The choice for a reseller or multi-sided platform

How to source our product?

The choice for a reseller or multi-sided platform

Master of Science, Supply Chain Management

University of Groningen, Faculty of Economics and Business Yanne de Lang

Student number: 2289431 Y.de.Lang@student.rug.nl

1

Supervisor: Dr. H. Broekhuis 2

Supervisor: Prof. dr. K.J. Roodbergen

March 13, 2017 Words: 11.682

Acknowledgements

First, I would like to thank my supervisor Dr. Manda Broekhuis for her guidance and comprehensive feedback. Secondly, I would like to thank Dr. Roodbergen for his detailed feedback on my research proposal. Moreover, special thanks to my company supervisor for helping me wherever possible so I could successfully complete this project. Furthermore, I would like to thank all the interviewees for their time and effort participating in this research.

Finally, I would like to thank my family and friends for their unconditional support during

these last months.

Abstract

This research aimed to explore how the need for perceived control and various product characteristics influence an e-retailer’s decision to source and sell a product via its reseller model or multi-sided platform. To investigate this issue, a multiple case study was conducted at a large e-retailer. This method was particularly suitable since it provided the opportunity to gather rich data about the relatively new trend among e-retailers to adopt both a reseller model and multi-sided platform to source and sell their products. The findings show that the e-retailer prefers to opt for the reseller model when it wants to exercise control over a product’s price, delivery and marketing activities. Additionally, this research’s findings suggest that a product’s characteristics in terms of seasonality, long-tail, the product’s life cycle and a product’s level of complexity influence the e-retailer’s model choice. This research adds to current literature as it allows for the examination of the interplay among the involved factors. Practitioners may benefit as this study’s findings might provide them with a guideline as which model to choose to source and sell a specific product.

Keywords: e-retailer; reseller model; multi-sided platform; perceived control

For confidential reasons, some data to describe the research context or that explicitly point

towards company ABC has been removed from this thesis.

Inhoudsopgave

1. INTRODUCTION... 4

2. THEORETICAL FRAMEWORK ... 7

2.1. THE RESELLER AND MULTI-SIDED PLATFORM ... 7

2.2. PERCEIVED CONTROL ... 8

2.3. PRODUCT CHARACTERISTICS AND THEIR IMPACT ON MODEL CHOICE ... 14

3. METHODOLOGY ... 18

3.1. RESEARCH CONTEXT ... 18

3.2. CASE SELECTION ... 18

3.3. DATA COLLECTION ... 19

3.4. DATA ANALYSIS ... 21

4. FINDINGS ... 22

4.1. RESELLER OR MSP: WHAT WAS THE DECISION? ... 22

4.2. SUPPLIER MANAGEMENT ... 23

4.3. WHAT FACTORS INFLUENCE THE CHOICE FOR THE RESELLER OR MSP MODEL: CONTINGENCY FACTORS ... 25

5. DISCUSSION ... 30

5.1. THEORETICAL IMPLICATIONS ... 30

5.2. PRACTICAL IMPLICATIONS ... 33

5.3. LIMITATIONS AND SUGGESTIONS FOR FURTHER RESEARCH ... 34

6. CONCLUSION ... 35

REFERENCE LIST ... 36

APPENDIX A: OPERATIONALIZATION OF VARIABLES ... 46

APPENDIX B1: INTERVIEW GUIDE (1

ROUND) ... 47

APPENDIX B2: INTERVIEW GUIDE (2

ROUND) ... 53

APPENDIX C1: CODING TREE – SUPPLIER MANAGEMENT AND CONTROL ... 54

APPENDIX C2: CODING TREE - CONTINGENCIES ... 57

1. Introduction

How to source our products? Almost any intermediary business is confronted with this question when it engages in business transactions (Hagiu & Wright 2013). Traditionally, most businesses buy products from their suppliers, which they may hold in stock and then sell them to their customers. This is referred to as the reseller model (George & Bock 2011). The increasing popularity of e-commerce has, however, given rise to a new business model, the so-called multi-sided platform (MSP). In this case, the intermediary offers a marketplace where suppliers can directly sell their products to the intermediary’s customers (Elaluf- calderwood et al. 2012). Instead of making an explicit choice in favor of either the reseller or MSP model, most e-retailers decide to operate as both (Gautier et al. 2015). Consequently, the question arises which products to source, and sell, via each model (Hagiu & Wright 2015).

Previous literature has already examined the dynamics of a MSP (e.g. Provan & Kenis 2007; Eisenmann et al. 2011; Abhishek et al. 2016) sometimes directly comparing them to those of a reseller (e.g. Alstyne et al. 2016; Hagiu & Wright 2015; Gautier et al. 2015).

Several researchers have studied the design of a MSP (Tiwana 2014; Staykova & Damsgaard 2015) and how to create value as MSP (e.g. Amit & Zott 2012; Simpson et al. 2001). This study treats the answers to those questions as a given, and attempts to explain how these factors influence the choice to source, and sell, a product as MSP or reseller. Previous literature depicts that the intermediary’s relationships with its suppliers significantly differs when it operates as MSP or reseller, which is suggested to influence the choice to source, and sell, a product via either model (Jiang et al. 2011; Hagiu & Wright 2015; Zott & Amit 2008;

Abhishek et al. 2016). Therefore, this research is also connected to the substantial amount of literature on buyer-supplier relationships (Poppo & Zhou 2014; Huang et al. 2014; Simpson et al. 2001b; Wacker et al. 2016; Kim & Choi 2015). Additionally, this research is related to earlier work about the (mis)match between a product’s characteristics (e.g. sales volume, pricing decisions, profitability, physical design and marketing activities) and the choice for either model (Hagiu & Wright 2013, 2015; Abhishek et al. 2016; Simpson et al. 2001; Jiang et al. 2011).

Neither model is argued to be universally better, as each has its strengths and weaknesses (Provan & Kenis 2007), but researchers seem to agree that the intermediary’s perceived level of control over the sales transaction is a crucial factor to distinguish the MSP from the reseller model (e.g. Hagiu & Wright 2013, 2015; Abhishek et al. 2016; Jiang et al.

