I
Technische Universität Berlin
Faculty VII – Economics and Management
Essays on Trust and
Reputation Portability in
Digital Platform Ecosystems
vorgelegt von
Dipl.-Math.
Maik Hesse
ORCID: 0000-0002-9116-9391
an der Fakultät VII – Wirtschaft und Management
der Technischen Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Wirtschaftswissenschaften
– Dr. rer. oec. –
genehmigte Dissertation
Promotionsausschuss:
Vorsitzender: Prof. Dr. Maik Lachmann
Gutachter: Prof. Dr. Timm Teubner
Prof. Dr. Rüdiger Zarnekow
Tag der wissenschaftlichen Aussprache: 12. März 2021
Berlin, 2021
II
1
Essays on Trust and
Reputation Portability
in Digital Platform Ecosystems
2
3
Abstract
Digital platforms increasingly shape product and service offerings around the world, change
consumer behavior, and disrupt a broad variety of traditional industries. In these digital
markets, to create trust and realize transactions, platform users and businesses alike rely on
their accumulated online reputation such as star ratings and reviews. With the multiplicity
of existing and quickly surfacing platforms and business models, reputation portability has
emerged as a nascent research topic in Information Systems and adjacent fields. From a
theoretical, empirical, and practical perspective, the possibility to transfer online reputation
across platform ecosystems and services is a complex and constantly evolving issue. This
cumulative dissertation provides a first holistic view on central dimensions of this novel topic
by investigating the foundations, the economics, and the design of reputation portability in
digital platform ecosystems. At the outset, general dynamics of trust-building, reputation,
and data portability are formally introduced and conceptualized. Subsequently, from a
user’s perspective, the economic effectiveness of porting reputation across electronic markets
is evaluated in a case study leveraging platform data and by applying survey-based and
experimental research. Thereafter, technological, and legal requirements, as well as
practical design guidelines to implement adequate reputation portability mechanisms, are
derived. Throughout, this work considers the broader societal implications for individuals’
digital sovereignty and identity management in online environments and, finally, concludes
by providing an outlook with avenues for future research.
Keywords: Reputation, trust, data portability, data sovereignty, digital identity, digital
platforms, platform economy, sharing economy, peer-to-peer, multi-method approach,
online experiment, multivariate statistics, interviews, thematic analysis, design guidelines,
signaling, purchase intentions, cognition, behavioral economics, GDPR, Personal
Information Management Systems, blockchain
4
5
Zusammenfassung
Digitale Plattformen prägen zunehmend Produkt- und Dienstleistungsangebote auf der
ganzen Welt, verändern Konsument*innenverhalten und disruptieren eine Vielzahl
traditioneller Branchen. Als Signal für Vertrauenswürdigkeit in digitalen Märkten und zur
Realisierung von Transaktionen, verlassen sich Plattformnutzer*innen und Unternehmen
gleichermaßen auf ihre angesammelte Online-Reputation – zumeist in Form von
Sternebewertungen und/oder geschriebenem Feedback (d.h. Reviews). Aufgrund der großen
Menge existierender und zahlreich neu entstehender Plattformen und Geschäftsmodelle, hat
sich die Übertragbarkeit von Reputation zu einem eigenständigen Forschungsgegenstand in
der Wirtschaftsinformatik und angrenzenden Bereichen entwickelt. Aus theoretischer,
empirischer und praktischer Sicht ist die Portierung von Online-Reputation zwischen
Plattformen und digitalen Diensten, ein komplexes und sich stetig veränderndes
Forschungsfeld. Diese kumulative Dissertation bietet einen ersten ganzheitlichen Blick auf
zentrale Dimensionen dieses neuartigen Phänomens, indem sie die Grundlagen, die
ökonomische Wirkungsweise sowie das Design von Reputationsportabilität in digitalen
Plattform-Ökosystemen untersucht. Eingangs werden in dieser Arbeit die allgemeinen
Voraussetzungen von Vertrauensbildung, Reputation und Datenportabilität formal
eingeführt und konzeptualisiert. Anschließend wird die ökonomische Effektivität von
Reputationsportabilität in elektronischen Märkten aus Nutzer*innensicht beleuchtet. Dazu
wird eine Fallstudie beruhend auf Plattformdaten und unter Anwendung umfragebasierter
und experimenteller Forschung durchgeführt. Schließlich werden technologische und
rechtliche Anforderungen sowie praktische Gestaltungsrichtlinien zur Implementierung
adäquater Mechanismen für Reputationsportabilität abgeleitet. Im weiteren Verlauf
erörtert die Arbeit gesellschaftliche Implikationen für die digitale Souveränität von
Bürger*innen und das Management von Online-Identität, um alsdann einen Ausblick auf
zukünftige Forschungsfragen zu geben.
Schlüsselwörter: Reputation, Vertrauen, Datenportabilität, Datensouveränität, digitale
Identität, digitale Plattformen, Plattformökonomie, Sharing Economy, Peer-to-Peer,
methodischer Pluralismus, Multi-Method, Online Experiment, multivariate Statistik,
Interviews, thematische Analyse, Design Guidelines, Signaling, Kaufbereitschaft, Kognition,
Verhaltensökonomie, DSVGO, Personal Information Management Systems, Blockchain
6
7
Acknowledgements
“No man is an island,” the English poet John Donne once said. This dissertation has benefitted
immensely from several people’s active input but also from the silent support of many others,
to all of whom I would like to express my profound appreciation. First, I thank Prof. Dr.
Timm Teubner for his outstanding advice in all aspects of a doctoral student’s life, especially
sound guidance and supervision through the ups and downs of academic research. Second, I
would like to thank everyone involved in shaping the work at hand: my co-authors (a special
thanks to Prof. Dr. Marc Adam), benevolent but also overly critical and blunt peer-reviewers
at journals and conferences, my colleagues at TU Berlin and the Einstein Center Digital
Future, and friends for providing actionable feedback and valuable outside-in perspectives.
Third and last, my deepest gratitude extends to my girlfriend and my family who carried me
during unexpected challenges.
Finally, I wish to dedicate this dissertation to the memory of my mother, Beate Hesse.
Berlin, December 2020
8
9
Contents
ABSTRACT ................................................................................................................................. 3
ZUSAMMENFASSUNG ............................................................................................................. 5
ACKNOWLEDGEMENTS ...........................................................................................................7
LIST OF FIGURES .................................................................................................................... 11
LIST OF TABLES ...................................................................................................................... 13
LIST OF ABBREVIATIONS ...................................................................................................... 15
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY .... 19
CHAPTER 1: TRUST-BUILDING AND REPUTATION .............................................. 21
MOTIVATION AND INTRODUCTION .............................................................................................................. 21
STRUCTURE OF WORK AND OVERARCHING RESEARCH QUESTIONS .............................................................. 23
METHODOLOGY ......................................................................................................................................... 27
UNDERSTANDING THE PLATFORM ECONOMY ........................................... 31
INTRODUCTION .......................................................................................................................................... 31
RELATED WORK ......................................................................................................................................... 32
HOW PLATFORMS BUILD TRUST .................................................................................................................. 34
Taxonomy for trust-building cues..................................................................................................... 34
Degree of Social Interaction (DoSI) .................................................................................................. 35
METHODS .................................................................................................................................................. 35
RESULTS .................................................................................................................................................... 37
Platform and provider perspective ................................................................................................... 37
Consumer perspective ........................................................................................................................ 39
DoSI perspective .................................................................................................................................40
DISCUSSION ............................................................................................................................................... 41
Limitations and future work ............................................................................................................. 42
Conclusion ........................................................................................................................................... 42
CHAPTER 2: CONCEPTUALIZING REPUTATION PORTABILITY ................. 45
INTRODUCTION .......................................................................................................................................... 45
TRUST AND REPUTATION IN DIGITAL PLATFORM ECOSYSTEMS..................................................................... 46
REPUTATION PORTABILITY: A CONCEPTUAL MODEL .................................................................................... 48
DATA PORTABILITY REGULATIONS: FRIEND OR FOE? ................................................................................... 51
EMPIRICAL FINDINGS ON REPUTATION PORTABILITY .................................................................................. 54
PRACTICAL APPROACHES TO PORTABLE REPUTATION .................................................................................. 56
The past and the present .................................................................................................................... 56
The future: decentralized technology on the rise ............................................................................. 58
OUTLOOK AND CONCLUSION....................................................................................................................... 63
Scenarios for reputation portability ................................................................................................. 63
Conclusion ........................................................................................................................................... 64
PART II: THE ECONOMICS OF REPUTATION PORTABILITY ....... 67
CHAPTER 3: USER PERCEPTIONS AND A MARKET DATA PERSPECTIVE .. 69
INTRODUCTION .......................................................................................................................................... 69
RELATED WORK ......................................................................................................................................... 70
Trust and reputation on platforms ................................................................................................... 70
Reputation portability ....................................................................................................................... 71
A SURVEY ON PORTABLE REPUTATION ........................................................................................................ 73
Study design and data ....................................................................................................................... 73
Results ................................................................................................................................................. 74
REPUTATION IMPORT ON BONANZA.COM .................................................................................................... 75
Study design and data ....................................................................................................................... 75
Results ................................................................................................................................................. 76
GENERAL DISCUSSION ................................................................................................................................ 78
10
CONCLUSION ............................................................................................................................................. 80
CHAPTER 4: TRUST MEASURES AND IMPORT THRESHOLDS .................... 81
INTRODUCTION .......................................................................................................................................... 81
REPUTATION PORTABILITY SCENARIO ......................................................................................................... 83
EXPERIMENT ............................................................................................................................................. 84
RESULTS .................................................................................................................................................... 85
The trust-building effect of imported reputation ............................................................................. 85
The interplay of on-site and imported reputation ........................................................................... 86
Account abandoning .......................................................................................................................... 87
Trust-maximizing strategies ............................................................................................................ 88
Control variable analysis and robustness checks ........................................................................... 88
DISCUSSION ............................................................................................................................................... 89
CONCLUDING NOTE .................................................................................................................................... 92
PART III: THE DESIGN OF REPUTATION PORTABILITY .............. 95
CHAPTER 5: SIGNAL DISCREPANCY AND COGNITION................................ 97
INTRODUCTION .......................................................................................................................................... 97
RELATED WORK AND HYPOTHESES ............................................................................................................. 98
Trust and reputation in platform ecosystems .................................................................................. 98
Discrepant ratings and economic prospects .................................................................................... 99
STUDY 1: ONLINE EXPERIMENT ON RATING DISCREPANCY AND USER INTENTION ...................................... 100
RESULTS .................................................................................................................................................. 101
STUDY 2: AN EYE-TRACKING STUDY ON ATTENTION TO CONFLICTING SIGNALS .......................................... 102
DISCUSSION AND CONCLUSION ................................................................................................................. 104
CHAPTER 6: DESIGNING REPUTATION PORTABILITY ............................. 107
INTRODUCTION ........................................................................................................................................ 107
BACKGROUND AND RELATED WORK .......................................................................................................... 108
Reputation portability in digital platform ecosystems ................................................................. 108
Designing reputation systems ......................................................................................................... 109
Hybrid approaches in developing design guidelines .................................................................... 110
RESEARCH METHODS ............................................................................................................................... 110
Deduction: analysis dimensions and stakeholders ....................................................................... 110
Induction: qualitative interviews and thematic analysis .............................................................. 111
FINDINGS .................................................................................................................................................. 114
Why: impact of portable reputation ................................................................................................ 115
What: features and functionality ..................................................................................................... 116
How: requirements for operationalization ..................................................................................... 119
Risks and challenges ........................................................................................................................ 120
Policy and regulation ....................................................................................................................... 122
DISCUSSION ............................................................................................................................................. 124
General discussion ............................................................................................................................ 124
Limitations and recommendations for future research ................................................................ 125
CONCLUDING REMARKS ........................................................................................................................... 127
PART IV: OUTLOOK ..................................................................... 129
CHAPTER 7: QUO VADIS, REPUTATION PORTABILITY? ............................ 131
RE-VISITING THE RESEARCH QUESTIONS ................................................................................................... 131
CONCLUSION, LIMITATIONS, AND FUTURE RESEARCH ............................................................................... 133
OUTLOOK ................................................................................................................................................. 134
APPENDIX .............................................................................................................................. 137
SUPPLEMENTARY MATERIAL PART I ......................................................................................................... 138
SUPPLEMENTARY MATERIAL PART II ....................................................................................................... 140
SUPPLEMENTARY MATERIAL PART III ...................................................................................................... 147
REFERENCES ......................................................................................................................... 151
11
List of Figures
Figure 1: A platform transaction with reputation portability (schematic) .............................. 22
Figure 2: Structure of this thesis .............................................................................................. 24
Figure 3: Overview of data collection and methods applied in this thesis .............................. 28
Figure 4: Rating distribution across platforms ........................................................................ 38
Figure 5: Activity concentration across platforms ................................................................... 38
Figure 6: Actual scores vs. trust thresholds ............................................................................. 39
Figure 7: Frequency of social cues vs. importance ................................................................... 40
Figure 8: Structured literature review methodology ............................................................... 49
Figure 9: A conceptual model of reputation portability .......................................................... 50
Figure 10: Future scenarios for reputation portability ............................................................ 63
Figure 11: Importance of portable reputation and preferred signal for import ....................... 74
Figure 12: Revenue difference composition ............................................................................. 77
Figure 13: Consumer trust in complementor by rating score .................................................. 86
Figure 14: Import and account abandoning thresholds ........................................................... 87
Figure 15: Trust-maximizing rating display ............................................................................. 88
Figure 16: Experimental setup and treatment distribution ................................................... 100
Figure 17: Analysis dimensions and stakeholder groups ........................................................ 111
Figure 18: Adapted thematic analysis approach based on Braun and Clarke (2006) ............ 113
Figure 19: Expanded thematic framework .............................................................................. 114
Figure 20: Online survey ........................................................................................................ 138
Figure 21: Landing page for online experiment participants................................................. 140
Figure 22: Blurred profile images of complementors on platform ........................................ 140
Figure 23: Mean trust scores per host name and profile image ............................................ 146
12
13
List of Tables
Table 1: Degree of Social Interaction (DoSI)............................................................................ 35
Table 2: Platforms’ and users’ use of trust-building mechanisms ........................................... 36
Table 3: Importance of trust-building mechanisms per platform type ................................... 37
Table 4: Empirical findings on reputation portability ............................................................. 55
Table 5: OLS regression models (score, price, demand) .......................................................... 76
Table 6: Parameter estimates (trust) ....................................................................................... 85
Table 7: OLS regression models (purchase intentions) ......................................................... 102
Table 8: Interview summary table .......................................................................................... 112
Table 9: Structured literature search ..................................................................................... 139
Table 10: Treatment matrix .................................................................................................... 141
Table 11: Balance of participant-specific controls across cases .............................................. 141
Table 12: Number of reviews across treatment variations ...................................................... 141
Table 13: Instrument variables .............................................................................................. 142
Table 14: Extension by (mean-centered) control variables ................................................... 143
Table 15: Interaction of rating score and number of reviews ................................................ 144
Table 16: Number of observations and average trust scores .................................................. 145
Table 17: Effects of gender (host/participant) on assigned trust scores ................................. 145
Table 18: Background of interview participants .................................................................... 149
14
15
List of Abbreviations
Application Programming Interface (API)
Area of Interest (AOI)
California Consumer Privacy Act (CCPA)
Chief Executive Officer (CEO)
Coronavirus Disease 2019 (COVID-19)
Curriculum Vitae (CV)
Consumer-to-Consumer (C2C)
Degree of Social Interaction (DoSI)
Distributed Ledger Technology (DLT)
Dependent Variable (DV)
European Data Protection Supervisor (EDPS)
European Union (EU)
Face Visibility (FV)
General Data Protection Regulation (GDPR)
Hypertext Markup Language (HTML)
Identity Document (ID)
Indirect Stakeholders (IN)
Information System (IS)
Intellectual Property (IP)
Know Your Customer (KYC)
Ordinary Least Squares (OLS)
Peer-to-Peer (P2P)
Personal Information Management Systems (PIMS)
Profile Image Availability (PIA)
Platform Operators (OP)
Platform Complementors (COMP)
Platform Consumers (CONS)
Self-Description Availability (SDA)
Social Network Sites (SNS)
Switzerland (CH)
User-Generated Content (UGC)
United Kingdom (UK)
Trust and Reputation Information (TRI)
United States (US)
User Experience (UX)
16
17
“A good reputation is
more valuable than money.”
