created by the innovative and value-adding complementors
that result in the formation of the inherent ecosystem (Parker et al. 2017; Wortmann and Flüchter 2015;
Hein et al. 2020).
Ecosystems are defined by the interacting actors, who depend on each other's activities to organize the value
creation based upon an IoT platform. Usually, a platform ecosystem consists of a platform provider,
multilateral complementors who use the platform functionalities, and the end customers, who can
contribute and share sensor data from their digitized products in IoT (Adner et al. 2006; Jacobides et al.
2018; Anke et al. 2020). Complementors are independent actors in the ecosystem. They develop platform-
based IoT solutions in a certain arrangement with the platform provider, creating additional platform value
for the end customers (Kude et al. 2012; Gawer 2014; Parker et al. 2017; Jacobides et al. 2018). The variety
of the possible complementor types in IoT ecosystems is wide and may include heterogeneous inter-
industrial entities. On the one hand, software companies are a typical complementor type. They develop
platform-based IoT solutions for their customers. On the other hand, non-IT companies such as automation
solution providers can make their hardware natively compatible with the IoT platform and thus contribute
a complement to the IoT ecosystem. For a platform provider, it is critical to understand which types of
complementors increase the value propositions for the platform customers in order to be successful
(Cusumano et al. 2019; Hevner and Malgonde 2019; Petrik and Herzwurm 2020b). From this point of view,
IoT platforms also act as multi-sided markets and correspond to the economic transaction platform
concept. However, in the case of IoT platforms, the transactions between actors processed via the platform
are of secondary importance since the added value primarily comes from the complementary innovation
(Gawer 2014; Cusumano et al. 2019). IoT platforms, therefore, need interfaces for distributed value
creation, which in IS research are facilitated by BRs (Ghazawneh and Henfridsson 2010; Eaton et al. 2015).
Achieving Complementor Satisfaction with Boundary Resources
Boundary resources represent a possibility to open a digital platform technically and organizationally, by
enabling a platform provider to integrate the heterogeneous expertise areas of the complementors into the
platform-based value creation (Karhu et al. 2018; Hein et al. 2019). Recognizing the tensions and potential
for conflict between the heterogeneous complementors or the complementor and the platform provider,
Ghazawneh and Henfridsson (2010) have conceptualized BRs to make sense of how a platform provider
could govern the complementors in the ecosystem (Ghazawneh and Henfridsson 2010; Eaton et al. 2015).
The boundary object theory, developed by Star and Griesemer (1989), builds the theoretical background of
the initial BR model. The model explains the combination of centralized platform control and pooling of
distributed expertise areas and knowledge via technical (TBR) and social (SBR) boundary resources, and
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
4
therefore shapes an overall platform strategy (Ghazawneh and Henfridsson 2010; Ghazawneh and
Henfridsson 2013). In more detail, TBRs can be characterized as application BRs (ABR) and development
BRs (DBR). ABR are resources to establish the interaction between the complements and the platform core.
DBR are the development resources for developers, determined by the platform (Dal Bianco et al. 2014).
Earlier research recognizes the extensive use of BRs in the IIoT domain and identifies several domain-
specific BRs (Petrik and Herzwurm 2019). Skog et al. (2018) recognize that BRs can also be applied to end
customers, although most BRs stay invisible to them, which is why BRs are predominantly used by
complementors to create complements within boundaries that are set by the platform provider.
In particular, BRs integrate the complementary developers with the platform functionalities, and by that
enable them to utilize platform functionalities effectively to leverage platform-based IoT complements
(Bender 2020). As this applies to potential and existing complementors, platform providers may compete
for the engagement of complementary developers on the quality of BRs. In the context of IoT, it can even
be observed that, due to the immature and fragmented state of the IoT platform market (Turck 2018), the
BR activities of platform providers are mainly focused on the sourcing process and the transfer of the design
capabilities to the complementors, instead of their strict control (Hein et al. 2019; Petrik and Herzwurm
2020a). Holding a focal position in platform-based ecosystems, platform providers can define the design of
the BRs and evolve them dynamically, while the complementors use the offered BRs and evaluate their
quality during the complement creation (Eaton et al. 2015; Petrik and Herzwurm 2020a). The perceived
quality of BR can initiate different reactions among the complementors. If complementors perceive poor
performance or quality issues with the BRs, they rate the quality of the platform as low. They may therefore
abandon a platform to mitigate the risks during the implementation of IoT applications. Since enterprise
IoT use cases usually integrate critical business processes to exchange data with the platform (Pauli et al.
2021), complementors may consider the poor quality of BRs to jeopardize the successful realization of IoT
projects. Despite the switching costs, poor quality BRs encourage multihoming as the IoT platform market
offers alternatives with comparable functionality for complementors willing to migrate (Rochet and Tirole
2003; Turck 2018; PTC 2019). Conversely, high perceived quality of BRs, positively impacts the intention
to actively use the platform among the complementors. Moreover, other complementors may get attracted
by the quality of BR and join an ecosystem to innovate (DeLone and McLean 2003; Mylläerniemi et al.
2018). Therefore, both consequences promote the network effects, especially in the enterprise application
of IoT. The quality of the BRs is of significant importance, which is why it should be in the interest of the
platform providers to guarantee the highest possible quality of BRs in order to create an attractive platform-
based IoT ecosystem in the highly competitive IoT platform market (Zhu and Iansiti 2007; Schüler 2020).
Prior work even recognizes the possibility for platform providers to achieve value-driven lock-ins through
a particularly beneficial design and configuration of BR, using an exceptional BR design to implement
governance (Ofe 2018). Especially in their focal role in the platform ecosystem, platform providers are able
to set the design rules for BRs, which shape the satisfactory access points to the platform or remove the
barriers (Marheine and Pauli 2020). Mylläerniemi et al. (2018) have proven how the design of BR affects
developers in their consumption choice of a software system. Aside from this work, papers on the concrete
guidelines of the BR-based platform design and platform quality are still scarce and build an immediate
direction for platform research (Hein et al. 2019). So, in this study, we take this research avenue and bring
forward the platform research on BR design in terms of their quality and the associated complementor
satisfaction.
Conceptualization of Complementor Satisfaction
An increase in customer satisfaction is recognized as a competitive strategic objective with an influence on
market share and corporate success (Matzler et al. 1996). Service and marketing research disciplines see a
strong connection between service quality and customer satisfaction, considering the construct of quality
as an antecedent for perceived satisfaction (de Ruyter et al. 1997). The theoretical foundation in this concept
draws on the disconfirmation of expectations paradigm, formalized as a model by Oliver (1980). According
to the model, customers form expectations of a product and evaluate them against its perceived quality. The
appropriate quality construct is user-defined quality. Accordingly, the extent to which products meet user
requirements defines the quality. Satisfaction measures how a product performs in meeting requirements
of complementors that work with the platform and develop complements with the help of BRs (Berger 1993;
Montgomery 2009; Hayes 2008; Hill and Alexander 2000). When the expectations are confirmed during
the product use, the customer develops repurchase intentions for the product; if not, he will not repurchase
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
5
it (Oliver 1980; de Ruyter et al. 1997). Bhattacherjee (2001) proves t he similarity between Oliver’s model
and the continuance decision of an IS user. Hence, the user’s satisfaction is positively associated with their
continuance intention (Bhattacherjee 2001).
Against this theoretical background, the quality of IS and software is a multidimensional concept, which
can be broken down into individual sub-attributes that address individual user requirements. These sub-
attributes represent different dimensions of quality that can be perceived and therefore evaluated by users.
Consequently, it is possible to assess the perceived quality via satisfaction surveys (Murine 1988; Berger et
al. 1993; Kekre et al. 1995; Hayes 2008). Therefore, assessing quality through user feedback is a valid
approach, whereby the subjective quality is based upon multiple events or incidents that were experienced
by the respondent (Hayes 2008).
