Digital Platform Leadership: Exploring the Role of National Factors, Platform, and Customer Type

Summary

Digital platforms are a ubiquitous phenomenon that threatens incumbent firms. Among digital platform providers, the competition is fierce. Platform giants such as Google, Amazon, or Microsoft strive for platform leadership. The question of how to achieve platform leadership is equally essential for platform companies and incumbent firms transforming into platform business models. This thesis aimed to provide a fundamental building block for the analysis of this question. In the first step, a taxonomy that conceptualizes platforms from a business model perspective was created. In the second part of this thesis, this taxonomy was used to analyze configurations of successful and leading platform business models. This thesis found multiple configurations of platform leaders within an ecosystem-level and firm-level model. The two models extended existing theory on the causal relation of factors on platform leadership.

Keywords:taxonomy, digital platform, configurational analysis, platform leadership, fuzzy-set qualitative comparative analysis

Table of contents

Summary II

Table of contents III

List of figures V

List of tables VI

List of abbreviations VII

1. Introduction 1

1.1 Motivation 1

1.2 Research questions 2

1.3 Research overview & philosophy 3

2. Theoretical background – digital platforms 5

2.1 Background, definitions & types of digital platforms 5

2.2 Perspectives on digital platforms 8

2.3 Platform ecosystems & dynamics 12

2.4 Determinants of platform leadership 13

2.5 Hypothesized configurational propositions for platform leadership 16

3. Research design – taxonomy development 18

3.1 Overview of the taxonomy development 18

3.2 Taxonomy development methodology 19

3.3 Grounded-theory literature review 22

3.4 Case database 24

3.5 Validation interviews 25

4. Results – Taxonomy development 27

4.1 Conceptual-to-empirical iteration 27

4.2 Empirical-to-conceptual iteration 34

4.3 Final iteration 37

5. Research design – fuzzy-set Qualitative Comparative Analysis 40

5.1 Introduction and overview of fsQCA 40

5.2 fsQCA analysis steps 42

6. Results – fsQCA 49

6.1 Ecosystem-level model 49

6.2 Firm-level model 53

6.3 Derived configurational propositions for platform leadership 58

7. Discussion 60

7.1 Contribution to theory and practice 60

7.2 Limitations 62

7.3 Outlook & further research 63

Appendix A. Taxonomy development 73

Appendix B. fsQCA 75

List of figures

Figure 1: Platform overview 7

Figure 2: Evolution of platform ecosystems 13

Figure 3: Platform leadership model 16

Figure 4: Design of the taxonomy development approach 19

Figure 5: Development of dimensions per iteration 27

Figure 6: Exemplary coding procedure 28

Figure 7: Derived dimensions in iteration 1 34

Figure 8: Development of dimension in iteration 2 37

Figure 9: Ecosystem-level model solution 50

Figure 10: Firm-level model solution 55

List of tables

Table 1: Iteration overview 21

Table 2: Hits for the defined search string and inclusion criteria 23

Table 3: Selection funnel for the initial set of literature 24

Table 4: Interview overview 25

Table 5: Meta-characteristics changes overview 32

Table 6: Final taxonomy 39

Table 7: Condition overview 43

Table 8: Membership allocation of conditions 44

Table 9: Exemplary coding guide 44

Table 10: Exemplary value assignment 46

Table 11: fsQCA condition allocation 46

Table 12: Ecosystem-level model solution sets 49

Table 13: Firm-level model solution sets 54

List of abbreviations

API Application programming interface

CtE Conceptual-to-empirical

EtC Empirical-to-conceptual

fsQCA fuzzy-set qualitative comparative analysis

IS Information systems

RBM Regression-based model

SDK Software development kit

WTA Winner-take-all

1. Introduction

1.1 Motivation

In 2007, Nokia was the dominant player in the market for mobile phones and leveraged its competitive advantages such as strong product differentiation and a trusted brand. At that time, Apple and Google were not active in this industry. Today, 13 years later, the dominance has dramatically shifted. Nokia and most of the incumbents from 2007 have lost their competitive battle against companies such as Apple and Google. Important follow-up questions after this shift include how exactly this has happened and potentially how other companies can replicate these trajectories.

What fundamentally differentiates the business models of Apple’s iOS and Google’s Android operating systems is the change towards a platform business model. With this platform business model, the users and app developers are brought together on the same platform. With technological developments and the evolution of their features, smartphones became central to our daily lives. In this regard, the magnitude of available apps and functionalities mostly determines the value of smartphones. Compared to Nokia, which used a linear set of activities, Apple and Google leveraged a platform business model to co-create value with third-party app developers and thereby increase the value for users (Van Alstyne, Parker, & Choudary, 2016b).

The success of this business model has not only been limited to (mobile) operating systems. In the hospitality industry, platform business models such as Airbnb challenges the traditional hotel-chain incumbents (Guttentag, 2015). Amazon, another thriving example of a platform business model, challenges incumbents in the retail (Hamilton, 2011), the cloud computing (Kenney & Zysman, 2016), and the film industries (Galloway, 2016).

In 2018, Apple shattered a barrier no company had ever broken down before: they achieved a market capitalization of $1 trillion (Gurman, 2018). The four companies – Alphabet, Amazon, Apple, and Microsoft – that had been poised to reach a $1 trillion market capitalization (Soper, 2018) have transformed their business model and are now operating multi-sided platforms. Interestingly, most of these platforms do not own or create any of the goods or services that are traded on them (Goodwin, 2015). Instead, they match supply and demand in a way that is mutually beneficial for all groups involved in platform-based exchanges.