– Publilius Syrus, (fl. 85 – 43 BC)
18
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
19
PART I: THE FOUNDATIONS OF
REPUTATION PORTABILITY
20
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
21
Chapter 1: Trust-building and reputation
Motivation and introduction
In the 21st century, digital platforms have become omnipresent. Today, electronic
marketplaces penetrate all levels of modern economic activity, shaping the way in which
individuals consume and provide services and products (Alt and Zimmermann 2014; De
Reuver et al. 2018; Hein et al. 2020). Following Hein et al. (2020, p. 90)’s definition, digital
platform ecosystems comprise three main stakeholder groups, where a platform operator (i.e.,
the owner of the platform) “implements governance mechanisms to facilitate value-creating
mechanisms on a digital platform between the platform owner and an ecosystem of
autonomous complementors and consumers.” The rise of these platforms, frequented by
billions of users, has been heralded as “one of the greatest success stories of the internet”
(Tadelis 2016, p. 322). In recent years, peer-to-peer (P2P) business models especially, became
increasingly important and their growth has been studied closely, including by popular science
(Boston Consulting Group 2017; Botsman and Rogers 2010; Sundararajan 2016). Matching
online supply and demand, well-known platforms exist, for instance, in domains such as
e-commerce (Amazon, eBay, Taobao), accommodation sharing (e.g., Airbnb, Couchsurfing,
Wimdu), mobility (e.g., BlaBlaCar, Lyft, Uber), or crowd work (e.g., Helpling, TaskRabbit). In
fact, providing essential infrastructure for mobility, food delivery and medicine supply in times
of crisis, some of these platforms were considered as “systemically relevant” during recent and
ongoing COVID-19 lockdowns (Bohn et al. 2020).
At the same time, certain platforms (such as Airbnb, Amazon, and Uber) have accumulated
enormous market power, creating high entry barriers for nascent and/or smaller platforms
(Gans 2018; Van Alstyne et al. 2016). In this context, an important issue is the so-called lock-in
effect users face on a respective platform, which represents a considerable obstacle for
switching to other platforms. This lock-in, inter alia, results from the process of building up
and maintaining a reputation history on a given platform – a process that involves effort and
is costly (Dunham 2011). Examples of this platform-based reputation are numerical ratings
(e.g., 5-star logic) and written customer feedback (e.g., text reviews). This online reputation is
indispensable for the realization of transactions and generally considered the strongest and
most frequently used signal to demonstrate trustworthiness towards other market participants
(Dellarocas et al. 2009; Mavlanova et al. 2012; Resnick et al. 2000; Tadelis 2016).
Now, reputation is always important for complementors, that is, for those market participants
who provide a service that consumers ultimately pay for. However, platforms exist where both
sides – including consumers – depend on reputation. Consider, for instance, guests on Airbnb
who can book apartments only if their reputation allows them to; otherwise hosts may refuse
their booking requests. The same applies to carpooling platforms and (to a limited extent) to
many crowd work scenarios. This thesis focuses mainly on complementors, that is, on product
and service providers. Players on this market side – the sellers, drivers, hosts, workers, etc. –
ultimately provide a service and receive a financial compensation for it. For them, reputation
is a crucial competitive factor on any platform.
However, this reputation is usually platform-bound, meaning that – as things stand today – it
cannot be transferred across platform boundaries. This is problematic both for complementors
and from a regulatory perspective. For one thing, many of these complementors are active on
Chapter 1: Trust-building and reputation
22
several platforms simultaneously (i.e., multi-homing). Hence, they must build a reputation
repeatedly, that is, one on each utilized platform (Dellarocas et al. 2009; Teubner et al. 2019).
Secondly, moving to another or new platform, complementors initially cannot prove their
trustworthiness to other market participants (cold-start problem; Kokkodis and Ipeirotis
2016; Wessel et al. 2017). And finally, because the resulting lock-in effects restrict competition
between platforms.
In this sense, reputation portability (similarly, reputation transfer or imported reputation)
refers to the idea of leveraging existing reputation from platform A and using it as a signal of
trustworthiness on another platform B (see Figure 1). It is important to delineate the concept
of reputation portability from the related but distinctively different notion of trust transfer. In
the latter, consumers’ trust in the platform itself inherently extends to complementors offering
their services on that platform (X. Chen et al. 2015; Teubner et al. 2019). In the case of
reputation portability, however, trust will be inserted exogenously from outside the target
platform, which the existing reputation is imported to. As an example, drivers on one mobility
platform, say Uber, could use their ratings to promote their services on other suitable
platforms, such as Lyft. To support reputation-building, the existing history would be
integrated as an additional source of information for potential passengers. Similarly, it is
conceivable that ratings could be ported across platforms in different domains. For example,
the rating track record of a trustworthy eBay seller may serve as a trust signal on Airbnb. This
immediately raises questions of required domain comparability for reputation portability (i.e.,
fit; Teubner et al. 2020).
FIGURE 1: A PLATFORM TRANSACTION WITH REPUTATION PORTABILITY (SCHEMATIC)
As of today, portability of reputation is neither common practice in digital platform
ecosystems nor is it technically supported by many platforms. The e-commerce platforms
TrueGether.com and Bonanza.com are noteworthy exceptions. Both allow their users to
import star ratings and/or text reviews from eBay and Amazon. In scientific literature,
however, calls for portable reputation have been prevalent over the past two decades. These
studies hypothesized on reputation portability’s potential to increase trust between
complementors and consumers. At the peak of the dot-com era, Resnick et al. (2000) already
COMPLEMENTOR
(no previous reputation)
PLATFORM B
PLATFORM A
…
CONSUMER A3
CONSUMER A2
CONSUMER A1
COMPLEMENTOR
(established reputation)
Reputation Portability
Service
Payment
…
CONSUMER B3
CONSUMER B2
CONSUMER B1
FIGURE 1: SCHEMATIC DEPICTION OF A PLATFORM TRANSACTION WITH REPUTATION PORTABILITY
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
23
noted that not being able to port reputation between platforms is an obstacle for sellers. In
fact, a few years earlier, Amazon had allowed its sellers to import their eBay ratings. Since eBay
considered these reviews to be proprietary content, the company threatened with a lawsuit,
and Amazon discontinued its import function. With the emergence of further platforms and
the sharing economy, Dellarocas et al. (2009) likewise advocated for the implementation of
portable reputation for multi-homing complementors. Since then, the idea of managing digital
reputation by linking profiles and creating meta-reputation scores across platforms or
enabling "identity portability" has been taken up frequently (Gans 2018; Puschmann and Alt
2016; Tadelis 2016).
Beyond such calls, the transferability of reputation across platforms has thus far received only
little scientific attention. Most studies deal either (a) with the effect of reputation on trust
within a closed platform environment (Resnick and Zeckhauser 2002; Tadelis 2016; Ter
Huurne et al. 2017; Teubner et al. 2017), or (b) with the general idea of data portability between
platforms (De Hert et al. 2018; Drechsler 2018; Engels 2016), especially against the
background of the European Union’s (EU) General Data Protection Regulation (GDPR). The
EU, for its part, is pressing for research into the mechanisms and potential benefits of
reputation portability (European Commission 2016, 2017, p. 131). In this thesis, I respond to
these calls and address a relevant gap in information system (IS) research. Today's platform
economy is on the verge of stronger interoperability between platforms. Some platforms
already offer import functions, and GDPR’s Article 20 stipulates the portability of data and
potentially, pending a legal consensus, also the portability of online reputation. The
overarching goal of this dissertation is therefore to investigate whether, from the perspective
of individual users and platforms, the establishment of reputation portability is beneficial and
– assuming that this is the case – how it should be designed and implemented across digital
platform ecosystems. This goal is operationalized in the following research agenda.
Structure of work and overarching research questions
Figure 2 depicts the structure of this cumulative dissertation
1
. Overall, it is subdivided in four
main parts: I) the development of a general understanding of trust-building cues and the
phenomenon of reputation portability, II) a detailed view on users’ perception of reputation
portability and its economic effectiveness, III) an exploration of the design, display, and
implementation of reputation signals from different sources, and IV) an outlook with
concluding remarks and avenues for future research.
1
This work is based on the results and contributions of seven major studies that have been published in
peer-reviewed journals, conference proceedings, and publicly accessible online repositories. Most of
these studies are parts of joint research endeavors together with my esteemed colleagues Timm Teubner,
Marc T.P. Adam, David Dann, Fabian Braesemann, and Otto Hans-Martin Lutz. They will be indicated
as such in the respective chapters of this dissertation.
Chapter 1: Trust-building and reputation
24
FIGURE 2: STRUCTURE OF THIS THESIS
This thesis is further partitioned into seven chapters. Chapter 1 is titled “Trust-building and
reputation” and consists of two sub-chapters. First, it motivates the need for research and
introduces the research agenda as well as the overarching structure and methodologies.
Second, laying the basis for further research, it provides a cross-platform comparison of
trust-building cues and examines the context-specific importance of these tools in the eyes of
platform users. This chapter draws on joint work together with Prof. Dr. Timm Teubner, Dr.
David Dann, and Dr. Fabian Braesemann, which was published in the proceedings of the
Hawaii International Conference on System Sciences (Hesse, Dann, et al. 2020). It builds on
the previous observation that two-sided markets are gaining increasing importance as
mediators of electronic commerce. As these businesses rely on transactions among users,
central aspects to virtually all platforms are the creation and maintenance of trust. The goal of
this first chapter is to understand how platforms can succeed in this task. While research has
considered effects of trust-building on diverse platforms in isolation, the overall platform
landscape has received much less attention. The first research question thus states:
RQ1: How do platforms build trust and how are these trust-building mechanisms perceived
by platforms users?
Having identified reputation signals as the most important trust cues on platforms, Chapter 2
(“Conceptualizing reputation portability”) introduces reputation portability as one aspect of a
broader dialogue on digital identity management. The article has been published in the journal
Electronic Markets and is based on research with Prof. Dr. Timm Teubner (Hesse and Teubner
2020a). Importantly, in this chapter, I delineate the relevant terms, actors, and concepts,
discuss common mechanisms, and present a conceptual overview to structure and summarize
prior research on the emerging theme of reputation portability. I thereby integrate literature
FIGURE 2: STRUCTURE OF THIS THESIS
Chapter 1
Trust-building
and reputation
Chapter 2
Conceptualizing
reputation portability
Part I: THE FOUNDATIONS OF REPUTATION PORTABILITY
Chapter 3
User perceptions and a
market data perspective
Chapter 4
Trust measures and
import thresholds
Part II: THE ECONOMICS OF REPUTATION PORTABILITY
Chapter 5
Signal discrepancy
and cognition
Chapter 6
Designing reputation
portability
Part III: THE DESIGN OF REPUTATION PORTABILITY
Chapter 7
Quo vadis, reputation
portability?
Part IV: OUTLOOK
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
25
on cross-platform transfer of reputation, which thus far has been fragmented, and currently
observed practical approaches. Deriving both the status quo and a forward-looking research
agenda for the nascent topic, the second chapter addresses the following research question:
RQ2: How can reputation portability be conceptualized along the existing literature and
current practical implementations?
Discussing the foundations of trust-building and reputation portability in digital platform
ecosystems, this thesis identifies a large gap in empirical research. Hence, the second part of
this work, consisting of Chapters 3 and 4, uses a quantitative approach to assess the economic
potential of reputation portability. Both chapters deal with the effectiveness of imported
reputation, measured as the impact on economic target variables. This is particularly
important for platform operators and complementors alike, since it helps to understand
whether reputation portability may generate a positive impact for these actors and, thus,
whether or not it is worth the effort of implementing it.
