The way digital platforms capture growth in a collaborative economy : an exploratory study on network effects on peer to peer platforms

(1)

(2)

2

Statement of Originality

This document is written by Witold Rosendaal who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Acknowledgements

“We must find time to stop and thank the people who make a difference in our lives.“ – John F. Kennedy This thesis is written as the final assignment for the Master Business Administration at the University of Amsterdam. The Thesis granted the opportunity to use the knowledge learned during this study and gain insight into the collaborative economy and platforms. During this research I improved my research capabilities such as writing skills, reviewing literature, conducting qualitative research and linking my findings to the literature. Furthermore, I have visited inspiring companies and expanded my network with interesting people such as platform specialists, managers and entrepreneurs.

I want to thank prof. Dr. P. van Baalen from the University of Amsterdam for his guidance. His knowledge on innovation, platforms and research knowledge is refreshing. His feedback was valuable and helped to create this Master These as result. I want to thank my father, Dr. B.W. Rosendaal for reviewing my English and providing smart research suggestions. Also, I would like to thank the anonymous second supervisor for reading and reviewing my Thesis. Furthermore, I am grateful to all P2P platforms, interviewees, respondents and others for participating in this research. Finally, I wish you as a reader pleasure with the experience this thesis provides.

(4)

4

Abstract

Current technologies enable mankind to embed platforms to decrease the idle time of cars, spare rooms, boats, other goods and make better use of people’s skills. Platforms can match demand and supply in a way which creates value that cannot be created by centralised networks. The literature on the collaborative economy is relatively scarce so far, while it is expected that the collaborative economy will grow to a $ 110 billion industry in the next years. In this exploratory research 120 documents are analysed and 13 platform specialists, -managers and -owners are interviewed. This led to a comparative case study analysis containing 13 platforms. This study provides answer to the research question: ‘’Which factors influence network effects for peer to peer platforms and what elements are enabling those factors?’’. Offering the P2P platform for free stimulates network effects and helps to build the initial user base. This research breaks down the superficial term ‘quality’ and suggests that the underlying quality dimensions’ trust, policies, usability, reliability, aesthetics, access and differentiation impact network effects. This study identified five different marketing channels that influence network effects. This research suggests that WOM, PR and events impact network effects the most. Additionally, advertising and referral marketing can stimulate network effects too, but are less effective and efficient. Partnerships, especially with mature and well-known firms, influence network effects. Furthermore, this research carefully suggests that mature P2P platforms can stimulate indirect network effects by opening up their software to third parties and facilitate the development of services and products complementary to the P2P platform.

Keywords: network effects, peer to peer, platforms, growth hacking, strategy, collaborative economy

(7)

7

1 Introduction

Many people own cars, houses or other goods that are idle for most of the time. Let’s take cars as an example. The average car costs $ 8.000 per year (Botsman, 2010). In North America and Western Europe, the average time of a car use is 8% of its time (Belk, 2014). Why not rent out the car the other 92% of the time and earn money on it? Or why not rent a car from your neighbour instead of spending $ 8.000 annually on owning one? The collaborative economy is a growing phenomenon, made possible by the rise of internet (Belk, 2014). Van de Glind (2014) states internet makes access an attractive alternative for ownership. Platforms can match demand and supply of existing capacities in an efficient and effective way which creates value that cannot be done by centralised networks (ShareNL, 2015). According to Cannon & Summers (2014), the sharing industry was valued at $ 26 billion in 2013 and is expected to grow to a $ 110 billion revenue market over the next years. Nielsen (2014) states 68% of the global consumers are willing to share their assets. Platforms enable people to offer and demand products or services from their peers. Examples of such P2P platforms are Airbnb, Peerby, and Bla Bla Car.

The collaborative economy concerns six trading forms (buying, renting, loaning, giving, sharing and swapping) that can be found in seven different markets (goods, space, mobility, money, energy, knowledge and services) (ShareNL, 2015). Schor (2014) states platforms can be business to consumer (B2C) and peer to peer (P2P). Chakravarty, Kumar & Grewal (2014) add business to business (B2B). In this study the trading forms buying, giving and swapping, the markets energy, knowledge and money and the platform shapes B2B and B2C are excluded, since a research in depth is preferred above breadth.

The value of a P2P platform increases with the growth of its customer base, which is called network effects (Katz & Shapiro, 2014; Salazar, 2015) and is a key to success (Gallagher & West, 2009). Following Metcalfe’s law, the value of a network equals the square of the number of users that are connected to it (n2_{) (Hendler & Golbeck, 2008). To understand how P2P platforms become successful, it is important to}

(8)

8

find the factors that stimulate network effects. The aim of this study is to identify the different ways P2P platforms create network effects and investigate what can be done to enable those factors. Therefore, the research question of this study is:

‘’Which factors influence network effects for peer to peer platforms and what elements are enabling those factors?’’

To frame this research, literature on multi-sided platforms, network effects and the collaborative economy is used. This study contains a comparative case study analysis of 13 cases. During this study 120 documents are researched and 13 interviews are held with platform specialists, -managers and -owners, in which factors relating to network effects for P2P platforms are investigated.

In this study the researcher seeks to gain new insights in the business side of P2P platforms and provides material for managers to rethink the way they want to grow their platform. The subject of this study is theoretically based on a combination of insights from authors within the field of the collaborative economy and digital platforms. Cohen & Kietzmann (2014) state the collaborative economy might be the next phase in the evolution of fundamentally reorganizing how economies work and they address the urgency of research to how collaborative economy business work. Evans (2008) argues that research is needed on the success and failure of the launching of platforms. Baden Fuller & Mangematin (2013) suggest multi-sided business models have received too little attention. Lamberton & Rose (2012) address an open question in their future research section on whether P2P platforms need an alternative business model that is sustainable for the long term. Since growth factors for P2P platforms received little attention in the academic field, it is an opportunity to make a contribution to this field of study.

Contributing to this academic field is also relevant for society and management of P2P platforms. Lamberton & Rose (2012) state that sharing elements like goods, services, or accommodations leads to access at lower cost, which is good for society. The generated efficiencies by these platforms bear new products, services and business growth (Owyang, Tran, & Silva, 2013). On the other hand, digital P2P

(9)

9

platforms make it possible for individuals to generate additional income by lending their house, cars, other goods and services. Extra income generated digital platforms stimulates wealth and the economy. Another benefit for society is the positive impact P2P platforms have on sustainability by stimulating consumers sharing goods instead of owning them (Cohen & Kietzmann, 2014). A research from ShareNL (2015) shows that one district owns on average 212 power drills, while seen the amount of drilling hours, 38 drills would be sufficient, meaning that 174 power drills do not have to be manufactured. Furthermore, the collaborative economy connects people in their neighbourhood during the pick-up of a good or delivery of a service, which is valuable for society too (ShareNL, 2015). While these contact moments are short, the collaborative economy contributes to the sense of security and solidarity (ShareNL, 2015).