2011; Eisenmann et al. 2011). Perceived control reflects the belief that one can command and

exert control over a process, and as a consequence its outcome (Collier & Sherrell 2010). As

reseller, the intermediary itself engages in a (contractual) relationship with its suppliers and end-customers, whereas as MSP the intermediary mainly acts as facilitator of the (contractual) relationship between its suppliers and end-customers (Hagiu & Wright 2015). Therefore, it controls and interacts differently with its suppliers for each model (Choi & Krause 2006;

Boudreau & Hagiu 2009; Alstyne et al. 2016). Generally, it’s believed that the intermediary’s perceived level of control over its suppliers, and hence the flows of goods and information, is higher when it operates as reseller than when it operates as MSP (Alstyne et al. 2016;

Abhishek et al. 2016; Jiang et al. 2011; Hagiu & Wright 2013). Previous studies have also shown a link between the choice for either model and certain product characteristics. For instance, scholars have already examined the link between the intermediary’s choice for the reseller or MSP model in relation to a product’s marketing activities (Hagiu & Wright 2015), its price (Abhishek et al. 2016) and innovative versus functional products (Zott & Amit 2008).

These studies suggest that the choice for the MSP or reseller model is determined by the extent to which a company wants to control factors that influence customers’ purchasing decisions.

All these studies, however, choose a product factor beforehand and then started to describe how this factor did or didn’t fit with either model’s level of control. They seem to have paid little attention to the question how the need for perceived control influences the intermediary’s model choice without relating it to a specific product factor (Abhishek et al.

2016; Hagiu & Wright 2015; Simpson et al. 2001; Dekker et al. 2013) nor how these factors might influence each other. This research tries to address this gap by studying how the interplay of the e-retailer’s perceived level of control and/or product characteristics influence an e-retailer’s decision to source and sell a product via its reseller or MSP model. Therefore, this research focuses on the question: how do the need for perceived control and product characteristics influence the decision for either a reseller or MSP model in an e-retail business?

Since the MSP is still a relatively new business model, insights that support managers to trade-off its pros and cons against those of a reseller model, is still scarce (Jiang et al. 2011;

Abhishek et al. 2016; Gautier et al. 2015). This research aims to contribute to existing

literature by focusing on how various factors interact and thereby influence the e-retailer’s

decision to opt for the reseller or MSP model. Practitioners may benefit from this study as it

provides them with a more complete overview of which trade-offs may be involved when

deciding to source and sell a product as reseller or MSP.

This paper continues as follows. First, the theoretical framework will introduce the reseller

and multi-sided platform and highlights the main differences between them. Thereafter, it’s

discussed how the e-retailer controls its suppliers via each model and how certain product

characteristics might fit better with one model or the other. Chapter 3 elaborates on the

methodology, by introducing the applied method, and the process of data collection and

analysis. The next chapter then introduces the findings concerning the factors that influence

the choice for the reseller model or its MSP. Chapter 5 places those findings in perspective by

linking them to existing literature and elaborates on the limitations and suggests directions for

further research. Lastly, some concluding remarks are provided.

2. Theoretical framework

2.1. The reseller and multi-sided platform

The Internet has transformed the retail industry as customers are increasingly shopping online (Doherty & Ellis-Chadwick 2006). Traditionally, e-retailers solely operated as reseller to source and sell their products. However, during the last decade, a new business model has emerged (Staykova & Damsgaard 2015). Several e-retailers have started to also operate as multi-sided platform (MSP) to source and deliver their products to their customers (Gautier et al. 2015). As a reseller, the e-retailer first purchases the goods from its suppliers, may hold them in stock and then sells them to its customers (Hagiu & Wright 2013), whereas as MSP, the e-retailer is only the facilitator of a transaction between the supplier and the e-retailer’s customers. The supplier directly sells its product(s) to the customer (Eisenmann et al. 2011).

Figure 2.1a and 2.1b illustrate this difference.

Figure 2.1a. Reseller model Figure 2.1b. Multi-sided platform

Another important difference between the reseller and MSP model, is the number of relationships the e-retailer needs to manage. The reseller model involves the establishment of a set of dyadic relationships between the e-retailer and its suppliers (Choi & Krause 2006).

Due to the limited number of ties, these relationships may result in close collaborative

practices between the parties involved (Kim & Choi 2015). This can include frequent

information sharing (Kim & Choi 2015), long-term contracts (Saccani et al. 2014) and a

feeling of mutual dependence (Autry & Golicic 2010). In contrast, a MSP represents a

network of numerous actors (i.e. e-retailer, suppliers, customers) that may all interact with

one another (Boudreau & Hagiu 2009). The more suppliers and customers connect via the e-

retailer’s platform, the higher its value (Alstyne et al. 2016). The e-retailer’s main

responsibility is to facilitate the interactions between suppliers and customers (Provan &

Kenis 2007). This model allows the e-retailer to establish thousands of ties with independent sellers that sell products via its platform (Jiang et al. 2011). Due to the higher number of potential relationships the e-retailer may establish, this model is more suitable for the development arm’s length ties that are often characterized by short-term contracts. limited interactions (Kim & Choi 2015) and a lower level of commitment of either party (Autry &

Golicic 2010).

The main benefits of the platform stem from its scale and evolvability (Olleros 2008).

Scalability is the ability to manage an increasing number of participants without an increase in its variable costs (Bondi 2000). Evolvability refers to the network’s ability to scale up and down to quickly anticipate changing customer demands (Mason & Spring 2011). As MSP, the company can rapidly expand its product portfolio’s variety or increase a certain product’s availability, without incurring significantly higher costs due to an increased number of products and/or suppliers that must be managed (Wang & Koren 2012). In contrast, the main advantage of the reseller model results from reduced transactional uncertainties (Srinivasan et al. 2011). For example, the e-retailer can reduce the negative impact of demand fluctuations or supply disruptions by holding inventories.

2.2. Perceived control

The concept of perceived control stems from the belief that people can deliberately influence their surroundings (Skinner 1996 in Hajli & Lin 2016). It reflects the change an individual believes he/she can achieve between the initial situation and that same situation’s end-state by undertaking a specific action (Kay et al. 2010). Therefore, perceived control can be defined as one’s perceived ability to command and exert control over a process and, as a consequence, its outcome (Collier & Sherrell 2010).

In inter-firm relations, a company can apply various governance mechanisms to increase the (perceived) level of control over a partner’s actions, processes and/or behavior.

Governance mechanisms can be broadly divided into two types: contractual and relational (e.g. Cao & Lumineau 2015; Lumineau & Henderson 2012; Woolthuis 2005; Zhou & Xu 2012). Contractual governance refers to the purposefully designed sets of structures, procedures, routines and processes to exercise control in interfirm transactions (Li et al.

