Who gets what : exploring the value appropriation among participants within the Sharing Economy : a multiple case study

(1)

Exploring the value appropriation among participants within the Sharing Economy

A MULTIPLE CASE STUDY

MASTER THESIS

André A. Klein 11377240

M.Sc. Business Administration – Digital Business / Strategy track Faculty of Economics and Business, Amsterdam Business School, UvA

Supervised by Merve Güvendik Second supervisor: Konstantina Valogianni

(2)

STATEMENT OF ORIGINALITY

This document is written by André A. Klein, 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)

I. Table of Figures

Figure 1: Vertical supply chain ... 15

Figure 2: Value creation and appropriation ... 16

Figure 3: Prerequisites for eligibility of services and platforms ... 23

Figure 4: Excerpt of SE-service selection process ... 24

Figure 5: Overview of eligible services and platforms ... 24

Figure 6: Overview of commercial comparison companies ... 29

Figure 7: SE-platforms "Accommodation sharing long-term" ... 35

Figure 8: Commercial companies "Accommodation sharing long-term" ... 35

Figure 9: Overview of industry categorization ... 37

Figure 10: Maps overview of origin of Sharing Economy startups ... 39

Figure 11: Overview graph of origin of Sharing Economy startups ... 39

Figure 12: Overview of global activity categorization ... 40

Figure 13: Overview of price intensity ... 41

Figure 14: Data-source method ... 43

Figure 15: Participant's share (appropriation) - first step ... 46

Figure 16: Participant's share (appropriation) - second step ... 47

Figure 17: Consumer’s share (appropriation) third step ... 50

Figure 18: Platform's share (appropriation) third step ... 52

Figure 19: Supplier's share (appropriation) third step ... 54

Figure 20: Overview of significant p-values ... 55

Figure 21: SE-platforms "Accommodation sharing short-term" ... 73

Figure 22: Commercial companies "Accommodation sharing short-term" ... 73

Figure 23: SE-platforms "Bicycle Sharing" ... 74

Figure 24: Commercial companies "Bicycle Sharing" ... 74

Figure 25: SE-platforms "Car sharing (no simultaneous use)" ... 75

Figure 26: Commercial companies "Car sharing (no simultaneous use)" ... 75

Figure 27: SE-platforms "Carpooling inner city" ... 76

Figure 28: Commercial companies "Carpooling inner city" ... 76

Figure 29: SE-platforms "Carpooling outer city" ... 77

Figure 30: Commercial companies "Carpooling outer city" ... 77

Figure 31: SE-platforms "Cloth sharing" ... 78

Figure 32: Commercial companies "Cloth sharing" ... 78

Figure 33: SE-platforms "Home sharing (swapping) ... 79

Figure 34: Commercial companies "Home sharing (swapping) ... 79

Figure 35: SE-platforms "Parking space sharing" ... 80

Figure 36: Commercial companies "Parking space sharing" ... 80

Figure 37: SE-platforms "Spare desk sharing" ... 81

Figure 38: Commercial companies "Spare desk sharing" ... 81

Figure 39: SE-platforms "Store-place sharing" ... 82

Figure 40: Commercial companies "Store-place sharing" ... 82

(6)

II. Acknowledgements

I want to thank everybody who helped me in realizing not only this thesis, but also in realizing my master in Amsterdam. Especially, I want to thank my mother and my father, as well as Madeleine for supporting me so much during these intense months. Also, I want to thank my friends for being so helpful. Thank you Andy, thank you Will, thank you Dani! Merve, thank you so much for your great supervision.

(7)

III. Abstract

The purpose of this master thesis is to investigate how the created value within the Sharing Economy is being appropriated. More specifically, the analysis includes a perspective on value appropriation for each of the identified participants within the Sharing Economy. The theoretical background is based on the model of value creation and value appropriation by Brandenburger and Stuart (Brandenburger & Stuart, 1996). The research data includes in total 330 observations from the Sharing Economy-platforms and for comparison-reasons 330 price points from their commercial “counterpart companies”, which were offering the same´ service. The observations are based on over 50,000 price points. All data was conducted online from Sharing Economy platforms within the three cities London, Sydney and San Francisco, which were selected because of their density of available SE-platforms. Prices of different platforms, services and cities were taken and compared with the prices of their commercial counterparts. The comparisons allowed an analysis of the three participants “consumer”, “platform” and “supplier” within the Sharing Economy. The overall value appropriation within the Sharing Economy, reflected through averaging of the 330 comparisons, is appropriated as follows: Consumer’s share: 20%; Platform’s share: 12%; Supplier’s share: 68%. Furthermore, multiple regressions revealed that the applied independent variables (for example industry, city and price intensiveness), explain roughly a third of the participant’s value appropriation. Thus, also other independent variables have significant influence on the value appropriation. Next to the finding on influencing independent variables, this research also explains how those variables influence the value appropriation. Finally, this research reveals possible opportunities for each participant to increase the appropriated value within the Sharing Economy.

(8)

1 Introduction

The act of sharing is totally natural to us. “We share particulars like names and lineages, ideas and experiences, kisses and embraces, as well as vital generalities like air and water, land and space. […] But what happens when sharing is put to profit?” (Schor, Walker, Lee, Parigi, & Cook, 2015, p. 12).

This is exactly what has happened in recent years. “Spurred by technological advancement, a number of decentralized peer-to-peer markets, now colloquially known as the Sharing Economy, have emerged as alternative suppliers of goods and services traditionally provided by long-established industries” (Zervas, Proserpio, & Byers, 2015, p. 1).

And not only has the Sharing Economy already reached a point of impact, it has also been estimated to grow rapidly. “As of 2015, the Sharing Economy is worth about $15 billion and it is estimated to grow to $335 billion within 10 years” (Habibi, Davidson, & Laroche, 2016, p. 114). This fact becomes more understandable when taking engagement rates into account. As an example, “44% of U.S. adults have participated in such transactions, playing the roles of lenders and borrowers, drivers and riders, hosts and guests. The number this represents (more than 90 million people) is greater than the number of Americans who identify, respectively, as Republicans or Democrats” (Steinmetz, 2016).

(9)

While recently conducted research focuses on how the value is created in the Sharing Economy (SE), this thesis aims to fill literary gap on how that value is appropriated. Or more loosely described: “Which participants gets what part of the cake?”.