This theoretical background enabled us to evaluate satisfaction in platform-based IoT ecosystems. Different
BRs, integrated by the platform companies into the design of IoT platforms, represent the mentioned
attributes that can be assessed through the feedback of complementors. Since BRs act as complementors’
access points with the platform, enabling the interaction with the platform core, they may improve the
platform-based development experience and support the use of specific platform functionalities. As the
complementors rely on and use these BRs to create IoT applications, they perceive their quality during their
use. Consequently, representatives in complementary companies, who actively use the BRs can retrieve
their experiences, evaluating the perceived quality of BRs. Following Bhattacherjee’s theory, if the quality
of the used BRs is perceived insufficient to develop an IoT application for productive use in an enterprise
environment, complementors may develop intentions to leave the platform (Tiwana 2015; Bhattacherjee
2001). Nevertheless, a high level of satisfaction prevents platform desertion and increases positive word-
of-mouth, attracting more complementors to the IoT ecosystem (Matzler et al. 2004). For the management
of platform companies, it is therefore preferable to ensure an appropriate satisfaction level of the
complementors within the ecosystem, as their innovations determine the added value of the platform
ecosystem (Saadatmand et al. 2019; Engert et al. 2020). Besides that, since BRs include both technical and
non-technical resources, embedded in the platform sponsoring activities (Eisenmann 2009; Ghazawneh
and Henfridsson 2013; Eaton et al. 2015), the described theoretical background fits for the IS and the
service quality perspective. Conversely, we consider BRs as an appropriate object of interest to study the
possibilities of achieving satisfaction in platform-based IoT ecosystems. Whether in software or services,
multiple system attributes have different influences on users’ satisfaction. Kano’s theory of attractive quality
with the coherent Kano model are recognized in multiple research streams (i.e., service science, service
quality, requirements engineering) to represent a valid construct to research sources of satisfaction,
measuring the satisfactory impact of different attributes. In addition, Kanos’ theory also proposes suitable
data collection techniques for empirical satisfaction surveys (Grigoroudis and Siskos 2010). Besides, Kano’s
survey technique was already applied to analyze attractive attributes of software systems (Mayer 2012).
Therefore, we chose Kano’s theory of attractive quality, the coherent Kano model, and Kano’s survey
methodology for the study design. Figure 1 illustrates the proposed research framework to determine the
study design, described in the next section:
Figure 1. Overview of the Research Process
Stage 1: Study preparation
Studying the satisfaction theory constructs
and related survey methods to analyze
information systems
Stages
Stage Description
Stage 2: Data collection Stage 3: Data interpretation
Searching for experienced developers from the
Azure IoT ecosystem and executing an online
survey based on Kano‘s methodology
Assessing the current satisfaction level with
Azure IoT TBR
Functional form:
How do you feel about Azure IoT having APIs?
Dysfunctional form:
Kano classification to assess the satisfactory
potential of TBRs
Recommendations for IoT platform
development through TBR considering
their satisfactory potential
Mean values for self-stated satisfaction level
with the TBRs to get the location on the x-axis
in the graphical representation of the Kano
model (Figure 2)
(survey excerpt on TBR “APIs”)
TBRs as
attributes
ofan IoT
platform
Dysfunctional
Like Expect Neutral Accept Dislike
Functional
Like Questionable Attractive Attractive Attractive One-
dimensional
Expect Reverse Indifferent Indifferent Indifferent Must-be
Neutral Reverse Indifferent Indifferent Indifferent Must-be
Accept Reverse Indifferent Indifferent Indifferent Must-be
Dislike Reverse Reverse Reverse Reverse Questionable
How was your experience using APIs?
Rating of user satisfaction: 1 = very dissatisfied ; 2 = rather dissatisfied ; 3 =
neither satisfied nor dissatisfied ; 4 = quite satisfied ; 5 = very satisfied
1 2 3 4 5
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
6
Study Design
Application of the Kano Model to Boundary Resources
At the beginning of the study we conceptualized complementor satisfaction by incorporating the integrated
confirmation-expectation models and Kano's theory of attractive quality, specifically by applying the Kano
model, to the concept of BR (Stage 1). As stated in the introduction, the satisfaction of complementors
depends on the quality of the BRs. Thus, for an IoT platform provider, it is advisable to create high-quality
BRs in order to attract developers from complementary firms, who aim to join the platform. We mainly
build this assumption on Kano’s theory for attractive quality, which describes the relationship between the
quality of a specific product attribute and customer satisfaction (Kano et al. 1984). In contrast to early
satisfaction research, which defined satisfaction as a two-factor construct, the theory for attractive quality
follows a three-factor model. The two-factor construct is considered outdated (Matzler et al. 1996; Matzler
and Hinterhuber 1998). The underlying assumption in explaining the relationship between distinctive
quality attributes and satisfaction recognizes non-linear and asymmetric dependencies. The current state
of the Kano model classifies factors influencing satisfaction in so-called basic factors, performance factors,
and attractive factors, also taking into account indifferent and reverse factors (Sauerwein et al. 1996;
Matzler & Sauerwein 2002). Table 1 gives an overview of Kano’s five classes and describes how the TBRs
can be classified according to the model (Kano et al. 1984; Berger et al. 1993):
Category
Description
Must-be (M)
A TBR classified into this category includes basic criteria. An inappropriate fulfillment, low quality or
unavailability of these criteria leads to an extreme dissatisfaction of a complementor. The fulfillment of these
basic criteria is taken for granted, and following does not lead to satisfaction. These facts lead to an under-
proportional relationship between criteria fulfillment and complementor satisfaction.
One-
dimensional (O)
A TBR classified into this category includes criteria that are explicitly demanded by a complementor. The
relationship between criteria fulfillment or quality level and complementor satisfaction is proportional.
Platform providers can use these TBRs to generate competitive advantages.
Attractive (A)
A TBR classified into this category includes criteria that are not explicitly demanded by a complementor.
Fulfillment or high quality of these criteria leads to strong satisfaction, whereas a non-fulfillment or low quality
does not lead to dissatisfaction. The relationship between criteria fulfillment and complementor satisfaction is
over-proportional.
Indifferent (I)
A TBR classified into this category includes criteria that have neither positive nor negative influence on the
complementor satisfaction; meaning they do not have any importance to the complementor.
Reverse (R)
A TBR classified into this category includes criteria that have a converse proportional relationship between
criteria fulfillment or quality level and complementor satisfaction. An appropriate fulfillment of these criteria
leads to an extreme complementor satisfaction, whilst an inappropriate fulfillment or unavailability leads to
an extreme complementor dissatisfaction
Table 1. Kano Classification Categories
To answer the RQ, we conducted an empirical survey to measure the level of satisfaction linked to specific
TBRs (Stage 2). Therefore, relying on the expectation-confirmation model, we combined the Kano survey
methodology with a Likert-based self-stated satisfaction questionnaire (LSSQ). The Kano method identifies
and classifies a satisfaction factor for each TBR and forms the central part of the survey. First, the
complementary developers are asked to rate their feelings if the TBRs are integrated in the IoT platform
(functional question). Second, they are asked to rate their feelings if the same TBRs are not integrated in
the IoT platform (dysfunctional question). However, it does not allow to rank the separate TBRs according
to their importance for the complementors. By conducting a LSSQ survey, the results of the Kano method
are complemented with satisfaction rankings for each TBR. (Kano 1995; Berger et al. 1993). We conducted
a pretest with senior researchers from the discipline of requirements engineering and selected practitioners
who have experience with Azure IoT. In doing so, we ensured that the structure and content of the
questionnaire are easy to understand. The survey comprising of ten pages was designed as follows. General
information on the survey participant (i.e., the number and context of Azure IoT projects participated) was
requested on the first page. Each of the remaining nine pages was dedicated to a specific TBR with the
following three parts: Overall, the survey elicits the qualification profile and the background of the
respondents and asks the following four questions for each BR to be evaluated: (1) TBR used or not?