The next section contains the literature review, including a conceptual framework and propositions to direct and scope the research. Then the methodology and results are described. Finally, the discussion and conclusion of the research, including the limitations of this study and suggestions for future research are presented.

(10)

10

2 Literature review

This section discusses the literature relevant to this research. First, literature on digital platforms, P2P concepts and collaborative economy is discussed. Second, the different strategies that platforms can apply are explored. Third, literature on factors that influence network effects, including propositions to guide and scope the research are given. Next, the conceptual model is presented.

2.1 Platforms in the collaborative economy

Platforms gain popularity and the concept of connecting two or more parties is frequently applied. Firms are moving towards a market where they enable others to deal and connect with clients directly via their platform, instead of serving clients themselves (Hagiu & Wright, 2015). Hagiu (2006) states that platforms have two major elements: decreasing search costs and reducing shared transaction costs among numerous customer types. The platform matches different groups of people to make a transaction, which reduces costs for searching. Chakravarty, Kumar, & Grewal (2014) define platforms as ‘’a market maker that brings different sides of a marketplace together for exchange’’. Hagiu & Wright (2015) state ‘’platforms enable two or more sides to interact directly and each side is affiliated with the platform’’, which is presented in Figure I. These definitions make clear that a platform facilitates interactions between two or more sides to exchange something.

Figure I: Platforms (Hagiu & Wright, 2015)

While the definitions in the previous sections, clarify how platforms enable a market place, they don’t tell anything about its emergence and functionalities. Therefore, some of the more technical

Direct interaction

Side A Side B

(11)

11

definitions are discussed too. According to Tiwana, Kosnynski & Bush (2010) a digital platform is ‘’ the extensible codebase of a software based system that provides core functionality shared by the modules that interoperate with it and the interfaces through which they interoperate’’. This definition clarifies the need for software code to enable the functionalities of the platform. Muffatto & Roveda (2002) describe platforms as ‘’an intentionally developed and planned set of sub systems and interfaces from which products can be developed’’. While this definition emphasises the sub systems and interfaces to stimulate product development, it does not clarify what these developed products could be. Gawer (2014) defines platforms as ‘’a building block, providing an essential function to a technological system – which acts as a foundation upon which other firms, loosely organized in an innovation ecosystem, can develop complementary products, technologies or services’’. This description clarifies that platforms can stimulate innovation by organizing their software in such a way that other firms can develop complementary assets. The modularity and interfaces clearly are part of this software organisation.

Technology and internet enable platforms to rise in the collaborative economy and make it a multi-billion-dollar industry. Stokes et al. (2014) state ‘’ the collaborative economy involves using internet technologies to connect distributed groups of people make better use of goods, skills and other useful things’’. Collaborative economy is stimulated by three market forces which are societal drivers (sustainability, desire for community, altruism and increasing population density), economic drivers (idle inventory, increase financial flexibility, access over ownership and influx or venture capital funding) and technology drivers (social networking, mobile devices and desktop platforms and payment systems) (Owyang, Tran & Silva, 2013). Van de Glind (2014) claims that technology drives collaborative economy and makes access an acceptable or preferable alternative for ownership. ShareNL (2015) defines the collaborative economy as ‘’economic systems of decentralised networks and market places that unlock the value of unused goods and services by matching demand and supply directly, making traditional institutional intermediary unnecessary’’. Meelen & Franken (2014) describe the collaborative economy as

(12)

12

‘’the phenomenon of consumers making use of each other’s unused goods’’. Whereas Meelen & Franken (2014) explicitly state it is about goods and not services since people cannot be ‘unused’, ShareNL (2015) recognises the confusion on collaborative economy including services or not, but argues that services is an element of it. However, a teacher working 26 hours per week has 14 hours left in which it can do work. This suggests people can be ‘unused’. ShareNL (2015) discusses two forms of collaborative economy. The sharing economy is about an economic system, in which consumers use each other’s unused goods, eventual in exchange for payment. Second, the on-demand economy, is an economic system in which jobs, tasks and services are connected to an executor, offering employment and services on demand, often in exchange for payment. Since Meelen & Franken talk about the on-demand economy too, it seems that they agree on the separate definitions (i.e. sharing economy and on-demand economy), but not on the on-demand economy being included in the collaborative economy. Belk (2014) defines the collaborative economy as ‘’people coordinating the acquisitions and distribution of a resource for a fee or other compensation’’. Instead of goods or services Belk talks about resources, which could be both. Burnett (2014) doesn’t distinguish the ‘sharing economy’ and ‘collaborative consumption’ and refers to them as ‘’any means of sharing goods and services among willing participants, often outside of traditional channels, and taking the form of leasing, loaning, borrowing, donating, bartering, buying and selling goods directly, as well as the facilitation of these activities by third parties, and often involving used or communally owned goods and services’’. Like the definition of ShareNL (2015), Burnett talks specifically about traditional channels being unnecessary. Except Meelen & Frenken (2014) all other authors state services is part of the collaborative economy. In this research the definition of ShareNL (2015) is adopted and the on-demand economy is viewed as an element of the collaborative economy.

The collaborative economy concerns six trading forms (sharing, lending, renting, buying, giving, and swapping) (figure II) that operate in seven different markets (goods, space, mobility, money, energy, knowledge and services) (figure III) (ShareNL, 2015). Frenken (2015) argues that buying, giving, and

(13)

13

swapping is not part of the collaborative economy, since consumers transfer the ownership of the good instead of giving temporal access. Therefore, these are excluded in this study. Examples of the first three trading forms are UberPool (sharing), Airbnb (renting) and Peerby (lending). Sharing assets makes use of the underutilization of assets. As stated in the introduction, a car is used on average for 8% of its time (Belk, 2014). A P2P platform, like Snappcar, makes use of the idle time and increases efficiency, by enabling peers to rent or offer a car via their platform. Sharing could be the next evolution in the way we view ownership (Lamberton & Rose, 2012). The fourth trading form, service, is about peers offering their fellow peers a service. Examples of such P2P platforms are Task Rabbit, Lyft and Butlerbuddy (Bellotti, et al., 2014). These platforms make it easy to generate an (extra) income. One can clean a house, be a taxi driver or fix furniture as a handyman in exchange for money. Maselli, Lenaerts & Beblavy (2016) argue that such platforms are used to generate complementary income or even replace jobs. They back this up by showing the growth of contingent workers in the EU from 27% in 2002 to 32% in 2014. The markets money, energy and knowledge are also excluded in this study, since research in depth is preferred above breadth.