Building upon the theory of Brandenburger and Stuart on value creation (1996), this thesis analyses each participant’s appropriated value within the Sharing Economy – namely “consumer”, “supplier” and “platform”. Subsequently, multiple case studies are conducted in order to collect quantitative, secondary data from companies within the Sharing Economy.

The remainder of this thesis is structured as follows: The second section discusses existing literature surrounding value appropriation and creation in the Sharing Economy, which leads to the aforementioned literature gap. The third section outlines the data and the collection-method, followed by results, discussion and conclusions.

(10)

2 Literature Review

2.1 Definition of the Sharing Economy

A study of recent literature has shown that the term “Sharing Economy” is used to describe a not-yet clearly specified range of subjects. “The truth of the matter is that the Sharing Economy is a floating signifier for a diverse range of activities” (Schor et al., 2015, p. 13). The Sharing Economy is also described as an “umbrella concept […], which endorses sharing the consumption of goods and services through online platforms.” (Hamari, Sjoeklint, & Ukkonen, 2016, pp. 2047–2048).

Recent literature identifies commonalities of emerging businesses with the approach of access to non-ownership consumption as “utilizing consumer goods and services” and “their reliance on the internet” (Belk, 2014), while other literature describes the Sharing Economy as a model of “web platforms that bring together individuals who have underutilized assets with people who would like to rent those assets short-term” (Horstkötter, Freese, & Schönberg, 2014, p. 32). Literature also defines the Sharing Economy as an “economic activity that is peer-to-peer, or person-to-person, facilitated by digital platforms” (Schor et al., 2015, p. 14). Consensus, therefore, is consumers do not own the consumed goods or services, or the connectivity with platforms to access them.

Moreover, literature argues that the Sharing Economy aims to maximize the utility of unused assets by sharing them with partners – and that this leads to a collaborative consumption with

(11)

“people coordinating the acquisition and distribution of a resource for a fee or other compensation” (Belk, 2014, p. 1597). Belk identifies consumers of such as collaborators, and points out that the model of the Sharing Economy occupies a middle ground between sharing and marketplace exchange, with elements of both. (Belk, 2014).

Other literature agrees on the monetary aspect, but also argues it is a key difference within the term Sharing Economy: “Some are genuinely collaborative and communal, while others are hotly competitive and profit-driven” (Schor et al., 2015, p. 13). This statement of differentiation within the industry is further explained: The Sharing Economy emerged through the so called consumption sharing, which is part of the peer production, making the Sharing Economy also part of the peer production. But while peer production initially yielded “products that are not created for money and are freely available (examples include open source software (Linux, Firefox)), citizen science, shared cultural content, and crowdsourced knowledge (Wikipedia)), […] growing platforms are those where providers earn cash and consumers get a good deal. These are the large, well-funded, for-profits getting most of the attention — Uber, Lyft, and Airbnb”. (Schor et al., 2015, p. 14). This is further explored in Belk’s finding, that the web facilitates sharing unused assets with partners to maximize the utility these assets generate.

Other recent literature also makes a distinction between models of the Sharing Economy to models which are not Peer-to-Peer, “such as Zipcar, which is Business-to-Peer, in that the company owns the assets (cars) and rents them to consumers” (Schor et al., 2015, p. 14). Thus, platforms, which are not Peer-to-Peer, are excluded from the term “Sharing Economy”.

(12)

2.2 “For-Profit”-Companies of the Sharing Economy

With respect to the “large, well-funded, for profits” companies of the Sharing Economy, which were mentioned in chapter 2.1, monetary aspects of the Sharing Economy have become especially interesting and a very important area of study.

Previous researchers conducted studies on the external impact of companies in the Sharing Economy on competing industries (Berger, Chen, & Frey, 2017; Zervas et al., 2015). Studies on internal factors with a special focus on the creation of value have been conducted. In addition, research has been conducted on the extension of the theory of value creation in the Sharing Economy (Reuschl, Bouncken, & Laudien, 2017). Thus, value creation is already an existing area of research focus.

However, little is still known about the appropriation of the created value in the Sharing Economy – even though the term “value” refers in e-business to the total sum of the appropriated value of a transaction, “regardless of whether it is the firm, the customer, or any other participant in the transaction who appropriates that value” (Amit & Zott, 2001, p. 503).

Research on value creation was also conducted in recent case studies, but again not on appropriated value. This included a recent case study, according to which Uber’s driver partners “earn at least as much as taxi drivers and chauffeurs, and in many cases more than taxi drivers and chauffeurs” (Hall & Krueger, 2015). This surplus in payment can be seen as value creation (which will be further explained in chapter 3 with the help of the introduced model of Brandenburger and Stuart (1996)). It remains unanswered if the drivers appropriate all of the

(13)

created value or to what extent this appropriated value reflects the total created value. How large is the percentage of appropriation by the participants providing the service or the goods (in this case the drivers) compared to the total created value of this service?

Another recent case study focusing on value created in the Sharing Economy and states that in general, market entry of new competitors leads to more intense market competition (which ends up in more diversified supply and cheaper prices) and expects Uber, as part of the Sharing Economy, to have the same impact in its operating markets (Haucap et al., 2015). Other recent case studies on the Sharing Economy focus on other companies, for example AirBnB (Zervas et al., 2015). But that focus is limited on the created value and again does not focus on the appropriated value. Do the lessened costs lead to a higher appropriation of value for the consumers of the platform in the Sharing Economy? If not, are the providers of services or goods, or the platform appropriating more value? In particular, the question of interest is: What part of the total created value does each participant appropriate in the Sharing Economy?

This will provide useful information to not only the direct, but all indirect participants of this rapidly growing industry, namely for example the government, as well as possible investors or future participants.

2.3 Summary

In summary, existing literature provides an explanation of a range of conditions and characteristics, which come under the umbrella term; Sharing Economy. It also differentiates between platforms with monetary efforts and those platforms pursuing the purpose of collective

(14)

sharing without the goal of profit. Lastly, literature provides a distinction between P2P platforms as part of the Sharing Economy and B2C platforms, which are excluded from the “umbrella”.

