Supervisor: prof. dr. H. Broekhuis Co-assessor: prof. dr. K.J. Roodbergen

(1)

“The influence of input & output control on value appropriation within a digital multi-sided platform.”

MSc Supply Chain Management

University of Groningen , Faculty of Economics and Business

Panagiotis Tzitzinos Tsanousas Student Number: S3116956

Supervisor: prof. dr. H. Broekhuis Co-assessor: prof. dr. K.J. Roodbergen

Word count: 12050 Date: 15.03.19

Acknowledgements

Firstly, I would like to thank my supervisor prof. dr. Manda Broekhuis for her valuable feedback and the overall guidance during the project. Moreover, I would like to thank prof. dr.

K.J. Roodbergen as well for the insightful feedback. Last, I need to thank the three interviewees who agreed to participate in my research project and provided me with all the information that I needed.

(2)

Abstract

Purpose: This study addresses the ways various input & output control policies of market- oriented online multi-sided platforms influence their ability to appropriate sufficient amounts of value.

Design/Methodology/Approach: A multi-method research approach was carried out for answering the research question. An explorative case of Bol.com, two scientific articles with in-depth case studies of Kickstarter and Wechat & Kakaotalk respectively and three semi- structured interviews with practitioners of the field from three other online platforms namely (Fithabits, Aspireiq & Taxi Pickups) were the main source of data.

Findings: The findings provided evidence that decisions about input control policies entail certain trade-offs between quality vs quantity and that loose or strict input control policies can influence value appropriation initiatives either positively or negatively based on the platforms’

strategic vision, maturity stage, cost analysis and market share. Additionally, output control mechanisms were found to affect value appropriation when their outcomes were directly linked with third-party sellers revenue performance.

Theoretical/Practical implications: Current literature will be enriched with a practical examination of different input and output control mechanisms and additional underlying mechanisms that may explain why these control mechanisms influence a platform’s value appropriation ability. Practitioners can use the implications derived from this study to design effective and suitable governance strategies for their platforms so that they can align them with the platform’s objectives and eventually appropriate higher amounts of value.

Originality/Contributions: This study will be one of the scant researches that focuses on the impact of input and output control mechanisms towards an online multi-sided platform’s performance and explore how and why this influence may occur.

Keywords: multi-sided platforms; value appropriation; input control; output control;

governance mechanisms

(3)

1. Introduction

Online multi-sided platforms (MSPs) such as Airbnb, Amazon, and Uber are reforming their industries with regard the way they create and appropriate value (Van Alstyne, Parker &

Choudary, 2016). Online MSPs create value for their different sides (third-party sellers, end- consumers) through connecting unmatched supply and demand (Täuscher & Laudien, 2018), orchestrating interactions and facilitating transactions between these sides (Cennamo &

Santalo, 2013). In return for the value created platform owners appropriate value by using either the commission fees model, the advertising model, the subscription model or the service sales model (Täuscher & Laudien, 2018). However, new online MSPs often struggle to appropriate sufficient value which harms their profitability and competitiveness (Laczko, Hullova, Needham, Rossiter & Battisti, 2018). In fact, the platform’s inability to appropriate value is cited as a common failure reason for many new online MSPs (Van Alstyne et al., 2016).

The fact that online MSPs create value for the whole ecosystem mostly because of the interactions between different sides, makes the need for regulating these interactions a top priority and a strategic decision (Hagiu, 2014). Orchestrating effectively the interactions among the different sides of the platform is cited as a fundamental success factor which is reinforced via suitable governance mechanisms (Schreieck, Wiesche & Krcmar, 2016). More specifically, platform owners need to manage and coordinate third-party sellers and their interactions with end-consumers and the platform itself (Wareham, Fox & Cano Giner 2014) as well as align stakeholder’s interests (Manner et al., 2012). Even though orchestrating the interactions within the platform can have a profound effect towards the platforms’ success, governing an online MSP is not an easy task. It requires balancing the control owned by the platform owner and at the same time managing the level of autonomy handed over to the platforms’ third-party sellers (Tiwana et al., 2010). This is labelled as the paradox of control, where the online MSP needs to centralize the control in order to ensure that platform interactions are conducted in a fair manner for all sides but also delegate control accordingly to its participants so that they can maximize value creation for the platform (Reuver, Sørensen & Basole, 2018).

Input & output control mechanisms are frequently mentioned as effective governance mechanisms (Tiwana, Konsynski & Bush 2010; Hein, Schreieck, Wiesche & Krcmar 2016).

According to Mukhopadhyay et al., (2016) establishing criteria to determine which third-party sellers can access the platform, what they are able to do within it once accepted (input control) and evaluate the final outcomes of the MSP (output control) can drastically reinforce platform owners’ initiatives to reach the platform’s objectives. Past literature acknowledges that there is no ‘one-size-fits-all’ approach in terms of how to use these input and output control mechanisms effectively as every online MSP entails a unique and complex ecosystem

(5)

(Schreieck et al., 2016) and even within the same industries governance mechanisms vary a lot (Hagiu, 2014). However, it is commonly accepted that control is a vital element of a platform’s governance strategy and more specifically input and output control mechanisms are valuable because they provide to platform owners the ability to manage the participants of the platform, measure the effects of the governance decisions and potentially adjust the governance and/or market strategy accordingly (Manner, Nienaber, Schermann & Krcmar, 2012). Prior literature has already identified and theoretically researched the various governance mechanisms used by online MSPs while an examination of their practical implementation still lacks. In addition, prior literature lacks empirical evidence of the potential influence that input and output control may have on the platform’s performance (i.e. value appropriation) (Cennamo & Santalo, 2013).

Thus, this paper addresses the following question:

“How do diverse input and output control mechanisms influence the ability of an online multi-sided platform to appropriate value?”