These trajectories and dynamics challenge organizations that lack guidance on how to design a platform business model suitably. The question of how to transform traditional business models into platform business models remains an open one for companies.

When it comes to the academic discourse about (digital) platforms, researchers (Gawer, 2014; Reuver, Sørensen, & Basole, 2017) identified several challenges that the research on (digital) platforms is confronted with. Reuver et al. (2017), for example, call for the provision of more precise definitions of the core concepts involved to establish clarity in platform research. Successfully achieving this accuracy demanded by the researchers also requires clarity on the unit of analysis and which aspect of a particular platform is included in studies. Other researchers share the awareness that a unified representation and description scheme is valuable for the academic discourse around platforms. Baldwin and Woodard (2009) were one of the first to work on analytical tools to study the representation of platforms. In this context, taxonomies help to provide a basis for fundamental research by guiding common domain language (Nickerson, Varshney, & Muntermann, 2013).

The question on the design of a platform business model is one of many that remain as postulated by Reuver et al. (2017). Tura, Kutvonen, and Ritala (2018) provide a design framework for platforms that is composed of four elements: (1) platform architecture (2) value creation logic (3) governance and (4) platform competition. Although this framework provides a starting point for guidance, it lacks concrete design elements from a business model perspective. To answer the question on how platform leadership can be achieved, the creation of design principles is required. This can be supported by a configurational analysis, which accounts for the complex character of digital platforms and the ambiguity of their development towards platform leadership. (Hein, Setzke, Hermes, & Weking, 2019)

1.2 Research questions

Current research on platforms is divided into two streams of literature: industrial economics and engineering design (Gawer, 2014). The derived concepts from these two streams can be further categorized into one of three perspectives: economic, technical, and governance (Gawer, 2014; Rolland, Mathiassen, & Rai, 2018). While providing valuable insights, current research lacks a holistic and comprehensive view of platforms (Reuver et al., 2017). The combination of the different streams and perspectives lays a foundation for further investigation on how platforms should be designed to achieve sustainable success and potentially even gain platform leadership (Gawer, 2014).

Following the proposal by Baldwin and Woodard (2009) to create an analytical framework to analyze platforms holistically, the first research question builds on the open question “How should digital platforms be designed?”by Reuver et al. (2017). Further, Tiwana, Konsynski, and Bush (2010) identified the lack of clarity on how the design and governance choices influence the evolution of a platform. Here, a taxonomy would help to structure the complex domain and provide a building block to these open questions. Accordingly, this is the first research question (RQ) for my thesis:

RQ 1: How can business models of (digital) platforms be classified in a taxonomy?

As a taxonomy enables researchers to study the relationships among concepts and theories (Nickerson et al., 2013), I extend past approaches that laid out basic principles(Boudreau & Hagiu, 2009) on how to design platforms. This framework is used to conduct configurational research on business models of successful and non-successful platforms. A taxonomy also helps to guide further research on platforms by providing clarity on core concepts and the respective unit of analysis, as proposed by Reuver et al. (2017).

In this regard, Boudreau and Hagiu (2009)argue that simply “getting the prices right” is not sufficient for platforms to achieve sustainable competitive advantages. Other non-price instruments must be applied to properly build successful platforms. Other considerations, such as where to place a platform on a continuum between open and closed (Gawer & Cusumano, 2008), are deemed to be essential in explaining the emergence of successful platforms.

To deep dive into platform leadership, I apply the taxonomy to investigate configurations of platform leaders in different settings and through different lenses. I use a fuzzy-set Qualitative Comparative Analysis (fsQCA) as a configurational research approach to generate insights on the configurations of successful platforms. The following three research questions are to be answered during the fsQCA and thereby further enhance research on platform leadership.

RQ 2: Are there configurational differences between successful innovation and transaction platforms?

RQ 3: Are there configurational differences between platform leaders in a B2B or a B2C environment?

RQ 4: Are there configurational differences between European and US-based digital platforms that are relevant to platform leadership?

I am particularly interested in the influence of the platform type, innovation, and transaction platforms, on platform leadership (RQ 2).Further, the differences between platforms in different customer settings have not yet been discussed. I, therefore, aim to answer RQ 3 to generate insights on platform leadership from this perspective. Lastly, data on platforms from different countries (P. Evans & Gawer, 2016) indicates that US-based platforms outcompete platforms from other countries. RQ 4 aims to investigate the influence of a platform’s origin on platform leadership.

Besides enriching the academic discourse around platform leadership, the results and expected findings should serve as a guide for the design process of business models of platforms.

1.3 Research overview & philosophy

To answer the research questions, I conducted two steps. In the first step, I developed a platform taxonomy that is used as an input framework for a fsQCA to investigate configurations of platform leaders in the second step.