Existing literature includes external impacts and firm-internal value creation of P2P platforms following monetary goals. However, missing from current literature is the next logical step: Research on the attribution of the created value. This illustrates a very interesting gap in existing literature. Although research exists on how value is created in the Sharing Economy, no literature exists on the value attribution by the firm and the involved participants.

Research in this field is therefore relevant for the scientific community, but also for practitioners, since the outcomes could be used to gain insights for a variety of different outcomes, for example for political discussions regarding regulations, for consumers and service providers in the Sharing Economy on how they relate financially to the other participants, as well as for entrepreneurial decisions regarding strategic focus of their company.

This leads to this master thesis’ goal of exploring the value appropriation among participants

within the Sharing Economy.

(15)

3 Method and Data

3.1 Underlying Theory

The outlined methodology is based on Brandenburger and Stuart’s model (1996) on value creation and appropriation (1996), which considers Porter’s vertical supply chain (1980) to define value creation. Furthermore, it also uses underpinnings of the cooperative game theory to then calculate the value appropriation. Brandenburger and Stuart’s theory follows the view of the total created value as the sum of the values appropriated by each party (Brandenburger & Stuart, 1996).

Figure 1: Vertical supply chain (Porter, 1980)

(16)

Figure 2: Value creation and appropriation (Brandenburger and Stuart, 1996)

In their model, the created value is calculated by subtracting the opportunity cost (OC) of the supplier from the buyer’s willingness-to-pay (WTP). The WTP emerges from the next cheapest alternative for the buyer, and the OC from the next best opportunity to generate the highest income for the offer.

The appropriated value of each participant is then calculated by simple subtraction: The buyer’s share is calculated by subtracting the “paid price” from the WTP; the firm`s appropriated value is calculated by subtracting the firm’s payment to the supplier from the received payment from the buyer; and the supplier’s appropriation is calculated by subtracting the OC from the received payment from the firm.

(17)

Considering the earlier introduced view of the term “value” as depending on the total value created – regardless from which participant, Brandenburger and Stuart’s theory fits very well and can be adopted (Amit & Zott, 2001).

Further literature takes other soft factors, such as perceived value, into account to illustrate the created value (Bowman & Ambrosini, 2000). Soft factors are not included in Brandenburger and Stuart’s theory (1996), which focuses on “hard” monetary facts. Including soft factors into the chosen model would generate an unclear path for the study. Furthermore, soft factors might also lead to a biased outcome. Therefore, this research will solely focus on the hard factors regarding monetary value attributed by each participant in a transaction and exclude all other factors.

To adapt Brandenburger and Stuart’s model (1996), each participant in a transaction in the Sharing Economy has to be allocated. Literature states that the role of participants increases in the Sharing Economy, since they can be both consumers and providers (Hamari et al., 2016). This leads to three participants in each transaction in the Sharing Economy: The in chapter 2.1 introduced “platform”, as well as the “consumer” and the “provider” of the goods and/or services on the platform. This again fits very well with Brandenburger and Stuart’s theory (1996) - the platform being the “firm”, the consumer being the “buyer” and the providers being the “supplier”. For a logical fit in the following, the terms “platform”, “consumer” and “suppliers” will be used to describe the three direct participants.

(18)

Brandenburger and Stuart’s model (1996) is well aligned with the outlined focus of this thesis and serves therefore as a grounding theory for the observations. Thus, this thesis follows a deductive approach.

3.2 Methodological Choice

This research follows an exploratory approach and does not require control of behavioural events. It also focuses on temporary events. Thus, this is a relevant situation for a case study to be used as the research method (Yin, 1994). This case study follows a “replication logic” and is supposed to reveal support for theoretically similar results in order to build theory – when conducted with the correct amount of examples (Eisenhardt, 1989; Yin, 1994). Thus, this research follows a multiple case studies approach.

Siggelkow (2007) states that “the goal of every author is to write a paper that readers (and reviewers) find convincing” (p. 20) and also states that even “a single case can be a very powerful example” (Siggelkow, 2007, p. 20). This means that it depends on each case and its uniqueness and thus leaves it open to the researcher what exact amount of samples is the correct one and how to come up with that number.

3.3 General Data-Framework

3.3.1 Source

In order to get a substantial overview of SE-platforms, the databases of the websites AngelList and TechCrunch were used. AngelList is a “matchmaking platform for founders and investors“

(19)

(Nanda & Kind, 2013, p. 9). AngelList further has “garnered substantial media attention and was used by many high profile angel investors and venture capitalists” (Nanda & Kind, 2013, p. 9). By 2013 it already “had approximately 100,000 start-ups and 18,000 accredited investors on its database” (Nanda & Kind, 2013, p. 9). “TechCrunch, founded in 2005, is a popular technology publication, dedicated to profiling startups, reviewing new products and breaking tech news daily. CrunchBase is TechCrunch’s open database with information about startups, investors, trends, milestones, etc. It relies on the web community to edit most pages” (Xiang, Zheng, Wen, Hong, & Rose, 2005, p. 607). Therefore, both websites offered a substantial overview for further research. Furthermore, detailed information on the platforms were conducted on each SE-platform’s websites, their offers in their Mobile-App (if available), data given by the company on a phone call and partly from price comparison portals. Thus, the data collection follows the quantitative collecting method.

3.3.2 Timeframe

In order to have a more comparable set of samples, the data of each company was observed within the predefined timeframe of April and May 2017, which was also built upon the timeframe of the master thesis given by the Amsterdam Business School.

3.3.3 Unit of Measurement

All observations are being made on a smallest unit basis. This means that the observations are being made on a single user basis, on a one time use of the service for the least possible length with a maximum of a single day length of use (one-day measurement).

(20)

Due to the differences in business and pricing models throughout the observations of all SE and comparison companies, some dissimilarities emerged, forcing to recalculate the initial measurement to “make it fit”. One example was the unit of measurement for the service “Cloth

sharing”. Each of the three observed SE-platforms offered their service only on a minimum

duration of above one day (from four up to a minimum of seven days). For comparison reasons, the prices of services with a minimum duration of above one day were simply divided by the number of days for the minimum duration. Thus, all prices matched the duration of a single use within a maximum of one day.