While value creation is customer-centric and refers to the benefits that third-party sellers and end-consumers derive from their overall platform experience, value appropriation stands for the economic profit that the platform owners will extract for their own interest (Mizik &

Jacobson, 2003). Even though value creation is extensively researched in exchange relationships, value appropriation is underexplored and poorly understood within an online MSP context (Ellegard, Medlin & Geersbro, 2014). Pricing regulations and bargaining with third-party sellers are the most frequently mentioned mechanisms when it comes to value appropriation (Schreieck, Wiesche & Krcmar, 2017). By collecting empirical evidence from practitioners of the field and other secondary data found online the current study will attempt to provide further clarity and understanding towards the topic of value appropriation within the context of a market-oriented online MSP. As mentioned earlier, the ability to appropriate value is a vital element of any online MSP arsenal, thus being able to understand how to appropriate higher amounts of value is a worth-studying area. Additionally, this research will enhance the current literature of input and output control mechanisms through observing different trade-offs that implementation may entails and other strategic decisions to be taken when designing a platform’s governance strategy (Schreieck et al., 2017).

In terms of practical relevancy, the current study will equip platform’s executives with empirical evidence on the impact that various input and output control mechanisms may have on appropriating value. This will eventually enhance their ability to identify how and which of the available input and output control mechanisms would be beneficial to implement within their own platforms. Consequently, this will lead to more effective decision making processes regarding the platform’s governance design, growth strategy and profitability.

(6)

The rest of the paper is structured as follows. First, the theoretical background of the main research constructs and their interplay is provided including a) the market-oriented online MSPs b) the value appropriation mechanisms being used c) the control mechanisms of input and output control and lastly d) the conceptual model of the research topic. Second, the research methods used for data collection and analysis are elaborated. Afterwards, the findings from the data analysis are presented. The final section concludes with a discussion of the results, their implications in practice and theory and some limitations which can become source of inspiration for future research initiatives.

2. Theoretical Background

Within the current chapter, the theoretical background of the topic will be further elaborated and analyzed. First, the online MSPs are defined and thoroughly discussed in order to understand in detail the research context. Next the concept of value appropriation within the online MSPs and the governance mechanisms of input and output control are further analyzed and explained. Lastly, the interplay between the aforementioned control mechanisms and value appropriation is examined.

2.1 Online Multi-Sided Platforms (MSPs)

Online platforms have been given various definitions in prior literature, depending on the perspective that each academic used to approach the topic (Wan, Cenamor, Parker & Van Alstyne, 2017). Schreieck et al., (2016) pinpoint that online MSPs have been studied from either a technology or a market oriented perspective. The technology-oriented perspective refers to software and hardware platforms such as App Stores and wireless networks respectively. On the other hand, the market oriented perspective refers to online marketplaces where goods (Alibaba), services (Airbnb) or information (Medium) are exchanged. In general, every online MSP possess attributes from both perspectives however technology-oriented platforms are focused towards increased innovation and number of third-party applications while market-oriented platforms care more about number of market sides and competitiveness (Schreieck et al., 2016). A more detailed view of the two different perspectives and which platforms they both include can be seen in the Table 1 below. Following the market-oriented perspective, the focus of this study is an online MSP or market intermediary as often mentioned which creates efficiencies for all of its sides by reducing searching costs while bringing together buyers with sellers (Hagiu, 2006).

(7)

Table 1: Technology vs Market Oriented Platforms

The basic structure of a MSP’s ecosystem comprises of four levels, the platform owner who is in control of the property and governs the interactions (i.e. Alibaba Group), the platform provider who serves as the interface point with the users (i.e. Aliexpress), the third-party sellers who sell their offerings in the platform and the end-consumers who purchase those offerings (Van Alstyne et al., 2016).

As it was mentioned before, executives are usually faced with the trade-off deciding the number of the sides that they will allow participating in the platform. For instance, Facebook has opened its platform to advertisers, app developers, game developers, and end-users. More sides result in increased network effects but also lead to much more complexity and potential interest conflicts among the participants (Hagiu, 2014). Thus, coordinating and regulating platforms’

participant’s interactions by using effective governance mechanisms is vital for creating value and eventually appropriating this value (Evans, 2012).

2.2 Value Appropriation

Although value creation and value appropriation are two distinguished processes because it is possible that a firm creates high amounts of value for its participants but fails to appropriate sufficient value for itself (Lepak, Smith & Taylor, 2007), a brief introduction of what value creation stands for will be provided. It is important to have at least an understanding of how platforms create value for their participants because without having value created within their platforms, owners are not able to appropriate value for themselves.

Value creation can be defined as “the invention or reconfiguration of assets and skills to create a usage value (new product, new service) subjectively viewed as new and relevant for potential

(8)

users” (Lepak et al., 2007, p. 182). In the context of digital MSPs, platforms use the information they possess and the interactions of third-party sellers with end-consumers as their main sources of value creation (Van Alstyne et al., 2016). Additionally, online MSPs create value for both their sides by generating indirect network effects (Pagani, 2013). In other words, the larger the number of end-consumers using an online MSP, the more valuable it becomes for third-party sellers to join this platform and sell their offerings. Similarly, the more options end-consumers have regarding third-party sellers, the more valuable the platform becomes for them to navigate and transact (King, 2013).

While value creation refers to the total net value (final outcomes minus input resources) that online MSPs create through their collaborative platforms, value appropriation has to do with the net value of revenue sharing that the platform owners successfully claim for themselves (Wagner, Eggert & Lindemann, 2010). In other words, online MSPs appropriate a portion of the value they co-create through their collaborative relationships with third-party sellers. As Pagani (2013) explained, online MSPs initially incur costs for building the infrastructure which enables them to serve all their sides and then strive to collect revenue from each of these sides for the service provided. However, one of the sides is usually subsidized in order to attract the opposite one. Additionally, a balance of value appropriation and relationship maintenance should be aimed, in order to avoid jeopardizing the platform’s valuable relationships with its third-party sellers (Ellegaard et al., 2014). Platform owners should strive to appropriate as much value as they possibly can for themselves however it should not be at the expense of the welfare of their relationships with the third-party sellers. Sellers and their offerings are the main source of value created for the platform ecosystem and by no means should they be treated inappropriately (Ritala & Tidstrom, 2014). The current study will follow the research of Lepak et al., (2007) with regard the mechanisms that an online MSP can implement in order to appropriate value which are pricing regulations, bargaining, co-selling, absorption and verticalization.