(yes/no); (2) functional question (Kano); (3) dysfunctional question (Kano); (4) satisfactory level with TBR
(LSSQ). The complete questionnaire is available at the following URL: https://bit.ly/3px7RGo.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
7
Case Selection
Considering the exploratory character of the study, we have selected complementors in the same ecosystem
as they all use the same TBRs. A single case study also created a specific setting. Selection complementors
from different platform ecosystems would decrease the comparability of the responses. Following that, we
have chosen the Microsoft Azure Cloud as a suitable case. In contrast to the prior BR-based satisfaction
survey (Petrik and Herzwurm 2020a) with industrial incumbents, we have chosen a platform provider that
has always followed an IT-based business model that has changed over the years from offering standard
software to providing a cloud platform. Hence, we have chosen a platform-native as a corresponding
counterpart (Gustavsson and Ljungberg 2019; Pauli et al. 2021). Multiple reasons support the suitability of
Azure IoT as a representative case study for exploring complementor satisfaction in IoT:
Besides Amazon Web Services (AWS) and Google Cloud, Microsoft is one of the largest IoT cloud
platform providers (Muhammed and Ucuz 2020; PTC 2019).
Azure IoT is described as an innovation platform with multiple functionalities to foster generativity and
convergence (Gustavsson and Ljungberg 2019). Microsoft maintains BRs for Azure IoT, so Gustavsson
and Ljungberg (2019) recognize the use of the BR concept during Microsofts’ platformization.
Although, they do not pursue the BR focus, the authors identify openness and generativity of the Azure
Cloud platform as one of the success factors of Microsofts’ platformization strategy. Based on these
findings, we examine the BRs of the Microsoft Azure Cloud for IoT applications.
Although Azure can be classified as Platform-as-a-Service from a cloud architecture perspective, it was
initially introduced as Infrastructure-as-a-Service by Microsoft. Compared to other IoT platforms (e.g.,
MindSphere) Azure is suitable for our research objective. Microsoft offers its complementors the same
set of TBRs, which we identified for MindSphere. Consequently, Azure IoT enables complementor
developers to implement similar use cases, even if it can be classified as more generic and not just
industry-oriented. For example, the analyst firm ForresterWave is also including Azure IoT in the
benchmark for IoT platforms (PTC 2019). Thus, we feel confirmed in our efforts to analyze satisfaction
construct for IoT platforms studying the Microsoft Azure IoT case.
Petrik and Herzwurm (2019) presented a list of TBRs that are representative for IoT and IIoT. In a
preliminary examination of the Azure IoT platform, we carefully studied the Azure website
(https://azure.microsoft.com/overview/) to discover the necessary information on the BR provided by
Microsoft. The Azure website contains links to the developer portal, which main task is to get the APIs
discovered by developers and to promote any relevant information on the technical resources of the
platform. The prior generated list of BRs used in IIoT (Petrik and Herzwurm 2019), derived from the study
of electronic documents on Siemens MindSphere, served as a checklist to discover TBRs, that Microsoft
offers its ecosystem participants. As a result of this preliminary study, we found a complete overlap among
the offered TBRs compared to the MindSphere IoT platform. Table 2 presents an overview of the nine TBRs
discovered and describes briefly their main aspects of benefit:
TBR
Brief description of the main benefit aspect
APIs
Platform interfaces that enable the third-party applications or other enterprise software to interact with the core
of the IoT platform.
Connectivity
libraries
They enable the connection between the IoT platform and devices or assets such as the programmable logic
controllers (PLCs) or other edge devices in the field.
Open
machine
protocols
Different protocols can be used in addition to the established TCP/IP protocol to meet the requirements of the
IoT use case, such as the real-time capability (e.g., process stop) or a low overhead (e.g., due to low bandwidth).
In addition to the HTTP protocol, Azure IoT natively supports MQTT and AMQP protocols.
Infrastructure
support
The possibility of operating a platform infrastructure agnostic can be an IoT requirement for various reasons.
Although Azure IoT is only offered on Microsoft's cloud infrastructure, the consideration of this question in the
Kano survey provides knowledge about whether IoT platform users demand for more flexibility from the platform
provider in supporting multiple cloud infrastructure providers.
DevOps
metrics and
demos
Provided metrics allow the monitoring of IoT applications in productive operation and help to maintain them
after their initial release. Azure IoT provides IoT Hub metrics to view user statistics and assess the overall integrity
of the IoT Hub service and the devices connected to it.
SDKs
Development tools providing a set of supported APIs in a specific programming language. SDKs standardize the
development routines and reduce the efforts to create IoT applications. Microsoft offers Azure IoT SDKs in
different languages. However, their use is optional for complementors.
Container
management
service
Container management systems enable the creation of ‘lightweight’ application modules and reduces the
developers’ efforts of deploying the application in the cloud. Especially with Kubernetes-as-a-service, Microsoft
implements the container approach for Azure IoT.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
8
App store
An app store makes the applications accessible to the entire ecosystem, fostering the visibility of the complements
and standardizing the monetization for the complementors. Microsoft offers its app store for all Azure-based
applications including the Azure IoT.
Low code
platform
Low-code development platforms describe a software development environment that enables the development
process using visual application designers and other graphical modeling methods for non-developers. Microsoft
has integrated Power Apps with Azure IoT.
Table 2. Overview of the examined TBRs from the Microsoft Azure IoT ecosystem
Data Collection
Between 4th of March 2020 and 24th of June 2020, we looked for development-savvy practitioners with
Azure IoT experience to participate in the online survey. To reach a wide variety of complementors from
the Azure IoT ecosystem, we shared the survey link in public LinkedIn groups as well as developer
communities. Furthermore, we explicitly searched for suitable respondents with IoT experience via
Hunter.io, so that we could contact 108 persons by e-mail at last. On top of that, we directly messaged 562
practitioners via LinkedIn. All contact persons were introduced into the case and asked to fill out the online
survey. In total, 83 participants (19 architects, 13 developers, 11 tech or project leads, 9 consultants, 8
software engineers, 2 evangelists, 1 test lead, 1 DevOps engineer; 18 respondents decided not to provide
information on their current position) answered the survey. Most respondents were technical experts and
had experience with more than 1 Azure IoT project. We determined 704 Kano ratings and 543 LSSQ
ratings. Having a detailed look at the different TBRs, one can observe a non-uniform distribution between
the number of answers (e.g., ‘APIs’ 74 LSSQ responds vs. ‘app store’ 33 LSSQ responses). Every answer was
optional so the participants could decide whether to rate their satisfaction with a specific TBR or not. This
option helps to filter false evaluations in case of a lack of experience of the respondents. Nevertheless, the
sample size is justifiable, as apart from looking at the pure numbers, one must be considered that the survey
addresses business-to-business enterprises and only questions active Azure partners, resulting in a smaller
target group than consumer-oriented ecosystems. We reason that the theoretical saturation can be
guaranteed because the last 10 responses did not significantly change the rating (Löfgren and Witell 2008).
Data Analysis
The data analysis aims to gain insights into the impact of TBRs on the complementors’ satisfaction level
with IoT platforms. It utilizes the theory of attractive quality acts as the theoretical foundation to explain
and analyze the collected data. According to our study’s research design, the data analysis is separated into
three analysis streams. These analysis streams are brought together by incorporating the theory of attractive
quality, specifically by applying the Kano model. The application of the Kano method allows to classify the
given TBRs and is considered to deliver valid and highly reliable results (Mikulic and Prebezac 2011).
Therefore, the answers to the functional and dysfunctional questions of each participant related to the single
TBRs are mapped according to the Kano evaluation table. This allows to derive one of the five classification
categories listed above. According to the Kano evaluation table, if there is no logical mapping possible, the
category ‘questionable’ (Q) will be derived (Berger et al. 1993). For every single TBR, a summarized
evaluation table for all respondents was generated to do further analysis. The simplest possibility is to
calculate the frequency distribution of the classification categories for each TBR. A TBR is then classified
into the category with the highest frequency (Sauerwein et al. 1996). If a clear classification is difficult
because the frequencies between ‘indifferent’ and satisfaction-critical classes (i.e., attractive, one-
dimensional, or must-be) are close together, Berger et al. (1993) suggest reducing the data noise by applying
the rule, where O, A, M, I, R, and Q depict the nominal frequency of the classification category:
if (O+A+M) > (I+R+Q), then Max(O,A,M) or if (O+A+M) < (I+R+Q), then Max(I,R,Q)
If a clear classification is difficult due to close frequencies between the satisfaction critical classes, Matzler
et al. (1996) suggest applying the following rule: ‘M before O before A before I’. By applying that additional
rule, platform providers can better prioritize one TBR over the others since prioritizing must-be attributes
helps to consider those attributes with the greatest possible influence on dissatisfaction (Matzler et al.