Another issue emerged due to needed paid memberships for some services. For example, the three platforms of the service “Accommodation sharing long-term” accessible “for free” – but in order to “unlock” all the available offers, the user had to pay for a membership. To account for this requirement, the membership-costs costs were also taken into account. If multiple options of the required membership were available, the lowest priced version, which still “unlocked” all offers, was taken into the calculation. In all cases, the memberships had a required minimum duration of above one day. Therefore, the membership costs were added to the costs of the service for the minimum duration of the membership – resulting in the total costs. Lastly, the total costs were divided by the required minimum duration of the membership (in days). Thus, all prices matched the duration of a single use within a maximum of one day – including membership costs.

Special offers like “first week free” were not taken into account, since this research goal is to reflect the general share for all participants – opposing a reflection of short term changes in

(21)

pricing. Furthermore, potential shipping cost were not included in the calculations in order to focus on the “real” cost of a product. All visible offers were included – also those offers, which were “in use” already and at the moment not available – since an exclusion of them would exclude those offers, which are actually being used by the consumers and therefore highly reflect the SE. Each initial unit is given in section 3.4.5.

3.4 Specific Observation-Framework

The purpose of this research is to develop theory, not to test it, “and so theoretical sampling is appropriate” (Eisenhardt & Graebner, 2007, p. 27). Siggelkow (2007) suggests that the author should rely on his/her personal opinion in order to determine a sufficient amount of data. This research aim was to include as many observations and thus platforms and services as possible. Therefore, all of the conducted services and platforms, which were “eligible”, were included in this research.

3.4.1 Eligibility

First, for time-efficiency reasons, this research was only conducted on data which was available online. Therefore, only SE-services, which were offered throughout an online-platform, are included in the observations of this research. Additionally, due to the model used by Brandenburger and Stuart (1996), this research can only be conducted on services, which include a financial transaction this research can account for. SE-services, which are offered “for free” are therefore not eligible. Third, the aim of this research is to conduct and compare SE-services not from only one, but from three major cities, in order to get a better understanding of the different dynamics based on the distinct locations. For this purpose, London, Sydney and San Francisco

(22)

were chosen for this research. Reasoning for this selection is further outlined in the categorization section under “cities” (section 3.4.7).

3.4.2 Definition

According to the arguments from the literature review of section two, the term “Sharing Economy” describes a yet undefined “umbrella” of activities. Recent literature gives only a “blurred” overview of possible distinctions in order to differentiate between SE and non-SE platforms and services. In order to be able to distinguish between platforms that are part of the SE and platforms that are not part of the SE, this research follows an own definition on the term “Sharing Economy”.

Decisive for the emerging definition is the peer-to-peer rationale and thus an exchange between two solely private parties, on which the SE is based on, leading to the exclusion of any professional participation on one of both “ends” of the transaction (supplier and buyer). The only involvement of a platform arises in enabling the connection of the peers in the form of the provision of an (online) platform.

The term “sharing” describes a specific form of transaction, in which one party owns a property, which is being shared with the other party. On the other hand, “selling” rather describes the opposite and is therefore distinctively excluded from the SE. Subsequently, purely service-based peer-to-peer transactions are not part of the SE, since they do not include any form of “sharing” but instead solely are based on “selling” time from the offering party. Instead, based on the author’s view, SE requires some sort of renting/lending of property. The renting-rationale

(23)

subsequently also leads to a fixed agreement, on when the property will be given back to its owner, leading to a time-wise limitation on the length of the transaction.

The described arguments led to the new definition of the Sharing Economy by the author:

The term "Sharing Economy" describes timely restricted and agreed private use of another

private person’s property, which sometimes includes an additional service.

3.4.3 Emerging Prerequisites for Data Collection

Subsequently, following prerequisites were found to hold a service “eligible” for further observation:

Figure 3: Prerequisites for eligibility of services and platforms

This lead to a process of selection for potential platforms. Ultimately, only those SE-platforms, which fulfilled all prerequisites were eligible for further research. The following figure reveals a part of this process. The full process is given in the appendix:

(24)

Figure 4: Excerpt of SE-service selection process

3.4.4 Selected Services and SE-Platforms

Out of all observations, the following variety of SE-platforms out of 11 services were found to be eligible for further analysis of the SE. For comparison-reasons, conducted was given a name by the author. In the following, each service is laid out in more detail:

Figure 5: Overview of eligible services and platforms

“Accommodation sharing long-term” refers the share of a privately own flat. The duration of the share exceeds the duration of a shared accommodation for a “short” trip, for example a holiday. This service may be best compared to a normal rent of an apartment. In this specification, all “sharing” is done between two private peers. This means that neither the owner of the apartment nor the renter is a commercial company, but both are private persons. Spareroom (London),

(25)

Flatmates (Sydney) and Easyroommate (San Francisco) were found to be most representative

(highest volume of offers) for further analysis and were therefore selected.

On the other hand, “Accommodation sharing short-term” refers to the private share of a flat on a short term and daily basis. Research revealed the leading market position of AirBnB in London, Sydney as well as in San Francisco with respect to the quantity of offered accommodations. Therefore, AirBnB was selected as platform for further research for all three cities. While very few accommodation-offers on AirBnB were found to also allow longer rents, the far majority was laid out for short term holidays or city vacations.

“Bicycle sharing” is self-explanatory. This term describes the share of bicycles of all types.

Spinlister’s business model applied to all mentioned prerequisites for the SE-eligibility.

Furthermore, Spinlister was also found to be present in all three conducted cities with a leading market position – and was therefore chosen for each city for further research.

“Car sharing” was found in a variety of ways within the Sharing Economy. One of them, “Car

sharing (no simultaneous use)”, refers to the “pure” share/rent of a car. Opposed to other found

“car-sharing” services, “car sharing (no simultaneous use)” does not include the presence of the owner of the car and simply means loosely speaking “I rent you my car for a specific amount of time”. In this scenario, for each city, several offering platforms were found. Easycarclub (London), Carnextdoor (Sydney) and Turo (San Francisco) were each respectively in their city the platform offering this service with the highest amount of offers on their website.

(26)

“Carpooling outer city” is another “car sharing” service within the SE. But in this scenario, the owner of the car does not only share his/her car, but instead drives the renter additionally from one city to the destination city – leading to a sort of “carpooling”. While Blablacar was found to be the platform within London with the highest number of offers, same applies for Coseats in Sydney and for Scoop in San Francisco. All data on Blablacar and Coseats was conducted from their respective websites. Data on Scoop was taken from their mobile App.