2.2.1 Pricing

Pricing structure is one of the most heavily researched mechanism when it comes to appropriating value as a platform owner. Because of the fact that online MSPs engage with different user-segments (third-party sellers, end-consumers, potentially additional sides) they could possibly charge all of them for participating which could lead to multiple sources of revenue and profit. However, majority of the platforms decide to offer their services for free or with reduced prices to the one side of the platform and profit on the other side (Hagiu, 2014).

Thus, pricing within online MSPs is usually highly asymmetric and it is based on the levels of cross-side network effect (Bakos & Katsamakas, 2008). In other words, the side with the lower

(9)

cross-side network effect usually enjoys a zero or a much lower price than the side which experiences high cross-side network effect (Bakos & Katsamakas, 2008). Subsidies could be also in terms of free information or technical support (Wan et al., 2017). An illustrative example here is a social network platform such as Facebook where end-users can join the platform for free while advertisers have to pay for reaching their audiences. Hagiu (2014) following the same logic suggests that if there is a monetary transaction between the sides then the side which extracts the more value from this transaction should be charged.

2.2.2 Bargaining

When it comes to revenue sharing, bargaining power is a significant variable which is tailored on the specific context of every MSP. In general, the higher the resources invested by an online MSP the more difficult becomes to be imitated (Oh, Koh & Raghunathan, 2015). This implies that the platform that is more difficult to be imitated has greater bargaining power over its third- party sellers. However, all sides are free to negotiate and if there is no agreement over the revenue sharing structures there will be a cost associated because either the platform provider or the third-party seller moves away which means less or even zero value created (Oh et al., 2015).

2.2.3 Co-Selling

Co-selling in another effective value appropriation mechanism that online MSPs leverage in order to increase the value they capture. Co-selling refers to the initiatives that an online MSP takes to actively help its third-party sellers increase their sales (Lepak et al., 2007). Despite the fact that third-party sellers will most likely increase their revenues only from the fact that they do participate in the platform, platform owners can further help them boost their sales which of course in return will increase platform’s revenues as well. Bundling third-party sellers’

offerings with other popular platform’s offerings is one popular technique that increases both sides revenue. Moreover, re-branding third-party sellers’ offerings under the platform’s brand and then selling it is a new product/service can potentially increase revenues. Lastly, providing third-party sellers with certifications regarding their offerings help them actively increase their sales (Lepak et al., 2007).

2.2.4 Absorption

When the platform owner acquires the copyrights of an existing product/service within the platform as well as a product/service that it was not previously offered through the platform is called an absorption initiative (Lepak et al., 2007). The idea behind this mechanism is that platform owners absorb already successful products/services offered via their platforms or not so they can integrate them even better with the platform and increase the revenue earned from these particular offerings.

(10)

2.2.5 Verticalization

Many times more specialized markets within a horizontal focused platform are difficult to seem appealing to their narrowed audiences. By horizontal platform is meant a platform which is not focused on a single market i.e. a specific niche rather a platform which can serve multi customer segments and niches. Therefore platform owners take the initiative to create vertical sub- markets within the horizontal platform (Lepak et al., 2007). That way they create a new sub- market within the already existing platform ecosystem which can attract easier audience interested on this specific vertical market. Eventually, increased interest and sales of this new vertical market will result to increased revenues for the platform itself.

2.3 Governance Mechanisms

Governance or control mechanisms constitute a single dimension of the overall platform governance concept (Mukhopadhyay et al., 2016). Manner et al., (2012, p. 8) defined them as

“the means to monitor the ecosystem, take action on changes in the platform’s ecosystem and to retain the owner’s predominance.” They are the tools that platform owners use for achieving their own goals while maintaining a healthy ecosystem for end-consumers and third-party sellers. Moreover, the platform owners should make sure that these governance mechanisms stimulate third-party sellers to provide offerings that serve the needs of a large and complex audience of end consumers, something that the platform would not be able to accomplish all by itself (Wareham et al., 2014). Boudreau & Hagiu (2008) suggested that platform owners should transition from a purely regulatory role where only pricing structures are considered as governance, to a more “shaping behaviours” and providing positive stimulus role.

With regard to the control and governance mechanisms, several conceptualizations can be found in the prior literature (De Reuver & Bouwman 2011). Tiwana et al., (2010) perceive platform control as a distinctive perspective of a platform’s overall governance and define it as the mechanisms that are used to ensure appropriate attitudes and actions from third-party sellers. Inspired by the main model of organizational control theory, Mukhopadhyay et al., (2016) approached platform governance mechanisms using the behavior, input and output control. However, according to Tiwana (2015) behavior or process control is rather impractical because usually platforms do care about the outcome not the way third-party sellers will achieve it. Third-party sellers are usually free to determine for themselves the detailed actions that need to be taken to reach these objectives (Goldbach , Benlian & Buxmann, 2018). Following the studies of Tiwana (2015) and Goldbach et al. (2018), behavior control is perceived as redundant and thus the current study will focus on the other two control mechanisms namely input and output for further investigating their impact towards the platform’s performance.

(11)

2.3.1 Input Control

Input control has received considerable attention in the prior literature, and it is one of the most vital decisions every online MSP or any platform ecosystem has to face (Boudreau, 2010). Input control refers to regulations that define (a) who is allowed to join the online MSP and (b) what are these participants allowed to do on the platform once they are accepted (Tiwana et al., 2010).

Who is allowed to join the online MSP is usually mentioned in prior literature as the platform’s openness strategy. A platform’s degree of openness can vary from open where participation takes place with zero restrictions to closed where each third-party seller experiences tight regulations in order to gain access to the platform (Wan et al., 2017). In general, opening an online MSP can be horizontally and/or vertically (Eisenmann et al., 2009). Horizontal openness refers to a platform’s interoperations with a rival platform while vertical openness refers to the degree a platform sets restrictions and prerequisites for third-party sellers with regard their participation in the platform (Wan et al., 2017). The current study will focus on vertical openness as the way online MSPs interact and manage third-party sellers is the focal point of this research. Input control can be seen as a gatekeeping mechanism to control platform participation based on the third-party seller’s skills, experiences, and resources which all these eventually affect the quality of their offerings (Tiwana et al., 2010). Other facets of input control are the establishment of service or product minimum requirements which can be checked via frequent inspections, and lastly an identity verification (Schreieck et al., 2016).