Chapter 3, titled “User perceptions and a market data perspective”, analyzes the effectiveness
of reputation portability based on consumers’ attitudes towards portable reputation and a case
study in a real platform environment. It is based on joint work with Prof. Dr. Timm Teubner
and was published in the proceedings of the Hawaii International Conference on System
Sciences (Hesse and Teubner 2020b). First, the study presents survey data on consumers’
perception of portable reputation in the platform economy to uncover domain-specific
nuances. That is, we ask potential platform consumers to critically assess the idea of reputation
portability. Second, the study analyzes actual market data from an e-commerce platform to
assess the impact of imported reputation on price and demand, two basic indicators of sellers’
economic success. It thereby systematically compares different types of complementors: those
leveraging reputation import versus those who do not and/or those with prior ratings on the
platform versus those without – including all potential combinations thereof.
Chapter 4, titled “Trust measures and import thresholds”, considers complementors’ ability to
increase consumers’ trust in them by importing reputational data from another platform. The
study is currently undergoing the third round of reviews at the international IS journal
Business & Information Systems Engineering and is based on a joint research project with
Prof. Dr. Timm Teubner and Prof. Dr. Marc Adam. Building on the previous chapter, this study
also investigates imported reputation’s impact, however, this time consumers’ trusting beliefs
are leveraged as a proxy for complementors’ economic success. To measure trust and derive
specific thresholds for reputation import, the potential lever is evaluated by means of an online
experiment testing specific combinations of on-site and imported star rating scores. This
could, for example, help complementors to evaluate whether the import of a specific star rating
score, say 4.0 stars, is worthwhile given their already existing rating – if one is available – on
the platform they plan to import to. In other words: importing 4 stars can make all the
difference if no prior on-site reputation exists or if the on-site rating is a meagre 2 stars,
whereas importing this rating may render potential consumers more cautious if the already
existing reputation shows a stellar 5.0 stars. Results from both individual chapters have been
conflated in a multi-method research approach and were published in the proceedings of the
European Conference on Information Systems with my co-authors Prof. Dr. Timm Teubner
Chapter 1: Trust-building and reputation
26
and Prof. Dr. Marc Adam (Hesse, Teubner, et al. 2020)
2
. The overarching research question
for Chapters 3 and 4 is summarized as:
RQ3: What is the economic effectiveness of reputation portability?
Having empirically established reputation portability’s effectiveness for trust-building, the
next part of this dissertation is dedicated to a detailed exploration of the design and
implementation of adequate mechanism for reputation signal import.
Chapter 5, which carries the title “Signal discrepancy and cognition”, pursues the evaluation
of reputation portability design from a user psychology perspective. This study is based on a
joint research project with Otto Hans-Martin Lutz, Prof. Dr. Marc Adam, and Prof. Dr. Timm
Teubner and the article was published in the proceedings of the International Conference on
Information Systems (Hesse, Lutz, et al. 2020)
3
. Importantly, the cognitive and economic
effects of simultaneously available signals from two different platforms – especially when the
information displayed to consumers is conflicting – have thus far received only little research
attention. For instance, it is unclear how users would react to two – or even more – different
reputation signals on a given platform and how the availability of such data would affect their
behavior and decision making. Hence, to inform the design of reputation portability and assess
to what extent consumers’ purchase intentions may be impaired by discrepant rating
information, the fourth research question states:
RQ4: How does reputation signal discrepancy affect cognition and purchase intentions?
Under the title “Designing reputation portability”, Chapter 6 provides practical design
guidelines for direct reputation import between platforms. This article has been published as
a single authorship working paper on ResearchGate (Hesse 2020). Since the design of
adequate solutions for reputation portability has thus far received no scholarly attention, this
paper addresses an important gap in the literature. Namely, to date no study exists that tries
to understand how exactly reputation portability should “look”, what type of information may
be displayed, and which minimum set of features and functionalities such a system requires.
Based on a hybrid approach of deductive and inductive reasoning, in this study I leverage
qualitative data from a series of semi-structured interviews that were conducted with platform
stakeholders to derive actionable design recommendations. The overarching research question
for this chapter thus states:
RQ5: How should reputation portability be designed and implemented in digital platform
ecosystems?
Lastly, Chapter 7 is titled “Quo vadis, reputation portability?” and concludes this work. It
summarizes and discusses the thesis’ contributions and provides answers to the research
questions as set forth in this introductory chapter. Furthermore, in the final chapter I take a
forward-looking perspective in considering limitations and deriving potential paths for future
research. As such, this chapter re-evaluates reputation portability’s ability to serve as an
important lever for individuals’ mobility and digital sovereignty in online markets.
2
Please note, however, that this short paper is not part of this dissertation as it represents a distillation
of the more fine-grained results presented in Chapters 3 and 4.
3
This contribution won the award for best short paper in the “Sharing Economy, Platforms and Crowds”
track and was nominated as best overall conference paper.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
27
Methodology
To answer the previously introduced research questions, this dissertation leverages a range of
data collection and analysis methods, thereby addressing the phenomenon of reputation
portability from a variety of angles. Generally speaking, research can be classified into two
broad epistemological philosophies: positivism and interpretivism (Hovorka and Lee 2010;
Weber 2004). While the former is often grounded in empiricism and quantitative data and
demands science be reproducible and objective, the latter actively promotes the researcher’s
subjective role in uncovering structure and meaning in the (often qualitative) data.
Nevertheless, simply allocating a particular study into one of these research paradigms may
not always be fitting. As a matter of fact, IS research has historically leaned towards one end
of the spectrum, that is, embedded in the positivist natural science toolkit of quantitative
measurement, statistical significance, and hypothesis testing (Mingers 2001). However, given
information systems’ fundamental importance for modern society, the IS discipline intersects
with and borrows from a wide range of fields, including the social sciences. While the practice
of mixed method research
4
is common in these areas, more recently IS research has also
started to embrace and actively advocate for methodological pluralism (Mingers 2001;
Niehaves 2005; Petter and Gallivan 2004; Walsham 1995). Most importantly, the argument
brought forward by Mingers (2001, p. 241) is that “a richer understanding of a research topic
will be gained by combining several methods together in a […] research program.” In this
sense, there is a phenomenal consideration to be made, which extends quite naturally to the
case of reputation portability. Namely, that by applying different methods, attention will be
drawn to different aspects of the study object. Hence, to address its full panoply,
methodological diversity is indeed necessary, if not mandatory (Niehaves 2005).
In line with this reasoning, for the work at hand I therefore apply a range of research methods
to investigate reputation portability and grasp in its entirety the multi-faceted and nascent
topic. Considered collectively, this cumulative dissertation hence follows a mixed-method
approach of combining positivist and interpretive research paradigms. However, even within
individual chapters (e.g., Chapters 1, 3, and 5) both multi and mixed methodology will be
proposed to collect and analyze data. Figure 3 provides an overview of these approaches, which
have been mapped to the respective parts of the thesis.
4
One may wish to distinguish multi-method research, which, in its simplest definition, employs more
than one method in a single study (e.g., either two quantitative or two qualitative) from mixed method
research, which combines both quantitative and qualitative methods within one study (Petter and
Gallivan 2004)
Chapter 1: Trust-building and reputation
28
FIGURE 3: OVERVIEW OF DATA COLLECTION AND METHODS APPLIED IN THIS THESIS
Both Chapters 1 and 3 rely on empirical platform data, which is publicly accessible online. In
Chapter 1, attributes of individual user profiles are retrieved to get a better understanding of
the trust-building mechanisms deployed across eleven different platforms, whereas in
Chapter 3, profile data from a single platform (allowing the import of ratings) is collected to
investigate reputation portability’s effectiveness. Additionally, both chapters leverage data
from a survey administered online to study participants, with the goal of understanding their
general preferences for trust-building cues (Chapter 1) and reputation portability specifically
(Chapter 3). This survey also collects open-ended verbal statements on the potential use of
reputation portability. In a content analysis, the qualitative data is then coded quantitatively
by categorizing statements thus, again, connecting different epistemological positions (Stol et
al. 2016).
Chapter 2 relies on a structured and integrative literature review as a research method in its
own rights. While each chapter by itself presents a brief overview of literature relevant to the
respective research question, Chapter 2 looks at the nascent body of literature as well as
practical examples with the aim to assess, critique, and synthesize existing work on reputation
portability and to create a first conceptualization of the emerging topic. Integrative literature
reviews have been deemed particularly useful for such purposes (Snyder 2019).
Within the domain of IS research, the use of surveys has a long-standing history. Especially in
the context of technology perception models, survey-based research has produced seminal
work. These studies greatly advance our understanding of the usefulness and adoption of
technology and e-commerce marketplaces as well as users’ trusting beliefs and purchase
intentions in online environments (e.g., Davis 1989; Gefen 2002; Gefen et al. 2003; Legris et
FIGURE 3: OVERVIEW OF DATA COLLECTION AND METHODS AS APPLIED IN THIS THESIS
Part I
Part III
Chapter 2
Chapter 4
Chapter 5 Chapter 6
Chapter 1
Chapter 3
Part II
Platform/ online data Survey Online Experiment
Structured/ integrated literature review Interviews Eye tracking
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
29
al. 2003; Venkatesh et al. 2016). Chapters 4 and 5 both leverage data gathered in an online
experiment, for which a total of 500 Internet users were recruited on Prolific.co (Palan and
Schitter 2018; Peer et al. 2017). Participants in these experiments were presented with a
stylized reputation portability scenario, that is, complementors’ profile pages exhibiting both
reputation on the platform itself but also imported ratings from another platform. The items
on the questionnaire are based on well-established constructs in the IS literature, borrowing,
inter alia, from the pioneering works introduced at the beginning of this paragraph. From the
data thus collected, different sub-samples and target variables are evaluated in Chapters 4 and
5, respectively. Moreover, Chapter 5 proposes a multi-method approach of combining this
survey data with an additional explanatory layer of data from an eye tracking experiment,
which was successfully tested and calibrated for application in future studies. Relying on the
positivist research framework, Chapters 1, 3, 4, and 5 all apply statistical analyses (e.g., linear
and multivariate regressions) to investigate the quantitative data, explain relationships
between the observed variables, and infer causal effects.
Finally, I supplement these positivist approaches with a different type of knowledge on
reputation portability, one that rather seeks to understand the context of the phenomenon
under investigation than to explain antecedent-consequent relationships between
independent and dependent variables (Hovorka and Lee 2010). Thus, in the penultimate
chapter (i.e., Chapter 6), a series of ten semi-structured interviews is conducted to collect
qualitative data in an interpretivist approach to research. Leveraging said data by means of a
reflexive thematic analysis (Braun and Clarke 2006, 2019), the chapter explores the potential
design of reputation portability in digital platform ecosystems based on the researcher’s (i.e.,
my own) subjective interpretation of study participants’ expressed views. I will provide more
details on each chosen method and the overarching methodology throughout the respective
chapters and in the Appendix.
30
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
31
Understanding the platform economy
To lay the basis for the subsequent discussion of reputation portability, in this second part of
Chapter 1, I present an overview of trust-building mechanisms across digital platform
ecosystems. First, this study introduces a taxonomy of trust cues leveraged on platforms.
Second, it investigates the context-specific importance of these cues as perceived by platform
users. The findings reveal that both platforms’ provision and users’ perceptions of
trust-building mechanisms are informed by a platform’s degree of social interaction.
Maik Hesse, David Dann, Fabian Braesemann, Timm Teubner
5
Introduction
Two-sided platforms have gained accelerating importance and research attention over the last
couple of years (Stummer et al. 2018; Van Alstyne et al. 2016). Examples from the consumer-
to-consumer domain include services for accommodation (e.g., Airbnb, Homestay) and
carsharing (e.g., Drivy, Turo), mobility (e.g., BlaBlaCar, Uber, Wingly), e-commerce (e.g.,
eBay, Gumtree), crowd work (e.g., Helpling, TaskRabbit), and many more. Such platforms
have caused significant changes in many incumbent businesses along with the reorganization
of a wide variety of markets, work arrangements, as well as value creation and capture
(Economist 2013; Kenney and Zysman 2016; Sundararajan 2016).
As platform-based business models rely on the realization of transactions among peers, a
central aspect to virtually all platforms are the creation and maintenance of peer-trust (Hesse
and Teubner 2020a; Mazzella et al. 2016; Tadelis 2016; Teubner 2018). While recent research
has considered the effects of trust- building mechanisms on different platforms separately
(e.g., how profile images and star ratings affect trust and booking intentions on Airbnb; Ert et
al. 2016), the overall platform landscape as a whole has received much less research attention
(Teubner and Dann 2018).
We suggest that a broader assessment of trust and reputation across platforms is urgently
needed. Comprehending how platform operators act to govern and guide user behavior and
usage patterns in digital platform ecosystems can allow for better understanding of resulting
behaviors and outcomes—and vice versa. By offering mechanisms to build trust and
reputation, platforms acknowledge that economic transactions are “socially embedded”
(Granovetter 1985). Ratings, reviews, and expressive user profiles function as “systems of
control” (Granovetter 2005) to counter opportunism in digital environments, as they inform
about the “identity and past relations of individual transactors” (Granovetter 1985, p. 491).
Based on this reasoning, we conjecture that platforms’ degree of social interaction (DoSI; i.e.,
the extent to which personal interactions between platform users become part of the overall
5
This study was published in the proceedings of the Hawaii International Conference on System
Sciences with the title “Understanding the Platform Economy: Signals, Trust, and Social Interaction”
(Hesse, Dann, et al. 2020); see https://doi.org/10.24251/HICSS.2020.631.
Chapter 1: Trust-building and reputation
32
value proposition and may enable positive social experiences; Hawlitschek, Teubner, et al.
2018) represents a key driver of the adequacy and use of trust cues.