Figure II: trading forms in the collaborative economy Figure III: markets in the collaborative economy

Six trading forms Sharing Lending Renting Buying Giving Swapping Seven markets Goods Space Mobility Services Money Knowledge Energy

(14)

14

Furthermore, platforms can take three different shapes (table I). Schor (2014) states platforms can be business to consumer (B2C) and peer to peer (P2P). Chakravarty, Kumar & Grewal (2014) add business to business (B2B). In this study B2B and B2C are excluded since, as explained in the introduction, a considerable research gap is identified in the field of P2P platforms (Plouffe, 2008). Benkler & Nissenbaum (2006) state that peer production is a socio-economic system of production which is evolving in a networked climate. With P2P platforms, peers exchange services and products via the internet (Plouffe, 2008). The characteristics of this socio-technical system are collaboration among large groups of consumers, who collaborate to exchange information or goods without counting on either market pricing or managerial structures to coordinate their collaboration, as Benkler & Nissenbaum (2006) argue. However, this last claim is questionable. In many peer systems consumers have to pay for the information or goods they use. While often these prices are set by their peers, it is a market system on its own. Schoder & Fischbach (2003) define P2P systems as those that enable ‘’two or more peers to collaborate spontaneously in a network of equals (peers) by using appropriate information and communication systems without the necessity for central coordination’’. This is in line with the suggestion of Benklar & Nissenbaum of exchanging information or goods without central coordination. Hagiu & Wright (2015) too emphasize the absence of central coordination. The platform facilitates the exchange, without central coordination.

Shapes platforms can take

P2P B2C B2B

Barqo, Peerby, Airbnb, UberPool Apple app store, UberBlack,

Werkspot FLOOW2, Alibaba, Ariba

Table I: shapes platforms can take (Chakravarty, Kumar & Grewal, 2014; Schor, 2014)

While nobody is questioning the positive impact of the collaborative economy on creating a sustainable economy, criticism is rising. Bardhi & Eckhardt (2015) criticise the collaborative economy by stating that it is not about sharing and social value, but it is about an economic change that creates value by giving access to goods or services at lower costs. The authors argue that consumers are interested in

(15)

15

the economic benefits and it is an access economy rather than sharing. Matofska (2014) adds that renting out a spare room via Airbnb is not sharing, but ‘filling pockets’. Killick (2015) affirms that the real winners are the platform owners, pirate capitalists that use technology to disrupt traditional industries. Tegenlicht (2014) and The Economist (2015) show adverse effects like privacy breach concerning personal data, new forms of exploitation, platform monopoly and unfair competition. The attractiveness of a platform depends on its user base, which can lead to monopolies with high profit margins (Frenken, 2012). Nevertheless, investment costs are moderate, which decreases entry barriers. The user data are commercially valuable, but may contain sensitive information too (Frenken et al., 2015). This is not only the case for platform companies, but also for other internet companies. The collaborative economy stimulates uncertainty and flexibility with their services, which leads in some cases to exploitation (Frenken et al., 2015). People who offer their service are dependent on the pricing, algorithms and reviews facilitated by the platform. The biggest criticism of the collaborative economy experiences is that it facilitates unfair competition (Frenken et al., 2015). Consumers that offer their service or goods via the platform are not subjected to all regulations the traditional market suppliers like hotels, hospitality and taxis have to meet. Where hobby chefs invite consumers into their home and can offer them a three course menu with wine, restaurants complain about unfair competition since these home restaurants don’t cope with regulations on alcohol, hygiene and safety (Simon, 2016). While the potential of the collaborative economy is undeniable, it is essential that the government leads these developments in the right direction (Frenken et al., 2015).

2.2 Platform strategies

According to Hagiu (2006) strategy defines the space in which platforms and its competitors are operating. The first issue platforms have to deal with is attracting multiple sides (i.e. A: peers that supply the service/product, B: peers that demand the product/service). In order to be valuable as a P2P platform the

(16)

16

demanding peer and the offering peer need to be both present (Evans, 2009). The literature refers to this issue as the chicken-egg problem.

Evans (2009) presents four strategies for platforms to deal with the chicken-egg problem. The first two strategies are about simultaneous entry of the different stakeholders. The first, the basic zig-zag strategy, is meant letting to grow the number of customers on both sides incrementally. In this case it is important to offer the customers certain value that is not available at other platforms (Hagiu, 2006). The second strategy is called pre-commitment on both sides (Evans, 2009). Both customer sides need to give a pre-commitment to the platform. A critical amount of customers on both sides is necessary (Hagiu, 2006). Evans (2008) states that obtaining the critical mass means a market that is substantial enough to create continuous growth. Achieving critical mass depends on the structure of the network effect (Evans & Schmalensee, 2010).

The other two strategies are sequential strategies, which try to get one of the customer groups on board and then go for the other side (Evans, 2009). In the third strategy, the single and double marquee strategy, the platform gains influential or famous customers on board first. Publicity is created to attract other customers. With the double marque strategy this is done on both sides, which is a simultaneous strategy. The two-step strategy is the fourth strategic option (Evans, 2009). At first customers are attracted on one side and then the other side is recruited. It only works when a platform can provide initial value without having the other side on board yet. Sequential strategies are easier to apply and platforms can analyse the stakeholders of their customer and select one of them as the second side for their platform. However, it is important to attract the most important customer group first (Hagiu, 2006).

Evans (2003) states that platforms should focus on a niche first, adjust operations and investments by using trial and error and then scale the platform. By focusing on a niche it is easier to acquire the first customers. Also, user-adoption is achieved quicker when the platform offers high quality (Duan & Zhong, 2007).