Once the online MSPs decide who is allowed to participate in the platform they have to provide guidance on what participants are allowed to do within the platform also referred as level of autonomy (Tiwana et al., 2010). Third-party sellers may have to follow specific community standards, shipping and return policies when selling physical products or predefined pricing policies which safeguard the platform’s identity and possibly limit their autonomy within the platform. Additionally, third-party sellers may be faced with platform or category exclusivity agreements (Eisenmann et al. 2009). A platform exclusivity agreement is when third-party sellers are obliged to affiliate only with this specific platform (single-homing) and not with any other rival ones (multi-homing). On the other hand, a category exclusivity agreement is when the MSP provides exclusively access to a third-party seller to conduct business within a specific product/service category of the platform. This way, the seller is motivated to invest more in their offerings and the platform has the right to claim higher fees (Eisenmann et al., 2009).

2.3.2 Output Control

While mapping an appropriate input control strategy is crucial for every online MSP, maintaining high output quality is also vital. Once third-party sellers have joined the platform

(12)

and are able to reach and sell their offerings to the platform’s audience, a form of output control is applied to ensure that end-consumers receive the best possible experience. This way the platform avoids the so called “lemons” market failure where sellers of low quality offerings turn away from the platform sellers with high quality offerings which eventually lead to the platform’s collapse (Boudreau & Hagiu, 2008).

Output control is a governance mechanism that refers to the “output requirements and performance targets as the prespecified objectives, which are then accordingly monitored, evaluated and rewarded” (Goldbach et al., 2018, p.272). Output control mechanisms differ for every platform usually because of their different business models i.e. objectives, priorities or product/service offerings. To illustrate, a social network controls its outputs via ratings, content engagement or advertising dashboards while a merchant will use reviews, statistics and third- party sellers profile feedback (Linnhoff-Popien, Schneider & Zaddach, 2017). Review systems are either a one-way ranking (Apple - App store) where the demand reviews the supply or a two-way ranking system (Uber) where third-party sellers and end-consumers rank each other (Hein et al, 2016). Boudreau & Hagiu (2009) refer to them as regulatory instruments aimed to ensure implementation of desired actions and mention a couple of them such as lock-out mechanisms, reputation systems, and quality reviews. Lock-out mechanisms are used to ensure that third-party seller’s offerings meet the platform’s quality criteria, an illustrative example is the one of Uber lock-out mechanism where drivers are automatically removed from the platform if they receive multiple times negative feedback from end-consumers. Reputation systems are labels or capabilities that the platform develops and attaches to third-party sellers in order to enhance their visibility within the platform ecosystem. That way, the sellers are rewarded for their performance and end-consumers are guided regarding proper seller selection.

Moreover, quality reviews refer to manually checks that the platform may perform in order to safeguard that the offerings of the sellers are according to the platforms’ standards and principles.

Once online MSPs have monitored the output quality of third-party seller’s offerings they need to have proper metrics in place in order to evaluate the outputs and act accordingly. Such metrics usually refer to number of buyers/sellers, repeat purchases, unmatched searches and lastly frequency & quality of transactions. In terms of analysing and interpreting the output results, online MSPs have as a common denominator the longevity and success of the complementary offerings that third-party sellers sell to end-consumers (Mattila & Seppälä,

2018). Output control is co-related with a non-linear relationship with the platform’s

monetization, i.e. the platform’s value appropriation, (Hein et al., 2016) however no detailed empirical evidence exist on how this relationship functions. A complete overview of the input and output control mechanisms that online MSPs use can be found in the Table 2 below.

(13)

Table 2: Input & Output Control Mechanisms

2.3.3 Interplay of Input & Output Control

In general, loose input control policy is likely to increase the effectiveness of a strict output control policy due to the fact that third-party sellers will perceive output control as fair and legitimate control mechanism which is not going to limit their autonomy levels (Kim & Tiwana, 2016). On the contrary, if strict input control policy is followed by strict output control policy it means firstly that resources invested towards controlling the inputs were wasted and also third-party sellers may resist the platforms’ strict output control mechanisms after experiencing strict input control policies (Kim & Tiwana, 2016).

2.4 Interplay of Control Mechanisms and Value Appropriation

It is possible that value creation and appropriation can be substantially decreased due to opportunistic behaviors, intellectual property conflicts or bad quality offerings (Gandia &

Parmentier, 2017). Collaborative relationships always entail the risks of opportunistic behaviors which must be identified and eliminated as soon as they recognized while intellectual property conflicts may arise during capturing value from the platform usually in relation with

(14)

the value creation activities (Gandia & Parmentier, 2017). Online MSPs can mitigate these negative effects through proper management of the platform’s interactions and interdependencies (Gandia & Parmentier, 2017). Thus, the platform owners need to have in place governance mechanisms which enable them to manage transactions and communications between different platform’s participants (Hein, et al., 2016) and to safeguard that all parties act appropriately and in a fair way (Grewal, Chakravarty, & Saini 2010). At the same time, the platform needs to balance the different or sometimes contrasting interests of the platform’s participants (Reuver et al., 2018). The access to wider audiences that the online MSPs provide to third-party sellers, reinforces the need for having proper governance mechanisms in place.

Transactions and communications between platform’s participants increase exponentially with the platform’s user base and so the platform owner has to ensure that interactions are governed in a beneficial way for itself (Hein et al., 2016).

As Eisenmann et al., (2009) suggest platform vertical openness decisions which are closely related with the input control mechanisms of the MSP should be seen as a trade-off between adoption and appropriability. Adoption pace of an open platform i.e. the ratio of new users joining the platform may be increased because of higher product/service variability offered by the third-party sellers. On the other hand, higher openness may lead to lower appropriated value levels because of increased competition among third-party sellers which result into seller’s dissatisfaction and eventually lower revenues (Eisenmann et al., 2009). This comes in line with the findings of Boudreau, (2010) that opening the platform to third-party sellers can boost platform’s growth and momentum while at the same time reduce the control the MSP can exert to participants or even reduce the platform’s ability to appropriate value. Having a strict input control policy may safeguard the quality of the platform’s offerings however it may also harm value creation in terms of limiting third-party sellers creativity as well as value appropriation in terms of increased operational costs (Schreieck et al., 2017). In general, not having a strict input control policy will lead to a higher user base and an increased product/service variety while at the same time can lead to decreased quality offerings and a less secure platform due to anonymity (Schreieck et al., 2016). Hein et al., (2016) support the idea of a maturity model where in the early stages of the platform development, platform owners follow a less strict input control policy in order to achieve user growth while later on they strive to retain this lost control of the platform.