2010). They explicitly specify which party, at which point in time, controls the (flows of) goods and information relevant to the exchange (Casadesus-Masanell & Ricart 2010).

Relational governance, on the other hand, can be defined as the extent to which the

relationship is governed by a sense of mutual trust, solidarity, fairness and informal rules and procedures (Cao & Lumineau 2015; Zhou & Xu 2012).

Whether contractual- and relational governance are complements or substitutes is still subject to debate, but most researchers agree that they can both help to control partners’

actions and behavior. Previous literature has identified a set of commonly used governance tools that are available to organizations. Those include: contracts (Cao & Lumineau 2015; Rai et al. 2012; Liu et al. 2010), (the development of) trust (e.g. Zhou & Xu 2012; Poppo & Zhou 2014; Rai et al. 2012), information sharing (e.g. Uzzi 1997; Frohlich & Westbrook 2001;

Nicolaou et al. 2011), operational alignment (e.g. Bennett & Klug 2012; Frohlich &

Westbrook 2001; Prajogo & Olhager 2012) and resource investments (e.g. Van Donk & Van Der Vaart 2005; ; Frohlich & Westbrook 2001).

Contract: A valid contract is “a promise or agreement that the law will enforce”

(Lysons et al. 2003, p.192). Often, it specifies the roles, responsibilities as well as liabilities for the parties (e.g. buyer, supplier, e-retailer) bound by this agreement (Griffith & Zhao 2015). An important issue the parties need to address is the level of contract specificity (Wuyts & Geyskens 2005). Contract specificity presents the contract’s degree of explicitness, specification and precision (Mooi & Ghosh 2010). The level of contract specificity can be viewed along a continuum ranging from a contract that only broadly outlines the terms of the exchange to one that precisely specifies each party's responsibilities and possible solutions in case of foreseeable issues (Wuyts & Geyskens 2005). Generally, the more detailed the contract, the higher the level of perceived control (Dekker et al. 2013).

As reseller, the e-retailer is expected to build a set of relatively close, long-term relationships with a limited number of suppliers. Broader contract guidelines provide the parties involved with more adaptation possibilities in case market conditions change (Weber

& Mayer 2011) and offer an incentive for bilateral adjustments that promote a cooperative relation (Poppo & Zenger 2002). In contrast, as MSP, the e-retailer may establish many short- term relationships with its suppliers. Due to the focus on short-term gains, the parties prefer to enforce and adhere to contractual terms (Kim & Choi 2015). To ensure the e-retailer can control the partner’s behavior and actions, a detailed contract would be desirable (Greenwood

& van Buren 2010).

Trust: The parties to a contractual agreement often also establish some kind of social

relationship (Wiseman et al. 2012) that could lead to a feeling of trust between the parties

(Srinivasan et al. 2011; Greenwood & van Buren 2010). Trust reflects the belief that the other party will act predictable, fulfills agreed obligations (Poppo & Zhou 2014), takes a bilateral approach towards problem solving (Poppo & Zenger 2002) and behaves fairly if opportunistic behavior is possible (Rai et al. 2012).. The involved parties might develop a mutual understanding of social norms that could provide guidelines as how to respond to specific circumstances (Greenwood & van Buren 2010; Dekker et al. 2013). However, this often requires a significant amount of time and effort to develop (Poppo & Zhou 2014).

It’s expected that the level of trust between the e-retailer and the suppliers of its MSP model is very limited. Both the e-retailer and supplier initially focus on short-term gains (Kim

& Choi 2015). Therefore, the time to develop a sense of trust is very limited. Secondly, platform participants often compete for the same customers that the e-retailer aims to serve via its reseller model (Alstyne et al. 2016). This doesn’t directly encourage the development of trust. As reseller, both the e-retailer and supplier often intend to collaborate for a longer time period. This offers them the chance to develop a shared understanding and sense of mutual dependence (Srinivasan et al. 2011). Over time, the partners can grant each other a status of ‘trustworthiness’ (Poppo & Zenger 2002), which reflects the belief that both parties work towards the same organizational objective (Sjoerdsma & van Weele 2015) and don’t purposefully pursue actions that will harm the other (Rai et al. 2012).

Information sharing: The exchange of data, information and/or knowledge between two or more organizations (Kembro & Näslund 2014). The organizations will need to decide what, and how much information they want to share with one another. The first, information’s quality, reflects the extent to which the shared information is timely, useful, appropriate, complete and accurate for either party (Goldenberg & Oreg 2007). The perceived quality of the information is subjective as it highly depends on the goal the involved parties aim to achieve (Lee et al. 2011). High quality information can lead to better forecasts (Lee & Whang 2000), supports the segmentation of customers (Sahin & Robinson 2002) and can help to build trust between partners (Lumineau & Malhotra 2011). The second, contact frequency, is the frequency with which the parties interact to exchange information (Habib et al. 2015). In general, the more information the partners share, the higher the level of (perceived) control (Nicolaou et al. 2011).

As reseller, the e-retailer can directly control the flow of information to and from its

suppliers. This high level of perceived control over the flow of information stimulates the

exchange of sensitive, firm-specific information to realize efficiency gains (Brandimarte et al.

2012). Additionally, the closer relationship between the e-retailer and its supplier often leads to a higher willingness to, more frequently share information with one another (Kim & Choi 2015; Wu et al. 2014). The primary aim of the information exchange is to exploit the information available to align their operations and identify opportunities to pursue together (Kim & Choi 2015; Palmatier et al. 2013). In contrast, as MSP, the e-retailer gives up a great deal of control, as suppliers and customers can directly interact with one another. Since suppliers and customers directly interact, the supplier and intermediary are expected to interact less frequently (Kim & Choi 2015). However, this is not necessarily bad. Due to the lower level of control of the e-retailer, the supplier has more freedom to interact with other partners to explore the market for new ideas and/or product innovations (Boudreau & Hagiu 2009).

Operational alignment: The extent to which logistics and operational activities are customized according to the other partner’s requirements (Frohlich & Westbrook 2001). It’s closely related to the e-retailer’s perceived level of control over the physical flow of goods. A high level of alignment implies that parties invest to integrate and align their logistics and operational activities and/or systems to better facilitate and control the flow of goods between them (Prajogo & Olhager 2012). Examples are changing the delivery frequency or adjusting packaging standards to establish a better fit with the other party’s operations. In contrast, parties can also decide to maintain their independent operations (Kim & Choi 2015). This is associated with a low level of operational alignment as both parties don’t intend to achieve a higher level of synchronization of their logistics and/or operational activities.