“Carpooling inner city” refers to a service similar to taxi driving, in which a private peer shares his/her car with another peer, while the car is being used to drive together to another destination, which is for the far majority within the same city. One widely known platform, which organizes this kind of service, is Uber. Uber was found to be the most used platform for this service in all three cities and therefore used for further research for London, Sydney and San Francisco.

“Cloth sharing” refers to platforms, which offer to organize peer-to-peer sharing of privately owned clothes. Research revealed that this scenario commonly focuses on designer-brands instead of “everyday” brands. Respectively for each of the three cities, Rentez-Vouz (London),

Designerex (Sydney) and Stylelend (San Francisco) were the chosen platforms for further

research. All data on those three companies was conducted from their respective website.

“Home sharing (swapping)” refers to online platforms, which organise a service similar to

AirBnB. But in this scenario, only whole apartments and/or houses were shared. Homeswap

appeared to be the most spread platform and active in all three cities. While Homeswap does not require a direct payment in order to rent an other’s peer home, “points” need to be collected

(27)

upfront by renting out one’s own home as well. Furthermore, a membership including a monthly payment is required in order to use this service.

“Parking space sharing” refers to the share of privately owned, but unused parking spaces between peers. The platforms Parklet (London), Parkhound (Sydney) and Spacer (San Francisco) represent the most spread platforms respectively in each city.

“Spare-desk sharing” refers to the share of office spaces in the form of unused spare desks.

Desknearme (London), Liquidspace (Sydney), and Officehub (San Francisco), represent the most

spread platforms respectively in each city.

“Store-place sharing” refers to the share of unused store-place peer-to-peer. Sharemystorage (London), Storemates (Sydney), and Spaceout (San Francisco) represent the most spread platforms respectively in each city.

3.4.5 Selected Commercial Companies

In order to work with the introduced model of Brandenburger and Stuart (1996) and to calculate the shares, the emerging WTP for the consumers have to be taken into account, which describes the maximum a consumer would be willing to pay, based on the next cheapest option for the consumer to get the service from somewhere else.

Therefore, each Sharing Economy-service was compared respectively with the best “non-Sharing Economy” service, which offered the same core service. Eligible were commercial companies,

(28)

which offered the same service on the same scheme, respectively not “selling” but also “renting out” property. The final selection was based on highest degree on comparability to the SE-company and subsequently also on the highest quantity of offers (and thus data points) given by each commercial company.

For each the 11 SE-services, one SE-company was compared in each of the three cities to one commercial comparison company based on 10 single observations, leading to 330 observations in total. If an observation of a Sharing Economy platform was directly comparable to an observation of a commercial comparison company, then this direct comparison was applied. For example, the service “car sharing (no simultaneous use)” compared similar cars with each other. Thus a direct comparison of the prices offered from the SE-platform and from the commercial comparison company was possible. This method is called “fixed observation” and in the following called “F-method”.

In some cases, the single observations of the SE-platform were not directly comparable to their commercial counterparts. For example, designer clothes in “Cloth sharing” differed completely from the offerings from their commercial counterparts due to distinct clothes and brand. Subsequently, the ten “main observations” were each based on the median respectively of all available prices from the SE-platform and the commercial comparison company. The result was then a sort of database, containing a large amount of prices, of which the median represented the ten observations. This method is called in the following “DB-method”. Instead of the mean, the median was taken in order to not be “distracted” by data-outliers, but to account for the data’s real centric meaning.

(29)

If possible, data on commercial comparison companies was conducted from a price comparison portal, in order to mirror the market as best as possible. In order to illustrate this process in more practical terms, the first service (“Accommodation sharing long-term”) will be explained in more depth in the coming section, followed by the 10 other services.

Following, the commercial comparison companies and the reason for their selection and the comparison are laid out:

Figure 6: Overview of commercial comparison companies

For the service “Accommodation sharing long-term”, commercial portals were conducted, on which companies instead of private persons rented out their accommodations. Respectively for the three cities, Easyroomlet (London), Furnishedproperty (Sydney) and Zeusliving (San Francisco) were found to be the most used companies and were therefore selected for the comparison to the SE-platforms.

“Accommodation sharing short-term” focuses on renting out accommodation on a short term

basis – and is therefore best comparable to hotels, because they offer the same service. Since hotel-rooms, as well as rooms rented out on AirBnB, vary widely, a direct comparison between rooms did not reflect the best option to compare this SE-service with its commercial counterpart. Instead, online hotel price-comparison platforms enabled not only one comparison, but multiple

(30)

comparisons with several different commercial companies. This option reflected a decent alternative approach for the comparison. For further research, Hotels.com was chosen as the price-comparison platform.

“Bicycle sharing” was compared to commercial companies which also rented out bikes on a

daily basis. For the comparison, Londonbicycle (London), Sydneybiketours (Sydney) and

BikrentalsSF (San Francisco) were chosen.

The service “Car sharing (no simultaneous use)” could be compared through simply comparing the prices of the offered cars on the SE-platform with the prices of those cars being offered by commercial companies. Therefore, in general, a direct comparability was given. Research revealed that not one single commercial comparison company offered at least ten same cars like their SE counterparts. In order to stick to the preferred direct comparison method, not one, but a variety of commercial comparison companies were taken into account per city. Thus, every SE-platform within this service was compared to different commercial companies. For example, ten observations from Carnextdoor in Sydney were compared to ten observations combined from

Drivenow (Sydney) and VroomVroomVroom (Sydney). London’s SE-platform Easycar Club was

compared to Rentalcars (London) and Easycar (London). San Francisco’s Turo was compared to

VroomVroomVroom (San Francisco) and Rentalcars (San Francisco).

“Carpooling outer city” referred to the service of being driven from one city to another city in

another private person’s car. A first comparability option were taxis. But whereas taxis also mirrored the service of being driven in another person’s car, taxis were too expansive and thus

(31)

not really used to get from one city to another city. In exchange, more used were trains and buses. Buses offered the next cheapest option and thus were the preferred option in order to derive the WTP. In London, ten observations were possible through bus-trips from London to another city within England (National Express). But in Sydney and San Francisco less than ten bus trips to other cities were offered – and thus also train trips were taken into account. Therefore, for Sydney Greyhound, Firefly Express and Murrays were selected, for San Francisco

Greyhound and Caltrain were selected.