(17)

17

When platforms solved the chicken-egg problem and built a sustainable business already, the next strategic decision is about deepening the basic functionalities for the existing stakeholders (depth) or expanding to radically new functionalities, which brings new sides on board (breadth) (Hagiu, 2006). Before a platform expands to new functionalities and new groups its should generate all value to be delivered in depth, so that competition can be defended. Expanding to new markets can decrease profits, whereas increasing the quality of the existing functionalities increases profits and strengthens the competitive position (Li, Liu, & Bandyopadhyay, 2010). After that, new functionalities can be added to the platform which might cause adoption by novel customer groups (Hagiu, 2006).

Gawer & Cusumano (2007) give other strategies to conquer a market. The authors refer to coring as the strategy to generate a new platform that does not exist yet. It makes its technology the core of an ecosystem. The second strategy, tipping, is applicable to win the platform battle from competitors. In case of tipping a platform takes a substantial lead and becomes the most acknowledged platform in its market, which gives an essential competitive advantage (Evans & Schmalensee, 2010). They warn however that tipping is overstated. The assumption that participation in a certain platform due to switching costs is debatable, since participation is not permanent and customers can switch.

2.3 Network effects

Firm performance can be measured and defined in many ways (e.g. profit, turnover, market share, customer growth, etc.). Shafer, Smith & Linder (2005) state that a firms chances to survive and grow is directly linked to its ability to capture and create value. Similarly, Evans & Schmalensee (2010) mention that the success or failure of a platform depends on the value it provides to its users. The authors also suggest that multi-sided platforms need to attract customers since their amount, or so called installed base, influences the platform performance. Salazar (2015), Haile & Altmann (2012) and Hagiu (2016) state that the growth in value of a platform depends on the degree that it attracts customers, which is described as a direct network effect. Metcalfe’s law explains, the value of a network equals the square of the number

(18)

18

of users that are connected to it (n2 _{= the value of the network), while costs grow linearly (Hendler &} Golbeck, 2008). Whether the costs grow linearly, might be the case for hardware based networks, like in telecommunication, but for our kind of platforms the growth may be less. Briscoe, Odlyzko & Tilly (2006) argue that networks grow in a much slower rate with the amount of users (n log(n)). Additionally, they state the value of a network cannot be calculated from its size alone, but depends on other factors to, like the type of network (e.g. television, fax, platform). While the authors disagree on the speed the network value grows, they do agree on significance of the size for networks. According to Qui & Srikant (2004) the size of the pool of potential users, so the market the platform targets, impacts the success of the platform too. When more peers offer their car on Snappcar, the value of the platform rises. Additionally, the usage level of a customer is positively associated with the usage level of its peers (Candogan, K., & Ozdaglar, 2010). Network effects can be indirect too. Lee et al. (2010) argue that the adoption of the platform stimulates production of complementarians and strengthens the value of the platform indirectly. The literature review on business models by Shafer, Smith & Linder (2005) show that value networks are important for value creation and besides complementarian partners consist of special relationships with suppliers too. For platforms it is important to establish network effects since it is a driver of growth in market share (Dubé, Hitsch, & Chintagunta, 2010). Network effects are a force that influences the adoption of a platform by the mass market (Gallagher & West, 2009). Network effect is key for the success of a platform (Salazar, 2015).

The theory on crossing the chasm, is related to network effects. Figure IV presents cut-off points for the adoption of a product or service by different types of user groups. Every group has a different demand which should be taken into account when entering a new stage in the user adoption process (Gallagher & West, 2009). Gallagher & West (2009) belief that shifting from one group to the next is largely caused by network effects. It is obvious that network effects and adoption by the next group of customers

(19)

19

are important to platforms. The next sub paragraphs elaborate on the drivers of network effects. Additionally, propositions are formulated to guide and scope the research.

Figure IV: Crossing the chasm (Moore, 1991) 2.3.1 Revenue models

Evans (2003) highlights the asymmetries between the different peer groups: if A generates more externalities for side B than vice versa, A tends to get a lower price. Optimal prices for numerous customer types should balance the demand and bring both sides on board (Rochet & Tirol, 2006; Rochet & Tirol, 2003; Evans, 2003). It could mean that only one customer group is charged (Gawer & Cusumano, 2007). Zhang & He (2013) state the optimal pricing strategy is one that increases the prices incrementally. However, before prices can be charged the platform should offer enough value. Haile & Altmann (2012) support this by stating costs impacts value negatively, meaning that the value should be higher than the costs.

Gallagher & West (2009) state that to enlarge adaption of the platform and increase chances of becoming a mass-market product prices can be reduced. Moeller & Wittkowski (2010) claim that decreasing costs has a positive influence on the tendency to share, meaning the price influences potential users. Another reason to lower the prices for customers is to decrease the attractiveness of the market for potential competitors (Hagiu, 2009). Evans (2003) even suggests a strategy wherein the platform pays the first customers to use the service or products. In this way a critical mass can be developed. The more

(20)

20

customers, the higher the value and this can turn the situation to a point where the platform can charge customers (Zhang & He, 2013). If platforms charge their end-users it can be done via fixed membership costs or fixed fees (Rochet & Tirole, 2006). Table II represent the different revenue models that are found in the literature.

To ensure the research is well scoped and guided in the right directions, proposition one is formulated as: ‘’Offering a peer to peer platform for free helps to build critical mass quickly.’’

Source Revenue model Explanation Example Teece (2010) Freemium/ Premium model

Offer service for free, gain many customers, then establish a premium paid service that creates more value than the free service.

Klusje.nl Udemy Teece (2010) Osterwalder (2004) Subscription model

Offer a service or product for a fixed amount on a monthly basis. Love Home Swap Picard (2000) Osterwalder (2004) Advertise-ment model

Free content is delivered and mixed with advertising. The content creator gets paid by the advertisers.

AirDnD

Osterwalder (2004) Fee model Buyers and sellers are brought together and transactions are facilitated, in which the platforms take a percentage.

Croqqer Barqo Teece (2010) Freemium Offer the platform for free, to gain many

customers.

Peerby Toogethr Table II: Revenue models for P2P platforms

2.3.2 Quality

Quality influences a companies’ market share (Salazar, 2015; Zhu & Iansiti, 2012). The quality of a platform is one of the enablers of network effects (Haile & Altmann, 2012; Gallagher & West, 2009). A distinguished platform can charge higher prices (Li, Liu, & Bandyopadhyay, 2010). Interestingly, Duan & Zong (2007) find in their simulation study, that quality is the key factor to achieve network effects and is even more important than the installed base. The authors suggest that companies should deploy their resources in such a way that it leads to better product quality and more product features. Zhu & Iansiti (2012) reveal

(21)

21

with their empirical study that in a market with a large competitor, a new platform with a small quality advantage still can conquer a dominant position. The authors recommend incumbent firms to attain a quality level that is at least comparable to that of the entrant. The expectation of the user about which platform will win the competition, influences their choice too (Zhu & Iansiti, 2012). However, they did not incorporate pricing in their research model which might influence the results of their research. Li, Liu & Bandyopadhyay (2010) state that competing undifferentiated firms can increase profits by improving their quality.