As it was mentioned before, every online MSP uses output control mechanisms in order to monitor, evaluate and punish or reward third-party sellers’ performance based on some prespecified criteria. When it comes to the one-way or two-way review systems the primary aim of the platform owner is to shift the responsibility of the quality inspection to the platform

(15)

users (Hein et al., 2016). This way, the platform owner can decrease operational costs which is translated into increased revenue for the platform with the fee of decreased control (Tan et al., 2015). Control mechanisms are the centrepiece of a platform’s governance and they have been identified as the means through which platform owners can influence third-party sellers’

behaviour and performance (Goldbach et al., 2018). At the same time, a more strict output control policy similar to a strict input control policy, can be perceived by third-party sellers as an attempt to limit their creativity which can affect the overall performance of the platform ecosystem and harm the value appropriation ability of the platform owner (Schreieck et al., 2017).

2.5 Conceptual Model

The Figure 1 below represents the conceptual model of the research question:

Figure 1: Conceptual model

3. Research Methodology

This chapter elaborates on the research methodology which was used in order to conduct the current study. It starts with the research design and why a multimethod study was appropriate for this research, then continues with the data collection and operationalization of the main constructs. It concludes with how the acquired data was analyzed.

3.1 Research Design

In order to explore the influence of input and output control mechanisms that online MSPs implement towards their ability to appropriate sufficient amounts of value a multimethod

(16)

research approach was followed. Moreover, to provide further clarity and a better understanding with regard the nature of the online MSPs analysed for this study, a classification proposed by (Täuscher & Laudien, 2018) was used which categorized online MSPs into six different clusters namely: a) efficient product transactions, b) product community, c) product aficionados, d) on-demand offline services, e) online services and f) peer-to-peer offline services.

The first stage of the research was to collect secondary data which was publicly available regarding Bol.com, an “efficient product transactions” platform facilitating mostly B2C and C2C transactions. To continue, practitioners from three much smaller online MSPs in terms of market share and user base were interviewed so as to collect rich primary data. Two of them were classified as “on-demand offline services” platforms facilitating solely B2C transactions and the third one as an “online services” platform facilitating C2C transactions. Last, two scientific articles analysing one in depth case each namely the “product aficionados”

crowdfunding platform Kickstarter and the “efficient product transactions” messaging platforms Wechat and Kakaotalk respectively, were used in order to collect extensive secondary data for answering the research question. Prior literature is scarce especially when it comes to empirical data with regard the influence of input and output control mechanisms towards value appropriation, therefore, an inductive research approach seemed more valuable for the scope of the study and its explorative nature (Voss, Tsikriktsis & Frohlich, 2002). Conclusions will be drawn based on the data collected from the three aforementioned sources. Collecting data from three different sources is important because increased generalizability of results could be achieved. Moreover, a comparison between established and new online MSPs will be carried out which otherwise would be infeasible as interviewing established and successful online MSPs seemed impossible during the time the research was conducted. Questions such as ‘how’

and ‘why’ consisted the base for the data collection process following the suggestions of Voss et al., (2002). The unit of analysis of the current study is an online market-oriented MSP which applies input and output control mechanisms in order to regulate participation, interactions and outcomes with regard to its third-party sellers.

3.2 Data Collection & Construct Operationalization

Once the appropriate research design was formulated, data collection was the next step which is explained in detail within this section. Additionally, the operationalization of the basic research constructs of input and output control mechanisms as well as value appropriation will be introduced.

Bol.com was selected because of its huge amount of publicly available data regarding the research topic and also due to the undoubtedly success that the platform has already achieved.

(17)

More specifically, analysing the input and output control mechanisms of a highly successful online platform which accommodates over 200.000 active third-party sellers seemed as a great chance to collect insights and draw conclusions based on decisions made within this platform.

The material selected for this explorative case was found on the platform’s website specifically developed for its sellers named “Partner platform”.

Taking into consideration the input and output control mechanisms presented in Table 2 above, selecting and interviewing three online MSPs was based on finding platforms implementing variations of these mechanisms. To illustrate, I perceive online MSPs with looser input control mechanisms when they do not pose any restrictions or requirements to potential third-party sellers regarding their participation into the platform. On the other hand, I define online MSPs with strict input control when they proceed with any of the aforementioned input control mechanisms or a combination of them. Therefore, the online MSPs which were interviewed had either a loose or a strict input control policy. Following the same reasoning for output control mechanisms, different variations of them were assumed to be always in place because every MSP controls its outputs one way or another. Thus, the desirable variation for output control mechanisms was also achieved. A theoretical replication logic between the cases that were discussed in this research was followed, and online market-oriented MSPs with different input & output control mechanisms were chosen to be investigated so as to produce different results for expected reasons and be able to draw meaningful conclusions for the phenomenon (Seawright & Gerring, 2008). Nonetheless, we do not label this research as multiple case research as the data that we collected per case was rather limited, and does not meet the criteria of in-depth data collection as required in case research.

The procedure for identifying suitable cases was done via the platforms’ websites, where I followed the process pretending that I wanted to participate and become a third-party seller in their platform ecosystem. The initial communication with the platforms was done via email in order to identify whether or not there was any interest for participating in the research. As Myers & Newman (2007) suggest, qualitative interviews were conducted as interviews are a way of getting rich data. I contacted approximately 25 online platforms, and ultimately three of them provided a positive response. After agreement the interviews were done via Skype sessions due to distance constraints and lasted on average 35-40 minutes each. The goal of the current research was to investigate the topic from the perspective of the online market-oriented MSP, thus employees from different but relevant (with the research scope) departments of the MSP were interviewed. Two of the three interviewees were the CEO & Founder of the platform while the third one was a key Account Manager. Further details regarding the online MSPs that participated in the research can be found in the Table 3 below.