Intensive information sharing practices and a long-term relationship are key to realize the integration of logistics (Prajogo & Olhager 2012). As reseller, the e-retailer and its suppliers fulfill both conditions. The alignment of their logistics and operational activities can be beneficial in terms of costs, speed, flexibility and reliability of the deliveries (Danese et al.

2013). In contrast, when it operates as MSP, the e-retailer’s level of control over the flow of

goods is limited. The supplier directly delivers the product(s) to the e-retailer’s customer

without the latter being physically involved. Additionally, the loosely-coupled ties and the

limited exchange of information don’t encourage investments aimed to achieve operational

alignment. Therefore, it is expected that the e-retailer and the MSP-supplier(s) prefer to

maintain their independent logistics and operational activities.

Resource investment: The degree to which the parties involved (i.e. supplier, e- retailer) make resource investments specifically for the relationship between them (Heide &

George 1990). To this end, scholars distinguish between dedicated and non-dedicated resources. Dedicated resources are those investments that are made for the relationship between the e-retailer and a specific supplier (Huang et al. 2014). Examples are, investments in a shared warehouse or a special IT system to share information. Under the assumption that only one of the parties involved invests in dedicated resources that will serve both of them, the investing party exposes himself to a significant risk if the other party behaves opportunistically (Williamson 1979). In contrast, non-dedicated resources can serve multiple suppliers in the e-retailer’s network (Van Donk & Van Der Vaart 2005). For instance, the e- retailer would have one IT system to which all suppliers can connect. Their main benefit is the owner’s flexibility in terms of switching options (Gimenez et al. 2012). If, for example, one supplier withdraws from the collaboration or market conditions significantly change, the investment can still be used for another party (or purpose) and won’t be lost.

Dedicated resources are the most tangible sign that a party is fully committed to a relationship and wants to make sacrifices for it (Ganesan 1994). However, a significant level of control over the other party might be needed to ensure they won’t engage in opportunistic behavior that decreases the value of the dedicated investment. Additionally, dedicated investments are often only worthwhile if the relationship is expected to last for a longer period of time as there’s a lot of time, money and effort involved (Brandimarte et al. 2012).

Therefore, dedicated resources are more likely to be made when the e-retailer operates as a

reseller. As MSP, the e-retailer has a very limited amount of control over its supplier(s) and

vice versa. This lack of control increases the likelihood that either party pursues actions that

harm the other (Ebers & Semrau 2015). Furthermore, when collaborating via the MSP model,

both the e-retailer and the supplier don’t necessarily have the intention to collaborate for a

longer period. For this reason, the MSP model is more likely to encourage non-dedicated

resource investments. Table 1 (page 13) presents an overview of the differences between the

reseller and MSP model.

Variable Sub dimension Reseller model Multi-sided platform

Reference

Contractual governance

Contract specificity

Broad guidelines Detailed contracts

Weber & Mayer (2011); Kim & Choi (2015

Relational governance

Development of relational trust

High Limited Kim & Choi (2015);

Srinivasan et al.

(2011) Information sharing Information

quality

Exploitation Exploration Kim & Choi (2015)

Contact frequency Frequent interactions

Limited interactions

Kim & Choi (2015)

Logistics responsiveness

Integration Autonomous Kim & Choi (2015

Resource investment

Dedicated Non-dedicated Brandimarte et al.

(2012) Table 2.1: Characterization of reseller model and multi-sided platform

So far, research has shown that the e-retailer’s level of control over the flow of goods and information differs when it operates as reseller or MSP (e.g. Alstyne et al. 2016; Jiang et al.

2011; Abhishek et al. 2016; Simpson et al. 2001a; Hagiu & Wright 2015) as the company differently controls its suppliers for each model (Choi & Krause 2006). However, insights on how this difference influences the sourcing decision are still scarce. Therefore, this research includes the question:

1. How does the need for perceived control influence the decision to source and sell

a product via the reseller or MSP model?

2.3. Product characteristics and their impact on model choice

Research has identified many factors that could influence the e-retailer’s decision to opt for either the reseller or MSP model to source and sell a product. Several authors explored the impact of the differences between a MSP’s dynamics and those as reseller (Alstyne et al.

2016; Wang & Koren 2012), others took a customer-oriented approach and looked at the influence of online reviews (Kwark et al. 2014), customer engagement (Chakravarty et al.

2014) and supplier behaviors (Simpson et al. 2001). However, multiple researchers agree that a product’s characteristics might also influence the decision to opt for either model (e.g.

Hagiu & Wright 2013, 2015; Abhishek et al. 2016; Jiang et al. 2011). Those authors examined the influence of a product’s price (Abhishek et al. 2016), marketing activities (Hagiu & Wright 2015), product innovativeness (Kapoor & Lee 2013; Zott & Amit 2008) and being a mid-tail product (Jiang et al. 2011). However, those authors agree that there are more product characteristics to explore. They suggested that the product’s sales volume, its complementary options and its level of complexity might also influence the e-retailer’s sourcing decision (Hagiu & Wright 2013; Jiang et al. 2011; Abhishek et al. 2016).

Product volume: The yearly average demand of a product (in units) (Hagiu & Wright 2013). In general, companies always aim for higher sales volumes to increase their profits.

Existing literature explains the two main ways in which this can be achieved; 1) focus entirely on increasing the volumes of the e-retailer’s current product range (Amit & Zott 2012) or 2) extend the current product range so the e-retailer increases the chance the customer can find the right product and purchases this resulting in higher total sales volumes (Xiao et al. 2014).

To realize the first option, the e-retailer often invests in promotion efforts (e.g. price discounts, marketing advertisements). Those promotions can significantly increase a product’s sales volume (van Heerde et al. 2002). As reseller, the e-retailer could exploit this opportunity as it has full control over the product’s price setting, marketing campaigns and advertisements. As MSP, the e-retailer can’t decide on its own how it would like to organize the marketing activities of the partner’s product(s). It would need to discuss the options with partner and see if both parties could agree to organize a joint marketing activity.