1996). The calculation of the customer satisfaction (CS+) and customer dissatisfaction (CS-) coefficient
allows to infer how strongly a TBR may influence complementor satisfaction or dissatisfaction. This
approach helps to distinguish the fine differences in the satisfaction effect within the same classification
category to assess the potential to influence satisfaction (CS+) and dissatisfaction (CS-) more precisely. If
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
9
CS+ approaches 1, there is a strong influence on the complementor satisfaction to be expected if a high
quality of the TBR will be fulfilled. If CS- approaches -1, there is a strong influence on the complementor
dissatisfaction to be expected if a high quality of the TBR will not be fulfilled. These coefficients are
calculated as follows (Berger et al. 1993; Matzler et al. 1996):
𝐶𝑆+ = (𝐴 + 𝑂)
(𝐴 + 𝑂 + 𝑀 + 𝐼) 𝑎𝑛𝑑 𝐶𝑆− = − (𝑂 + 𝑀)
(𝐴 + 𝑂 + 𝑀 + 𝐼)
Likert scales are the most widely used scales in survey-based studies. They enable the measurement of an
agreement level with a predefined statement (Hayes 1992; Grigoroudis and Siskos 2010). In the LSSQ part
of the survey, the statements represent the perceived satisfaction levels with a specific TBR used during the
implementation of IoT applications. The survey contains six-point Likert scales for each of the nine
surveyed TBR. The starting value equals zero if the respondent does not have any experience with the given
TBR or did not use it to implement IoT applications in the past. Other evaluation points lie between 1 and
5 with an interval of 1. Accordingly, the scales do not contain a neutral score for experienced satisfaction.
The scales were developed following the recommendations of the scientific literature, which states that
satisfaction scales with more than five measurement levels do not offer any additional accuracy (Hayes
1992). In addition, the discovery of the current satisfaction level with the TBR provided the obtained LSSQ
values nurture the Kano model and enable the formation of a correct sequence along the x-axis for TBR in
the graphical representation of the Kano model according to their satisfaction ratings (see Figure 2). All the
detailed calculations are also accessible online: https://bit.ly/3px7RGo
Results
Table 3 presents the Kano survey results including absolute and relative frequencies for the evaluated TBR:
Kano
classification
TBR
Attractive
One-
dimensional
Must-be
Indifferent
Reverse
Question-
able
Total
abs.
rel.
abs.
rel.
abs.
rel.
abs.
rel.
abs.
rel.
abs.
rel.
APIs
14
17%
24
29%
21
26%
16
20%
4
5%
3
4%
82
Connectivity libraries
5
6%
26
32%
20
25%
23
28%
2
2%
5
6%
81
Open machine protocols
4
5%
37
45%
19
23
15
18%
0
0%
7
9%
82
Infrastructure support
8
10%
2
2%
1
1%
32
39%
29
35%
10
2%
82
DevOps metrics & demos
7
9%
22
28%
20
26%
25
32%
0
0%
4
5%
78
SDKs
4
5%
40
49%
11
13%
19
23%
3
4%
5
6%
82
Container management serv.
7
9%
21
27%
9
12%
35
45%
0
0%
6
8%
78
App store
5
7%
11
16%
11
16%
36
52%
1
1%
5
7%
69
Low code platform
9
13%
9
13%
5
7%
41
59%
0
0%
6
9%
70
Relative share
9%
27%
17%
34%
6%
7%
Table 3. Kano Evaluation Table
As described in the previous section, we initially processed the results of a Kano survey by using the derived
frequencies for each answer. It becomes evident that the assignment of categories for some TBRs cannot be
done unambiguously since the scores of the Kano classes for the APIs, the connectivity libraries, the
infrastructure support, and the DevOps metrics & demos are relatively close to each other. Hence, to
improve the classification, we applied further analysis rules developed by Matzler et al. (1996) and Berger
et al. (1993). The difference in API classification between a ‘one-dimensional’ and a ‘must-be’ attribute is
only 3%. Applying the rule suggested by Matzler et al. (1996), APIs should be instead classified as a ‘must-
be’ TBR of an IoT platform. Looking at the supported machine protocols and the DevOps metrics, the
frequency of the ‘indifferent’ category exceeds the most significant satisfaction-critical classes just by 4%.
Besides, this rule gives a preference for the narrow category difference, considering the DevOps metrics &
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
10
demos. Table 4 gives a complete view of the defined categories for the TBRs used in the Azure IoT
ecosystem. The calculated values for positive complementor satisfaction coefficients (CS+) indicate how
close the APIs are to the classification as a ‘must-be’ attribute. Additionally, we notice that the satisfaction
potentials are slightly above the value of 0,5. Consequently, this TBR has a significant satisfactory impact,
but it is not exceptionally high as the values are far away from a value 1 (Berger et al. 1993; Matzler and
Hinterhuber 1998). Consistent with the Kano classification, TBRs categorized as ‘indifferent’ seem to have
little influence on satisfaction. Therefore, investments in their quality improvement are not target-oriented
to improve the satisfaction of the ecosystem participants. Conversely, the calculated CS- values indicate
which attributes would cause greater dissatisfaction if their quality fulfillment is not met. Table 4 shows the
calculated arithmetic mean values for the obtained levels of self-stated satisfaction, allowing us to create
satisfaction rankings for the surveyed TBRs. Surprisingly, the container management services (‘indifferent’)
are ranked third. Apparently, the technical quality of this TBR is particularly good at Azure IoT, although
this TBR is only an assistive resource, not essential to implement an IoT use case. Based upon the evaluated
classification categories and LSSQ values, we can complete the graphical representation of the Kano model,
applied on the analyzed TBRs in Figure 2.
TBR
Kano classification
CS+
CS-
Mean values for Self-
stated level of satisfaction
Position in
the Ranking
APIs
Must-be
0,50
-0,6
3,94
4
Connectivity libraries
One-dimensional
0,41
-0,62
3,85
6
Open machine protocols
One-dimensional
0,54
-0,74
4,01
2
Infrastructure support
Indifferent
0,23
-0,06
3,67
8
DevOps metrics & demos
One-dimensional
0,39
-0,56
3,57
9
SDKs
One-dimensional
0,59
-0,68
4,14
1
Container management services
Indifferent
0,38
-0,41
4
3
App store
Indifferent
0,25
-0,34
3,90
5
Low code platform
Indifferent
0,28
-0,21
3,69
7
Table 4. Complementor Satisfaction Survey Results for Kano and LSSQ
Figure 2. Graphical representation of the Kano Model Application on Azure IoT TBR
APIs
Must-be
attributes
One-dimensional /
performance
attributes
Attractive
attributes
Indifferent
attributes
Complementor
satisfaction
Complementor
dissatisfaction
Quality
expactations
not fulfilled
Quality
expactations
fulfilled
Connectivity
libraries
DevOps
metrics &
demos
Open
machine
protocols
SDKs
Container
management
service
App
store
Low
Code
platform
Infrastructure
support
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
11
Discussion of Key Results
Summing up the theoretical contribution, the study confirms a clear but not disproportionate satisfactory
potential of TBRs that may be considered in future research on the critical determinants in ecosystem
development. From a practical perspective, the results contribute to the corpus of knowledge on developing
IoT platform ecosystems. First, the results provide a basis for decision-making for IoT platform providers
to optimize development activities considering the multiplicity of resources . Secondly, the
results provide an opportunity to decompose an IoT platform into individual BRs, which supports a
complementary orientation of the platform development. Both, the BR-oriented view of IoT platforms and
the Kano method can be used in the field by platform providers during the platform development in line
with the interests of the ecosystem participants. The analysis of system determinants with the help of Kano's
principles is a technique commonly used in practice in system design. Especially the compatibility of Kano's
model and quality function deployment is helpful for a user-oriented (i.e., complementor-oriented) design
of IoT platforms (Bijan et al. 2013; Sauerwein et al. 1996). Such a platform development approach can
counteract multihoming in the competition for innovative complementors in IoT. The satisfactory potential
of BRs may additionally be taken into account when selecting the right platform to leverage IoT use cases.