Our research objective, thus, is threefold. First, based on actual market data crawled from the
Internet, we provide a cross-platform overview and propose a categorization for the different
mechanisms platforms employ to build and maintain trust between users. In addition to
uncovering this full palette of trust-building tools available to platform users, we also highlight
the frequency of utilizing said mechanisms. Second, based on self-reported data from a
consumer survey, we analyze how the available mechanisms are actually perceived and
evaluated—and thus shaped—by the specific platforms’ users while also investigating which
additional trust cues they would like to see. In this sense, we juxtapose the availability
(platform perspective) with the usage frequency (user perspective: provider) and perceived
importance (user perspective: consumer) of trust-building on two-sided markets. These
perspectives help us to better understand which of the many striking observations are specific
to certain platforms—and which point to more fundamental phenomena of the platform
economy as a whole. Third, thanks to the multi-platform perspective presented here, we are
able to compare the use and perceived importance of trust-building mechanisms with respect
to a platform’s degree of social interaction.
To evaluate our reasoning, we draw on two independent and unique data sets: To start with,
by considering eleven platforms from various domains and assessing several characteristic
statistics, including rating scores, activity concentration, and profile images from a total of
over 42,000 user profiles. Next, by combining this empirical market data with quantitative
and qualitative survey data that we collected from 187 participants. They evaluated the
importance of different trust-building mechanisms across several platform types (e.g.,
accommodation, mobility, etc.).
In a nutshell, our results show that most of the investigated platforms make use of a variety of
means for trust-building—yet that there exist marked differences in how users actually utilize
them. Both, usage and perceived importance depend on context: the higher a platform’s degree
of social interaction, the more trust cues are used and deemed important.
Related Work
Typical users of consumer-to-consumer (C2C) platforms are non-professional individuals with
neither an established brand image nor global recognition. Consumption on these platforms
requires different levels of trust (Zarifis et al. 2019), yet many C2C transactions yield high
economic, social, and physical exposure (Hawlitschek et al. 2016). Users of BlaBlaCar (rides)
and Wingly (flights), for instance, literally put their lives into the hands of their respective
driver or pilot. Trust (into the prospective good/service provider) is hence of utmost
importance, as the degree of interaction between users is very high (Möhlmann 2015). As
recent research shows, high levels of trust can be achieved without prior in-person encounters
(Mazzella et al. 2016). We will now review which levers specifically create trust and propose a
taxonomy for them.
Research from the domains of economics, information systems, and electronic commerce has
studied the role of online reputation for trust-building, realization of transactions, and prices
for almost two decades (Bolton et al. 2013; Resnick et al. 2000; Tadelis 2016). The positive
economic effects of transaction-based trust cues (e.g., star ratings and text reviews; see (Hesse
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
33
and Teubner 2020a) for a definition) have been broadly investigated across a range of
platforms: from earlier work on realized price premiums for providers on eBay (Ba and Pavlou
2002; Resnick et al. 2006) to more recent investigations of effects on sharing economy
platforms such as Airbnb (Qiu et al. 2018; Teubner et al. 2017) or BlaBlaCar (Otto et al. 2018).
In view of the ever-increasing importance of platform business models and the emergence of
C2C platforms with private individuals in the role of providers, online reputation and user
representation are charged with particular meaning. Personal aspects and individual
characteristics are essential to many C2C platforms and transactions (Möhlmann 2016), and
that is for at least two reasons. First, contingent on the specific platform type, social
interactions between users do take place. Certainly, the extent of these encounters can vary—
from mere online communication to co-usage sharing. In the latter case, personal interactions
between users are a prerequisite for the transaction to materialize. Second, scenarios that
involve personal interactions may create additional, social value for consumers. This implies
that providers themselves are part of the overall value proposition of the platform (pleasant
and locally versed Airbnb hosts or entertaining BlaBlaCar drivers; Dann et al. 2018;
Hawlitschek, Teubner, et al. 2018; Tussyadiah 2016a).
Research has thus focused on the role of social experience and benefits beyond economic and
product-related considerations for platform usage. In fact, social motives are frequently
reported as a driver of consumption on C2C platforms (Bellotti et al. 2015; Botsman and
Rogers 2010; Tussyadiah 2015). Moreover, consumers’ intention to use said platforms is
shown to be positively influenced by social utility (Barnes and Mattsson 2017; Bucher et al.
2016; Oyedele and Simpson 2018; Schaffner et al. 2017). In consequence, providers can benefit
from users’ expectations of social value and accordingly market their product and themselves
as a centerpiece of the overall usage experience (Harris and Rae 2011; Tussyadiah 2016b).
Thus, the prospect of interpersonal trust from social contact enables these providers to become
their own brand (Mazzella et al. 2016). However, this works only, if the transacted product or
service is related to personal interaction. If the interaction is limited to the exchange of
standardized goods on an e-commerce platform, for example, there is little need in providing
“brand-building” trust cues.
In cases where the degree of social interaction is high, platforms provide dedicated means to
create expressive user profiles. Trust towards a prospective interaction partner, therefore,
hinges on what is conveyed through the platforms’ user interfaces—including supposedly
incidental clues such as personal preferences, facial expressions, wording, etc. Previous
findings suggest that high-quality sellers display a wider array of signals (Mavlanova et al.
2012) and recent research has begun to examine this particular role of user representation
explicitly (Abramova et al. 2017; Ert et al. 2016; Fagerstrøm et al. 2017; Ma et al. 2017).
Moreover, the proliferation of trust mechanisms has raised interest in view of regulation as
there arise new potential harms, including new forms of discrimination (Edelman et al. 2017)
as well as possibilities for strategic manipulation and market failure (Watt and Wu 2018).
On a different note, a user’s verified identity—validated and vouched for by the platform or an
external institution—can signal authenticity to other participants and platform operators have
developed a range of mechanisms (Mazzella et al. 2016; Sundararajan 2014). Such identity
verification was shown to increase trust, for instance in online dating (Norcie et al. 2013). In
the context of peer-to-peer accommodation sharing (e.g., Airbnb), the availability of verified
Chapter 1: Trust-building and reputation
34
identity is a popular research topic (Dolnicar 2017; Guttentag 2015; Jung and Lee 2017). While
some studies state a positive effect of verified identity on prices (Wang and Nicolau 2017)
others find no significant effects (Teubner et al. 2017; Xie and Mao 2017). Another means of
identity verification is the practice of linking social network accounts across online platforms
(Y. Chen et al. 2014; Silvestri et al. 2015).
Moreover, platforms deploy additional mechanisms to promote trust in service providers. The
use of badges as a sign of achievement has been described in the context of social networks,
but can generally be extended to all types of digital platforms (Kwon et al. 2015). Empirical
analyses show the effects of these badges on Airbnb (“superhosts”; Ert and Fleischer 2019;
Neumann and Gutt 2017; Teubner et al. 2017; Xie and Mao 2017) but usage is prevalent on
other platforms as well (e.g., Uber; Scheiber 2017). Further implicit information includes the
number of transactions/reviews and users’ membership duration (Qiu et al. 2018; Teubner et
al. 2017; Xie and Mao 2017).
Beyond mechanisms to increase trust (e.g., through ratings), platforms also employ means to
mitigate risks (e.g., fraud) and hence reduce users’ trust thresholds. Such “e-commerce
institutional mechanisms” include insurances and warranties, escrow payment services, and
privacy assurances (Fang et al. 2014). In addition to trust into prospective transaction
partners, also trust into the platform itself represents a prerequisite for transactions to
materialize. Importantly, a platform’s trustworthiness is suggested to rub off on the providers
on the platform (“trust transfer”; X. Chen et al. 2015; Möhlmann and Geissinger 2018;
Teubner et al. 2019).
While extant research has mainly considered platforms and trust cues in isolation, with this
paper, we extend this body of literature by providing a cross-platform trust cue “panorama”
and classification, which allows to take the varying degrees of social interaction as a driver of
the usage and adequacy of trust cues into account.
How platforms build trust
Taxonomy for trust-building cues
Overall, platforms use a variety of mechanisms for reputation management and trust-building.
Based on our review of related work and in situ exploration across a broad range of platforms
from different contexts, we propose the following categorization:
(1) Transaction-based trust cues refer to evaluations provided by prior transaction
partners. These may be numeric (e.g., star rating scales, positive/negative) or
written assessments (e.g., text reviews).
(2) Socially rich and Expressive User Profiles are created and organized by users
individually. Typical content includes photos/profile images, self-descriptions,
video messages, or references to external resources. We hypothesize that such trust
cues’ importance increases with a platform’s degree of social interaction.
(3) Next, Identity Verification refers to the platform examining a user’s personal
identity and authenticity, for instance, by email or phone number confirmation.
Moreover, depending on context, identity verification may include the provision of
an ID card, licenses, or linkage to online social network accounts.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
35
(4) Last, Implicit Information and distinctions such as status badges or the display of a
users’ membership duration, number of transactions, or average response time are
provided by the platform and accentuate a user’s particular status, behaviors,
achievements, or merits.
Degree of Social Interaction (DoSI)
Table 1 introduces the notion of platforms’ degree of social interaction—defined by the level of
personal interaction and social experiences on a given platform (Möhlmann 2016). A
platform’s classification is based on two factors: (1) the necessity for and degree of physical
encounters for service provision and (2) the extent to which personal interaction with the
service provider is part of the overall value proposition—in addition (and as opposed to)
focusing solely on the underlying product or service. Hence, where personal experience
becomes a constituent part of the overall value proposition, Expressive User Profiles are
particularly well-suited to convey information to prospective consumers. One would expect
this to be reflected in both the availability of expressive trust cues as well as their perceived
importance.
Proposition: A platform’s degree of social interaction is positively related to (1) the usage
of Expressive User Profile elements by providers and (2) the importance of such cues in the
eyes of consumers.
Methods
Data Collection 1: To examine how platforms build trust (empirically) and to evaluate our
proposition, we draw on actual and recent market data. In a first step, we inspect the outlined
platforms and trust-building mechanisms manually. Note that within the scope of this paper,
we focus on trust into the providers (i.e., hosts, drivers, workers, lessors, sellers). Naturally,
providers’ trust into consumers is essential as well since providers face considerable levels of
economic exposure and, after all, usually have the last say in whether a transaction will actually
be realized. Using web scraping (implemented in Java), we then retrieve samples of user
profiles and the associated data such as ratings, descriptions, profile images, and so forth for
each platform. Overall, we queried 11 platforms collecting data points of more than 42,000
user profiles. Data was collected between August and September 2018 and is available from
DoSI
Rationale
Low
No physical encounters between seller and buyer; very limited online communication;
focus entirely on the sold product; interaction purely transactional (e.g., e-commerce)
Moderate
Physical encounter with provider possible (e.g., brief interaction for service handover);
focus on underlying commodity (e.g., apartment, car); personal interaction may add to
the value proposition (e.g., accommodation, carsharing)
High
Physical encounter with provider prerequisite to service provision; consumers entrust
physical safety, access to property to provider; focus on service and provider; personal
interaction part of value proposition (e.g., accommodation, mobility, crowd work)
Note: Within in the same type of platform, nuances may occur for the DoSI and are reflected here (e.g., staying
in a shared apartment (high DoSI) vs. renting an entire home (moderate DoSI) on an accommodation platform)
TABLE 1: DEGREE OF SOCIAL INTERACTION (DOSI)
Chapter 1: Trust-building and reputation
36
the authors upon request. Table 2 summarizes the web-crawling results on the use and
characteristics of different trust cues on the respective platforms (Airbnb, Homestay, Wimdu,
EasyCarClub, Drivy, Turo, BlaBlaCar, Wingly, eBay, Gumtree, TaskRabbit).
Data Collection 2: As a second step, we launched an online survey
6
. Participants were
recruited via Prolific.ac (Palan and Schitter 2018) with a total of 204 participants from across
the EU, the US, Canada, and Israel. We controlled for online shopping frequency (i.e., more
than once per month) and age (i.e., 18-49; mean=33.7, 66% female) and equally divided study
participants among five platform types (accommodation, carsharing, mobility, e-commerce,
crowd work). A total of 187 valid responses remained after attention and completeness checks.
Based on the empirical findings, we asked participants to rate the overall importance of the
different trust-building mechanisms as well as the importance of individual artifacts and
properties (such as face visibility) on 7-point Likert scales. We added open-ended questions
that were manually coded into categories, for example, other drivers of trust and distrust, and
general drivers of platform use. The data was collected in May 2019, reviewed, and coded
independently by two researchers. Table 3 summarizes the survey results on the importance
of specific trust-building mechanisms per platform type.
6
See Appendix for further details.
(1) Transaction-based
(2) Expressive User Profiles
(3) Identity Verification
(4) Implicit Info
Platform
Rating
Score
Text Review
Profile
Image
Self-
Description
Email
Phone
ID
SNS
#T
#R
Since
Mean
Skew
Length
Sent.
Frequ.
Face
Frequ.
Length
Airbnb
92.34
.81
189
.85
1.00
.61
.58
234
×
×
×
×
×
×
Homestay
4.71
.85
216
.86
.99
.60
.54
108
×
×
×
×
(×)
Wimdu
9.16
.79
265
.76
.89
.45
.62
140
(×)
×
EasyCarClub
4.91
.95
122
.89
.50
.32
.18
175
×
×
×
Drivy
4.83
.91
76
.81
.66
.51
.26
137
×
×
×
Turo
4.86
.93
110
.87
.93
.56
.08
139
×
×
×
×
×
BlaBlaCar
4.81
.90
69
.87
.86
.81
.62
121
×
×
×
×
×
×
Wingly
4.99
.99
221
.81
.90
.46
.55
234
×
×
×
×
eBay
99.55
.99
213
.81
.16
.01
.02
136
×
×
Gumtree
4.21
.62
×
×
TaskRabbit
95.52
.89
92
.86
1.00
.92
1.00
166
(×)
×
×
Note: Sent. = sentiment ([0, 1], English texts only); text lengths in median values; SNS = social network sites;
#T = number of transactions; #R = number of reviews
TABLE 2: PLATFORMS’ AND USERS’ USE OF TRUST-BUILDING MECHANISMS
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
37
Results
Platform and provider perspective
Table 2 summarizes the platforms’ use of trust-building mechanisms and how these are being
utilized by their users. As can be seen, many of the identified artefacts such as rating scores,
text reviews, and profile images are present on most platforms, while others are less common
(e.g., social media linkage). Moreover, there are marked and platform-contingent differences
in how they are being used.