Following the introduction, in Chapter 2, I present the theoretical foundation for my thesis. This chapter gives an overview of the field of (digital) platforms, covers the status quo of several research streams, and presents current research challenges associated with platforms. This chapter also presents the key definitions that are relevant. Furthermore, I introduce platforms and their associated concepts and characteristics using the proposed clustering into three different perspectives (Gawer, 2014; Rolland et al., 2018). In this chapter, I also cover the topics of ecosystems and their evolutionary dynamics. Chapter 2 concludes with a summary of the determinants of platform leadership and an overview of configurational propositions that are the subject of study in my thesis. In Chapter 3, I present the employed methodology for the taxonomy development process as the first part of the thesis. In Chapter 4, I present the results of the iterative taxonomy development process. Since the results from the taxonomy development inform the methodology of the configurational analysis, I introduce the methodology of the fsQCA in Chapter 5. I present the results of the fsQCA in Chapter 6 and conclude with a discussion of its limitations, the contribution, and an outlook for further research in Chapter 7.

Research projects implicitly include certain assumptions about the nature of knowledge and how it can be derived, either consciously stated or unconsciously developed by the conducting researchers (Easterby-Smith, Thorpe, & Jackson, 2015). These assumptions fundamentally shape the design of a particular research project (Saunders, Lewis, & Thornhill, 2009). Therefore, it is necessary and essential for any research project to provide transparency about the ontological and epistemological position to explain and justify the chosen research design (Easterby-Smith et al., 2015).

In this research, I tried to understand which configurations of platforms have a positive impact on platform leadership. This required the assessment of current knowledge and the development of new knowledge. The chosen research methodology determines the explanatory value of the result. The methodology, in turn, is determined by the underlying research philosophy, respectively.

This thesis follows the proposal by Henfridsson and Bygstad (2013) to adopt critical realism as the underlying philosophical standpoint for this study. Critical realism is recognized for its emphasis on building theory (Easterby-Smith et al., 2015). Critical realism combines an interpretive epistemology with a realist ontology (Archer, 1998) and is increasingly recognized as a promising philosophical standpoint to bridge the gap between objectivism and relativism (Smith, 2010). It accounts for differences in the value of theories in approximating reality and thereby makes the assessment of current knowledge valuable. Following Henfridsson and Bygstad (2013), I adopt critical realism to align present prospects on digital platforms. Hence, I chose a multi-method research design (Mingers, 2001), which is introduced as a condition for research incorporating critical realism (Wynn & Williams, 2012).

2. Theoretical background – digital platforms

2.1 Background, definitions & types of digital platforms

The economic study of platforms has been a research topic since the 1980s. The economists' Rochet and Tirole (2003, 2006) were among the firsts to operate in this field. Other influential economic work has followed in the early 2000s (Gawer & Cusumano, 2002). With the rise and success of online companies such as eBay, Amazon, or PayPal, the idea of platforms became popular for both researchers and professionals (Spulber, 2018). It garnered the attention of researchers, specifically within information systems (IS) research (Reuver et al., 2017). Since then, platforms have been a research object for academics from various research streams (Gawer, 2014).

As already introduced, research on platforms is divided into two different streams of literature. The economic stream defines platforms as markets and focuses on the related dynamics within and across platforms. The engineering design stream focuses on platforms as technological designs that foster innovation using modular components. (Gawer, 2014, p. 1240) Due to the heterogenic literature on platforms, researchers suggest unifying the different streams and perspectives (Gawer, 2014; Spulber, 2018). Gawer, in particular, proposed to view platforms as evolving organizations (2014) rather than conceptualize platforms either as markets (Rochet & Tirole, 2003) or modular technological architectures (Baldwin & Woodard, 2009).

This fragmented landscape of research also results in various definitions of platforms and associated concepts. Existing definitions suffer from either being too specific or too vague. This results in disagreement within platform literature about what constitutes a multi-sided platform (Hagiu & Wright, 2015). In the context of digital platforms, Reuver et al. (2017) argued that all literature, while providing a valuable partial explanation, lacks a holistic understanding of platforms. The authors, therefore, call for greater clarification of the core concepts.

According to Reuver et al. (2017), research in the field of digital platforms comes with three challenges. First, in current research, many conceptualizations of digital platforms exist. The authors raise the concern that part of this research does not consider the specific characteristics of digital technology. Therefore, they recommend providing explicit definitions of concepts used in research projects, such as digital platforms or ecosystems. Secondly, the unit of analysis of digital platforms changes over time because of the vibrant character of digital platforms. Cross-platform development, for example, unbundles applications such as Spotify from one single platform such as the Apple App Store and iOS, respectively. The comparison of platforms naturally comes with a high degree of complexity due to the intertwined nature of ecosystems. The authors, therefore, recommend emphasizing on the respective scope of a research project when studying and comparing different platforms. Third, they argue that from a methodological point of view, the difficulty in isolating the unit of analysis and the evolutionary dynamics of digital platforms makes research on digital platforms increasingly complex.

Picking up the first recommendation by the authors, I start by providing a clear definition of platforms and digital platforms, respectively.

Armstrong (2006, p. 668) defined platforms as “markets involving two groups of agents interacting via platforms where one group’s benefit from joining a platform depends on the size of the other group that joins the platform”. When a platform mediates more than two different user groups, Boudreau and Hagiu (2009) considered a platform to be multi-sided.

Hagiu and Wright (2015) proposed two key features that help to distinguish platforms from other business models. First, platforms allow direct interaction among two or more different parties, and second, each side makes platform-specific investments.

Therefore, platforms are particular kinds of markets that facilitate exchange between different types of users that otherwise could not interact or transact with each other. Gawer (2014) and several authors share the understanding of a platform as being at least two-sided to be considered a platform (Eisenmann et al., 2006; Tiwana, 2014).