Accordingly, knowledge of the satisfactory potential of the different BRs and their distinctive quality levels
enables systematic platform evaluation and may support the development of a scientifically supported
evaluation process for digital platforms used in IoT and other application domains.
The applied Kano model helps to understand the particular role of the APIs in the satisfaction perception
of the complementors. Although the classification is close to the threshold between the ‘performance’ and
‘must-be’ both classes confirm the dangerous frustration potential in API quality. The results reveal the
need to implement automated quality assurance for APIs, to avoid dissatisfaction. The API classification
confirms the classification conducted by Petrik and Herzwurm (2020a) for the IIoT. In the given study, the
narrow difference in the survey results can be explained by a generally high level of technical quality with
the TBR provided in the Azure IoT ecosystem and in the use of different measuring techniques (e.g., Kano
survey instead of critical incident conversion to build Kano classes). In contrast, the quality of the
MindSphere APIs received more critics, and therefore, the data supporting the classification was more
definite. After all, the classification of APIs as a ‘must-be’ attribute corresponds to their essential function
to connect complements to the platform, thus enabling the ecosystem participants to carry out
complementary innovation activities. However, since complementors take APIs for granted, their stable
high quality does not evoke extraordinary satisfaction. Thus, APIs cannot be considered as a source
of competitive advantage with regard to their satisfactory potential. In turn, high-quality and value-
added SDKs can significantly increase satisfaction because they have been classified as ‘performance’
attributes. Previously, SDKs have already been identified as the most important attribute in the domain of
app development for smartphones (Koch and Kerschbaum 2014). By adding the satisfaction construct, we
confirm this result for IoT platforms. Regarding the supported protocols, we see a difference to the
MindSphere study, where this TBR was classified as an attractive attribute. We explain this difference by
the fact that Microsoft has been natively supporting certain protocols for Azure for a long time (i.e., AMQP
or MQTT), while the support for the proprietary Siemens protocol S7 (i.e., high penetration rate in some
industries) and OPC UA was left to the complementors, requiring integration efforts. Consequently, the
classification achieved can be explained by the scoping of the IoT platform. Hence, Microsoft’s scoping for
the Azure IoT did not focus on the industrial protocols, missing the chance to achieve high satisfaction with
this TBR. Connectivity that is as simple and reliable as possible is highly important for industrial actors.
That is why the exclusively industrial users of the MindSphere platform experience the maintenance of the
high quality of this TBR as ‘attractive’ compared to the Azure IoT ecosystem. Hence, solving connectivity
issues can be a source of competitive advantage for IoT platform providers if implemented with
sufficient quality. Most indifferent ratings can be explained by the fact that this TBR is relatively new in the
BR portfolio. Despite the achieved classification of DevOps metrics, the quality and scope of this TBR are
critical to the satisfaction of the majority of respondents. After applying the evaluation rules of Berger et al.
(1993), a linear relationship between the satisfaction effect and quality of DevOps metrics can be derived.
In comparison to the MindSphere case (Petrik and Herzwurm 2020a), in which no complementor was able
to evaluate this resource, the increasing importance of this TBR can be predicted if the future platform
updates are to be released according to the continuous paradigm. Hence, we propose for platform
providers to invest in DevOps metrics as it can be equally a source of frustration and satisfaction
depending on its quality of implementation.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
12
Practical contributions. Our study addresses the noticed lack of guidance for platform providers in the
highly competitive IoT domain (Pauli et al. 2021). Analyzing one of the market -leading and mature IoT
platforms, the satisfaction study of TBRs offers various implications for platform providers. It suggests
which types of TBR impact satisfaction, and support the balance in the BR portfolio in a challenging
setting of IoT. The Kano classification not only shows the current status of quality-related satisfaction but
also helps, in particular, to see the satisfaction potential of the different BR offered, to orchestrate
investments in the BR portfolio considering developers’ preferences. TBRs with high frustration potential
include all ‘one-dimensional’ and ‘must-be’ attributes. Hence, to keep complementary developers satisfied,
it is recommended to monitor the quality level of these TBRs. As for all performance attributes , according
to Kano, it applies that a company is compared with its competitors via the performance attributes. Besides,
knowledge about the relevance of TBR for satisfaction has implications for evaluating one's platform against
competitors (Matzler et al. 1996). Additionally, the overall comparison with the satisfaction potential of BRs
of the MindSphere study offers another valuable insight. Our results indicate that ‘attractive’ TBRs are
missing completely. For the most part, this result is included in the MindSphere study (Petrik and
Herzwurm 2020a), and our study confirms the finding that SBRs offer an option to create overproportional
satisfaction among complementary developers, supporting the ecosystem development. Furthermore, we
add insights to the results of Koch and Kerschbaum (2014), which can be explained partly because of the
measurement technology used and partly because of the domain specificity (e.g., lower maturity of IoT
platforms). In this respect, a proposition derived for platform providers is to invest in the development
of non-technical BRs (i.e., partner programs or developer events) in addition to the TBRs as they can
greatly enhance the overall attractiveness of an IoT platform.
Research contributions. Our study drives initial findings into the existing open fields of platform quality
and complementor dynamics (McIntyre and Srinivasan 2017) in the IoT domain. More particularly, the
results may support the complementor dedication research (Hurni et al. 2020) that was recognized as a
platform governance objective. The results shed light on the satisfactory impact of concrete BR to increase
the dedication of the ecosystem participants. The fact that none of the surveyed TBRs was classified as an
attractive attribute of the IoT platform brings forward the research on ecosystem development. Obviously,
non-technical BRs create greater satisfaction, which fosters the attractiveness of the IoT ecosystem to a
greater extent than the technical characteristics of the platform. Consequently, if a platform provider can
shape the ecosystem attractiveness (Jacobides et al. 2018), the research should keep closer attention to non-
technical activities. The different satisfaction ratings of technical and non-technical BRs provide a
foundation to the research on the possible formats of complementor communication with the
ecosystem, which is an important research avenue in digital platform research (Hein et al. 2019). Further
research could also address the contrast between satisfaction and restrictive governance rules for IoT
platforms. In this context, it should be discussed whether a high level of satisfaction caused by a high
level of quality of TBRs can inadequately regulate and counterbalance restrictive governance structures,
perceived input control, and compensate the perceived vendor lock-in, which is critical in the
enterprise IoT (Hurni et al. 2020; Croitor et al. 2020; Pauli et al. 2021). At least, this phenomenon can be
assumed for the Apple ecosystem with its restrictive rules for iOS (Eaton et al. 2015).
From a high-level perspective, our study adds to the research stream on the competition in platform-based
ecosystems. Greater consideration of the satisfactory potential of BRs may work for platform providers in
general, as well as IoT platform providers in particular, as a mechanism in the ecosystem competition.