Rating Scores—All considered platforms use some sort of numerical rating score. The
mechanisms vary in terms of scales, aggregation, granularity, and display. While some
platforms display a five-star rating scale rounded to half stars (e.g., Airbnb, Turo), other
platforms offer a more fine-grained display (e.g., .10 stars; BlaBlaCar, Gumtree). We observe
rather skewed rating score distributions where a majority of users on most platforms exhibit
highly positive ratings. This is not unexpected since rating scale skewness represents a
common phenomenon across platforms and domains (Schoenmueller et al. 2018; Zervas et al.
2015). Figure 4 shows the cumulative (log-scaled) fraction of users with ratings smaller or
equal to a specific rating score (from min to max). Most platforms exhibit rather similar
patterns of rating score skewness, while there exist outliers in both directions. Table 2 reports
a quantification of the platforms’ rating score skewness (determined similar to the Gini index
with 0 skewness 1). For a recent study on the antecedents of such extreme distributions on
review sites and platforms, we refer to (Schoenmueller et al. 2018).
Profile Images—Faces create trust (Ert et al. 2016; Teubner et al. 2014). It is hence not
surprising that almost all platforms offer the possibility to upload a profile photo. Note,
however, that there occur marked differences in how this option is exerted by the users. For
instance, users may upload a profile photo both on eBay and Airbnb but, while virtually all
hosts on Airbnb have uploaded a photo (99.8%), only a minority has done so on eBay (15.8%);
a finding that is in line with our proposition. In addition, facial recognition analysis (using
Microsoft’s Cognitive Services; Microsoft 2018a) reveals that the fraction of users with a
“proper” profile photo, that is, a photo for which a face is detected by the software, ranges from
1% (eBay) to 92% (TaskRabbit).
(1) Transaction-based
(2) Expressive User Profiles
(3) Identity Verification
(4) Implicit Info
Platform type
Rating
Score
Text
Review
Profile
Image
Self-
Description
Email
Phone
ID
SNS
#T
Since
Imp.
Face
Accommodation
6.59
6.36
5.05
4.49
3.69
6.49
6.44
6.56
4.41
6.15
5.85
Carsharing
6.31
5.94
4.39
4.44
3.83
5.89
6.31
6.53
3.64
5.75
5.69
Mobility
6.22
5.41
6.16
6.43
2.89
5.32
6.14
6.70
3.03
5.59
5.54
E-commerce
6.15
5.85
2.98
2.40
2.18
5.53
4.93
5.18
3.43
5.85
5.03
Crowd work
6.11
5.89
4.63
4.86
4.69
5.86
5.94
6.00
3.89
5.57
5.23
Note: n=187; values on Likert scale [1,7]; Imp. = importance; SNS = social network sites; #T = number of transactions;
= importance > 6.0; = importance < 4.0
TABLE 3: IMPORTANCE OF TRUST-BUILDING MECHANISMS PER PLATFORM TYPE
x
x
Chapter 1: Trust-building and reputation
38
FIGURE 4: RATING DISTRIBUTION ACROSS PLATFORMS
Text—Moreover, the way people write about themselves and others differs between platforms
(e.g., in terms of text length and sentiment). While on eBay, only 2% of sellers provide a (rather
short) personal description about themselves, the saliency of textual self-descriptions is
significantly higher on platforms characterized by a higher degree of social interaction (e.g.,
Airbnb: 58%). Similarly, reviews vary considerably in length between platforms. Also, there
exist subtle platform-specific differences of how users write about each other. Text sentiment
analysis (Microsoft 2018b) shows that a vast majority of text reviews is highly positive (ranging
from 0 to 1).
Identity Verification—Mechanisms for identity verification are less widely used across
platforms. While Airbnb and BlaBlaCar display a wide range of identity verifications (email,
phone, id, social network sites), other platforms (e.g., eBay, Drivy) do not make any use of it.
User Activity—Another property for platform distinction roots in the concentration of user
activity as indicated by the number of completed (and reviewed) transactions. Figure 5 shows
the cumulative fractions of user activity against the user samples. As can be seen, concentration
varies markedly.
FIGURE 5: ACTIVITY CONCENTRATION ACROSS PLATFORMS
Rating Score
Fraction of users score (log scale)
.000
.001
.010
.100
1.000
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
easyCar
Fraction of users
Activity (cumulative)
.00
.10
.20
.30
.40
.50
.60
.70
.80
.90
1.00
.00 .10 .20 .30 .40 .50 .60 .70 .80 .90 1.00
easyCar
Rating Score
Fraction of users score (log scale)
.000
.001
.010
.100
1.000
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
easyCar
Fraction of users
Activity (cumulative)
.00
.10
.20
.30
.40
.50
.60
.70
.80
.90
1.00
.00 .10 .20 .30 .40 .50 .60 .70 .80 .90 1.00
easyCar
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
39
Consumer perspective
To better understand platform users’ preferences, we asked survey participants to rate how
important they personally deem the different trust cues across platform contexts (Table 3;
7-point Likert scales labelled as (1)=“not important at all” to (7)=“very important”). Several
observations are remarkable:
Rating scores—Within transaction-based trust cues, rating scores unanimously score highest
in importance when deciding to book/buy a service from a prospective provider (i.e.,
consistently rated above 6 out of 7). This appears natural given ratings represent the most
commonly deployed signal. Users rely on it more than on text reviews in each investigated
platform type though the latter also receives generally high ratings for importance.
Figure 6 depicts both actually observed rating scores from across the 11 platforms and the
category-specific trust thresholds. These thresholds are based on survey respondents’ stated
star rating score (on a scale of 1.0-5.0 stars) above which they would generally consider a
provider as trustworthy. All platform rating scores have been normalized to the 1–5 stars
interval accordingly. A striking observation is that the average thresholds are much lower than
the vast majority of actual ratings with only very few outliers falling short of the thresholds.
FIGURE 6: ACTUAL SCORES VS. TRUST THRESHOLDS
Profile images—We observe marked differences in the perceived importance of profile images.
The results seem to be mirroring our market data findings and corroborate our reasoning on
the degree of social interaction. Especially low ratings within the e-commerce category (2.98)
are well in line with actual usage of profile images on these platforms (only 16% and no option
to upload, respectively); if users do not find it important, as their interaction with transaction
partners is limited, then there is not much reason to upload one. However, there is a high
demand for profile images on mobility platforms and visible faces (each >6). As we have seen,
not all providers use profile images on the investigated platforms but may benefit from doing
so.
Self-description—Participants were explicitly asked to rate the availability of self-descriptive
elements such as hobbies, occupation, or personal background. Overall, these seem to play a
minor role for building trust. Only in the crowd work category did participants rated them as
Mobility Carsharing E-Commerce Crowd work Accommodation
Platform type
Rating score
easyCar
4.0
4.2
4.4
4.6
4.8
5.0
≙Trust threshold
Chapter 1: Trust-building and reputation
40
important. When prompted to comment on particularly helpful pieces of personal information
“skills” and “experience from previous jobs” occurred frequently. Other than that, users seem
to be somewhat indifferent towards extensive use of such trust cues.
Identity verification—We notice clear support for verified provider identity, especially on
platforms with the potential for higher degrees of social and in-person interaction such as
mobility, crowd work, and accommodation but also for carsharing. In e-commerce these
elements play a less important role. This makes sense given its rather transactional-focused
and impersonal nature. It is noteworthy, however, that the linkage of social network accounts
is not deemed as of high importance.
When directly asked for additional trust cues users would like to see, the provider’s age came
in on top with 11% of all participants mentioning this aspect explicitly (17% and 14% for
carsharing and mobility, respectively). Albeit demanding this piece of information can be
considered discriminatory, long years of experience and a certain maturity seem to play a role
for services that involve driving or renting out vehicles. Other elements that have received
mentions are legal background checks for providers on mobility (27%) and accommodation
platforms (5%). Detailed policies on this topic are opaque though at least some platforms seem
to be conducting background checks in some jurisdictions (e.g., Airbnb, Uber).
DoSI perspective
To evaluate our proposition, Figure 7 combines both platform and user perspective. It plots the
empirically observed frequency of trust cues from the Expressive User Profiles category (i.e.,
profile image availability (PIA), face visibility (FV), self-description availability (SDA); Table 2)
for the 11 platforms against the stated importance of these elements by survey participants per
platform type (Table 3; n=33). Drawing on our reasoning from Table 1, the color-coding
represents the platform’s respective degree of social interaction
7
. This value seems to explain
both frequency and importance of “soft” trust cues well. In addition, the fitted regression
indicates a positive relation between actual usage of these mechanism and perceived
importance (Figure 7).
FIGURE 7: FREQUENCY OF SOCIAL CUES VS. IMPORTANCE
7
Low: e-commerce; Medium: carsharing, accommodation; High: mobility, crowd work
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
41
First, usage frequency and importance are in fact correlated positively (Pearson r=.711,
p<.001). Moreover, two OLS regression analyses provide support for the conjectured relations
between a platform’s degree of social interaction (DoSI; coded numerically as 1 to 3) and
providers’ use of different trust cues on the platform (b=.362, p<.001; R²=.53; n=33) as well
as between platform domain’s DoSI and consumers’ assessment of the importance of trust cues
within that domain (b=1.129, p<.01; R²=.50; n=15).
We acknowledge the fact that nuances in the DoSI may occur depending on the type of
transaction within a particular platform (see Table 1 for an explanation). Hence, we varied the
DoSI value for accommodation platforms as a sensitivity check for our model. For instance,
alternating the DoSI score from “moderate” to “high” for accommodation sharing yielded
almost identical estimates for the regressions.
Discussion
The rise of the platform economy has led to the emergence of multi-sided markets “that shape
the terms on which participants interact with one another” (Kenney and Zysman 2016). Given
the paramount role of platforms’ design choices for many of today’s social and economic
interactions, this paper sets out to provide a practical overview and empirical insight into the
ways platforms build trust between their users and, in turn, how those measures are perceived
by platform participants.
As outlined, for many of today’s platforms dealing with experience products, a blurring occurs
between trust-building, marketing, and product/service descriptions. Beyond the role for
trust-building, personal information about providers can play a substantial role in the offer’s
value proposition as many people explicitly look for social value such as authenticity, good
conversations and company, or cultural experiences (Hawlitschek, Teubner, et al. 2018;
Tussyadiah 2016a). In this sense, providers on peer-to-peer platforms have themselves become
part of the product they are offering and will make use of the full breadth of the trust cue palette
to advertise their services. Our analysis yields some key take-aways along those lines:
Most of the studied platforms employ all four basic types of trust-building mechanisms as
introduced within this work (Transaction-based cues, Expressive User Profiles, Identity
Verification, Implicit Information). Despite several similarities, there exist marked differences
in how these mechanisms are being utilized and how their importance is perceived by the
platforms’ users. Specifically, user profiles are particular expressive on platforms that are
characterized by a high degree of social interaction. This is in line with the finding that “social
aspects and individual characteristics become more relevant in this particular context”
(Möhlmann 2016, p. 26). Contrary, expressiveness is much lower on platforms for renting and
selling, i.e. those platforms on which user interactions are limited. This justifies the labeling of
elements from the Expressive User Profiles category as soft trust cues—when deciding to book
or buy from a particular provider, users, in many cases, deem “hard” trust cues (especially
Transaction-based cues and Identity Verification but also Implicit Information) of higher
importance.
Now, when accepting the premise that there exists a link between a platform’s degree of social
interaction and its use of expressive trust cues, one may draw on any platform’s use of such
cues to infer its degree of social interaction, providing a convenient and efficient parameter to
compare platforms. It strikes the eye that especially e-commerce platforms such as Gumtree
Chapter 1: Trust-building and reputation
42
and eBay (and their users) do not make use of the full breadth of user profile design and
identity verification. While it appears plausible that the rather anonymous and impersonal
nature of these platforms’ transactions does in fact not necessitate the use of such means, it
seems natural to ask whether they may nevertheless benefit from doing so.
Limitations and future work
Our approach does not come without its caveats and limitations and we delegate it to future
work to explore the causes, effects, mechanisms, and limits of trust-building within the
platform landscape in greater detail. A first suggestion is to extend our assessment to analyze
the trust-building categories beyond Expressive User Profiles to uncover platform (type-)
specific drivers of their usage in-depth. Further, the focus of this study could be broadened to
encompass trust into consumers, that is, the opposite market side. Also, the rapid emergence
and vanishing of platform-based businesses will yield new players, trust requirements, and
mechanisms, all of which future research will have to take into account. One technical
limitation roots in the use of Microsoft’s cognitive services for text sentiment analysis and facial
recognition. The latter appears to be rather conservative in flagging faces, resulting in many
false negatives. Thus, the estimates for face visibility in Table 2 should be seen as lower
benchmarks. Future work may want to apply more sophisticated tools and methods to our data.
On average, 51% of survey participants had used one or more platform from their assigned
segment (i.e., accommodation, carsharing, mobility, e-commerce, crowd work). Yet, there
occurred considerable category-specific differences. While 93% of respondents had previously
used e-commerce platforms, only 14% stated so for carsharing. Hence, these results must be
interpreted with some caution as a considerable fraction of respondents in some categories
does not have actual insider experience. Future research may want to draw upon more
experienced subjects. Moreover, there exist other ways of trust formation beyond the
mechanisms considered here. Examples include pre-purchase communication such as typical
for Taobao (Huang et al. 2017) or the formation of social capital through repeated interactions
(X. Chen et al. 2017).
Eventually, in view of the multiplicity of platforms and the many parallel, unconnected
reputation silos (“isolated islands”; Mazzella et al. 2016), recent research has set out to
consider the transfer of reputation between platforms (Hesse and Teubner 2020a; Otto et al.
2018; Teubner et al. 2020). While platforms could thus build trust by allowing their users to
import reputation from other platforms, none of the considered platforms is currently offering
such a functionality. Neither are third-party solutions to reputation portability (as of yet) well
known (Teubner et al. 2019). In fact, when asked about additional desirable trust cues, only
one participant explicitly mentioned the use of an “identity verifying or reputation service”.