However, the MSP model might provide the e-retailer with a good opportunity to realize the second option. As MSP, the e-retailer can possibly connect with (many) more suppliers that can all provide new additions to the e-retailer’s existing product line (Sashi &

O’Leary 2002). A major advantage is that the e-retailer doesn’t need to coordinate and control

many more flows of goods (and information), since the delivery of the products to the

customer can be made a responsibility of the partner. Additionally, the relationships with these suppliers are often short-term oriented and loosely-coupled (Kim & Choi 2015).

Therefore, the e-retailer doesn’t need to commit a lot of time, effort and resources to control these relations. In contrast, as reseller, more suppliers often imply significantly higher costs, as the e-retailer needs to invest a lot of time, money and effort to control the additional flows of goods and information.

Product complementarity: Two products are complementary to one another if the availability of one product increases the customer value of the other product (Hall & Soskice 2001). Complementary products can be a valuable source of revenue to the e-retailer, as some products have a higher value to the customer when sold in combination than when offered separately (Hagiu & Wright 2013). The e-retailer can benefit from this through product bundling or by offering a wider product range to its customers.

The concept of product bundling refers to a company’s decision to bundle two or more products in one package and sells it as such to the customer (Stremersch & Tellis 2002).

Often this package is cheaper than when the customer would have bought the products separately (Gedenk et al. 2006). As reseller, the e-retailer can first purchase the products it wants to combine from its suppliers, bundle them, set the price for the package, design and plan activities to promote it and ensures same-time delivery of the package’s products. As MSP, the e-retailer lacks control over the assortment, the product’s pricing, promotion and delivery since these decisions and responsibilities are mainly determined by its partner (Abhishek et al. 2016; Jiang et al. 2011; Hagiu & Wright 2015). The e-retailer would need to convince multiple partners to bundle their products, agree on its price and ensure it would be delivered as on package to the customer. Therefore, it’s more complicated to bundle products as MSP than as reseller.

To achieve product complementarity through a wider product range, the e-retailer

assumes the customer knows best which products are complementary to one another. An

extended product range can have similar effects as product bundling as it increases the chance

that the customer can find products he/she believes are complementary, and subsequently

buys them together via the e-retailer’s website (Boudreau & Hagiu 2009). As explained

earlier (see product volume), the MSP offers the e-retailer a relatively easy and cheap channel

to offer a wider product range to its customers, whereas as reseller this would come at a

significant cost.

Product complexity: The professional expertise, brand credibility and time required to be able to compete in the marketplace (Hagiu & Wright 2013b). Customers who shop online can’t physically see, smell or touch the product. They often search (online) for information about the product before making their purchase decision (Lee et al. 2008). In general, the higher the product’s technical complexity, the more information the e-retailer would need to provide to convince the customer about the product’s quality level (Brown et al. 2012). The e-retailer’s professional expertise is the extent to which the e-retailer possesses the knowledge and skills to provide the customer with a satisfactory amount and quality of product information (Van Der Vegt & Bunderson 2005). Generally, the e-retailer has less professional expertise than its suppliers or MSP-partners as they are the experts regarding their products.

As reseller, the e-retailer itself is responsible to deliver a certain level of customer service whereas as MSP the e-retailer can re-direct the consumer to its partner. Therefore, the e-retailer needs to ask itself if it possesses, or can easily acquire, the specialized knowledge that is required to properly serve its customers (Baihaqi & Sohal 2013). If not, it might be wise to estimate the costs of acquiring this knowledge versus the costs of using an external seller on its MSP to serve the customer before making the decision in favor of either model.

Nonetheless, seeking information is costly and a customer can only process a limited amount of information. Customers, typically, associate brands with a certain quality level (Keller & Lehman 2006). They tend to use the brand image of the product’s producer as a substitute for missing product information (Benedicktus et al. 2010). As MSP, the customer and the direct supplier of the product interact, whereas as reseller the e-retailer is an additional party in the chain. Logically, both the suppliers and the e-retailer have a certain brand image in the eyes of the customer. When the supplier’s brand image is better than the e-retailer’s brand image, the MSP might be more effective (i.e. higher chances to complete the sales), while if the e-retailer’s brand image is better than that of the supplying partner, the reseller model might be a more profitable option.

In conclusion, existing literature suggests that the product’s volume, its complementary options and its level of complexity influence the e-retailer’s decision to source a product via its reseller or MSP model (Hagiu & Wright 2013a; Jiang et al. 2011; Abhishek et al. 2016).

However, it is still unclear how these product factors exactly influence the decision to opt for

the reseller or MSP model when making sourcing decisions.

For this reason, this research explores the following research questions:

2a. How does a product’s sales volume influence the decision to source and sell a product via the reseller or MSP model?

2b. How do a product’s complementary options influence the decision to source and sell a product via the reseller or MSP model?

2c. How does a product’s complexity influence the decision the decision to source and sell a product via the reseller or MSP model?

The proposed relationships between the different elements (i.e. perceived control, product factors, model choice) are shown in the conceptual model in figure 3.

Figure 2.2. Conceptual model

3. Methodology

To investigate how the need for perceived control and product characteristics influence an e- retailer’s choice to opt for either the reseller or MSP model to source a product, this study adopts a multiple-case study approach. Since literature on this issue is scarce, this research question focuses on exploration and theory building (Saccani et al. 2014). Case research allows for the gathering of rich data that helps to better understand phenomena that involve a high number of variables (Yin 1994), as is the case with the decision to choose for the reseller or MSP model. Additionally, this type of research provides a valuable method to study emerging concepts (Voss et al. 2002), thereby well suiting the new trend among (e-)retailers to adopt both a reseller and MSP model to source and sell their products .

Products aren’t sourced per unit, but often per category (i.e. buy 100 laptops from a supplier at once). Therefore, the unit of analysis will be the purchase and sales of a product category. A product category is defined as a group of products that scores similar on a set of predetermined dimensions (Kushwaha & Shankar 2013).

3.1. Research context

The research context is the e-retail industry. IT has led to a situation in which suppliers, customers and the intermediary don’t necessarily have to be at the same place nor do they need to communicate at exactly the same time (Gunasekaran et al. 2002). It offers a unique opportunity to reach many suppliers and customers via an online channel. This is one of the important reasons that e-retailers became the first adopters of the hybrid business model that combines a MSP and reseller model to source and sell products to customers. Therefore, it’s expected that e-retailers can offer the most complete insights on this issue. For this reason, this study is conducted at a large e-retailer who has adopted both the reseller and MSP model to source and sell its products. The company collaborates with many partners to offer its customers a wide variety of different products.