Reuver et al. (2017) argued that there are fundamental differences between digital and non-digital platforms. A non-digital platform refers to the fundamental underlying concept of a platform, such as multi-sidedness and the match of supply and demand. Digital platforms are rooted in digital technologies, which in turn imply a variety of characteristics that differentiate them from non-digital platforms. Digitality involves a layered architecture of software where the homogenization of data, editability, reprogrammability, distributedness, and self-referentiality characterize these technologies. These characteristics create the possibility to separate form from function and to postpone individual design decisions. (Reuver et al., 2017; Yoo, Henfridsson, & Lyytinen, 2010)

Researchers do not necessarily agree when it comes to defining digital platforms. Proposed and used definitions differ between different authors. Ghazawneh and Henfridsson (2015, p. 199), for example, defined digital platforms as “software-based external platforms consisting of the 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”.

Contrary to this view, Constantinides, Henfridsson, and Parker (2018, p. 1) defined digital platforms as platforms that “are created and cultivated on top of digital infrastructures” and “that allow multiple stakeholders to orchestrate their service and content needs”. The difference in these definitions lies in the main transaction. In the former definition, the main transaction is supposed to be digital. In the latter definition, it can also be offline if the platform itself is technology-enabled. Given the case of Uber or Airbnb, where the main transaction is offline, the classification of whether these platforms are digital platforms differs depending on which definition is used.

Based on the request to present clear definitions, postulated by Reuver et al. (2017), I refer to the latter definition of digital platforms as proposed by Constantinides et al. (2018). I define digital platforms as a platform where the platform and the underlying processes are digital. This definition can also include platforms where the main transaction is offline.

Figure 1: Platform overview

Own illustration based on (Boudreau & Hagiu, 2009; Eisenmann, Parker, & Van Alstyne, 2006)

The ubiquitous existence of platforms in different contexts needs classification and differentiation of different types of platforms. Although such a classification would be helpful in our understanding of configurational differences between different types of platforms, current research and proposals to cluster platforms lack a common understanding and stringent perspective.

Gawer and Cusumano (2014) differentiated between internal, company-specific platforms and external, industry-wide platforms. Internal platforms are a collection of assets organized in a shared structure that enables a company to efficiently create and build a range of derivative products. Industry-wide platforms build on the former but form the basis on which external complementary companies can develop their products or services. Baldwin and Woodard (2009) also differentiated between platforms as product lines, platforms across firms as multi-product systems, or platforms as multi-sided markets. I refer to platforms as industry-wide platforms.

These industry-wide platforms are further classified into different types of platforms. D. S. Evans (2003) classified platforms following their coordination mechanism into market-makers, audience-makers, and demand coordinators. In a later publication, the authors (2008) proposed to distinguish between exchanges, advertiser-supported media, transaction devices, and software platforms.

To overcome complexity, Cusumano, Gawer, and Yoffie (2019) presented two basic types of platforms depending on their primary function: (1) transaction platforms and (2) innovation platforms. The authors also noted that some platforms operate as hybrid platforms at the intersection between these two types. Transaction platforms mainly enable transactions for people and organizations to share, buy, sell, or access a variety of goods and services. On an innovation platform, the owner and complementor can make use of the central architectural (technical) building blocks that are accessible via interfaces to create new products or services on the platform.

It does, however, make sense to distinguish platforms in other dimensions. They can further be clustered according to their transaction type – if digital or offline – and their transaction content – if product or service – (Täuscher & Laudien, 2018). Further, platforms can be differentiated if the platform exists on its own (e.g., eBay or Uber) or is a complement to a physical product (e.g., Apple OS X) (Kim, Prince, & Qiu, 2014).

2.2 Perspectives on digital platforms

Adding to the two different streams of literature, as touched upon in the previous chapter, platforms are analyzed from three different perspectives (Gawer, 2014; Rolland et al., 2018; Tiwana et al., 2010). The perspectives are mainly based on the respective research stream they originate from (Gawer, 2014). Current research lacks an appropriate bridging between the various research streams (Gawer, 2014) to explain the evolutionary dynamics of platforms (Tiwana et al., 2010). Both streams have developed separately and conceptualize platforms either as markets (Rochet & Tirole, 2003) or modular technological architectures (Baldwin & Woodard, 2009). In the following, I present these perspectives and fundamental concepts.

Economic perspective

The economic perspective views platforms as markets that facilitate an exchange between different user groups (Gawer, 2014) and that disrupt traditional markets through efficient interaction between these user groups (Rolland et al., 2018)

Network effects might be the most fundamental consideration of platforms. Network effects refer to the increase in value for one user when brought together with another user or user group (Cusumano & Gawer, 2002; D. S. Evans, 2003; Reuver et al., 2017). Network effects can be direct (same side) if the value of a user depends on the number of users in the same user group or indirect (cross-side) if the value of the user depends on the number of users in other user groups (Gawer, 2014; Reuver et al., 2017). Network effects – both same-side and cross-side – can also be negative or turn negative. One example would be the effect of advertisers on a social media platform. At first, the value for users increases with advertisers joining the platform, assuming valuable advertising content. Although, at a certain point, this effect turns in the opposite direction because the increasing number of ads overloads users with information and leads to a perceived lack of independent advice. (Reuver et al., 2017)

Gawer (2014) classified direct network effects as demand-side economies of scale and indirect network effects as demand-side economies of scope.