However, a lot of people would value imported reputation as, on average, the concept scored
5.7 on the 7-point importance scale. High-DoSI platforms (i.e., crowd work, mobility) exhibited
strongest desirability of the feature and text reviews imported from other platforms were
appreciated slightly more than imported star ratings (Hesse and Teubner 2020b).
Conclusion
Trust and reputation among users of two-sided platforms are imperative for flourishing
markets. The process of building and maintaining trust is hence of utmost importance to
platforms. As we have shown, given the specific nature of a platform type, the choice of
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
43
adequate means is contextual. For instance, nuances in the degree of social interactions on
platforms impact both the provision of trust cues by platforms as well as the perceived
importance in the eyes of their users. Our results with regard to the close relation between the
degree of social interaction and the importance of trust-building mechanisms may be
instructive for platform managers and related business models that involve the handling of
online trust and reputation: the more a platform’s value proposition hinges on social
interactions, the more important it is to provide expressive trust cues. By providing an
empirical basis and first cross-platform insights into the use of reputation systems and
trust-building mechanisms (both by platforms and users), we hope to contribute to making the
discussion more explicit, if not even more objective.
44
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
45
Chapter 2: Conceptualizing
reputation portability
Having introduced general dynamics of trust and reputation-building on digital platforms,
I now focus on developing a conceptual model for reputation portability. The present study
thereby analyzes both the emerging literature on the nascent topic and currently observed
practical approaches. It delineates relevant terms, actors, and mechanisms while also
considering external influencing factors, such as novel regulation and decentralized
technology. In an outlook, I present archetypical scenarios for reputation portability’s future
development and derive a forward-looking research agenda.
Maik Hesse, Timm Teubner
8
“A good reputation is more valuable than money.” – Publilius Syrus, (fl. 85 – 43 BC)
Introduction
With any activity on the Internet users leave a digital footprint, a data trail that lingers online,
which is why the management of digital identity has become crucial for individuals and
companies alike (World Economic Forum 2018). Establishing and curating online reputation
is becoming more important and inherent in day-to-day life – examples include social media
(e.g., Facebook, Twitter, Instagram), professional networks (e.g., LinkedIn, GitHub), e-
commerce (e.g., eBay, Amazon), and sharing platforms (e.g., Airbnb, Uber). All these
environments allow people to create online identities and build reputation through their
actions. This includes, collecting likes and followers on social media, receiving testimonials on
professional networks, or accumulating ratings on digital marketplaces. Importantly, for many
of today’s platforms, users also rely on their reputation (Dellarocas et al. 2009; Tadelis 2016),
especially for trust-sensitive applications. This is particularly true for the sharing economy (Ert
et al. 2016; Mauri et al. 2018; Qiu et al. 2018).
Reputational data emerges in several (platform-bound) contexts. Building and maintaining
this reputation as a signal of trustworthiness and enabler of transactions is associated with
effort and costs (Dunham 2011). Importantly, the majority of users is active on multiple
platforms (Teubner et al. 2019) and users have to manage each reputation separately (Dakhlia
et al. 2016). This raises the question whether and, if so, how reputation acquired in one context
may represent a valuable signal in another.
Until now, a plethora of research has focused on either a) the role of reputation within a given
(but enclosed) platform environment (Resnick et al. 2006; Teubner et al. 2017) or b) the
general idea of data portability between platforms (De Hert et al. 2018; Drechsler 2018; Engels
8
This study was published in Electronic Markets – The International Journal on Networked Business
with the title “Reputation Portability – Quo Vadis?” (Hesse and Teubner 2020a); see
https://doi.org/10.1007/s12525-019-00367-6.
Chapter 2: Conceptualizing reputation portability
46
2016). However, little scholarly attention has focused on the question of reputation cross-
platform portability. We hence set out to pursue the following research objectives:
• Develop an understanding and conceptual overview of the overarching dimensions of
reputation portability.
• Assess whether reputation acquired by users on one platform serves as a valuable
signal for trustworthiness when imported on another platform.
• Analyze the mechanisms and technologies supporting cross-platform reputation
portability and elucidate platform and third-party strategies for their implementation.
With this work, we seek to portray the status quo of the emerging theme of reputation
portability as one aspect of a broader dialogue on digital identity management. In doing so, we
contribute to the IS literature in two ways. First, we delineate the relevant terms, players, and
concepts, discuss common mechanisms, and present a conceptual overview to structure prior
research – thus touching upon the what, why, and how of reputation portability. To the best
of our knowledge, this study is the first to develop such a comprehensive model for this timely
topic. Second, based on the rigorous exploration of our model’s dimensions and summarizing
the nascent body of empirical findings, we derive and discuss gaps and directions for future
research. Throughout, current practical approaches are thoroughly integrated into the analysis
and implications pinpointed.
The remainder of this paper is organized as follows. Sections 2 to 5 outline the state-of-the-art
of reputation portability along the scientific literature while Section 6 takes a practical
perspective. Specifically, Section 2 introduces the basic terms and concepts and motivates the
idea of cross-platform reputation portability. In Section 3, we introduce our methodology and
develop a conceptual model for reputation portability, illustrating the most relevant actors,
mechanisms, data types, and external influencing factors. Section 4 considers the legal
perspective by examining the impact of recent regulatory advances in the Western world. In
Section 5, we then provide an overview on empirical findings on reputation portability. Section
6 considers existing and potential practical approaches and critically assesses blockchain
technology as an enabler of reputation portability, decentralized platforms, and digital identity
management. In this section, we complement the practical discussion by highlighting
corresponding scientific work. Both views converge in Section 7, which concludes by deriving
potential future scenarios and implications for academia and practice.
Trust and reputation in digital platform ecosystems
It is safe to say that electronic markets are pervasive in nature and penetrate all levels of
modern economic activity (Alt and Zimmermann 2014). Today’s digital commerce makes up
for one out of every ten dollars spent globally and this number is expected to double over the
next few years (Forbes 2018). Contributing to this development, the rise of digital
marketplaces has been heralded as “one of the greatest success stories of the internet” (Tadelis
2016, p. 322). Since its early days, digital platforms, such as eBay, Amazon, Taobao, and
others, provide users and businesses with the means to sell and buy goods and services online.
While this constitutes a classical market model, more recently the so called “sharing economy”
burgeoned – built upon collaborative peer-to-peer consumption, that is, the marketization of
spare time, goods, and assets (Puschmann and Alt 2016). These consumer-to-consumer
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
47
transactions are mediated through platforms such as Uber, Airbnb, and TaskRabbit, and have
quickly gained traction as a new form of economic activity attributed with the potential to
disrupt entire industries (Economist 2013; Sundararajan 2016). Indeed, the sharing economy
is experiencing unprecedented growth with the overall market in the EU estimated at €26.5bn
annual consumer spending (European Commission 2018a) and compound growth rates
exceeding 35% year-on-year (PwC 2015).
Success of digital platform businesses depends on transactions among users. In the sharing
economy, an important antecedent to realizing such transactions is the establishment and
maintenance of peer-to-peer trust (Tadelis 2016; Teubner and Hawlitschek 2016). After all,
exemplary business models entail renting out ones home to complete strangers and entrusting
ones physical safety to a driver when ridesharing. The importance of trusting prospective
transaction partners has made reputation systems, aggregating transaction-based feedback
into individual scores, a popular research object since their inception on eBay (Resnick et al.
2000) and other early e-commerce marketplaces. And while third-party assessments still
represent the most commonly used signal to engender trust in online markets (Mavlanova et
al. 2012; Teubner et al. 2019), today all major platforms – from Amazon to Zalando, from
Airbnb to Uber – employ a whole variety of trust-building mechanisms to support users in
managing and leveraging their reputation (Jøsang et al. 2007; Tadelis 2016; Teubner and
Dann 2018). We refer to Bolton, Greiner, and Ockenfels (2013), Dellarocas et al. (2009), and
Resnick et al. (2000) for more extensive discussions on the general mechanics of reputation
building.
Many reputational elements have been found to positively affect perceptions of
trustworthiness (Ter Huurne et al. 2017). Examples include, (star) ratings and number of
reviews on Airbnb (C. Chen and Chang 2018; Qiu et al. 2018), reputation scores on eBay
(Resnick et al. 2006; Resnick and Zeckhauser 2002), or text reviews from transaction partners
in peer-to-peer rental markets (Cheng et al. 2019; Zloteanu et al. 2018). Thus, many trust-
building and reputational cues function as reliability signals towards future transaction
partners on the respective platforms (Dellarocas et al. 2009).
9
In this sense, Internet users’
digital footprint has been described as “a cumulative record of how well we collaborate and if
we can be trusted.” (Botsman and Rogers 2010, p. 217). In the spirit of Syrus’ quote, a good
digital reputation might soon be more valuable than money, becoming our single most
valuable asset in the reputation economy (Fertik and Thompson 2015).
Therefore, it comes as no surprise that calls for reputation portability are frequent in the
literature. In their seminal work, Resnick et al. (2000) mention the lack of portability from
system to system as a difficulty. The authors highlight the case of Amazon which, back in the
1990s, initially allowed users to import their ratings from eBay but discontinued the services
when eBay claimed the ratings as proprietary content.
10
Similarly, Dellarocas et al. (2009)
make a strong case for the advantages of portable reputation, especially if users are active on
several platforms (i.e., “multi-homing”).
9
Besides transaction-based reputation mechanisms, the list of such artifacts includes profile images,
self-descriptions, identity verifications (incl. social media presence) and implicit information, such as
badges, membership duration, and number of transactions (Teubner and Dann 2018).
10
There is an ongoing debate among legal scholars about the ownership of transaction-based feedback
in the form of reviews; we touch upon this aspect in Section 4.
Chapter 2: Conceptualizing reputation portability
48
A few years later, Botsman (2012) describes a world in which our aggregated online reputation
becomes a constituent part of the reputation economy and raises questions of ownership and
portability. Namely, if a platform user has been working towards, say, an impeccable
reputation for hospitality as a host on Airbnb or achieved a sublime chauffeuring score with
Uber – shouldn’t they then be able to use that hard-earned reputation to get a head start for
offering services on another sharing platform?
In this light, scholars from the field of electronic markets have recently called for research on
how consumers can manage their online identity across platforms and connect existing
profiles on a variety of sharing services (Puschmann and Alt 2016). Building on this thought,
others propose a broader notion of “identity portability” that could encompass “ratings and
reviews used by platforms to form reputations” (Gans 2018, p. 15). It is remarkable, however,
that – despite two decades of theoretical arguments in favor of reputation portability – to this
day, little empirical evidence has been presented to substantiate its hypothesized value.
Reputation portability: a conceptual model
To identify the relevant dimensions of a conceptual model for reputation portability, we
conducted a structured literature search. We chose Google Scholar to perform this task for
multiple reasons. First, it represents the most comprehensive tool for academic and
bibliographic search (Gusenbauer 2019). Second, it is rapidly expanding compared to other
services (De Winter et al. 2014), includes full-text search, books and conference papers, and is
now essentially regarded a superset of other databases “with substantial extra coverage”
(Martín-Martín et al. 2018, p. 1). As a first step, we queried specific terms (and combinations
thereof) which have previously been used to describe what we refer to as “reputation
portability”
11
. We included results after 2010 (i.e., the advent of sharing platforms) and
screened the first 50 results’ abstracts, selecting articles for in-depth review. We continuously
extended the list of search queries based on new findings and included a forward and backward
search. Next, we developed a draft framework with the aim to capture the most relevant aspects
of reputation portability: actors, mechanisms, data types, and external influences. We further
broke these down into sub-categories and challenged our conceptual overview in discussions
with senior scholars from the fields of IS and economics. Additionally, we conducted
workshops (internally and at an external research institution) to validate our proposal. In these
moderated discussions, attended inter alia by senior scholars of different tenure, we presented
intermediate results of the literature review and our initial proposal for the model’s
dimensions. Based on the feedback obtained, we refined the conceptual model and investigated
its dimensions along the identified literature. Throughout, we added relevant research and
practical findings as we dug deeper into specific aspects. Figure 8 gives an overview on the
process:
11
See Appendix for a detailed list of considered search terms. In total, 42 terms (excl. combinations)
were derived. These include “reputation transfer”, “cross-platform signaling”, “review portability”,
“reputation aggregation”, “digital identity management”, “reputation economy”, and various
combinations of terms such as reputation, signal, rating, review, platform, aggregation, portability,
transfer. To specify the search in an additional step, we used combinations with platform names (e.g.,
eBay, Airbnb, and Uber).
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
49
FIGURE 8: STRUCTURED LITERATURE REVIEW METHODOLOGY
Figure 9 sketches out the conceptual model for reputation portability12. It illustrates the most
relevant dimensions, including digital platforms and users, feedback providers, data transport
(e.g., through linking, embedding, or importing), Personal Information Management Systems
(PIMS), as well as external influencing factors such as (novel) regulation and the emergence of
decentralized technology.
First, the PLATFORM USER is at the very center. Users can be active across platforms, acting
both as a consumer and/or provider of content, goods, and services.
Second, for exposition, a selection of DIGITAL PLATFORMS – acting as online intermediaries of
data, content, and services – is shown at the top of the figure. We distinguish between a) social
& professional networks (e.g., Facebook, Twitter, LinkedIn), b) e-commerce marketplaces
(e.g., eBay, Amazon, Shopify), and c) sharing & collaboration platforms (e.g., Airbnb, Uber,
TaskRabbit).
Third, FEEDBACK PROVIDERS are prior transaction partners of the platform user. They play a
pivotal role by providing either a written statement (i.e., text review) or a numerical (star)
rating. These form the basis of a platform’s aggregated (transaction-based) reputational data
on a user. There are certain implications regarding ownership and copyright of this received
reputation, which we will touch upon in our discussions of legal and technical aspects of
portable reputation (in Sections 4 and 6 respectively).