The service of being driven in another person’s car within the area of a city (“Carpooling inner

city”) was best comparable to the service of simple taxis. Therefore, ten routes were selected and

the prices of the SE-platform and taxis for those routes compared. In order to establish a more meaningful comparison, the routes differed in length (short, medium, long). The website Ride

Guru generated the prices for the taxis and for Uber as well for the routes in all three cities.

“Cloth sharing” referred to renting designer clothes. While also a variety of commercial

companies offered the same service, the clothes themselves were not matching at all. In some cases, not even the offered designer-brands were matching between the SE-platforms and their commercial counterparts. Thus, no direct comparability was possible. Instead, the DB method was applied. Selected commercial companies were Chic by Choice (London), Yourcloset (Sydney), and Renttherunway (San Francisco).

The comparison of “Home sharing (swapping)” through Homeswap led to some difficulties. For one, Homeswap did not charge any fees for the rent of a house itself, but instead charged a

(32)

monthly fee. On top of that, a “points-system” was applied by Homeswap: A customer had to use points in exchange to rent out other person’s homes. Those points could only be collected when offering his/her own home to another member. In order to compare Homeswap’s service, both the monthly fee and the points system had to be taken into account. The points system’s “costs” were reflected by the missed financial income by offering his/her home to another member “for free”. Thus, the missed financial income reflected the “costs” of generating the points, which needed to be collected. In order to calculate those costs, the median price of a house/apartment on AirBnB was calculated and then used to reflect the missed financial income for the points. The comparison was done on a month-basis. Thus also the monthly fees were simply added to the calculation. Since no two home were directly comparable, again the DB method was applied. For that reason, best comparability was given by a hotels-comparison website. Thus, hotels.com was chosen as the commercial comparison company for all three cities.

“Parking-space sharing” was also simply compared to commercial companies, which also

offered car parking spaces. Since now two parking spaces matched exactly for a comparison (varieties for example in location, quality and space), again the DB method was applied. Therefore, Divvy was selected as the commercial comparison company for Sydney, while

Parkopedia was selected not only for London, but also for San Francisco.

“Spare-desk sharing” SE-platforms were compared to commercial companies also offering free

tables in offices. Some spaces were only available on a monthly basis. Therefore, the overall comparison took place on a “one month” (30 days) calculation. No two office spaces (desks) were directly comparable. Therefore, the DB-method was applied in this case. Within all three

(33)

cities, WeWork was the commercial company having the highest amount of offers of the same service on their website. Thus, WeWork was selected.

The SE-service “Storeplace sharing” was compared to commercial companies also offering free store places. Since no two store places were directly comparable due to differences for example in location and quality, again the DB method was applied. Some offers were only available for a minimum duration of four weeks. Therefore, the overall comparison was calculated for a duration of 28 days. Selected commercial companies were Storage Mart (London), Keltic

Selfstorage (Sydney) and Sparefoot (San Francisco).

All other information on each comparison, including for example the exact unit of comparison and the data of data collection, are illustrated in the two figures created for each service – one for the SE-platforms and one for the commercial comparison companies. While all other figures can be found in de appendix, as an example the two figures for “Accommodation sharing long-term” is laid out below.

3.4.6 Data-Comparison of SE-Platforms and Commercial Companies

For a better understanding of the comparison process, first, the process for one of the eleven SE-services is being outlined. The data on the comparisons of the other ten SE-services can be found in the appendices. In the example, the SE-platforms of the service “accommodation sharing

long-term” were compared to their commercial counterparts in order to calculate the WTP and thus

the shares of each participant involved in the SE-transaction. The SE-platforms were: London’s

Spareroom, Sydney’s Flatmates and San Francisco’s Easyroommate. Research revealed three

(34)

comparability with the services of the SE-platforms. The comparison was based on the DB-method. For a more meaningful comparison, the analysis included data of five distinct areas within each city. To ensure the intended collection of ten final observations, each median of an area was taken into account twice. The prices accounted for different durations depending on the SE-platform, but all exceeded the “one-day measurement” (see 3.3.3). Furthermore, a membership had to be paid (with a minimum duration of seven to ten days) in order to use the services.

In order to account for those variables, first all prices on each platform were conducted from their website (on the date of the 4th of May in 2017), then sorted by the city areas and then calculated for each city-areas’ median. Then, the calculated medians (five per city due to the five areas) were divided by the amount of days the price accounted for (i.e. seven days). Furthermore, the membership prices (if needed) had to be taken into account as well. In all cases, the memberships allowed a usage of above one day. So taking the total membership costs into account for the “one-day usage” would exceed the price for the single use. Therefore, the previously calculated sum was multiplied by the minimum duration of the membership (in days). Then, the costs for the membership were simply added, leading to the final observation. The same logic was respectively applied to each commercial counterpart, leading to ten observations for each of the three counterparts.

Both the supplier as well as the consumer had to pay each eleven British Pound for the membership, leaving for example the supplier of a furnished room in London in Leytonstone with a median of 124.33 British Pound – and the platform with 22 British Pound for a single

(35)

transaction. In this step, the data was “just laid” out – the final opportunity costs and thus the percentages were calculated in section 4 (“results”) of this thesis. The information “timeframe” accounted for the exact date and duration of the use of the service. But the “timeframe” was in this case not of interest, since the prices did not vary by day/time of usage.

Figure 7: SE-platforms "Accommodation sharing long-term"

Own illustration based on (“SpareRoom,” 2017), (“Flatmates,” 2017), (“EasyRoommate,” 2017)

Figure 8: Commercial companies "Accommodation sharing long-term"

(36)

3.4.7 Opportunity Costs for Suppliers

From the author’s point of view, all conducted SE services did not simply offer another way for the suppliers to monetizing their property, but really created those opportunities in the first place. No other comparable opportunity was available – and thus no opportunity costs were existent. An example further illustrates this: “Cloth Sharing” allowed the “supplier” (the offering peer) to monetize the owned clothes by renting them to another peer, which was prior not possible. Furthermore, no other option existed prior to monetize the clothes by renting them out instead to another party. The only other option was instead to sell the clothes – which in turn is not comparable and thus does not reflect an opportunity cost, since this option involves a permanent “loss” of the item – instead of a timely restricted rent to another peer. Other SE platforms within “Cloth Sharing” might also be available and might offer a second option to rent the clothes out, but this research aims at comparing the SE to the next best “not-SE” option. Following this logic, no other opportunity is available for the supplier to monetize his/her property. Opportunity cost are therefore for all SE services “zero”.