Satisfactory BR foster the transition of a production ecosystem towards a consumption ecosystem, by
enabling the possibility to become a digital hybrid, and increasing the share of the exploited network effects
compared to the servitization of own products (Subramaniam et al. 2019; Hein et al. 2019). Furthermore,
our study makes a methodological contribution. Since only a few studies used Kano’s survey method to
evaluate satisfactory attributes of IS in the past (Meyer 2012), we confirm the suitability of the Kano
method in combination with the BR concept for studying IoT platforms and encourage other
researchers to replicate the study design to examine multidimensional and interdependent nature of IS
success in the complex domain of IoT (DeLone and McLean 2003). Although at the beginning of the study,
it was not our goal to reinvent the DeLone and McLean IS success model, we are convinced that the results
obtained offer a new perspective on how to adopt the model in the context of IoT platforms since BRs act
as satisfactory contact points between complementors and platform providers, thus helping to assess and
determine the system and service quality of IoT platforms. The model can break down complex
multidimensional issues such as platform ecosystem development into individual constructs and make
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
13
them better measurable. Given the lack of holistic models in the context of IoT platforms, and because it
takes the purposeful continuation of platform use into account, the model embeds satisfaction in the
management of platform ecosystems in IoT in a purposeful way. Future research can specifically examine
the explanatory connections between BR and existing dimensions of success or even uncover new IoT-
specific dimensions based on BR. The applied survey techniques help to measure the user satisfaction
variable in particular. Therefore, our results provide a basis for further research on the
instantiation of the IS success model - in the spirit of the model creators (DeLone and McLean 2003).
Besides, replicating the study design in other IoT-affected domains such as smart buildings, cities, or
logistics may provide further valuable insights for the research to facilitate the formation of IoT ecosystems
(Kapoor 2018; Jacobides et al. 2018; Baiyere et al. 2020), relying on the satisfaction construct and the BR
as mechanisms for ecosystem design.
Limitations and Outlook
Research on the enterprise IoT and digital platforms may build upon the classification results for some
TBRs as ‘indifferent’. Cloud-native development increasingly uses containerization. Studies indicate that in
the future, various other practical advantages can be expected from the use of Kubernetes (Taherizadeha &
Grobelnika 2020; Rossi et al. 2020). Therefore, it is surprising that Azure IoT support for container
management has been classified as indifferent. Hence, future research could tackle the general role of the
containerization approach in the IoT. Furthermore, the classification of the app store as ‘indifferent’, is a
motivation for further research efforts to fundamentally examine the app store concept for its suitability in
the IoT environment. Although, according to Kano, new product attributes must diffuse and are initially
classified as indifferent by users (Löfgren and Witell 2008), the support in the Azure Marketplace for IoT
applications was enabled in 2018. For further research, a possible correction is the conceptual use of app
stores in the IoT domain because the satisfaction-critical effect of app stores in these domains could not be
proven so far. This finding extends the knowledge on the role of app stores in the setting of IoT platforms
(Schreieck et al. 2016). Finally, it is interesting to note that the integration of low code platforms (e.g.,
Microsoft PowerApps) also has no significant satisfaction effect. This raises the same question as the Kano
evaluation of the app store: how valuable are such development environments for the implementation of
IoT applications? Based on this result, we propose another opportunity for future research. Despite the
general demand in practice for rapid development, and Microsoft’s serious position in the benchmarks
(Rymer and Koplowitz 2019), we assume that IoT applications are complex. Therefore, implementing IoT
applications using PowerApps brings no efficiency advantages than conventional development as
professional developers, who usually do not rely on low code platforms, are nevertheless vital to create IoT
applications due to the complexity level of IoT scenarios. Interesting are also the results of the segment-
specific analysis of the survey data on TBRs that could not be categorized so clearly. For this purpose, we
have divided the survey results into three sub-clusters according to the IoT project experience: 1-4 projects;
5-9 projects; and >=10 projects. This analysis shows the typical Kano fading of the satisfaction potential of
TBRs over time. While the respondents with the least project experience classified the APIs and connectivity
libraries as ‘one-dimensional’, the most experienced respondents classified them as ‘must-be’ TBRs. The
same applies for the ‘DevOps metrics & demos’, where the satisfaction potential changes with experience
from ‘indifferent’ to ‘must-be’. The data for the segment-specific analysis is available online:
https://bit.ly/3px7RGo.
We also sense limitations in our paper that need to be considered. First, the generalizability of the results
may be limited. Since we studied the TBRs used by a single IoT platform, the transfer of our findings should
be made cautious since other platform ecosystems, and other domains may offer different BRs. Future
research could nevertheless investigate multiple platforms. Nevertheless, we are convinced that the results
obtained apply to other IoT platforms since Azure IoT is one of the leading IoT platforms in benchmarks,
covers a variety of IoT cases, is industry agnostic, and offers a heterogeneous complementor ecosystem,
which is why we assume that survey sample is representative for other IoT platforms. Second, the design
decisions made by the platform provider and the selected development scope for BR can affect the Kano
classification. It is possible that in other IoT platforms, certain BRs are developed, for example, in
cooperation with complementors or exclusively by complementors. Such collaborative platform
participation gains momentum (O’Mahony and Karp 2020) as it reduces the perceived ‘vendor lock-in’,
which can have an impact on satisfaction with the platform design, e.g., in the case of IIoT platforms, against
the backdrop of the criticality of the processed data (Pauli et al. 2021). Third, the level of details chosen for
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
14
the survey focusing on individual TBRs may lead to several ‘indifferent’ classifications. To address this
threat to validity, we plan to enrich the derived results with the sentiment reports obtained through the
critical incident technique (CIT). This future step will improve the validity of the results through data
triangulation since CIT adds qualitative statements and assesses the causes of the derived satisfaction
levels. This technique offers insights on a granular level, uncovering which characteristics of every single
TBR are responsible for creating satisfaction or dissatisfaction among complementors. This future step
enables a mixed-method approach, enriching the quantitative Kano method with qualitative data (Hayes
2008; Mikulic and Prebezac 2011) to generate further implications for practice and research.
References
Adner, R. 2006. “Match your innovation strategy to your innovation ecosystem,” Harvard Business Review
(84:4), pp. 98-107.
Anke, J., Poeppelbus, J., and Alt, R. 2020. “It Takes More than Two to Tango: Identifying Roles and Patterns
in Multi-Actor Smart Service Innovation,” Schmalenbach Business Review (72), pp. 599-634.
Baiyere, A., Topi, H., Venkatesh, V., Wyatt, J., Donnellan, B. 2020, Internet of Things (IoT) – A Research
Agenda for Information Systems, Communications of the Association for Information Systems.
Baldwin, C., and Woodard, J. 2009. “Platforms, markets and innovation,” in The Architecture of Platforms:
A Unified View, A. Gawer (ed.), Cheltenham: Edward Elgar, pp. 19-44.
Bender, B. 2020. “The Impact of Integration on Application Success and Customer Satisfaction in Mobile
Device Platforms,” Business & Information Systems Engineering (62:6), p. 515-533.
Berger, C., Blauth, R., Boger, D., Bolster, C., Burchill, G., DuMouchel, W., Pouliot, F., Richter, R., Rubinoff,
A., Diane, S., Timko, M., and Walden, D. 1993. “Understanding Customer-defined Quality,” Center for
Quality Management Journal (2:4), pp. 3-36.
Beverungen, D., and Kundisch, D. 2020. “Transforming into a platform provider: strategic options for
industrial smart service providers,” Journal of Service Management.
Bhattacherjee, A. 2001. “Understanding Information Systems Continuance: An Expectation-Confirmation
Model,” MIS Quarterly (25:3), pp. 351-370.
Bijan, Y., Yu, J., Stracener, J., Woods, T. 2013. “Systems Requirements Engineering – State of the
Methodology,” Systems Engineering (16:3), pp. 267-276.
Boudreau, K. J., and Lakhini, K. R. 2009. “How to Manage Outside Innovation,” MITSloan Management
Review (50:4), pp. 69-77.
Ceccagnoli, M., Forman, C., Huang, P., and Wu, D. J. 2012. “Cocreation of Value in a Platform Ecosystem:
The Case of Enterprise Software,” MIS Quarterly (36:1), pp. 263-290.
Croitor, E., Adam, M. and Benlian, A. 2020. “Perceived input control on digital platforms: a mixed-methods
investigation of web-browser platforms,” Journal of Decision Systems (30:1), pp. 50-71.
Cusumano, M., Gawer, A. and Yoffie, D. 2019. The business of Platforms: Strategy in the Age of Digital
Competition, Innovation and Power, New York: Harper Business.