Fourth, the REGULATORY FRAMEWORK represents an important dimension as current
legislation and policymaking greatly influence both the platform landscape as well as
individual user rights. Political foundations have recently discussed the desirability of
aggregated and portable ratings to avoid losing one’s reputation when migrating between
platforms (F. A. Schmidt 2017). Moreover, several EU reports see the lack of cross-platform
portability as a critical issue, stating that monopolistic situations impeding data portability of
profiles ought to be monitored (European Commission 2016, 2017). In Section 4, we
investigate current policies in the Western world – with a particular focus on the EU given the
12 Note that within the context of research in the field of electronic markets, we focus predominantly on
the economic dimension of reputation portability (incl. its legal and technological triggers and
implications). Other aspects such as societal, ethical, and cultural dimensions are an important part of
the general topic, though, for the sake of focus and brevity, we chose to only briefly touch upon these
items where applicable.
Reputation portability:
▪What?
▪Why?
▪How?
RESEARCH
OBJECTIVES
LITERATURE
SEARCH &
SCREENING
DRAFT
FRAMEWORK
CHALLENGE &
FEEDBACK
CONCEPTUAL
MODEL
Research database:
▪Since 2010
▪First 50 abstracts per
search screened
▪Papers for in-depth
review selected
▪Forward and backward
search conducted1)
Important aspects of
reputation portability:
▪Actors
▪Mechanisms
▪Data types
▪External influences
1) Extending beyond 2010 where applicable
35+ terms &
combinations
1:1 discussions
& workshops
Refined proposal:
▪Senior IS and
economics scholars’
perspective included
▪Additional literature
identified
▪Sub-categories
finalized
Final proposal:
▪Most relevant
dimensions
▪In-depth description
and analyses
▪Inclusion of further
relevant literature
Chapter 2: Conceptualizing reputation portability
50
recent developments around data portability laws13 – and assess their applicability for and
support of reputation portability.
FIGURE 9: A CONCEPTUAL MODEL OF REPUTATION PORTABILITY
Fifth, the arrows represent the transport of (reputational) data with a focus on the LINK/EMBED
and IMPORT functionalities between digital platforms. Several types of data are associated with
users’ online and offline activity:
• Personal data is the basic information that pertains to the individual user, including
identification documents, citizenship, educational records, email address, and
phone/mobile number. This category can also be extended to include information such
as biometric data, health/patient data, or even credit and insurance records, which
become relevant in a broader digital identity management.
• On social and professional networks, users build and cultivate their own profiles. Their
activity on the respective platforms results in rich data sets, for instance, by gathering
connections (e.g., friends, colleagues, followers), acknowledging other users’
contributions via “likes”, or by providing “user-generated content” (UGC, e.g., text,
images, video/audio; Daugherty et al. 2013; Van Dijck 2009). We refer to the entirety
of the data generated in these networks as profile and activity data.
• As previously introduced, transaction-based data occurs typically on e-commerce
marketplaces and sharing platforms. The most common types are third-party ratings
(numerical) and reviews (textual) as well as the resulting agglomerated scores (Teubner
and Dann 2018).
13 Note that laws such as the EU’s General Data Protection Regulation (GDPR) or the California
Consumer Privacy Act (CCPA) apply to platforms universally, that is, irrespective of their home
jurisdiction, so long as they are active in the European Union (or in California, and hence the United
States (US), respectively).
(e.g., profiles, activity, connections)
Reputation Portability as part of Digital Identity Management
REGULATORY
FRAMEWORK
Social & professional networks E-commerce marketplaces
DIGITAL PLATFORMS1)
DECENTRALIZED
TECHNOLOGY
▪Passport / ID / citizenship
▪Biometric information
▪Email / phone
▪Educational records
▪Credit score / insurance
▪Health / patient data
▪…
Sharing & collaborative platforms
Personal data
Profile & activity data Transaction-based data
IMPORT
(e.g., reviews, ratings, scores)
LINK/EMBED
1) Exemplary –overlaps exist
2) Personal Information Management Systems Transport of (reputational) data; red arrows indicate service provision ( )
Copyright or trademark rights for logos is with the respective company
Sources: Wikimedia Commons, company website
PIMS2)
Reputation…
▪…validation
▪…certification
▪…aggregation
▪…dashboards
PLATFORM USER
FEEDBACK
PROVIDERS
FEEDBACK
FEEDBACK
FEEDBACK
…
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
51
Figure 9 shows the mechanisms of linking/embedding and importing data from one platform
to another. Linking/embedding refers to the common practice of integrating one’s profile
information from social or professional networks on another platform to function, inter alia,
as a signal of trustworthiness (Y. Chen et al. 2014; Silvestri et al. 2015). Other scholars have
investigated the effects of mutual friends on social networks as an external trust signal between
service consumer and provider in the context of ride sharing (Otto et al. 2018).
For the present work, we are particularly interested in the import of transaction-based
reputational data, such as star ratings and reviews. Even though Amazon had to discontinue
its import of eBay ratings early on, examples of such practices exist. For instance, the e-
commerce platform Bonanza.com allows its users to import feedback directly from eBay,
thereby translating their positive, neutral, or negative eBay ratings into Bonanza’s 5-star rating
scheme (Bonanza 2019a). Similarly, TrueGether.com allows buyers and sellers to import their
reputation from eBay, Amazon, and Shopify as well as to link/embed their Facebook and
LinkedIn profiles (TrueGether 2019). Notwithstanding these practical advances, research on
the effectiveness of cross-platform reputation is scarce. We provide an overview of the few
empirical findings on this matter in Section 5.
Sixth, PIMS (Personal Information Management Systems) aggregate and verify (reputational)
data from disparate sources (e.g., transaction-based data such as ratings/reviews from across
the web) and combine it with verifications from social and professional accounts (profile &
activity data) or even government-issued IDs (personal data) to build individual trust and
reputation profiles, widgets, and dashboards. Since there is typically no technical integration
across platforms and users have to manage separate reputations, such systems allow for a
reduction of transaction costs and opacity (Botsman 2012; Teubner et al. 2019). Specifically,
PIMS such as Deemly and Traity offer their services to peer-to-peer marketplaces, sharing
economy platforms, and individual users (Deemly 2019a; Traity 2019a). We consider practical
approaches to reputation portability and the role of PIMS in Section 6.1.
Last, DECENTRALIZED TECHNOLOGY promises to be of importance for reputation portability
and the broader concept of digital identity management. High expectations have recently
evolved around blockchain technology
14
as a tool that could empower individuals with more
control over their digital identities (CBInsights 2019). Unlike today’s siloed approach to
reputation management, distributed ledger technology is expected to enable alternative
concepts and business models for cross-platform reputation portability (Catalini 2017). We
discuss the technology’s implications in Section 6.2.
Data portability regulations: friend or foe?
Given the power of reputation, users’ benefits from portability between platforms become
obvious – be it to avoid lock-in effects (European Commission 2017), overcome the inherent
“cold-start” problem when first using a platform (Wessel et al. 2017), or to realize price
premiums (as providers) and increase chances of being able to book services (as consumers)
(Otto et al. 2018; Qiu et al. 2018; Resnick and Zeckhauser 2002). Consequently, platforms
could indirectly benefit from imported trustworthiness, increased sales numbers, and higher
prices and, in turn, fuel their own business model.
14
Or, more generally, distributed ledger technology (DLT) of which blockchains form a subset of
(Burkhardt et al. 2018). We use both terms interchangeably.
Chapter 2: Conceptualizing reputation portability
52
Thus, enabling (or even enforcing) reputation portability has emerged as an interesting option
for regulation (European Commission 2017). In this regard, many scholars have recently
referred to Article 20 of the European Union’s General Data Protection Regulation (GDPR;
European Union, 2016b). It aims to foster and enable unprecedented competition in online
markets by breaking the monopoly of incumbents and large platforms. The Article in fact
compels platform operators (referred to as “controllers”) to grant their users (referred to as
“data subjects”) the right to data portability. Specifically, a user “shall have the right to receive
the personal data concerning him or her, which he or she has provided” to the platform “and
have the right to transmit those data to another controller without hindrance” (European
Union 2016a, p. 45). In a similar vein, data privacy laws are being introduced around the globe.
Most notably, the California Consumer Privacy Act (CCPA) follows suit in incorporating EU
GDPR’s data portability rights and will become effective for entities doing business in the
largest US state economy on January 1, 2020 (Barrett 2019). In the following, we analyze
GPPR’s implications for reputation portability as a proxy and de facto standard for comparable
data regulations worldwide.
Legal scholars have mostly welcomed the European legislation as a laudable step (Kathuria
and Lai 2018) and strong tool that could enable increased competition between digital
platforms and give users more control over personal data (De Hert et al. 2018; Engels 2016).
Yet, since authorities have not provided clear instructions on the interpretation and
boundaries of “data portability” in the sense of GDPR (De Hert et al. 2018; Ursic 2018), it
remains unclear, which type of data can be ported from one platform to another. This leads
many authors to conclude that the wording and vague scope of GDPR’s Article 20 might
hamper cross-platform portability of data altogether (Drechsler 2018; Vanberg and Ünver
2017).
It is indeed quite unlikely that this provision applies to reputational information such as rating
scores and text reviews (i.e., transaction-based data) in any meaningful way. To start with, the
Article explicitly refers to information which data subjects have provided themselves (i.e.,
comments, posts, likes, photos, etc.; profile & activity data), effectively limiting the scope to
data users contributed directly to the platform (Gans 2018). As many authors have pointed out,
this data – provided by users themselves – is not the most interesting when it comes to
reputation portability (Drechsler 2018; Graef et al. 2013).
Instead, reputation data is usually provided by others, namely prior transaction partners (i.e.,
the “feedback providers” in our conceptual model). Based on these third-party assessments,
platforms calculate aggregated scores. The exact scope of data encompassed by the regulation
is debatable. But since ratings are provided by the data controllers’ reputation systems and not
the data subjects themselves, GDPR’s Article 20 will most likely not cover this type of
information (Graef et al. 2013; Vanberg and Ünver 2017).
Further, when looking at textual reviews provided for platform users by previous transaction
partners, neither existing intellectual property rights nor the GDPR seem to support
cross-platform reputation portability in a plain sense. In fact, since the data object’s rights
“shall not adversely affect the rights and freedoms of others” (European Union 2016a, p. 45),
Kathuria and Lai (2018) conclude that from the perspective of traditional IP law, “today’s legal
reality is […] that ownership will lie in either the authors or platforms” (p. 6). Platforms,
however, may demand a waiver of rights, so that the feedback becomes a public property and
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
53
can be displayed at the platform’s will. The authors go on to infer that review portability, if
achieved, aims at informing a public interest (i.e., facilitating information about one’s
reputation to other consumers). Yet, GDPR’s Article 20 is inapplicable when data processing
is necessary for this particular purpose (European Union 2016a, p. 45). Martinelli (2019, p.
139), however, states that interpreting the right to data portability less strictly and extending
it to “non-personal” or “observed” data would ensure that “comments of the buyers in the
reputational feedback system could be ported too”. Others claim that a user’s “reputation or
feedback score is related to the data subject, even though this data was not given by the
individual, and should therefore fall under the scope of data portability as personal data” (Van
der Auwermeulen 2017, p. 14). These arguments, when considered jointly, indicate that the
ownership and copyright situation for personal reputation on a platform does not obviously
favor the individual platform user. At best, the debate is still up to more concrete clarifications
and guidelines.
Despite this, let us now assume that reputational data (e.g., reviews) could be exported or
downloaded technically, or, as formulated specifically “in a structured, commonly used and
machine-readable format” (European Union 2016a, p. 45). Several reasons come to mind why
this data could lose its credibility and hence trust-building potential as soon as it leaves the
platform it originates from (e.g., in the form of a data dump file):
• It could be subject to post hoc manipulation by users.
• The reputation system of a target platform onto which the data is supposed to be
imported might not allow for an import.
• Even if users can upload text reviews or individual numerical ratings onto a target
platform, they might be only a weak signal if not available in aggregated form.
Given that most reputation systems’ effectiveness relies on the fact that they are hard to tamper
with, such data would either require some sort of encryption or digital signature or be
processed in a way that allows for full technical integration of aggregated reputational data
between platforms. Neither of these are warranted by EU’s regulation nor is it likely that the
major platforms are willing to implement such measures any time soon.
For one, “forced” interoperability (i.e., a requirement for controllers to configure systems so
that data portability between platforms is supported) did not become part of the GDPR,
although originally envisioned (Drechsler 2018). This could have required incumbents to offer
a direct data import/export functionality, which in turn, would significantly increase
competition in online markets given network effects and high switching costs for users once a
platform has acquired scale (De Hert et al. 2018; Engels 2016; Kathuria and Lai 2018;
Martinelli 2019). However, the proposed passage was taken out of the regulation’s draft after
“heavy criticism” from the private sector (Drechsler 2018). Instead, GPPR now states that data
shall be portable directly between platforms “where technically feasible” (European Union
2016a, p. 45). Additionally, Recital 68 GDPR adds that users’ rights “should not create an
obligation for the controller to adopt or maintain processing systems which are technically
compatible” (European Union 2016b, n. 7). Given the leading platforms’ large user bases and
their efforts to create lock-in effects, it is easy to imagine how the GDPR’s wording may lead
them to “contend that such a transfer is technically infeasible” (Vanberg and Ünver 2017, p. 4).
Chapter 2: Conceptualizing reputation portability
54
Hence, Krämer (2018) expects platforms (that invested in user lock-in) to design their
information systems in a way that renders data portability difficult. What is more, it could lead
to silent agreements between incumbents to not offer import functions for data from their
respective competitors. The author likens this situation to the well-known prisoner’s dilemma
in business and economics: even though each platform would profit from winning over users,
it eventually weakens its own position by increasing overall competition.
Despite calls for more collaboration among platforms to define norms and mandatory
interoperability standards (Drechsler 2018; Vanberg and Ünver 2017) it seems somewhat
unlikely that incumbents will deploy mechanisms aimed at enabling portability of reputational
data. The concept, however, is likely to yield economic value both for users and entrant
platforms. It remains to be seen whether and, if so, how legislative bodies around the globe will
do them justice.
Empirical findings on reputation portability
As outlined in previous sections, prior research underpins the economic value of on-site
reputation, that is, the reputation acquired and used within the confines of a given platform.