Platform adoption influences the success of a platform. With existing network effects, the critical question for platform operators is how to bring all sides on to the platform. In platform literature, this is referred to as the "chicken-and-egg problem" (Caillaud & Jullien, 2003) and relates to the challenge for a platform that certain user groups only join the platform if the other user group is already present. An adequate pricing strategy that potentially involves subsidizing one side of the platform can help to overcome this challenge (Gawer, 2014, p. 1241). Authors defined the subsidy side of platforms as the side that is highly valued by the money side when enticed in high numbers (Eisenmann et al., 2006; Parker & Van Alstyne, 2019).

The pricing of a platform is different compared to pricing done in a traditional industry. In these industries, prices are, for example, determined by the marginal costs or customers’ willingness to pay. In contrast, for the pricing of a platform, the platform owner has to set prices for both sides, considering the impact of the chosen price on the respective growth of users. (Eisenmann et al., 2006)

Eisenmann et al. (2006) introduced the idea of winner-take-all (WTA) dynamics where a market is likely to be served by one single (or few) platform(s) if certain conditions are fulfilled. The authors argued that three conditions have to be met for a market to be a WTA market: (1) At least one side of the platform encounters high multi-homing costs, (2) network effects are positive and strong, and (3) no side of the platform has strong preferences for special features. Multi-homing costs refer to all costs for user groups (adoption, operation, and opportunity costs) that are incurred with the presence on multiple platforms (Rochet & Tirole, 2003).

The economic view is a valuable source for understanding pricing and market dynamics. These topics need to be extended by other topics such as innovation dynamics and platform generativity. Generativity, in the context of a digital platform, results from the postponing of certain design decisions on product features (Reuver et al., 2017). This is enabled by the layered modular architecture of digital technology, which will be discussed in more detail in the next section of this chapter (Yoo et al., 2010). Generativity, therefore, refers to the possibility to recombine modules (Constantinides et al., 2018), which enables an infinite space for product possibilities (Boudreau, 2017). Svahn and Henfridsson (2012), in this context, note that certain design decisions allow a distributed innovation process with users and complementors.

Further, Eisenmann, Parker, and Van Alstyne (2011) referred to two options to obtain substantial market share for platform entrants: First, due to network effects and switching costs, new platforms have to offer functionality that is superior to competitor’s offerings. A second path is to combine its functionality with functionality that has proven to be relevant and successful and thereby leverage the shared user relationships. This strategy is called platform envelopment, where a platform envelops functionality or complete platforms, for example, by an acquisition. Platform envelopment aims to foreclose or limit the incumbent’s access to its users and can be a powerful strategy as examples such as Microsoft’s envelopment of the streaming media platform RealNetworks with its solution, the Windows Media Player shows.

These strategic moves relate to the topic of competition with complementors on the platform. The example of Microsoft and other examples show that these strategies are a relevant consideration. Wen and Zhu (2019), for example, showed that competition of the platform operator with its complementors, in a mobile operating system environment, results in a decrease in updates and an increase in prices of complementary products. They also showed that the threat of competition alone triggers discouragement among complementors. Hence, the authors also noted that this problem is multi-folded and that platform competition can also be beneficial for complementors if it increases the overall value of the platform. Foerderer, Kude, Mithas, and Heinzl (2018), for example, showed that the introduction of the Google Photos app leads to an increase in the number of apps in adjacent areas. The introduction also had a positive effect on the section of photo apps.

Technological perspective

The technological perspective views platforms as technical architectures in which the design hierarchy of the different modules of the platform is the object of reflection (Gawer, 2014). Platforms employ a modular architecture with a stable core and variable peripheral components (Rolland et al., 2018) and offer the opportunity for distributed development and innovation through modularization (Baldwin & Clark, 2000).

The platform’s core and periphery differ in its evolution. The core of the platform remains stable to provide central functionality for the core interaction (Parker, Van Alstyne, & Choudary, 2016). The core builds the foundation for all other products developed on the platform. In contrast, the peripheral part of the platform, which is composed of its complements and accessed via interfaces, is subject to change and development (Baldwin & Woodard, 2009).

The architecture of a platform aims to reduce both structural and behavioral complexity to make it manageable and support innovation activities on the platform (Tiwana, 2014). Architecture is the lever for a platform to balance the need for autonomy of complementors and the integration of the complementor’s work into an interrelated ecosystem. A platform can, therefore, be described as a “purposefully designed complex system” (p. 84) with an underlying construct that determines how complementor and user act, work, and develop (Baldwin & Woodard, 2009). Tiwana (2014) defined four desirable properties of platform architectures. The properties are interrelated, and platform designers have to consider trade-offs when making design decisions:

·Simple: The architecture should be simple so that the major subsystems, the reused functionality, and the interactions are comprehensible

·Resilient: The connection of complements and the platform should be weakly coupled using interfaces so that the platform is resilient to an app malfunction

·Maintainable: The use of interfaces and the partitioning into subsystems should ensure cost-effective maintenance of the platform without destructing depending complements

·Evolvable: Interfaces must be designed and developed so that the platform can evolve into a configuration that it was not originally intended to