3.2. Case selection

The selected cases represent different types of product categories that are sold by the e-retailer under scrutiny. The selection criteria were the volume of the product category, the extent to which complementary product options are offered, and the level of complexity of the product.

Numbers about the product’s volume were retrieved from an internal database and the level of

complexity was based on the product specifications that were provided on the e-retailer’s

website. Additionally, two company experts provided their advice on the level of product

complexity for the available product categories based on 1) after-sales service required, 2)

number of product components and 3) the e-retailer’s familiarity with the product. Based on literature, the factor of product complementarity was initially included as well. However, the company experts indicated that this factor doesn’t play a role to choose between a MSP or reseller model at this e-retailer due to the limited capacity of its IT system. Therefore, this factor has been removed. Table 2 presents an overview of the included cases.

Case Volume

Product complexity - Brand image

Product complexity - Professional knowledge

Product category

A Low Low Low Large furniture

B Low High High Hoover board

C Low Medium High Laptop

D High High Low Toys core

E Low Medium Low Suitcase

Table 3.1: Overview included cases 3.3. Data collection

Data was collected during the months January and February in 2017. Data was collected by means of face-to-face interviews, the product’s sales information that was provided on the e- retailer’s website and a draft of the e-retailer’s mathematical model that was developed to support buyers and market managers to make the decision whether a product should be source and sold via the company’s reseller or MSP model.

The interviews were conducted in two rounds. The first round of interviews was mainly aimed at exploring the factors that influenced the choice for the reseller or MSP model, and how the involved parties were managed and controlled by the e-retailer. Next, a second round of interviews was conducted to gain more in-depth insights in how and why certain factors influenced the decision for either model. All interviews included semi- structured questions, and the first-round interviews also included a set of closed questions.

The interviews were mainly focused on the semi-structured questions, because this allowed

1

Cut-of value has been removed for confidential reasons

2

High: brand image is important determinant when customers purchase a product; medium: some

customers value a specific brand, while others find other factors more important; low: brand image is

of very limited influence when customers purchase a product

the interviewees to express their explanations and opinions about the subject in their own terms (Blumberg et al. 2008). This increases the richness and depth of the data (DiCicco- Bloom & Crabtree 2006). In total 13 interviews were conducted of which nine interviews were part of the first round, and four were conducted in the second round. The interviewees were selected based on their position in the company as this was an important determinant of what they knew about the (differences between) reseller and MSP model and their knowledge about what factors influenced the model choice for a product category (see table 3.2 for an overview). In addition, one interview was conducted with one of the company’s sourcing specialists who draws the guidelines for the sourcing decisions of any market manager or buyer. The interviews were conducted in Dutch and recorded with a recording device. These recordings, along with the notes the interviewer made during the interview, supported the correct transcription of the interviews, hence increasing the accurateness of the data obtained (Yin 2009).

The interview protocol was developed by the researcher under guidance of a supervisor. The operationalization of each variable, as described in the theoretical background, is included in appendix A, and the entire interview protocols in appendix B1 and B2. The interview protocol evolved during the data collection to more in-depth address specific topics, which were considered relevant to answer the study’s research question.

Table 3.2: Interview details

Case Interviewee Length of interview

A A1: Market manager/buyer (double function) 32 minutes B B1: Market manager

B2: Buyer

56 minutes 36 minutes C C1: Market manager

C2: Buyer

C3: Category manager

58 minutes 47 minutes 33 minutes D D1: Market manager

D2: Buyer

56 minutes 36 minutes E E1: Market manager

E2: Buyer

E3: Category manager 1 E4: Category manager 2

75 minutes

43 minutes

45 minutes

35 minutes

3.4. Data analysis

After the interviews were transcribed, the data was coded. The software ATLAS.ti was used for the coding process. Before the start of the coding process, a set of initial codes had already been developed, but this set was extended during the coding process as the first analysis of the data showed several new concepts that were also relevant to answer the main research question. This resulted in 39 codes that covered three topics: (a) how the e-retailer manages and controls its suppliers (15), (b) factors the e-retailer wants to control (12) and (c) contingencies that influence the e-retailer’s decision to opt for either the reseller or MSP model to source and sell its products (22). Those 39 codes subsequently formed 17 aggregate dimensions. Those dimensions enabled the researcher to make inferences about the data and draw conclusions to provide an answer to the main research question. The coding trees can be found in appendix C1 and C2.

4. Findings

4.1. Reseller or MSP: what was the decision?

The findings show that Case A is only sourced and sold via the e-retailer’s MSP. Most products that belong to case C (70 percent) and D (75 percent) are sourced via the e-retailer’s reseller model, while products of case B (90 percent) and E (60 percent) are mainly sourced via the MSP. Table 4 presents the most important reasons for each product category to opt for either model.

Case Reseller Multi-sided platform

A Not applicable Logistics constraints (e.g. warehouse space, delivery requirements)

B Profit margin per unit sold High sales volume

Product complexity Demand uncertainty C Supplier history

Market maturity Brand image

Add unique assortment

D Supplier history Brand image

E-retailer’s market position

Suppliers’ attitude towards online market platform

Add unique assortment

Additional product availability Seasonality

E Profit margin per product Brand image

Inventory costs

Suppliers’ attitude towards online market

Table 4.1: Reasons to opt for reseller or MSP model for each product category

4.2. Supplier management

How the e-retailer uses the different governance mechanisms (i.e. contracts, trust, information sharing, logistics responsiveness, resource investments) is described in the following sections

.

Contract and managing the formal contract: For all cases, regardless of model choice, contracts are the main tool to control the partner’s actions and behaviors and the flow of goods and information. In all cases “the contract is the most important tool, and forms the basis for our collaboration” (category manager, case E). However, the cases show that the e- retailer’s willingness to negotiate and adjust contractual conditions differs for its reseller and MSP model. For case A, B, D and E, the e-retailer “controls everything through its standard conditions that are specified in the contract (market manager, case E) for its MSP-suppliers.

Case E, however, shows that the e-retailer might be willing to compromise on these conditions when the seller is considered essential (brand name, price level, unique product) to the e-retailer’s business; […]. “Except for suppliers that are very important to us, then we’re more flexible and might be willing to adjust them” (market manager, case E). In contrast, the buyers of case C, D and E state that the e-retailer always negotiates with each reseller- supplier. “We negotiate with each single supplier about the contractual conditions” (buyer, case C). The goal is to identify and maximally exploit the opportunities together. “We really try to leverage the supplier’s potential” (buyer, case D).