(18)

Table 3: Online MSPs Characteristics

A semi-structured interview method was used and the complete interview guide was structured in such a way that could enable the collection of the data needed, including these four aspects:

general information regarding the interviewee as well as the online market-oriented MSP (see part A of the interview guide), value appropriation mechanisms in the MSP context (see part B of the interview guide), governance of platform’s input & output (see part C of the interview guide), and lastly the interplay of those governance mechanisms with value appropriation (see part D of the interview guide). The interview guide included open questions that were linked to each of the variables of the conceptual model. All the research constructs were operationalized based on past literature findings as described in the theoretical background section, research on the platform’s official websites and by using specific questions which could potentially provide measurable outcomes. The detailed operationalization for each variable can be found in Table 4 in Appendix along with the complete interview protocol. The interviews were conducted, recorded and transcribed by the researcher, always asking for the consent of the interviewees.

Last, two scientific articles that describe an in-depth case each about the B2B crowdfunding platform Kickstarter (Thies, Wessel & Benlian 2018), and the messaging applications Wechat and Kakaotalk which turned into B2C online MSPs (Staykova & Damsgaard 2016), were chosen to explore the research question. Kickstarter is an online crowdfunding platform where ambitious entrepreneurs publish their prospective projects looking for capital from funders.

This capital helps entrepreneurs bring in life their projects and funders get rewards in return once the projects are completed. Wechat and Kakaotalk were initially free messaging applications which attracted very fast huge audiences and they were then turned into MSPs connecting supply with demand. The selection criteria for these particular articles were that

(19)

they should describe control mechanisms of already successful online platforms and take into account their whole trajectory i.e. from being a start-up to possessing a leadership position within their industries. Therefore, any changes across the maturity stages of the platforms would provide meaningful insights on the impact of the different control mechanisms towards the ability of the platforms to appropriate value.

3.3 Data Analysis

Analyzing the data collected from the explorative case, the two in-depth case studies and the semi-structured interviews was done via a thematic analysis. Regarding the explorative study of Bol.com all the data acquired from the platform’s website was accumulated and analyzed by allocating descriptive codes on all the data quotes and eventually use the ones more relevant with the research topic. As for the three interviews, initially the data was transcribed and the goal was to get a general understanding of the data which was accomplished by reading thoroughly the transcripts. To continue, data reduction was the next step, which was the result of repeatedly reading the data acquired. Descriptive codes were allocated to all the text segments following the same way mentioned before. By getting to read and thoroughly understand the data collected, replies from the interviewees that were not useful for the overall study scope were eliminated which helped reducing the complexity of the data coding process.

Finally similar codes were then grouped into same categories when needed, so as to conclude with research ‘themes’. Last, with regard the in-depth case studies data was acquired via reading thoroughly the whole articles and then focusing on their findings sections. After reading and understanding the findings sections, all the data relevant with the research topic was also accumulated and analyzed following the same way mentioned before. Once the theoretical

‘themes’ were identified then pattern recognition was done, in order to get a clear overview of the topic. The exact coding tree can be found in Appendix, Table 5.

4. Findings

This chapter aims to present all the findings related with the research topic which were derived from the analysis of the primary and secondary data.

4.1 Retail Case (Bol.com)

Background: Bol.com is the biggest “efficient product transactions” online MSP within The Netherlands & Belgium with over 8 million active customers, 200.000 active third-party sellers accounting for more than 40% of the company’s sales and over 17 million products to offer. It is a product focused online marketplace where B2C and C2C transactions take place. Its focus is on facilitating the connection between supply and demand in the fastest and most effective possible way.

(20)

4.1.1 Bol.com – Main Variables

Value Appropriation: The platform generates most of its revenue by charging its third-party sellers a commission fee when a product is being sold as well as selling its own products. The commission fee differs per (sub) category and consists of a fixed amount per product sold and a percentage of the sales price. Additionally, the platform implements co-selling mechanisms to increase the value that appropriates such as volume discount capability offered to its sellers or assigning golden labels to the top performing sellers signifying their high quality products to prospective customers. Last, the platform generates part of its revenue by providing logistics and warehousing services to its third-party sellers.

Input Control: Bol.com accommodates two different types of third-party sellers namely the Business sellers and the Private sellers. A business seller is “any legal entity that is registered with the Chamber of Commerce in the Netherlands or with the trade register in Belgium, has a VAT number, has created a Sales Account in the platform, has accepted the bol.com Terms and Conditions and has been accepted as a seller by bol.com”. On the other hand, private seller can become anyone who “owns a private customer account at bol.com, is at least 18 years old, has a Dutch or Belgian postal address, has a payment account with a bank in the Netherlands or Belgium, can be reached by e-mail, can accept returns from buyers, sells only original items”. The only input control that the platform implements in order to grant access to the platform is verifying sellers’ identity and thus entering the platform and becoming able to sell is quite easy. However, the seller’s autonomy within the platform is limited in terms of pricing, customer service, shipping and product selection policies. Third-party sellers need to follow specific community standards that the platform has set if they want to be able to sell within the platform. “Bol.com is solely entitled to determine which product categories and/or subcategories may be offered on the platform and / or added to the platform, as well as from which time this becomes possible” (Bol.com, 2018) . As mentioned before, Bol.com has the right to determine the price range that the products should be sold (overpriced products are not shown to the platform’s visitors), to modify any product information that is considered incorrect or misleading or can harm the platform’s reputation and last requires always from sellers to offer free returns to their customers.