Dal Bianco, V., Myllarniemi, V., Komssi, M., and Raatikainen, M. 2014. “The Role of Platform Boundary
Resources in Software Ecosystems: A Case Study,” in Proceedings of the Software Architecture
IEEE/IFIP Conference. Sydney, Australia.
DeLone, W. H., and McLean, E. R. 2003. “The DeLone and McLean Model of Information Systems Success:
A Ten-Year Update,” Journal of Management Information Systems (19:4), pp. 9-30.
De Reuver, M., Sørensen, C., and Basole, R. C. 2018. “The digital platform: A research agenda,” Journal of
Information Technology (33:2), pp. 124-135.
De Ruyter, K., Bloemer, J., and Peeters, P. 1997. “Merging service quality and service satisfaction: An
empirical test of an integrative model,” Journal of Economic Psychology (18:4), pp. 387-406.
Eaton, B. D., Elaluf-Calderwood, S., Sorensen, C., and Yoo, Y. 2015. “Distributed tuning of boundary
resources: the case of Apple's iOS service system,” MIS Quarterly (39:1), pp. 217-243.
Eisenmann, T. R., Parker, G., and van Alstyne, M. W. 2009. “Opening Platforms: How, When and Why?,”
in The Architecture of Platforms: A Unified View, A. Gawer (ed.), Cheltenham: Edward Elgar, pp. 131-
162.
Endres, H., Indulska, M., Ghosh, A., Baiyere, A., and Broser, S. 2019. “Industrial Internet of Things (IIoT)
Business Model Classification,” in Proceedings of the 40th International Conference on Information
Systems, H., Krcmar, Fedorowicz J., Boh, W. F., Leimeister, J. M. and S. Wattal (eds.), Munich,
Germany.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
15
Engert, M., Hein, A. and Krcmar, H. 2020. “Partner Programs and Complementor Assessment in Platform
Ecosystems: A Multiple-Case Study,” in Proceedings of the Americas Conference on Information
Systems B., Anderson, Thatcher, J. and R. Meservy (eds.), Salt Lake City, USA.
Floetgen, R. J., Novotny, M. A., Urmetzer, F. and Böhm, M. 2021. “Connecting the Dots of Digital Platform
Ecosystem Research: Constructs, Causal Links and Future Research,” in Proceedings of the European
Conference on Information Systems, F., Rowe, R., El Amrani, M., Limayem (eds.), Marrakech, Morocco
Gawer, A. 2014. “Bridging differing perspectives on technological platforms: Toward an integrative
Framework,” Research Policy (43:7), pp. 1239-1249.
Gawer, A. 2020. “Digital platforms’ boundaries: The interplay of firm scope, platform sides, and digital
interfaces,” Long Range Planning.
Ghazawneh, A., and Henfridsson, O. 2010. “Governing Third-Party Development through Platform
Boundary Resources,” in: Proceedings of the 31st International Conference on Information Systems,
St. Louis, MO, USA.
Ghazawneh, A., and Henfridsson, O. 2013. “Balancing Platform Control and External Contribution in Third-
Party Development: The Boundary Resources Model,” Information Systems Journal (23:2), pp. 173-
192.
Grigoroudis, E., and Siskos, Y. 2010. Customer Satisfaction Evaluation. Methods for Measuring and
Implementing Service Quality, Heidelberg, New York, London: Springer.
Gustavsson, M., and Ljungberg, J. 2019. “Platformization of a Cloud Service,” in Proceedings of the 40th
International Conference on Information Systems, H., Krcmar, Fedorowicz J., Boh, W. F., Leimeister,
J. M. and S. Wattal (eds.), Munich, Germany.
Guth, J., Breitenbücher, U., Falkenthal, M., Fremantle, P., Kopp, O., Leymann, F., and Reinfurt, L. 2018.
“A Detailed Analysis of IoT Platform Architectures: Concepts, Similarities, and Differences,” in Internet
of Everything, Di Martino, B., Li, K.-C., Yang, L.T. and A. Esposito (eds.), Singapore: Springer, pp. 81-
101.
Hayes, B. E. 2008. Measuring Customer Satisfaction and Loyalty. Survey Design, Use, and Statistical
Analysis Methods. Milwaukee: Quality Press.
Hein, A., J., Weking, M., Schreieck, M., Wiesche, M., Böhm and Krcmar, H. 2019. “Value co-creation
practices in business-to-business platform ecosystems,” Electronic Markets (29:3), pp. 503- 518.
Hein, A., M., Schreieck, T., Riasanow, D., Soto Setzke, M., Wiesche, M., Böhm and Krcmar, H. 2020. “Digital
platform ecosystems,“ Electronic Markets (30), pp. 87-98.
Hevner, A., and Malgonde, O. 2019. “Effectual application development on digital platforms,” Electronic
Markets (29), pp. 407-421.
Hill, N., and Alexander, J. 2000. Handbook of Customer Satisfaction and Loyalty Measurement,
Burlington, VT: Gower.
Hurni, T., Huber, T. L., Dibbern J., and Krancher, O. 2020. “Complementor dedication in platform
ecosystems: rule adequacy and the moderating role of flexible and benevolent practices,” European
Journal of Information Systems (30:3), pp. 237-260.
Iansiti, M., and Levien, R. 2004. “Strategy as Ecology,” Harvard Business Review (82:3), pp. 68-78.
Jacobides, M. G., Cennamo C., and Gawer, A. 2018. “Towards a theory of ecosystems,” Strategic
Management Journal (39:8), pp. 2255–2276.
Kano, N., Seraku, N., Takahashi, F., and Tsuji, S. 1984. “Attractive quality and must-be quality,” Journal of
the Japanese Society for Quality Control (41), pp. 39–48.
Kano, N. 1995. “Upsizing the organization by attractive quality creation,” in Total Quality Management,
Proceedings of the first world congress, G. K. Kanji (ed.), Dordrecht, Netherlands.
Kapoor, R. 2018. “Ecosystems: broadening the locus of value creation,” Journal of Organization Design
(7:1), pp. 1–16.
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. 253-
523.
Karhu, K., and Rittala, P. 2020. “Slicing the cake without baking it: Opportunistic platform entry strategies
in digital markets,” Long Range Planning.
Kekre, S., Kirshnan, M. S., and Srinivasan, K. 1995. “Drivers of Customer Satisfaction for Software Products:
Implications for Design and Service Support,” Management Science (41:9), pp. 1456-1470.
Koch, S., and Kerschbaum, M. 2014. “Joining a smartphone ecosystem: Application developers’ motivations
and decision criteria,” Information and Software Technology (56:11), pp. 1423-1435.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
16
Kude, T., Dibbern, J., and Heinzl, A. 2012. “Why do complementors participate? An analysis of partnership
networks in the enterprise software industry,” IEEE Transactions on Engineering Management (59:2),
pp. 250–265.
Leminen, S., Rajahonka, M., Wendelin, R., and Westerlund, M. 2019. “Industrial internet of things business
models in the machine-to-machine context,” Industrial Marketing Management (84), pp. 298-311.
Löfgren, M., and Witell, L. 2008. “Two Decades of Using Kano’s Theory of Attractive Quality: A Literature
Review,” Quality Management Journal (15:1), pp. 59-75.
Marheine, C., and Pauli, T. 2020. “Driving Generativity in Industrial IoT Platform Ecosystems,” in
Proceedings of the 41st International Conference on Information Systems, George, J., Paul, S., and
Rahul, D. (eds.), Hyderabad, India.
Matzler, K., Hinterhuber, H. H., Bailom, F., and Sauerwein, E. 1996. “How to delight your customers,”
Journal of Product & Brand Management (5:2), pp. 6-18.
Matzler, K., and Hinterhuber, H. H. 1998. “How to make product development projects more successful by
integrating Kano’s model of customer satisfaction into quality function deployment ,” Technovation
(18), pp. 25-38.
Matzler, K., Bailom, F., Hinterhuber, H.H., Renzl, B. and Pichler, J. 2004. “The asymmetric relationship
between attribute-level performance and overall customer satisfaction: a reconsideration of the
importance-performance analysis,” Industrial Marketing Management (33), pp. 271-277.