A last example is illustrated with the service “Accommodation sharing short-term”, which is reflected by the platform of AirBnB. Clearly, peers were also prior able to rent out their privately owned rooms to other peers throughout online platforms. But in fact, these transactions exactly matched the prior given definition of the SE and thus do not reflect a new opportunity, but simply another, less spread platform than AirBnB within the SE. Following this logic, again no other opportunity is available for the supplier to monetize his/her property.

(37)

3.4.8 Categorization

While the model of Brandenburger and Stuart (1996) does give the option to compare the three participants within one observation, the model does not specifically enable a comparison between distinct categories on a balanced, since the total amount of value appropriation differ widely between services and platforms and industries. But by relating the appropriated value of each observation to a percentage based-calculation, all observations can, next to simple averaging in order to derive to the solution of this research’s centric goal, also be compared with respect to a variety of different variables. Subsequently, all observations were categorized based on following variables to allow a richer comparison and analysis.

3.4.8.1 Industry

All services were categorized in three separate industries from the North American Industry Classification System, based on their business model:

(38)

3.4.8.2 City

In order to analyse differences and similarities in value appropriation based on locations, another distinction was done between three separate cities. To ensure availability of rich data within the compared locations, only cities with a sufficient amount of active companies within the Sharing Economy were eligible. Due to bans and hard regulations for example within countries of Europe for specific companies within the Sharing Economy (Pfaffenbach, 2015), some cities are more useful than other.

The aim was to find three cities, with a high density of SE-platforms, in order to allow a rich comparison. Recent literature did not provide sufficient data on distinctions regarding cities with a high impact on the Sharing Economy. Thus, I collected data from online research by from websites dealing with this topic. A website called “JustPark” offered a comprehensive list of companies within the SE, including the SE platforms origin. The authors state that the data was collected in 2015, is based on datasets from AngelList.com and Crunchbase.com and was crosschecked with the company-appendix of the book “The Business of Sharing” (Justpark, 2015).

The following graphic includes the number of foundations of Sharing Economy platforms within specific cities in 2015 – according to Justpark’s research. It reveals that North America, Europe and Australia provide strong density of SE-platforms.

(39)

Figure 10: Maps overview of origin of Sharing Economy startups (Davidson, 2015)

Figure 30 illustrates that the three continents of North America, Australia and Europe have the highest density of SE-platforms. Within those three continents, London, Sydney and San Francisco are observed to be the strongest cities (figure 31). The three cities together provided the chance of gathering a high amount of available data form three separate continents within the given timeframe. Therefore, London, Sydney and San Francisco were chosen for this research.

Figure 11: Overview graph of origin of Sharing Economy startups (Davidson, 2015)

(40)

In order to allow a direct comparison between all three cities, only those SE-services were included, which were offered in London, Sydney and in San Francisco. Research revealed that there is a lack of data to reflect every service more than once in each involved city. Some services are offered by more than one SE platform within one city. For example, the service “Carpooling inner city” not only represented by Uber, but also by Lyft in San Francisco. But this scenario only applies in this case for San Francisco – and not for London or Sydney. To avoid comparison issues, every service is only represented once in each city for this research.

3.4.8.3 Globality

The distinction separated SE-platforms which were present in all three examined cities and had in each a leading market position (most number of offers online), and those SE-platforms, which were not. If one SE-platform represented the service in London, Sydney and in San Francisco, globality was given – otherwise not. This comparison is interesting, because it might reveal differences in appropriation for SE platforms differing in size and success.

(41)

3.4.8.4 Price

All observations were sorted by their median-price. The upper half of the median prices was marked as price intensive; the lower part was marked as not price intensive. This distinction is helpful in order to analyse if a higher pricing of a service results in a different or similar result in value-appropriations.

Figure 13: Overview of price intensity

3.4.8.5 Observed duration of service for measurement: Daily or more

Initially, all services were supposed to be measured on the basis of a duration of use of one day. This categorization was made in order to distinguish between those services, which were compared on a single day basis and those services, which were based on a “above one day” measurement. Research revealed, that some services were only available for a minimum use of above one day. Additionally, inclusions of memberships forced some calculations to be based on a duration of above one day. The next two distinctions (“Membership” and “Minimum service time: Daily or more”) account for both reasons separately. This way, also the reason for a measurement of above one day was analysed.

(42)

3.4.8.6 Membership

This categorization is helpful to analyse differences for value appropriation solely due to a required membership. For all services, a required membership exceeded the length of one day. Therefore, the comparison had then to be done on a basis of above one day. Thus, “membership” offered one possible explanation for a reasoning for an observation on a basis of above one day (3.4.8.5).

3.4.8.7 Minimum Service Time: Daily or More

In some cases, the SE-platform simply required to use the service for a duration of above one day (e.g. seven days or one month). Thus, this requirement also increased the duration of the observation. Therefore, this requirement represented the second and last reason for an observation of a service of above one day (3.4.8.5).

3.4.8.8 Control Variable: Direct Comparability

If observations of the SE-platform and its commercial counterpart was directly comparable, then a direct comparison was applied. This method is called “F”. But as mentioned in section 3.4.5: If observations were not directly comparable to their commercial counterparts, comparability was constructed by comparing the median of all available data (price points) of each the SE-platform and the commercial counterpart. This method is named “database” (DB). This control variable was implemented in order to account for the two distinct applied methods for the comparison.

(43)

Figure 14: Data-source method

3.4.8.9 Usage of Price Comparison Portals

Also mentioned in section 3.4.5, some SE services’ observations were not directly comparable with their commercial counterparts. In order account for this fact, but to still allow a comparison, as much data as possible was gathered from that particular SE company and its commercial counterpart. But in few cases, the commercial counterpart was even mirrored by collecting data from a price comparison portal – in order to gather even more data. This method was helpful from the author’s point of few to generate a better mirroring of the “real” market. This categorization is useful to analyse, if this method created differences in value appropriation.