Output Control: Apart from controlling what the sellers are able to do within the platform, Bol.com has strict regulations in terms of controlling their outputs. The platform uses a two- way ranking system where customers are able to evaluate their experience and sellers are able to reply to these reviews. Moreover, Bol.com implements a lock-out mechanism based on the service levels the platform has established. Sellers are informed from the beginning that

“Bol.com uses the so-called 'service levels' and that Bol.com reserves the right to block their

(21)

sales account if the they do not meet these service levels. Bol.com reserves the right also to change service levels”. More specifically, “Every week we check whether partners meet the service standards of 'Seller cancellations', 'On time delivered', 'Rating' and 'Track & trace number'. On the seventh time that they do not comply with the service standard, their sales account will be permanently closed.” Apart from monitoring and evaluating third-party sellers’

performance, Bol.com also rewards top performers of the platform. The platform sets its acceptable performance targets and then incentivizes third-party sellers to reach those goals with multiple ways. First, by assigning them the golden label which helps sellers stand out from the rest of their competitors leading to increased visibility i.e. increased sales. Secondly, by providing them new features such as the ability to offer volume discounts to the customers and last, by presenting their product in the buy block which is according to the platform the best offer and it is presented first in the search results. These rewards are only offered once the sellers reach the required performance levels.

4.1.2 Bol.com – Interplay of Control Mechanisms and Value Appropriation Bol.com does not place any significant difficulties for joining the platform and becoming able to sell within it, however due to the regulations that the platform enacts, the sellers’ autonomy level once they’ve joined the platform can be characterized as low. This can be explained due to the increased market share and power that the platform possesses. Bol.com is an established platform in the Netherlands with a huge audience holding a leadership position in its market (E-commerce Report Netherlands, 2018). Therefore, the platform leverages its market power and leadership position to control everything sellers can do within the platform without facing high resistance or complaints from their side. Third-party sellers are aware that in order to continue leveraging the platform’s audience they have to follow these specific community standards, otherwise they will be removed. This way the platform believes that can achieve more controlled interactions, higher service levels, quality products and satisfied end- consumers. All these outcomes then lead to increased top-line revenue for the platform i.e.

increased value extracted.

Regarding the output control policy, the platform directly links the seller’s performance with the revenue that they can generate through the platform. Nurturing its third-party sellers to care about their performance in order to increase their sales eventually increases the platform’s revenue which is getting paid a mediation fee once a product is being sold through the platform. Obviously, Bol.com is able to appropriate higher amounts of value by helping its sellers to increase their revenue and ensuring a great experience for the end-consumers. An overview of the control mechanisms used by Bol.com and their influence on value appropriation can be seen in the Table 6 below.

(22)

Table 6: Bol.com Findings

4.2 Three Cases based on empirical data (Interviews)

4.2.1 Fithabits

Background: The first of the three cases is about Fithabits, an “on-demand offline services”

online MSP operating in the fitness industry based in Greece connecting professional fitness studios with people looking to exercise. The majority of the platform transactions are B2C and efficiency gains in terms of connecting conveniently supply with demand is the value proposition of the platform for both sides (sellers, end-consumers). It is a fairly new business endeavor founded in 2017 with a quite low but steadily increasing amount of end-users. End- consumers use the platform for booking their offline workout sessions to any of the available fitness studios that the platform has to offer.

4.2.1.1 Fithabits – Main Variables

Value Appropriation: The fitness studios (third-party sellers) are able to list in the platform all their available offline services and leverage its audience. In return, the platform charges a 15%

commission fee to these sellers’ for every workout session booked through the platform. At the

(23)

moment, this is the only way the platform appropriates value for itself. Even though, the CEO

& Founder admitted that not too many users have already booked their workout sessions through the platform, this number is steadily increasing. As the CEO mentioned, they also plan to add additional revenue streams such as advertising capability for other firms once the user base grow significantly. Additionally, entrance fee payments for third-party sellers will be required who join the platform for free at the moment.

Input Control: According to the CEO & Founder of the platform, the input control policy with regard the sellers selection process of the platform could be currently characterized as strict.

“When we reach out to new fitness studios we make appointments in their places, we also check their infrastructure to ensure that it is in line with our standards”. Checking manually the fitness studios is there to ensure that service providers meet the quality criteria. This way, the platform owner can guarantee a pleasant experience for the end-consumers who book their workout sessions through it. However, this manual check of the fitness studios is only feasible at the moment because of the low number of third-party sellers participating in the platform.

“When the requests from studios will be increased at a pace that we cannot do that for every one of them manually, we will still for sure make an identity verification”. The long-term goal of the platform is to “dominate the fitness market, thus we will not be able to exclude someone and all the quality control will be shifted to end-users who will be able to choose based on the reviews”.

With regard to the autonomy levels that sellers have once they are allowed to join the platform, the CEO & Founder believes that the sellers are fully autonomous on how they conduct business within the platform. The only requirement that the platform sets is that “the price for the platform’s customer should be equal or lower to the price offered to their customers off the platform.” Other than that, sellers are free to decide which of their services can offer within the platform even though they are encouraged by the platform owner to include all the services they are offering offline.

Output Control: Controlling its outputs is also important for the platform. That is why apart from asking end-consumers to review their experience with the seller (one-way ranking system) they do run additional quality tests. “We have a team of 20 people visiting the fitness studios, pretending they are customer and living the whole platform’s experience from the beginning.

In the end, we ask them to fill out a questionnaire and we get feedback from their own true experiences”. The platform does not have in place specific performance requirements based on which punishments/rewards are derived and they still act per case. As usual, a punishment is the permanent removal of a seller from the platform. On the other hand, the platform rewards

(24)

the very communicative and engaged with the platform sellers by promoting them more through the platform’s social media pages. That way, they congratulate the seller by providing higher visibility to their services while at the same time encouraging other sellers to embrace the same attitude towards the platform.

4.2.1.2 Fithabits – Interplay of Control Mechanisms and Value Appropriation The platform is in its early stages where building a reputation and creating a loyal customer base is more important than profiting. Therefore, input control mechanisms are chosen in such a way to ensure that the sellers who join the platform will attract end-consumers and service them in a satisfactory way. Granting access for participation in the platform is a strategic importance issue for the platform owner. Implementing a strict input control policy with proper fitness studios selection is the way the platform owner decided to ensure the high quality offerings that they are looking for. Therefore, the input control policy of the platform influences its value appropriation initiatives indirectly, through providing the means for proper selection of the fitness studios. This in return, can and will enhance the overall positioning of the platform within the marketplace which will lead to increased end-user and fitness studios interest in joining the platform. Building a highly engaged community around the fitness industry will allow the platform owner later on to appropriate higher amounts of value through adding additional revenue sources mentioned before.