It is obvious, that quality influences network effects and is important for a platform to become successful. Nevertheless, none of the authors in the previous paragraph, define what quality is. Zeithaml, Parasuraman & Malhotra (2002) found in their literature review on service quality through web sites that quality is a multi-dimensional aspect. They state that the literature agrees on four dimensions that are part of quality, including the ease of use of the website, privacy, reliability and the aesthetics/site design. However, the study misses empirical evidence. In a survey in the US done by Yang & Jun (2002) they find four dimensions with a significant relation to the perceived e-service quality. The dimensions are reliability of the service, personalization, ease of use of the website and access for help. Madu & Madu (2002) produced a long list of e-quality dimensions that affect only virtual operations. Except comparable dimensions, like ease of use, access, reliability and aesthetics other dimensions are found too, such as product/service differentiation, trust, serviceability and web policies. Whereas the article provides a large list of dimensions that are linked to e-quality, empirical evidence and a ranking in the importance of the different dimensions is missing. In this research the following quality dimensions are used: access, aesthetics, differentiation, personalisation, policies, reliability, trust and usability/ease. These are defined in attachment I.

To ensure the research is well scoped and guided in the right directions, proposition two is formulated as: ‘’The quality of the peer to peer platform has a positive impact on user growth.’’

(22)

22

2.3.3 Marketing activities

Moeller & Wittkowski (2010) state that familiarity with the platforms increases the tendency to share, suggesting a positive impact of marketing on user adoption. Gallagher & West (2009) show in their study that marketing influences the adoption by users. Platforms can influence potential customers’ decision on participating in the sharing program, by promoting the cost and benefits of sharing versus ownership (Lamberton & Rose, 2012). Gallagher & West (2009) clarify the role of marketing in user adoption by using the 3P method of marketing: price, place and promotion. They state that chances to become a mass-market product increases when the price is low (see 2.3.1 Revenue models). As for place, products/services gain larger end-user adoption if it is available everywhere (Gallagher & West, 2009). This is not the same for online platforms as for offline products. However, place for online platforms can be interpreted differently. Platforms can grow quicker if they are available in more languages, countries, on different devices (e.g. tablet, smartphone and PC) and operating systems (e.g. iOS android, Windows). The third P, promotion, influences the adoption rate for end-users by creating awareness of the product and promote it to potential customers (Gallagher & West, 2009).

Bampo et al. (2008) state that person to person promotion, or the so called word of mouth (WOM), is another important marketing factor that even can result in viral marketing. The authors built an experiment based theory, which suggests that influential peers function as a hub for further promotion and that incentives stimulate promotion done by peers. Similarly, Peres, Muller & Mahajan (2010) state that word of mouth leads to market penetration and drives growth. However, Yoo, Sander & Moon (2013) find that intrinsic motives (e.g. the desire to help somebody) have a larger impact than extrinsic motives (e.g. monetary rewards) to tell other peers. Hill, Provost & Volinsky (2006) strengthen the theory on the importance of WOM. They find evidence for a three to five times greater adoption rate if potential customers are notified by customers than groups selected by best practices of the marketing team. Nielsen (2014) finds in their research among 29.000 consumers from 58 countries, that recommendations by

(23)

23

friends or family is by 84% of the people seen as the most reliable source for information. Nevertheless, WOM is difficult to control. Thorson & Rodgers (2006) describe electronic WOM (eWOM) as positive or negative statements made about a product, company, or media personality that are made widely available via the Internet. Evans (2009) state positive eWOM fastens network effects and will spread quicker when users are similar, when the innovators have many connections when the innovators friends have many connections. Therefore, it is beneficial for platforms to gain influential users, with many dense connections.

While it is clear that WOM and marketing activities have a positive impact on network effects, none of the authors talk about what marketing activities exactly are. The word ‘marketing activities’ suggests that it has different elements. WOM could be one of them. Besides WOM, Keller (2009) defined six other marketing channels: advertising, direct marketing, events and experience, personal selling, public relations and sales and promotion. Since marketing is a broad concept, these seven dimensions are used during the research. The seven marketing dimensions are defined in attachment I.

To ensure the research is well scoped and guided in the right directions, proposition three is formulated as: ‘’Marketing activities have a positive impact on user growth.’’

2.3.4 Partnerships

According to Singh & Mitchell (2005) collaboration can lead to increased sales and growth. This impact is larger when a small firm is collaborating with a large firm than when two small firms are collaborating (Singh & Mitchell, 2005; Stuart, 2000). This makes sense since the incumbent firm has its existing procedures, products, routines and clients. Not a partnership as such increases the performance, but the performance improves depending on the characteristic of the partnership (Stuart, 2000). Stuart (2000) states that partnerships can be an exchange of resources and know-how, but also give social status, when a partnership is established with a firm that has a good reputation.

(24)

24

Gawer & Cusumano (2007) and Haile & Altmann (2012) talk about products complementing the platform and state that the value of a platform and the potential for network effects is influenced by such complementarians. Platforms function as an ecosystem where the different stakeholders collaborate to promote value co-creation and sharing (Salazar, 2015). For example, the development environment of Uber enables companies like Pebble and Microsoft to integrate and co-create their products with that of Uber. Amit & Zott (2001) found in a depth study of 59 firms that the complementarians are important value creating aspects and are combined with the platform more valuable than separately. It is beneficial for companies to join a platform and offer complementary products/services since it often increases sales and chances on an Initial Public Offering (IPO) (Ceccagnoli, Forman, Huang, & Wu, 2011).