A platform can be decomposed into smaller subsystems – a step called modularization. Across subsystems, the interdependence is reduced by design. (Henfridsson & Bygstad, 2013) Through modularization, it is intended to balance coordination costs and the increased imitation risk (Baldwin & Woodard, 2009). Modularization also needs a balance between openness and secrecy to share information about available interfaces while keeping the proprietary specifics of the platform secret. This balance, in turn, is achieved through decoupling and the standardization of interfaces. Decoupling refers to the process of decomposing a system into a set of smaller subsystems that are independent of each other. Through the standardization of interfaces, platforms aim to successfully reintegrate independently developed subsystems such as applications. (Tiwana, 2014) In the design process, the platform owner must make certain decisions on which functionality is kept within the platform and which is kept outside the platform. A simple heuristic for this decision is to ask whether the functionality is low or high in reusability, and if the use is unique to a few apps or shared by many apps (Tiwana, 2014). Modularization can further be extended by the idea of layered modular architecture, which extends the modularization of physical products with four loosely coupled layers of digital technology: device, network, service, and content. (Yoo et al., 2010)

The specifics of interfaces further specify the decision of what to keep inside and outside the platform. Boundary resources are the interfaces for third-party complementors to overcome the boundary line between what is inside and what is outside the platform, access the platform, and create value (e.g., in the form of applications) on the platform (Ghazawneh & Henfridsson, 2013). The important strategic considerations entail the provision and the design of technical boundary resources such as Application Programming Interfaces (API) or Software Development Kits (SDK) as well as social boundary resources such as incentive schemes (Yoo et al., 2010). Hence, boundary resources are the mechanism for a platform to maintain control over the platform while enabling complementors to contribute and participate (Eaton, Elaluf-Calderwood, Sørensen, & Yoo, 2015). These boundary resources are subject to shift the focus of the platform from developing or innovating itself to providing third-party complementors with the needed capabilities to generate complementary value on the platform. Hence, the design of boundary resources raises the tension between control over the platform and stimulating third-party innovation. (Ghazawneh & Henfridsson, 2013)

Governance perspective

The governance perspective is defined by several authors (Gawer, 2014; Rolland et al., 2018) and is related to the interaction and tension of platform innovation and competition. This organizational perspective combines the economic and technological aspects of platforms and highlights the design of a platform as an organization (Gawer, 2014) to foster the organization and coordination of innovation practices through technical mechanics and social precautions (Rolland et al., 2018). Further, platform governance defines decision rights on the platform and deals with the influence of these decisions on the evolution of modules within and the ecosystem of a platform (Tiwana et al., 2010). For platforms, next to the technical configuration, one of the most fundamental governance questions is the choice of boundaries, to concurrently open the platform to external complementors while maintaining control and coordination over the platform (Boudreau 2017). In this context, Tiwana et al. (2010) defined three distinct governance perspectives, that contain certain decisions a platform must make: (1) decision right partitioning, (2) control, and (3) proprietary vs. shared ownership.

From a governance perspective, a platform is located on a continuum between being open or closed (Gawer, 2009). The decision to open a platform involves certain trade-offs that have to be considered in the process. In general, there is a trade-off between adoption and appropriability (West, 2003) or put in a different way between innovation and control over the platform (Boudreau, 2010).

Boudreau (2010) distinguished between granting access to a platform, enabling the development of complementary components, and giving up control over the platform itself. The different approaches of opening a platform, in turn, have different impacts on the innovation on platforms.

The level of restrictions that are placed on the participating actors, both users, and complementors, further determine the degree of openness. While users can be restricted in joining the platform, complementors can be restricted in the development or the commercialization process. In this context, restrictions can further be evaluated in terms of reasonability. (Eisenmann, Parker, & Van Alstyne, 2009)

Further, the openness of the sponsor and provider level determines the degree of openness of a platform. The sponsor refers to the instance that controls the participation on the platform, and that exercises property rights. The provider refers to the first point of contact for the different user groups. On both levels, a platform can be open, which means that several organizations can be sponsor or provider, or closed, which in turn means that the role of sponsor and/ or provider is obliged to one organization only. (Eisenmann et al., 2009)

Digital infrastructure evolving at unprecedented speed blurs organizational boundaries as it invariably triggers the creation of new control points. There is also a tension between flexibility through digital technology and the constraints created by former investments and design decisions. Generativity is associated with a paradoxical duality of change and control so that infrastructure is stable and flexible on the one hand and controlled and autonomous on the other hand. (Reuver et al., 2017; Tilson, Lyytinen, & Sørensen, 2010) This paradox of change is characterized by “the opposing logics of stability and flexibility that operate across infrastructural layers and components” (Tilson et al., 2010, p. 6). The paradox of control further considers the impact of specific strategic actions on the mode of control regarding the change (Tilson et al., 2010).

2.3 Platform ecosystems & dynamics

Jacobides, Cennamo, and Gawer (2018) define ecosystems as “[..] a set of actors with varying degrees of multilateral, nongeneric complementarities that are not fully hierarchically controlled”. An integral classification of ecosystems is that they support the coordination of interrelated but autonomous organizations mostly through a modular layered architecture (Baldwin & Clark, 2000). Some authors (Jacobides et al., 2018) claim that Modularity is one condition for the emergence of an ecosystem and that rules, standards, and interfaces support in resolving evolving coordination issues

From a business network perspective, an ecosystem is a multi-entity system composed of independent groups of companies or organizations that foster relationships between involved companies. These relations are characterized by a balance between competition and collaboration that involve complementarity and therefore promote coevolution. (Aulkemeier, Iacob, & van Hillegersberg, 2019)

Tiwana (2014, p. 3) defined an ecosystem as the combination of the platform itself, the apps that are connected to the platform via interfaces, and the shared technical infrastructure the platform is using. He described ecosystems as “new blueprint for competition – one that puts ecosystems in head-to-head competition”. Van Alstyne et al. (2016b) share this notion and referred to platform ecosystems as the combination of the platform itself as well as the producers and consumers that interact on the platform. Gawer and Cusumano (2008) referred to a platform leader and the associated complementors on the platform as an ecosystem of innovation. Other authors (Cusumano & Gawer, 2002; Tiwana et al., 2010) included the platform and the modules that are specific to the platform into a platform’s ecosystem. The notion of ecosystems extends my view on platforms.