At the moment Fithabits charges its sellers a 15% commission fee on every workout session booked via the platform, however “The ultimate goal though, is to reinvest this 15%

commission earned from the fitness studios to coupons/discounts for our customers i.e.

provide lower prices to our platform’s users which will lead to not profiting that much at the time but which will make end-users more familiarized and used to book through the platform, it will also help us create this network of sports studios and end-users, which will eventually provide us the ability to monetize afterwards this loyal customer base.” In other words, the platform in the beginning decided to appropriate value through charging sellers low commission fees, with the plan to generate additional revenue sources once the platform grows and creates a passionate community around fitness. These additional revenue sources for the platform will be revenue from advertising initiatives via relevant fitness product placements and participation fees for the fitness studios who join the platform for free at the moment.

As the CEO & Founder claims, the real reviews on the platform at the moment are not that much, therefore this is a reason why the platform deploys additional internal resources to test the quality of the sellers’ services. Deploying additional resources for testing the sellers quality means that the platform is able to appropriate less value. Nevertheless, as it was mentioned

(25)

before, value appropriation is not the most important goal of the platform at the current stage.

The platform is now focusing on increasing indirect network effects which will expand its user and seller base growth. This in return, will allow the platform owner to appropriate higher amounts of value. An overview of the control mechanisms used by Fithabits and their influence on value appropriation can be seen in the Table 7 below.

Table 7: Fithabits Findings

4.2.2 Aspireiq

Background: The second case is about Aspireiq an “online services” B2B Influencer Marketing platform which connects content creators with brands looking for exposure and additional content. The creators are considered as the sellers of the platform and the brands as the end-consumers. To illustrate, when a brand is looking to increase its brand awareness they can use this platform to find and collaborate with content creators who will help them achieve this goal.

(26)

4.2.2.1 Aspireiq – Main Variables

Value Appropriation: The platform subsides the creators and let them be for free in it because they would find it difficult to convince them join otherwise. However, the brands do pay on a monthly basis for the software that the platform offers to them as well as the connection with a large number of creators. As the account manager said, “We leave the creators using the platform for free because we emphasize on extracting value from the brands by providing them the connections with the creators as well as the CRM tool to manage these relationships.”

Input Control: Controlling the inputs of the platform is not a high priority as there is not even an identity verification taking place before creators can join the platform, however the platform has a community manager responsible to check whether the creators’ profiles are inappropriate in terms of type of content and originality of their audience. As the account manager who was interviewed said “We do not have strict selection criteria but it happens to reject applications from accounts that are radical and inappropriate and do not align with our platform terms of use. But it’s not very common that this happens, and usually when a brand wants to work with a creator we will probably let them join the platform.” Additionally, once creators join the platform they do have full autonomy to present their work portfolio and all the information that will make it easier for brands to trust them and work with them. However, they can only apply to certain campaigns that the platform shows to them.

Output Control: A one-way review system is applied within the platform, where the brands are able to evaluate the performance of the creators once the collaboration is completed. The platform does not have pre-specified performance metrics for evaluating creators because brands have different KPIs when it comes to evaluating creators’ performance. Therefore, performance of creators is judged per case and based on the completion of the terms agreed upon with the brands. The platform rewards the creators who receive positive reviews from the brands, by providing them higher visibility within the platform and allowing them to apply for more campaigns.

4.2.2.2 Aspireiq – Interplay of Control Mechanisms and Value Appropriation Aspireiq creates value for its customers (the brands) by connecting them with a huge variety of creators and by providing them a CRM tool to manage all their relationships with them.

However, they decided not to control its inputs and in general if a brand wants to work with a creator they let them join the platform. Therefore, input control mechanisms are loose and the only reason a creator can be denied access to the platform is because their profile is inappropriate and does not match with the platform’s community standards. It is not their focus

(27)

to deploy resources for securing the inputs of the platform. Nevertheless, account managers consult the brands on selecting the correct creators and on proper strategies for leveraging the platform to its maximum. The main value proposition of the platform for the brands is the software it provides to them by paying a monthly subscription as well as the connection with a large amount of creators. It is not in the platform’s strategic interest to guarantee high quality creators for the brands rather to provide them with plenty of options. Thus, a clear platform’s decision on quantity vs quality of creators is being made and is achieved via the platform’s input control policy. Eventually, achieving the platform’s strategic vision i.e. connecting brands as efficiently as possible with as many creators as possible will lead to higher amounts of value appropriated.

Using the output control mechanisms to identify which of the creators perform well and then make them more visible to the brands, gives the ability to the platform to provide brands with well-performing creators to work with i.e. creates value for the brands which are then more likely to continue using the platform i.e. paying the monthly subscription to the platform. In other words, measuring and evaluating the creators’ performance and then rewarding them based on this performance is a win-win situation for all the platform’s participants and the platform itself which lead to happier customers and increased value appropriated for the platform itself. An overview of the control mechanisms used by Aspireiq and their influence on value appropriation can be seen in the Table 8 below.

Table 8: Aspireiq Findings

Supervisor: prof. dr. H. Broekhuis Co-assessor: prof. dr. K.J. Roodbergen

MSc Supply Chain Management

University of Groningen , Faculty of Economics and Business

Panagiotis Tzitzinos Tsanousas Student Number: S3116956

Supervisor: prof. dr. H. Broekhuis Co-assessor: prof. dr. K.J. Roodbergen

Word count: 12050 Date: 15.03.19

Abstract

Table of Contents

1. Introduction

2. Theoretical Background

2.1 Online Multi-Sided Platforms (MSPs)

2.2 Value Appropriation

2.3 Governance Mechanisms

2.4 Interplay of Control Mechanisms and Value Appropriation

2.5 Conceptual Model

The Figure 1 below represents the conceptual model of the research question:

3. Research Methodology

3.1 Research Design

3.2 Data Collection & Construct Operationalization

3.3 Data Analysis

4. Findings

4.1 Retail Case (Bol.com)

4.2 Three Cases based on empirical data (Interviews)