To ensure the research is well scoped and guided in the right directions, proposition four is formulated as: ‘’Partnerships have a positive impact on the user growth of peer to peer platforms.’’ When speaking about complementary products or services in terms of integrated software, like the Pebble smart watch which can be used to convoke an Uber taxi, the way the software of the core platform is designed is important. According to Olleros (2008) the value of a platform is created by providing opportunity space to complementarian parties to evolve a platform ecology. In order to open up markets for complementarians, access to the platform and the software is required (Boudreau, 2010). This enables external parties to innovate or connect complementary products and improve scale and scope. In-house development is an option too, but requires resources and skills, which makes these partnerships an attractive alternative. However, providing access to the platform has disadvantages too. There is less control over the platform and there might be conflicting interests between stakeholders (Boudreau, 2010). The compatibility designates the interface between the standard and complementary service or product (Gallagher & West, 2009). Deciding on the compatibility and openness is crucial for the future of the platform and will have a large impact on the adaption (Salazar, 2015). Additionally, Gawer (2014) argues that the degree of openness also impacts the way innovation is facilitated, since others get

(25)

25

more space to develop software complementary to the platform. Boudreau (2010) finds in an empirical research of 21 computing systems that granting access to complementarians fiercely increases the rate to which new devices are developed, with fivefold, whereas giving up full control (beyond granting access to their platform) only increases the innovation rate with 20%. Despite the interesting outcomes, it must be taken into account that this empirical study was focused on hardware complementarians. For software complementarians it might have different results. Furthermore, it was tested in a stable period which makes it difficult to generalize it to a hyperactive market. While complementarians can increase the value of the platform they have to be attracted at first. When attracting complementarian parties, the platform owner should keep the quality of its platform and complementary products in mind (Hagiu, 2006). Haile & Altmann (2012) argue that the quality of the software is important to get adoption by users and complementarian parties. They address the quality of software as functionalities of the platform, its speed, interoperability in which systems can exchange information and the efficiency of transactions that the platform facilitates. While Salazar (2015), Boudreau (2010) and Gallagher & West (2009) all state openness and compatibility of the platform influence the adoption by complementarian parties, Olleros argues that the open versus closed perspective in the literature is limited (Olleros, 2008). Tiwana, Konsynski & Bush (2010) anticipate on this limitation and argue that the evolutionary dynamics of a platform ecosystem depends on the software architecture, governance and environment. While the authors have developed a broad and applicable framework which gives proper insights into the influential factors of software for platform ecosystems, the empirical back up is still missing.

To ensure the research is well scoped and guided in the right directions, proposition six is formulated as: ‘’The openness of software of the platform influences the adoption of peer to peer platforms by complementarian firms positively.’’

(26)

26 2.4 Conceptual framework

In order to clarify the scope of the research, a framework is needed. The framework needs to be simple, and operationally feasible (Morris, Schindehutte, & Allen, 2005). In the previous paragraphs the elements that influence network effects were discussed. Figure V presents a framework of the influential factors on network effects. Each proposition is visible in the framework to present the directions of the research clearly. The propositions help to guide the data collection and analysis (Kohlbacher, 2006). Each factor is described in more detail in attachment I. This description increases the feasibility of the research and provides direction to the operationalisation. The operationalisation is explained further in the next chapter, Methodology.

Figure V: Conceptual model

Software openness Partnerships Freemium vs. Paid Quality: - Access - Aesthetics - Differentiation - Personalisation - Policies - Reliability - Trust - Usability (In)direct network effects Marketing activities: - Advertising - Direct marketing - Events & experience - Personal Selling - Public relations - Sales & Promotions - Word of mouth

(27)

27

3 Methodology

This section presents the research design, followed by a representation of different research methods. Within each method the data collection method is clarified. Next the validity and reliability of the research is discussed.

3.1 Research design

In this research there is chosen for a multiple case study. Thomas (2011) states that the comparison of different cases is interesting to research and not particularly the dynamics of the individual cases itself. Therefore, the cases shall be compared comprehensively. The research is explorative and inductive, meaning that observations will lead to elements for a new theory (Eisenhardt, 1989). In order to guide the data collection and analysis, propositions are formulated based on literature (Kohlbacher, 2006). Propositions are inferences about the relations between concepts from the literature. They can’t be tested like hypotheses, since for testing quantitative research is needed. Yet propositions can be used in our case as landmarks. They guide the research and can be seen as a porch for future empirical hypothesis testing research. Propositions also improve the scope and feasibility of the research (Baxter & Jack, 2008). This study contains a multiple case study, containing 13 cases/platforms. To identify interesting findings in this research, cross case comparison is done by viewing the data in different ways (Yin, 2013). For example, comparing marketing activities among high performing and low performing cases. Other comparisons are based on industry. Furthermore, similarities and differences between cases are analysed. Since a limited amount of cases can be studied, theoretical selection of extreme or transparent observable cases is most obvious (Eisenhardt, 1989). To get different insights the 13 that are chosen differ in market (goods, mobility, space and service) and size (1-20, 20 – 50, 50 – 200, 200 – 500 and +500 employees) (attachment VI).

(28)

28 3.2 Data collection

This paragraph describes the data collection methods. Document research is used to find general data of the platforms. Interviews functioned as a primary resource for information. Additionally, seven questionnaires are completed to assess certain dimensions of quality.

Document research. Over 120 different types of documents, such as video’s, articles, reports, investment sheets and webpages are studied. These documents function as an extra source for information and found on the internet.

Interviews. Within this research 13 semi-structured interviews are conducted (interview questions: attachment III). This approach allows participants to be the specialist (Leech, 2002). Guest, Brunce, & Johnson (2006) state that data saturation appears within the first 12 interviews. So the 13 semi-structured interviews make it possible to assess the data and draw conclusions. Interviews are performed with two P2P platform specialists, eight founders, and three managers of P2P platforms. The selection increases the chance that information is provided by people that really know about the firms’ business model and influential factors on customer base growth. In total 25 potential interviewees are approached by mail and phone. Variety in market and size was given attention during selecting potential interviewees (list of participants: attachment II). In order to ’seduce’ the respondents, the researcher did offer them the results of the study.

Questionnaires. For certain dimensions of quality, defined in attachment I (e.g. access, aesthetics, usability) a questionnaire makes it more reliable and feasible to review them (the questionnaire: attachment IV). Therefore, a questionnaire is compiled and filled in by seven persons. These people vary in age and gender and all could be a customer of the platform. It takes about one and a half to three hours to fill in the questionnaire. Obviously no statistical analysis or far reaching conclusions can be drawn from these questionnaires. However, some careful judgements can be made about the specific quality dimensions on each case.