Figure 2: Evolution of platform ecosystems

Own illustration based on (Tiwana, 2014)

In the context of evolutionary dynamics, Tiwana et al. (2010) raised the question of how the internal fit of the platform’s technological architecture and governance influences the evolutionary dynamics of ecosystems and modules. This question also confirmed the proposal by Gawer (2014) to bridge the different perspectives on platforms. The presented perspectives in Chapter 2.2 together create a holistic picture of platforms that can explain evolutionary dynamics (Figure 2). The combination of the endogenous choices by the platform and the exogenous dynamics influence the evolution of a platform (Tiwana et al., 2010).

2.4 Determinants of platform leadership

The question of platform leadership is linked to the work of Utterback and Abernathy (1975). The researchers investigated the emergence of a dominant technology in a certain industry and characterized this as the emergence of a dominant design. This research initiated a stream of literature (Suarez, 2004).

The battle for platform leadership not only determines the fate of the involved actors and their technologies, but in the case of platforms, it also determines the outcome of a whole array of complementors and their offerings (Suarez, 2004).

According to Leong, Pan, Leidner, and Huang (2019), the number of connected actors, the level of market dominance (D. S. Evans & Schmalensee, 2016), and the platforms switching costs (Eisenmann et al., 2011) are usual measures to determine platform leadership. Gawer and Cusumano (2014) argue that platform leaders are in the position to influence the development of the overall technological and business system. Further, platform leadership can be characterized by the ability of a company to drive innovation in a platform’s ecosystem and the leveraging of positive network effects (Cusumano & Gawer, 2002). In summary, platform leadership or a platform leader is defined as the leading force behind industry-wide innovation (Cusumano & Gawer, 2002).

In a longitudinal case study on Intel, Microsoft, and Cisco Gawer and Cusumano (2002) uncovered insights regarding the path to achieving platform leadership. In this context, the authors referred to platform leadership as the common objective “to drive innovation in their industry” (Gawer & Cusumano, 2002, p. 6). The idea of this objective is to create a platform that evolves into the foundation upon which other companies build their products or offer their services. Therefore, strategic decisions mainly contain considerations regarding the integrity and evolution of a platform. (Gawer & Cusumano, 2002, 2008)

Gawer and Cusumano (2002) developed a framework with four levers for platform leadership that should guide strategies for and the design of platforms:

(1) Scope of the firm: refers to the decision of the platform what to keep and do inside the platform and what complementors are encouraged to do outside

(2) Product technology (architecture, interfaces, intellectual property): refers to decisions regarding the openness and modularity of the platform

(3) Relationships with external complementors: refers to the desired relation to complementors that create value on the platform and if this relation is more collaborative or competitive

(4) Internal organization: refers to the organizational design of a platform and determines how the platform deals with complementors and the evolution of the platform

Suarez (2004) argued that independent of the market-size and the technology type, two groups of factors influence the battle for leadership: (1) firm-level factors, and (2) environmental factors. This view also confirms the hypothesis by Tiwana et al. (2010) that the environmental fit of firm-level factors determines the evolutionary dynamics of a specific platform. The environmental factors have two effects on technological dominance or platform leadership. The first effect is direct, where the factors have an immediate effect on the outcome of the leadership battle. The second effect is indirect as it mediates the firm-level factors to some extent. Suarez (2004) named four factors for each of the factor groups. For the firm-level factors, the author lists the following elements:

(1) Technological superiority in relation to competing alternatives

(2) Complementary assets and credibility to profit from the technological innovation

(3) Installed base to leverage network effects and increase adoption of users

(4) Strategic maneuvering, including entry timing, pricing, relationships with complementors

For the environmental factors, the author lists the following elements:

(1) Regulation within the home market to operate a certain technology

(2) Network effects and switching costs

(3) Regime of appropriability to capture the value created by a certain technology

(4) Characteristics of the technological field, especially in terms of which technology outcompetes other technologies

Suarez (2004) specifically pointed out the argument that no single factor is strong enough to determine leadership. Instead, it is the result of the interplay of several firm- and environmental-level factors that determine the outcome.