How the e-retailer managed the formal contract varied for each model. In all cases, the performance of supplier is evaluated based on a set of predetermined KPIs (e.g. prices, delivery speed, delivery reliability, service level), but they’re managed differently. In all cases, parties that sell products via the e-retailer’s MSP, are mainly controlled via standardized and automated systems. The market manager of case C explains: “The performance department (VPM) is responsible for monitoring and controlling the partner’s performance. [...]. After a few warnings, their accounts will be closed automatically”. For cases B, C, D and E, the suppliers that deliver goods (and information) via the reseller model, are managed on a more personal basis. The buyer of case D states: “You always first discuss the issue with the partner and, together, we try to find a solution”. However, case B, C, D and E also show that there are limits to this problem-solving approach: “Worst case scenario would be that we stop ordering products from them” (buyer, case C).

3

The market manager oversees products sold via the MSP; the buyer oversees products sold via the

reseller model; the category manager guides the buyer and market manager of one product category

Trust and managing the relationship: Generally, the e-retailer and its suppliers don’t seem to only rely on the contract to manage their relationship. First, for products that belong to case B and C and are sourced via the reseller model, the e-retailer believes that also the supplier has a certain responsibility to ensure all products in stock are eventually sold.

“Both of us are responsible for selling the product. In case a product doesn’t sell well, we expect the supplier to agree to a swap (i.e. e-retailer returns the products that don’t sell and receives products that sell at a high rate in return). This seems to suggest that in these cases, the e-retailer sufficiently trusts its reseller suppliers to rely on agreements that go beyond what’s explicitly stated in the contract. In contrast, case A and E show that the e-retailer barely trusts its MSP-parties nor believes there’s some shared responsibility to sell a product.

“Once the supplier has signed the contract, its his/her responsibility to live up to the contractual terms. […]. Everything is formally covered in the contract, in case something goes wrong” (market manager, case E).

Information sharing: The findings show that information practices differ for each model. In case B, C, D and E the e-retailer shares relevant information daily with suppliers that deliver products via its reseller model. “Although, we purchase the product from our suppliers, we still discuss many issues with the partner” (buyer, case C). Once or twice a month they schedule a face-to-face meeting to evaluate and discuss the performance (e.g.

generated sales and revenues, inventory, joint marketing activities, supplier’s assortment range at the e-retailer) of the past period. Additionally, for these cases, the-retailer might be willing to share sales forecasts and weekly sales- and inventory reports upon the supplier’s request. The buyer of case E confirms this idea and adds that information sharing for reseller suppliers also helps to build trust and ensure long-term commitment. “We build a close relationship with those suppliers, we negotiate and constantly monitor the (additional) conditions, agreements and future opportunities. This is meant to be a long-term investment”.

On the contrary, the market managers of case A, B, C, D and E interact, on average, twice a month via telephone or e-mail and plan every 2-3 months a face-to-face meeting.

During those meetings, the e-retailer shares sales- and revenues numbers. However, most

information is communicated to the partner via his/her personal webpage on the e-retailer’s

website.

Operational alignment: In all cases, regardless of model choice, the interviewees indicated that the level of operational alignment was very limited. For example, the category manager of case E states “We have one way of working and our partners must adjust if needed”. However, some suppliers of case C, D and E had to adjust their data systems. Case D and E showed that this was compulsory for their reseller-suppliers; “[…]. They must connect to all our systems to quickly share relevant data” (buyer, case E). According to case A and E this was optional for their MSP suppliers; “They have the possibility to connect their database(s) to ours, but due to the significant costs some just decide to manage everything via their account on the website” (market manager, case E).

Resource investment: Except for suppliers’ investments, as described in the previous section on operational alignment, in none of the cases, the e-retailer invested in resources to support the collaboration with a specific partner. In all cases, it was stated that “in general, we don’t make any special investments” (market manager, case C). However, in cases D and E, the e-retailer sometimes incorporates advice of an important partner when it builds new tools to better anticipate new market trends; “The main reason was the market conditions, but the pressure of the partner was certainly an additional reason to invest in this new tool”

(market manager, case E).

4.3. What factors influence the choice for the reseller or MSP model: contingency factors

This research identified nine factors that seem most relevant when the e-retailer’s needs to decide whether to source and sell a product category via its reseller or MSP model, namely:

assortment range, delivery process, perceived inventory risk, price, company profit, marketing activities, the product’s perceived level of complexity, the quality of potential partners and the e-retailer’s logistics capacity.

Assortment range: Perceiving control regarding the assortment range includes two aspects, namely 1) the number and kind of products the e-retailer intends to offer to its customers and 2) the guarantee that the assortment is available when the customer wants to purchase a product. The first is labelled ‘assortment choice’, the second as ‘assortment availability’.

Assortment choice: For case C, D and E the most important factor to source a product

as reseller is the product’s brand name. The buyer of case D explains “Most important is that

we have the most important brands ourselves. [...]. We want to have those products, that bring customers to our web shop, on our own shelves”. Case C and E show that the market situation is an important factor to determine the importance of the brand. For case C, there are only a limited number of market brands; “There are only 6-7 brands in the market, so we must have those in stock” (buyer, case C). In contrast, the market of case E’s products consists of many market players, so “we need to make choices since we can’t do all those brands ourselves” (category manager, case E). Other products add value to the e-retailer’s assortment because of their uniqueness (i.e. long-tail products). Case D and E show that this level of product uniqueness influences the e-retailer’s model choice. As stated by the category manager of case E: “We only source the most important brands ourselves. […]. Our MSP model offers a great opportunity to extend the total assortment choice that is offered to the customer”.

Assortment availability: Case C, D and E suggest a relation between assortment choice and assortment availability. For these cases, the e-retailer prefers to source the most important brands as reseller, because this increases the perceived control over their ability to deliver on time. “For products we deem most important, we do it ourselves because you don’t want to depend on others whether it’s available or not” (category manager, case E). This seems to be related to the customer’s expectations about delivery speed and – reliability which will be explained next.