(29)

29 3.3 Reliability and validity

Validity is the accurateness in which the research results presents what the data actually measured (Golafshani, 2003). The variety in the selected cases sharpens the external validity. This research uses multiple data collection methods which strengthens the construct validity (Eisenhardt, 1989). Using multiple methods, leads to higher credibility of the research too (Tracy, 2010). Internal validity is built by formulating hypotheses for future research and comparing the outcomes to previous literature afterwards (Eisenhardt, 1989). This too strengthens the position of theories in this study. Reliability presents the accuracy of the measurement and the extent to which the results can be reproduced (Golafshani, 2003). The reliability of the study is enlarged by providing complete information on the cases that are researched and participants that are interviewed, so that other researchers can reproduce the study.

3.4 Data coding and analysis

Each interview is recorded and transcribed by the researcher, followed by coding the documents and interviews. Coding is described as grouping concepts that contain similar elements together (Ryan & Bernard, 2003). A method, explained by Yin (2005) as relying on theoretical propositions, is used to create categories, meaning that they originate from the literature and propositions formulated during the literature review (attachment I & attachment V). However, when using this method important information could go lost, since outcomes might not match the theory. Since it is allowed to use more coding strategies than one, inductive coding is used too (Ryan & Bernard, 2003). Herewith, categories that were not in the deductive coding scheme, are captured too. Data are coded into categories of: marketing activities, revenue model, incentives, partners & complementarians, software openness, strategy and market. To increase precision of assigning concepts to certain categories, each category is defined and presented in attachment I. The software program NVivo 11 is used to insert all codes, facilitate coding, identify the frequency of appearance of categories and creating data displays (attachment V).

(30)

30

4 Results

This section presents the results from the interviews and document research (list of interviewee’s: attachment II; cases and codes: attachment I and V and; the cross case analysis table: attachment VI).

4.1 Revenue model

Freemium model. In five cases platforms have offered their platform for free. Bla Bla Car and Peerby stated they had offered the platform for free to build critical mass (van den Oever, 2014; Peerby, 2016). This implies offering the platform for free helps to grow quickly. The Country Manager Benelux of Bla Bla Car, told that when people don’t pay via the platform, they are more willing to change the departure time or destination than when the payment is made in advance. While offering the platform for free is a smart way to grow the user base quicker, she tells it has side-effects too. Peers behave less reliably since there are no consequences, like losing money if they do not show up. When the platforms introduced the fee model, the media and users had negative reactions. Anne van Arkel, community manager at Thuisafgehaald, said in the interview that the negative reactions lasted for a short period and once introduced new users do not question it Nevertheless at some point revenues need to be made. The founder of Thuisafgehaald, told that they underestimated the costs and challenge a platform has, after launching it for free they needed to generate revenues and introduce a fee (Tielbeke, 2015). Whereas offering the platform for free did help to build critical mass in some cases, the platforms Spull and Toogethr offer their platform for free too, but did not achieve great successes (Tielbeke, 2015).

Fee model. It is difficult for P2P platforms to become profitable. Kim Tjoa, founder and CEO of Heel Nederland Deelt, states:

‘’At peer to peer markets it is generally known that the revenue model is a very big problem. You need to have so many users, especially with the fee model, to become profitable.’’

(31)

31

Although making profits with the fee model is hard, eight out of the 13 platforms that are researched, apply the fee model as their primary revenue stream. The range of the fee is between 10% and 30% of the transaction amount. Some platforms with low transaction value seem to experience difficulties with becoming profitable. It’s easier to become profitable, if the transaction value is higher, like with Airbnb or Barqo. Barqo’s margin per transaction is five times larger (15% of an average transaction of € 280) than for example that of Croqqer (20% of an average transaction of € 50). The founder of Toogethr, a carpool platform that ended up having no revenue model said:

‘’If your platform offers a service which presents products with a low transaction value, what’s left over is negligible. For example, suppose the transaction value is € 5. After paying 21% tax, the financial service provider (e.g. PayPal) and the driver, what is left over? Nothing. Besides if you charge people they expect proper service too.’’

While this makes sense, five of the P2P platforms have found a solution to reduce the tax they need to pay, by setting up a foundation. Rob Langendijk, the founder and CEO of AirDnD which charges a fee of 10% explains how it works:

‘’The foundation controls the money of the peers that have paid. 90% of the money goes from the foundation to the customer that has the rights to receive the money, because it’s their revenue. 10% of that foundation goes to the platform. That’s our revenue. We then decently pay tax over that 10%, but that is of course a much lower amount than if we had to pay 21% tax over the full 100% of the transaction.’’

Advertising. Obviously it is hard to become profitable with the fee model, but what is an alternative? Some platforms sought a way to become profitable by additional incomes via advertising. Rob Langendijk from AirDnD, a platform for hobby chefs and home restaurants, states:

(32)

32

‘’Eventually you have your operating expenses, personnel costs and there are tons of euro’s necessary to cover that costs. That money will not come from that 10% fee. Next we went brainstorming on for who our platform is interesting. That is for super markets, food markets and kitchen manufactures.’’

They identified interesting stakeholders and made deals with food markets and kitchen manufacturers in order to grow their revenues. In exchange for a large sum of money AirDnD targets its hobby chefs with advertisements on food or kitchen equipment from their advertisement partner. Subscriptions. A third revenue stream for P2P platforms is the subscription model. Kim Tjoa said:

‘’If your platform is so successful and everyone wants to be part of it, then it is not illogical to charge users a subscription.’’

This suggests the platform needs a proper user base and success before it can ask users for a subscription. Another interesting subscription model, explained by platform specialist Martijn Arets, is that of Werkspot.nl, a handyman-jobs platform. On the demand side users can post jobs for free. The supply side can react and apply on the jobs. However, in order to react to several jobs, the handyman needs to subscribe and pay a monthly fee. Martijn Arets said:

‘’This platform just connects handyman and people that need a handyman and they facilitate the communication. They leave the rest for what it is and get paid for facilitating communication.’’ Martijn Arets suggests that platforms should consider to create additional revenues from providing training to users. For example, AirDnD could give cooking lessons to its hobby chefs. It requires creativity from platforms to establish several revenue streams.

(33)

33 4.2 Quality

In the interviews and documents quality and marketing activities were most often referred to (attachment V). While, it seems obvious that quality impacts user growth, we need to know more about the specific mechanism underlying quality.