In a later work, Gawer and Cusumano (2014) highlighted four key factors that platforms need to consider to achieve platform leadership:

(1) A vision for the platform and appropriate communication of this vision within the platform’s ecosystem

(2) A sufficiently open and/ or modular architecture to foster and orchestrate third-party innovation and complements

(3) Mutually beneficial relationships with complementors and other participants

(4) An approach to continuously evolve the platform and associated ecosystem

In recent work, Pan and Lin (2019) identified three main buckets of characteristics, with several factors within each bucket, that positively influence platform leadership. These following buckets confirm the existing elements from previous work:

(1) Entrepreneurial model of network systematism, including business rules and patterns with continuous innovation

(2) Embedded collaborative open innovation, including the openness of technical interfaces and the availability of data

(3) Platform resources, including valuable technical resources

Contrary to the work on platform leadership, scholars also worked on common reasons and explanations for platform failure (Van Alstyne, Parker, & Choudary, 2016a):

(1) Failure to optimize openness

(2) Failure to engage developers

(3) Failure to share surplus

(4) Failure to launch the right side

(5) Failure to put critical mass ahead of money

(6) Failure of imagination

Yoffie, Gawer, and Cusumano (2019) investigated and analyzed 250 platform cases and uncovered mistakes of platforms that can be clustered into four categories:

(1) Mispricing on one side of the market

(2) Failure to develop trust with users and partners

(3) Prematurely dismissing the competition

(4) Entering too late

I synthesized the existing work on platform leadership into a model, as illustrated in Figure 3. The beforementioned work shares the differentiation between firm-level and ecosystem-level factors. On the firm-level, all authors agree that a sufficient open product technology is crucial to foster third-party innovation, which in turn has a positive impact on platform leadership (Gawer & Cusumano, 2002, 2014; Pan & Lin, 2019; Suarez, 2004). Authors also agree that a mutually beneficial relationship with complementors (Gawer & Cusumano, 2002, 2014), as well as an entrepreneurial plan for the evolution of the platform and the ecosystem, is needed (Gawer & Cusumano, 2014; Pan & Lin, 2019). Further, appropriate pricing is crucial for platform leadership and furthermore impacts the relation with complementors (Suarez, 2004). On the ecosystem-level, authors share the assumption that high switching costs and positive network effects have a positive impact on platform leadership (Suarez, 2004).

Figure 3: Platform leadership model

Own illustration based on (Gawer & Cusumano, 2002, 2014; Pan & Lin, 2019; Suarez, 2004)

The work on platform leadership is criticized for its ex-post view, which potentially limits the practical application. Suarez (2004) argued that a more valuable source for the understanding of these leadership dynamics comes from an ex-ante view on industry dynamics.

2.5 Hypothesized configurational propositions for platform leadership

In this chapter, I present configurational platform propositions of platform leaders. Based on the developed platform leadership model and the presented concepts in Chapter 2.2, I develop propositions for platform configurations that are tested in the fsQCA. In this context, I refer to platform leadership, as presented in Chapter 2.4.

The analysis of the origin of a platform showed that the market capitalization of US-based platform companies is significantly higher than the one of Asia- or Europe-based companies. In fact (P. Evans & Gawer, 2016) showed that from the combined market capitalization of the 176 studied platforms in the Global Platform Survey, 73% of those come from US-based platforms. Interestingly, this is not a matter of the number of platforms, as only 36% of the platforms in the sample come from the US.

Based on these results, I develop the first proposition. US-based platforms have a competitive advantage solely through being based in the US. In other words, it is likely that out of two platforms that are identical except for their origin, a US-based platform is expected to outcompete an EU-based platform. Accordingly, the first proposition for this thesis is:

P1: US-based platforms tend to achieve a high degree of platform leadership than non-US platforms.

Further, the link between openness of and innovation on a platform has been touched upon by several researchers. Boudreau (2010) found that opening a platform by granting access to outside complementors had a positive effect on innovation by complementors. Hence, the study also showed that maximizing openness by giving up control over the platform has a decreasing impact on innovation. The adoption of complementors is found to be influenced by the openness of a platform (availability of interfaces) and appropriate documentation of these interfaces (J. Song, Baker, Wang, Choi, & Bhattacherjee, 2018). Innovation on a platform is believed to have a positive effect on platform leadership (Gawer & Cusumano, 2014; Pan & Lin, 2019). This leads to the second proposition: Openness of a platform in terms of available interfaces as well as documentation of these has a positive impact on innovation on the platform, which in turn has a positive impact on platform leadership. The second proposition is defined as:

P2: Platforms with a high degree of platform leadership are characterized by a certain degree of openness.

As mentioned before, not only the degree of openness but also the governed relation with complementors is assumed to be crucial for a platform’s success (Gawer & Cusumano, 2014). In that regard, the documentation of interfaces, the utilized pricing schemes, or the competition with complementors are crucial aspects. Platforms in this context are believed to emphasize creating mutually beneficial relationships with complementors. In that sense, the third proposition for this thesis is:

P3: Platforms with a high degree of platform leadership are characterized by emphasizing mutually beneficial relationships with complementors.

Following the notion of WTA dynamics (Eisenmann et al., 2006), strong network effects, as well as high multi-homing costs, are assumed to have a positive impact on platform leadership. Network effects work as the main lever for the beforementioned factors. Multi-homing costs work as a lock-in mechanism to keep users and complementors on a platform. Hence, this lock-in is, to some extent contradicting to the point of mutual beneficial relations as made before. The fourth proposition for this thesis is:

P4: Platforms with a high degree of platform leadership can leverage strong network effects and high multi-homing costs.

A final proposition is closely related to the topic of innovation on a platform. Several researchers emphasized the importance of the evolution of a platform. Gawer and Cusumano (2014), as well as Pan and Lin (2019), highlighted the need for continuous innovation of a platform. This includes the development of the ecosystem surrounding the platform. The final and fifth proposition for this thesis is the following:

P5: Platforms with a high degree of platform leadership are characterized by efforts to further develop the platform.