Delivery process: The delivery process includes multiple aspects. The e-retailer’s proposition is as follows: “We promise the customer that we deliver within a pre-specified timeframe”. This includes two aspects, namely delivery speed and delivery reliability. The speed refers to the promised delivery window, whereas the reliability refers to the e-retailer’s ability to deliver within this period. It follows that assortment choice and assortment availability are closely related to delivery speed and reliability. The buyer of case D states

“We’re one of the most important market players. This implies that we must ensure that the

customer can always order the most important products at our shop and that we can deliver

them within the pre-specified timeframe. Therefore, we want to manage the inventory of those

products ourselves”. This statement shows that, in case D, the e-retailer’s own market

position influences its willingness to comprise on its delivery proposition. It seems that this’s

one the reasons why most products that belong to case D are sourced via the reseller model.

Inventory risk: This guaranteed delivery time window comes, however, at significant costs due to the need to hold inventory. Case B, C, D and E suggest that a product’s sales volume affects the model choice, because this factor influences how long the product will need to be held in stock. The market managers of these cases agree that “when products don’t sell at such high rates, they’re not worth the time, money and risks associated with holding inventory”. […]. We prefer to source them via our MSP since this implies we don’t need to hold the inventory ourselves” (market manager, case B).

Another factor that might increase the e-retailers perceived inventory risk is seasonality. Some products that belong to case D and E are subject to seasonality, which results in “high product volumes within a limited time frame” (market manager, case D). In case D and E, those products are preferably sourced via the MSP model, because “it’s very hard to accurately predict at what time the customer demands a lot of which product”

(market manager, case D). As MSP, the e-retailer doesn’t have to hold those products in stock, so it doesn’t carry the inventory risk. “We don’t have to worry about getting rid of products that become outdated when people stop buying them” (market manager, case E).

Case B and C suggest that also the product’s life cycle influences the e-retailer’s perceived inventory risk, and therefore the model choice. Case B is early in the product life cycle. The e-retailers prefers to source this product category’s products via the MSP. “For such products, that are new to our market, we prefer to wait so others explore market conditions first” (market manager, case B). Case C is at the maturity phase of the product life cycle and the e-retailer has already gathered a lot of data about customer demand. “Over the years, we’ve gathered a lot of data about customer preferences and each product’s sales volume. […]. The inventory risk is limited, because we know how much we need to purchase of each model as we know what models sell at high rates, and which don’t” (buyer, case C).

Price: Case B, C, D and E suggest that a product’s price is often a major factor when

luring customers to the e-retailer’s web shop. In all cases, the buyers agreed that as reseller,

the e-retailer can control to set and adjust product prices. Most products that belong to case

C and D are sourced via the reseller model, because the e-retailer wants to control their sales

price. “When we operate as reseller, we can set the price ourselves and adjust it when

needed” (buyer, case D). According to case C and D this enables the e-retailer to be more

competitive in the marketplace. “These are our most important brands. […]. Enables us to

quickly anticipate price changes of competitors.” (buyer, case C). So, those cases also seem

to suggest that control over prices is more relevant for more important brands. For all cases,

the e-retailer has limited control over sales prices when the product’s sourced and sold as MSP. “We can’t control the product’s sales price. […]. The e-retailer can only provide advice, but the partner decides” (market manager, case A).

Company profit: This different level of control over the product’s prices influences the company’s profit. In all cases, the interviewees stated that the realized profit margin is a major factor to opt for the reseller or MSP model. “As reseller, the e-retailer earns, for each unit sold, the profit margin as in sales price minus purchase price, while as MSP it earns a fixed commission fee for each unit sold. […]. We generally earn more when we source and sell a product as reseller than we do it via our MSP” (category manager, case C).

Additionally, the data about sales volumes of the different products within the product category of case B and C show that the conversion rates are generally higher when the e- retailer sells it via the reseller model than via its MSP. Generally higher conversion rates translate in higher profits.

Marketing activities: In this context, marketing activities comprise both promotional activities as well as the web shop’s layout. Case B and E show that some products within their category are sourced and sold via the reseller model so they can properly organize their web shop’s layout. “One of reasons why we source and sell products ourselves is that we want to offer a relevant and nicely organized web shop to our customer (buyer, case B).

Regarding perceived control over promotional activities, case B and C show that products belonging to their category are sourced via the reseller model, because “it’s easier to arrange everything. […]. You’re more in the lead when it comes to the design and roll out of the activities when you source as reseller” (buyer, case B). Case A and E state that products that for the products that belong to their category and that are sold via the MSP, the partner often doesn’t want to participate in marketing activities, because “they often see us as one of their competitors. […]. They don’t want us to benefit too much” (category manager, case E).

Product complexity: Case B, D and E show that the higher the level of specific knowledge and after-sales service required the higher the chance the e-retailer prefers its MSP over its reseller model to source and sell the product. “If more specific knowledge and/or after-sales service is required, we prefer to collaborate with a specialized MSP- partner. […]. They possess the knowledge about all (technical) details of the product”

(market manager, case B). However, case C deviates from this perspective. Case C sells

products that can generally be classified as products with a high level of product complexity.

Nevertheless, most products are sourced via the reseller model. “Our specialists can answer all technical questions, […]. We might be different since we’ve the ability to negotiate good return- and service level agreements” (buyer, case C).

Quality of potential reseller or MSP partners: Occasionally, the market situation implies the choice for either model isn’t even present. Case E states that there are suppliers that aren’t willing to serve the online market. “Some major brands are pretty conservative and they try to avoid the online market altogether” (category manager, case E). In this case, the MSP might be the only option for the e-retailer. In cases B and E, some suppliers can’t serve the B2C market due to the different requirements compared to the B2B market. “The suppliers are used to omni-packs, they aren’t able to send just one piece to a single customer” (buyer, case B). In this case, the e-retailer’s only option the reseller model.

Additionally, the buyers of case C and D state that the e-retailer’s history with its supplier(s) might influence the decision to opt for the reseller model. “This category is older than the MSP model. […]. It’s difficult to convince them to switch to our MSP model when you have a long relationship with a supplier” (buyer, case D).

Logistics capacity: Sometimes model choice isn’t a question simply because the e-

retailer can’t hold the product in stock even if it wanted to. Case A and B source and sell

some products within their category via the MSP-model due to the e-retailer’s logistics

capacity, because “Our warehouse practices aren’t suitable for handling all these separate

items” (market manager, case B) or “the products are too big, we can’t store such large

products in our warehouse” (market manager, case A).