Aesthetics. Following respondents of the questionnaire, platforms such as Airbnb, Peerby, AirDnD, Barqo and Bla Bla Car, offer a beautiful website. While AirDnD and Barqo started recently, the others are high performers and are active in over 10 countries and acquired millions of funding. The reactions in the questionnaire state these platforms have a one-pager as website, look professional, use nice colour combinations, pictures and fonts. Contrary, the respondents state that the website of Spull, a book sharing platform that has no revenue model, looks amateurish, uses inconsistent font types and sizes and has a bar with different menu options instead of a one-pager as website. About Toogethr, respondent four said:

‘’The layout is boring. The structure is simple. The colour and font of this part of the platform looks not fresh at all.’’

This is a huge difference with a successful platform. For example, respondent three stipulated the following about Airbnb:

‘’Many pictures of accommodations. The homepage is a one pager, and does not contain many menu options. The icons are used consistently. The font and colour are modern and clear. The pictures are nice. The site looks professional.’’

Usability. Respondents state Spull is poorly structured which makes it difficult to navigate through the website. Again the platforms Airbnb, Peerby, AirDnD, Bla Bla Car and Barqo offer great online experience with easy navigation. Respondent four remarked the following on the website of Peerby:

‘’Very easy to use the website, search, find and rent stuff.’’ Respondent one states about Spull:

(34)

34

‘’The website is not user friendly. It is not easy to understand. The way the website works is not explained and every time you search for a book you end at a buying page of Bol.com. Spull presents itself as a study book sharing platform, while many fiction books are offered. There is no homepage button to go back to the homepage which is annoying.’’

Furthermore, Airbnb discovered that a proper working app at the end of 2013 and found this is important since 50% of the Airbnb members access the platform via their mobile and hosts respond three times faster than via desktop, which means bookings happen eight times faster (Brown, 2014).

Differentiation. Most P2P platforms are innovative and unique compared to traditional market parties. For example, AirDnD is the only firm in The Netherlands with a concept of matching hobby chefs to people that want to go to a home restaurant. Toogethr and Bla Bla Car have the same core concept. However, with Bla Bla Car you cross the border, while Toogethr only offers national rides. Since the cost advantage of international rides is larger than national rides, Bla Bla Car differentiates itself from Toogethr. Another aspect Bla Bla Car distinguishes from carpool competitors, is the smart functionality of presenting the extra minutes that it takes when the driver would make a detour for its rider when their destinations differ (Breij, 2014). The larger platforms have the advantage of a large user base with supply and demand which makes it difficult for potential new entrants.

Access. Another quality dimension is the access users have to customer service, user support and help documentation. The less successful platforms Spull and Toogethr and the platform that just launched, Butlerbuddy have no proper help documentation with answers to frequently asked questions. Help documentation assists users in finding what they need easily and saves time for employees. Anne van Arkel told in the interview that before they had help documentation they were called continuously on how the website worked. Furthermore, it seems to be important to be accessible for customers. Floris van Hoogenhuyze, co-founder of Barqo, states:

(35)

35

‘’By tracking what clients communicate via a life chat, we know what they want. We learn a lot from that information and show to people we are involved.’’

Barqo is the only platform of all cases that is accessible via chat. Six of the 13 platforms are accessible via phone. All other platforms are accessible via a contact form or email. Airbnb is only accessible via phone for emergencies.

Personalisation. High performing platforms personalise the user profiles and show active users extra appreciation. For example, Bla Bla Car has five user levels. Users start as newbie and can become ambassador when it satisfies criteria, like having registered more than a year ago and received more than 12 reviews. Ambassadors are invited for diners and sometimes receive small presents. The ambassador level is also visible for other users. Manager of the Benelux of Bla Bla Car states:

‘’That piece of appreciation is very important for our users. If somebody becomes ambassador, it takes 12 hours before it appears on their profile, they contact us and ask why their ambassador certificate isn’t visible on their profile.’’

A small competitor from Bla Bla Car is Toogethr. Bla Bla Car is active in 22 countries and got over € 300 million funding, while Toogethr is only active in the Netherlands and got no funding (Alba, 2015). Toogethr has no user levels or personalisation of the profile. From the 13 platforms that are researched, only four have user levels. These are the larger and successful platforms.

Policies. The Country Manager Benelux of Bla Bla Car tells in the interview that they introduced a payment policy in May 2015. The implementation of a cancelling policy, in which car poolers only get 50% of their payment back when they cancelled within 48 hours before their planned ride, resulted in a reduction of the no show rate of 90%. Another policy that successful platforms have is insurance. Airbnb has an insurance for accommodations up to $ 1.000.000 (Brown, 2014) and Spinlister for some bikes up to $ 10.000 (Smith, 2015). Boat platform Barqo, launched in March 2015, recently partnered with an

(36)

36

insurance company to be able to guarantee users insurance (Boogert, 2016). From the five platforms that lend or rent goods, Spull is the only platform that doesn’t offer an insurance. For platforms that offer a service, an insurance is less applicable. The exchange doesn’t involve goods that can be damaged.

Trust. Offering insurances enables trust, which is the next quality aspect to address. Rob van der Star, founder and CEO of Croqqer, a platform that connects handy neighbours with jobs, said:

‘’We enable trust with three things. Firstly, users can review each other with a five-star rating and text. Secondly, every new member is screened by the community manager from that area. Thirdly, we stimulate users to complement their profile with the right personal information.’’

Except Spull and Toogethr, all platforms offer the functionality to give reviews. Reviews or completing the information on a profile are stimulated by notifications by email or on the platform when logged in. Some platforms, like Butlerbuddy and Croqqer, obligate giving reviews by progressing it into the transaction process. The payment to the user is done after the review is given. Airbnb found out that people find it difficult to give bad reviews and therefore implemented a review process in which one review isn’t visible on the platform until the other user has given the review too (Gebbia, 2016). A lot about trust can be learned from Airbnb. Co-founder Joe Gebbia states in a TED talk (2016) that they enable trust by designing certain sizes of text boxes and giving an example of what users can say when requesting an accommodation. Professional pictures of the accommodations also generate trust (Brown, 2014). Another functionality that Airbnb uses to create trust, are social connections (Brown, 2014). This enables users to see via Facebook how they are connected with other users and searches similarities (e.g. you both have studied at the University of Amsterdam). Butlerbuddy is building a community of trust by only accepting users to their community when they have been invited by another user. How this effects the user growth is difficult to tell since they just launched. Bla Bla Car, Spinlister and Airbnb embedded verified profiles, which shows when users have voluntary performed verified elements such as, phone, email and credit