3.5 Summary on Gathered Data

This thesis aims to answer how created value within the Sharing Economy is distributed among its three participants (“consumer”, “platform”, “supplier”). The calculation is based on a model of Brandenburger and Stuart (1996) on value creation and appropriation and therefore takes the consumer’s willingness to pay (WTP), the price the consumer has to pay, the platform’s cost - equalling the amount paid to the supplier, and the supplier’s opportunity cost into account. The observations are based on three cities, each including eleven services. Per city and service, ten

(44)

observations were gathered, which led to a total of 330 SE observations. Each observation consists of up to 53,904 single transaction price-points.

No values were missing. Each SE-platform was compared to a commercial company, which offered the same type of service. If the single observations of the SE-platform were directly comparable to the observations of the commercial counterpart, ten direct comparisons were taken.

In some cases, a direct comparison was not possible. One example occurred in all three cities for the SE-service “cloth sharing”, in which the clothes offered by the SE-platforms did not exactly match the clothes offered by their commercial counterparts (Rentez-Vouz, Designerex and

Stylelend). In such cases, the comparison was done by calculating the median of all available

prices of each, the SE-platform as well as the commercial counterpart. The median was taken instead of the mean in order to not be “distracted” by data-outliers, but to account for the data’s real centric meaning. Opposed to the direct comparison through the observations, here simply the comparison of the two medians of the SE-platform and its commercial counterpart accounted for the ten comparisons. Thus, even though a direct comparison was not feasible, still a comparison was realized.

Each comparison therefore included the information on prices of the SE-platform, the WTP of the consumer, as well as the supplier costs. Since the opportunity costs were already derived to be zero, each comparison could calculate the following dependent variables: The created value, the consumer’s share, the platform’s appropriation and the supplier’s share. All shares were

(45)

being provided as a number and also as a percentage of the total created value, while only the latter was important in order to calculate and compare the appropriation. Additionally, the created value was also not of further interest after it derived the appropriations of each observation.

(46)

4 Results

4.1 First Step

The final analysis consists of three steps for each of the three participants (consumer, platform, supplier): First, in order to derive a preliminary answer to the thesis question on how the created value is appropriated within the SE, the average of the 330 observations was taken for each of the three participants. This derived the value appropriation in percent of each participant in average in the SE in a single transaction.

Figure 15: Participant's share (appropriation) - first step

4.1.1 Consumer’s Share – First Step

The average consumer’s share over all 330 observations ranges at 20%.

4.1.2 Platform’s Share – First Step

The average platform’s share over all 330 observations ranges at 12% and thus ranges lower than the consumer’s share.

4.1.3 Supplier’s Share – First Step

(47)

4.2 Second Step

Secondly, each participant’s average value appropriation was then calculated for the independent variables in order to account for the various segmentations/categorizations given throughout the data collection. Furthermore, the two control variables were taken into account as independent variables. The average of each participant within each independent variable differed “significantly”, if it deviated more than (+/-) 5% from the overall average. For example, if the average supplier share was 20%, only average supplier-shares in each segmentation above 25% and below 15% were differing “significantly”. This step was taken in order to avoid taking too small of a change into account.

(48)

4.2.1 Consumer’s Share – Second Step

Significant higher average consumer’s shares were observed within “Global” and “London”. This implies a higher share (appropriation) for consumers when using a platform, which operates in all three cities and also when using a platform specifically in London (compared to San Francisco).

Lower shares are observed for “i-T&W”, “San Francisco” and “Membership”. The control variables did not relate to any significant differences in this step for the consumer shares.

4.2.2 Platform’s Share – Second Step

With respect to the averages of each independent variable, no differences were observed

4.2.3 Supplier’s Share – Second Step

Significant higher average supplier’s shares were observed within “Not global” and “San Francisco”. This implies a higher share (appropriation) for supplier’s when using a SE platform, which does not operate in all three observed cities. Also, this implies a higher supplier’s share in San Francisco, compared to the other two cities.

No significant platform shares, which are lower than 63% were observed with respect to the averages of each independent variable.

(49)

4.3 Third Step

Finally, in order to examine rigor, the data was then subject to multiple regressions. Multiple regression is a method to test, if there is a correlation between independent variables and one dependent variable. The execution of it therefore gave an explanation for coherences connected to the “distribution” of the value. Through regression, every independent variable was tested on its effect on each dependent variable - thus, 18x3 hypotheses were tested. Each null-hypothesis (“no significant correlation”) was rejected when its p-value was below 0.05 (statistically significant/ statistically highly significant when p-value is below 0.001); meaning the correlation of independent with dependent variable was significant. Those cases are given in each of the “results” sections of each participant.

Dependent variables consist of consumer’s, platform’s and supplier’s shares as a percentage. For each independent variable (except “initial days”), dummy variables were used to calculate the correlation (1-yes, 0-no). Independent variables are determined by the industry: Transportation and Warehousing (“i-TW”: 1-yes, no), Real Estate and Rental and Leasing (“i-RRL”: 1-yes, 0-no) and Accommodation and Food Services (“i-AS”, reference category); the globality of a platform (present in all three cities): “Global” (1-yes, 0-no); the city: “London” (1-yes, 0-no), “Sydney” (1-yes, 0-no) and “San Francisco” (as reference category); “price”:(price-intensity high: 1-yes, 0-no); “Initial days” (given in measured days of use of service); “membership” (need of a membership in order to use the service: 1-yes, 0-no) and “min usage above 1d” (required use of service above one day 1-yes, 0-no). Furthermore, two additional independent control-variables account for possible changes in the final result only due to the way the data was collected: Data collection: “fixed” (direct comparability of respectively each observation

(50)

between SE company and commercial counterpart: 1-yes, 0-no) and “portal” (Commercial data was conducted from a comparison-portal: 1-yes, 0-no).

The following regressions were executed and expanded/modified based on the performance of the results.

4.3.1 Consumer’s Share – Third Step

The multiple regression’s adjusted R-Square is 0.27, closely related to the R-Square, which ranges at 0.29, indicating that the model can explain more than a fourth of the variance of the consumer’s share, and that individual independent variables just slightly “interrupt” the calculation. Further regressions removing independent variables did not result in more significant results. Thus, the first regression was used for final analysis.