Mayer, J. H. 2012. “Using the Kano Model to Identify Attractive User-Interface Software Components,” in
Proceedings of the 33rd International Conference on Information Systems, Orlando, USA.
Mineraud, J., Mazhelis, O., Su, X., and Tarkoma, S. 2016. “A gap analysis of Internet-of-Things platforms,”
Computer Communications (89-90), pp. 5–16.
McIntyre, D. P., and Srinivasan, A. 2017. “Networks, platforms, and strategy: Emerging views and next
steps,” Strategic Management Journal (38:1), pp. 141–160.
Mikulic, J., and Prebezac, D. 2011. “A critical review of techniques for classifying quality attributes in the
Kano model,” Managing Service Quality: An International Journal (21:1), pp. 46-66.
Montgomery, D. C. 2009. Introduction to Statistical Quality Control. New York, NY: John Wiley & Sons,
Inc.
Murine, G. E. 1988. “Integrating software quality metrics with software QA,” Quality Progress (21:11), pp.
38–43.
Muhammed, A. S., and Ucuz, D. 2020. “Comparison of the IoT Platform Vendors, Microsoft Azure, Amazon
Web Services, and Google Cloud, from Users’ Perspectives,” in Proceedings of the 8th International
Symposium on Digital Forensics and Security, Beirut, Lebanon.
Myllärniemi, V., Kujala, S., Raatikainen, M., and Sevon, P. 2018. “Development as a journey: factors
supporting the adoption and use of software frameworks,” Journal of Software Engineering Research
and Development (6:6), pp. 1-22.
Ofe, A. H. 2018. “Building a Platform’s User Base: An Investigation of Value-driven Lock-ins in Digital
Service Innovation,” in Proceedings of the Americas Conference on Information Systems, New
Orleans, USA.
Oliver, R. L. 1980. “A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions,”
Journal of Marketing Research (17:4), pp. 460-469.
O'Mahony, S., and Karp, R. 2020. “From proprietary to collective governance: How do platform
participation strategies evolve?” Strategic Management Journal.
Parker, G., Van Alstyne, M., and Jiang, X. 2017. “Platform Ecosystems: How Developers Invert the Firm,”
MIS Quarterly (41:1), pp. 255-266.
Pauli, T., Fielt, E., and Matzner, M. 2021. “Digital Industrial Platforms,” Business & Information Systems
Engineering (63:2), pp. 181–190.
Petrik, D., and Herzwurm, G. 2019. “iIoT Ecosystem Development through Boundary Resources: a Siemens
MindSphere Case Study,” in Proceedings of the 2nd ACM SIGSOFT International Workshop on
Software-Intensive Business, Smolander, K., Grünbacher, P., Hyrynsalmi, S. and S. Jansen (eds.),
Tallinn, Estonia, pp. 1-6.
Petrik, D., and Herzwurm, G. 2020a. “Boundary Resources for IIoT Platforms – a Complementor
Satisfaction Study,” in Proceedings of the 41st International Conference on Information Systems,
George, J., Paul, S., and Rahul, D. (eds.), Hyderabad, India.
Petrik, D., and Herzwurm, G. 2020b. “Towards the IIoT Ecosystem Development – Understanding the
Stakeholder Perspective,” in Proceedings of the 28th European Conference on Information Systems,
Rowe, F., El Amrani, R. and M. Limayem (eds.), Marrakech, Morocco.
Satisfaction Potential of Technical Boundary Resources
Forty-Second International Conference on Information Systems, Austin 2021
17
Porter, M. E., and Heppelmann, J. E. 2014. “How Smart, Connected Products Are Transforming
Competition,” Harvard Business Review (92:11), pp. 64-88.
PTC 2019. Forrester’s 2019 Industrial IoT Platform Evaluation. URL:
https://www.ptc.com/en/resources/iiot/report/forrester-wave (accessed April 24, 2021).
Rochet, J.-C., and Tirole, J. 2003. “Platform competition in two-sided markets,” Journal of the European
Economic Association (1:4), pp.990-1029.
Rossi, F., Cardellini, V., Presti, F. L., and Nardelli, M. 2020. “Geo-distributed efficient deployment of
containers with Kubernetes,” Computer Communications (159), pp. 161-174.
Rymer, J. R., and Koplowitz, R. 2019. “The Forrester Wave: Low-Code Development Platforms For AD&D
Professionals, Q1 2019,” URL: https://wwwcdn.spanishpoint.ie/wp-content/uploads/2019/04/The-
Forrester-Wave%E2%84%A2_-Low-Code-Development-Platforms-For-ADD-Professionals-Q1-
2019.pdf (accessed April 30, 2021).
Saadatmand, F., Lindgren, R., and Schultze, U. 2019. “Configurations of platform organizations:
Implications for complementor engagement,” Research Policy (48:8), p. 103770
Sandberg, J., Holmström J., and Lyytinen, K. 2020. “Digitization and Phase Transitions in Platform
Organizing Logics: Evidence from the Process Automation Industry,” MIS Quarterly (44:1), pp. 129-
153.
Sauerwein, E., Bailom, F., Matzler, K., and Hinterhuber, H. H. 1996. “The Kano Model: How to Delight your
Customers,” in Proceedings of the International Working Seminar on Production Economics,
Innsbruck, Austria.
Schreieck, M., Wiesche, M., and Krcmar, H. 2016. “Design and Governance of Platform Ecosystems- Key
Concepts and Issues for Future Research,” in Proceedings of the 24th European Conference on
Information Systems, Istanbul, Turkey.
Schüler, F. 2020. “Loyalty on Industrial Internet of Things Platforms: An Empirical Study Integrating
Network Effects, Human-Computer Interaction and Agency Theory to Explore Platform Ecosystems,”
in Proceedings of the European Academy of Management 2020 Conference, Dublin, Ireland.
Skog, D., Wimelius, A. H., and Sandberg, J. 2018. “Digital Service Platform Evolution: How Spotify
Leveraged Boundary Resources to Become a Global Leader in Music Streaming,” in Proceedings of the
51st Hawaii International Conference on System Sciences, Waikoloa Village, USA.
Star, S. L., and Griesemer, J., R. 1989. “Institutional ecology ‘Translations’ and boundary objects: Amateurs
and professionals in Berkeley’s Museum of Vertebrate Zoology, 1907-39.,” Social studies of science
(19:3), pp. 387-420.
Subramaniam, M., Iyer, B., and Venkatraman, V. 2019. “Competing in digital ecosystems,” Business
Horizions (62:1), pp. 83-94.
Suominen, A., Seppänen, M., and Dedehayir, O. 2019. “A bibliometric review on innovation systems and
ecosystems: a research agenda,” Journal of Innovation Management (22:2), pp. 1-35.
Taherizadeha, S., and Grobelnika, M. 2020. “Key influencing factors of the Kubernetes auto-scaler for
computing-intensive microservice-native cloud-based applications,” Advances in Engineering
Software (140:2), pp. 102734
Thomas L. D. W., Autio, E., and Gann, D., M. (2014). “Architectural Leverage: Putting Platforms in
Context,” The Academy of Management Perspectives (28:2), pp. 198-219.
Tiwana, A., Konsynski, B., and Bush, A. A. 2010. “Platform Evolution: Coevolution of Platform Architecture,
Governance, and Environmental Dynamics,” Information Systems Research (21:4), pp. 675-687.
Tiwana, A. 2015. “Platform Desertion by App Developers,” Journal of Management Information Systems
(32:4), pp. 40-77.
Turck, M. 2018. Growing Pains: The 2018 Internet of Things Landscape. URL:
https://mattturck.com/iot2018/ (accessed March 12, 2021).
Wortmann, F., and Flüchter, K. 2015. “Internet of Things - Technology and Value Added,” Business &
Information Systems Engineering (57:3), pp. 221-224.
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.
Zhu, F., and Iansiti, M. 2007. “Dynamics of platform competition: Exploring the role of installed base,
platform quality and consumer expectations,” Proceedings of the 28th International Conference on
Information Systems, Montreal, Canada.