The results are, for instance, increased trustworthiness, price markups, and higher chances to
book or buy (Qiu et al. 2018; Resnick et al. 2006; Tadelis 2016). Given this value, there have
been attempts to develop theoretical models and reputation systems that allow for aggregation
and usage of reputation across online communities (Grinshpoun et al. 2009; Pingel and
Steinbrecher 2008) and subsequent evaluation of such models based on empirical data (Gal-
Oz et al. 2010). Furthermore, scholars conducted research to predict user trustworthiness and
performance based on prior reputation data from other contexts (Kokkodis and Ipeirotis 2016;
Venkatadri et al. 2016). In this context, it was demonstrated that social media data (e.g., from
Facebook or Twitter) can be leveraged to infer users’ trustworthiness on newer networks
(Venkatadri et al. 2016). Similarly, gig workers’ future performance at a given task is indicated
well by prior reputation earned at tasks from different categories (Kokkodis and Ipeirotis
2016).
However, although this research hints at the concept of reputation portability, neither of those
projects informs entirely about the economic value of transaction-based data (i.e., reviews and
ratings) when imported onto a different platform. Specifically, Venkatadri et al. (2016) focus
on profile & activity data and Kokkodis and Ipeirotis (2016)’s analysis remains within the
confines of a single platform. Yet there have been prominent calls for applied research to
“combine economic thinking and analyses with human-computer interaction” (Tadelis 2016,
p. 338) and “how consumers [might] connect different identities on different sharing platforms
towards a cross-platform identity management” (Puschmann and Alt 2016, p. 98). This
motivates our review of empirical findings on reputation portability. Only recently, though,
first work-in-progress has been made in empirically assessing the effectiveness of portable
reputation from a user’s psychological perspective (Otto et al. 2018; Zloteanu et al. 2018).
Otto et al. (2018) conduct a laboratory experiment where they present users on BlaBlaCar,
Europe’s largest ride sharing platform, with imported ratings from Airbnb as well as mutual
friends on Facebook. This allows them to make inferences about the trust-building potential of
these signals for users without prior rating on BlaBlaCar. They find that imported star ratings
function as a driver of people’s trust in prospective service providers, whereas mutual friends
alone have no significant influence.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
55
In three consecutive studies, Zloteanu et al. (2018) explore the impact of trust and reputation
information (labeled as “TRI”) on users’ perceptions and intentions to book on a hypothetical
accommodation platform. Unanimously, the studies uncover a positive impact of TRI on users’
perception of a host’s trustworthiness and increased tendencies to book. As one type of TRI the
authors consider “online market reputation,” that is, ratings imported from other platforms
(here: Airbnb and Uber). Based on the data presented, it is somewhat difficult to assess which
type of TRI users value most, though it seems that online market reputation falls somewhere
in the middle.
15
Teubner et al. (2019) underline the potential for “reputation transfer” presenting survey data
on multi-platform usage as an indication for its effectiveness. They find that most participants
(65%) use more than one platform. However, they also illustrate that practical solutions to
reputation portability (PIMS such as Deemly and Traity) are hardly known (1.4%). The authors
go on to discuss future research opportunities, including aspects of platform strategy, legal
considerations, boundary conditions of reputation transfer, and user interface design.
Table 4 summarizes the nascent body of empirical findings on reputation portability where
transaction-based data is considered. We focus explicitly on this data type, even though some
of the included papers also consider on-site reputation (Zloteanu et al. 2018) or imported
profile & activity data (Otto et al. 2018). All studies find supporting evidence for either positive
effects or the potential of imported reputational data.
15
The other TRI-elements used were: host verification, guest reviews, host reviews, social media
presence, number of reviews, and star rating.
Authors
(Year)
Platform Scope
Signal
Signal
Range
Method
Sample
Theory
Terminology
Cross-
platform
findings
Target
Source
Kokkodis
and
Ipeirotis
(2016)
oDesk
Star
rating
[0, 1]
Empirical
>1
million
trans-
actions
—
Reputation
transferability
Increases
accuracy of
performance
predictions
Otto et al.
(2018)
BlaBlaCar
Airbnb
Star
rating
4.0
Survey
141
Signaling
theory
Incorporation of
external trust
signals
Star ratings
from other
platform
increase trust in
providers
Zloteanu
et al.
(2018)
Fictional
accommod.
platform
eBay,
Uber
Star
rating
{4.0,
4.5,
5.0};
{90, …,
100%}
Surveys
120; 117;
189
—
Online market
reputation (≙
cross-platform
reputation)
Online market
reputation not
within top-three
selected items
Teubner
et al.
(2019)
28 peer-to-peer
platforms
—
—
Concept;
Survey
494
Signaling
theory
(Cross-platform)
reputation
transfer
Evidence for
multi-platform
usage but low
user familiarity
with PIMS
TABLE 4: EMPIRICAL FINDINGS ON REPUTATION PORTABILITY
Chapter 2: Conceptualizing reputation portability
56
Practical approaches to portable reputation
As outlined above, there are many good reasons to employ mechanisms geared towards
reputation portability. Yet most large platforms are thus far reluctant to offer any import
functionality, presumably to deliberately create lock-in effects or because they assess its
benefits as too low or dread the organizational effort. In this section, we give an overview on
past, current, and potential future approaches to reputation portability in practice
16
.
The past and the present
When Amazon allowed its users to import their ratings from eBay in the 1990s, it appeared as
though this might become common practice between online marketplaces. However, legal
claims around the proprietorship of reputational data quickly put a stop to the service
(Dellarocas et al. 2009; Resnick et al. 2000). It is still possible, though, to “extract” ratings
from several platforms and despite the reluctance of incumbents to offer reciprocal reputation
import, newer e-commerce market players do advertise such services to their customers. For
instance, Bonanza and TrueGether, two popular eBay alternatives, both allow users to directly
import their existing eBay and/or Amazon ratings onto the platform (Bonanza 2019a;
TrueGether 2019). While Bonanza supports the import of users’ eBay ratings (i.e., positive/
neutral/negative) and translates them into their own 5-star rating scheme, TrueGether does
the same but displays next to it both the corresponding text comment and the feedback
provider’s username on eBay. This enables sellers who are new to these platforms to
circumvent the “cold-start” problem by importing countless ratings and reviews. Such a
practice promises to be advantageous for both seller and platform (Wessel et al. 2017).
Personal Information Management Systems
Such examples are not limited to e-commerce platforms. A range of third-party applications
and services have specialized on accessing ratings on users’ behalf. They retrieve data from
several sources
17
to validate, certify, and aggregate it – either via integration into reputation
dashboards or by calculating proprietary reputation scores. These services form part of a
broader category, referred to as Personal Information Management Systems (PIMS). PIMS
provide trust-based data management tools and technical infrastructures to put consumers in
control over their own data and digital identity (European Data Protection Supervisor 2016).
First, users allow PIMS to access their data (e.g., personal data, profile & activity data,
transaction-based data). Then they set preferences and grant consent on the extent to which
the data can be used and shared with online service providers, businesses, and authorities.
Thus, PIMS can be understood as intermediaries connecting “individuals offering their data
for (re)use […] and organisations wishing to (re)use this data” (European Data Protection
16
Examples were selected in a structured way: as they occurred in the academic literature that was found
in the search process (see Section 3) or based on a list of previous startups working on solutions to gather
and aggregate reputational data from across online services (Botsman 2012). It is quite remarkable that,
at the time of writing, none of the companies is still in existence. However, the list served as a starting
point for our search of current players. An additional search of non-scientific work around reputation
portability was conducted, for instance, on company websites or – in the case of blockchain technology
– in (technical) whitepapers.
17
For example, Airbnb offers such an “export functionality” in that it enables users to copy an
HTML-code of reviews they have received and embed this badge on personal websites or blogs:
https://www.airbnb.com/users/badges (only accessible when logged in).
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
57
Supervisor 2016, p. 11). The services are typically free to users while online service providers,
platforms, or other third parties usually pay fees to access the PIMS’ technology and interfaces.
Current examples of providers with a particular focus on reputation portability are Deemly.co
and Traity.com. Traity, for instance, has been described as “general-purpose portable
reputation platform” (Sundararajan 2016, p. 65). The company generates a holistic profile of
users based on ratings from sharing platforms, several social network accounts, official identity
documents, and recommendations from friends (Traity 2019a). Similarly, Deemly connects a
range of sources to create trust profiles for users, thus, allowing them to make use of their
digital reputation and trustworthiness across peer-to-peer marketplaces (Deemly 2019a). Both
services ask individuals to provide their credentials (i.e., username and password) for the
sources they wish to connect. Albeit a one-time procedure to pull relevant ratings and reviews,
it still requires a leap-of-faith from users and may inhibit these services’ more wide-spread
adoption (Greiner et al. 2018). Indeed, historically, these third-party reputation aggregators
have not fared well and there is a long list of failed startups and inactive services (e.g.,
famust.com, erated.co, trustcloud.com, peertru.st, tru.ly). Research also shows that companies
such as Deemly and Traity are still largely unknown to sharing economy users with recall rates
between 1 and 2 percent (Teubner et al. 2019). Deemly’s trust score can currently only be used
for a few smaller Danish and Dutch sharing platforms. And Traity, seven years since its
founding in 2012, is focused on providing alternatives for traditional credit scores (e.g., with
banks and insurance companies). Yet, considering GDPR and the recent rise of PIMS, both
companies might be in a pole position to help with implementing reputation portability for
platform users. At least, both startups claim to be fully compliant and providing the golden
standard of GDPR (Deemly 2019b; Traity 2019b).
It remains to be seen if third-party providers can provide the technical means and tools to
practically establish reputation portability and if users are willing to adopt these services.
However, if platforms themselves fail to come up with satisfying solutions, PIMS may have a
competitive edge and seem to be the regulator’s favored option. The European Data Protection
Supervisor (EDPS; an EU body) suggests in a recent report (2016) that pure data aggregators
are best suited to implement by design the right to data portability – a view that is supported
by legal scholars (Kathuria and Lai 2018). The EDPS (2016) argues that for PIMS to effectively
penetrate a market dominated by large platforms, the EU will have to create incentives for
platform operators. These should support platforms in establishing interfaces with data
protection compliant PIMS for direct export of personal and reputational data. PIMS in turn,
should take an active role in addressing crucial interoperability requirements and
standardization.
Presumably, PIMS are among the most promising efforts to facilitate a broader digital identity
and reputation management for individuals. Moreover, the practice of aggregating a variety of
(digital) reputation elements into trust profiles, dashboards, or reputation passports could take
traditional (credit) scoring to a new level (Heiland 2018). It might certainly pave the way to
more holistic solutions for digital identity management, as has been hypothesized for some
time (Botsman 2012; Botsman and Rogers 2010). Especially the transfer of (aggregated)
reputation from online to offline context bears potential. For instance, as a reference on a CV
Chapter 2: Conceptualizing reputation portability
58
for job applications
18
or to allow access to innovative financing solutions (Traity 2019b;
Trustbond 2019). Some platforms also offer such services directly: Uber in East Africa has
partnered with local banks to grant high-reputation drivers access to vehicle financing (Stanbic
Bank 2017). To qualify as eligible, drivers must have completed over 500 rides at an average
rating of above 4.6 out of 5 stars. A few hundred drivers (more than 8% of Uber’s active fleet
in Kenya) have already participated in the deal and receive interest rates that are substantially
lower than those normally charged for car loans by commercial banks (Business Daily Africa
2018)
19
.
State-Controlled Reputation
One might ask how far practices of aggregated and portable reputation may reach when
state-controlled trust scores come into play. China is currently implementing a national
reputation scheme labeled “Social Credit System” that penalizes and rewards individuals and
businesses based on their activity across a range of dimensions (Hatton 2015; Kobie 2019). The
system explicitly includes online behavior such as electronic shopping and social media
conduct. By 2020, it will assess every citizen nation-wide with a unique, standardized, and
publicly ranked score. Interestingly, the Chinese government is currently testing several
systems in pilot projects launched across the country’s biggest digital platform ecosystems
20
,
(Botsman 2017a) which also provide parts of the technology stack and software. Western
observers and governments by and large criticize such plans for big data-powered scoring
algorithm as dystopian, Orwellian, or chilling (Campbell 2019; Human Rights Watch 2017).
However, many also opine that with an abundancy of digital platforms, predictive algorithms,
and scoring mechanisms, the Western world is closer to the Chinese system than one might be
willing to admit (Botsman 2017a; Forbes 2019a). Specifically, Wong and Dobson (2019) invoke
China’s plans as a looming vision of what to expect for democratic countries should strict data
policies for the digitization of identity and reputation not be applied. Hence, scholars suggest
closely studying the Chinese system, assessing its compatibility with more liberal forms of
Western society, and transferring learnings to inform discussions about aggregated reputation
and law-making (Síthigh and Siems 2019).
The future: decentralized technology on the rise
While state-controlled surveillance of citizen’s reputation and trustworthiness might (still)
appear as an extreme scenario, it is true that a substantial share of what is considered our
online reputation emerges within the boundaries of centralized institutions, that is, digital
platforms. As a possible alternative to such central and state-like platforms, blockchain
technology promises to empower users with more control over their data – irrespective of
intermediaries (Drescher 2017; Narayanan et al. 2016). Blockchain first gained prominence as
the underlying technology of the cryptocurrency bitcoin, a cryptographically secured, public,
18
As happened with reputation earned on Stack Overflow, a question and answer page for programmers;
eventually headhunters screened the platform for high-scoring developers with specific skill sets
(Botsman and Rogers 2010).
19
We had the opportunity to talk to Uber’s CEO for East Africa, Alfred Msemo; the company plans to
extend these services throughout the region and looks into more innovative ways to make use of drivers’
reputation scores.
20
Incl. Alibaba Group, WeChat owner Tencent, and the leading Chinese ride-sharing and online-dating
services.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
59
append-only, and distributed database (or “ledger”), which facilitates tamper-proof
transactions between parties in a peer-to-peer network without the need of a central authority
(Nakamoto 2008; Risius and Spohrer 2017). Werbach (2018, p. 55) introduces the technology
as the new architecture of trust and likens its possibilities to Google and other platforms as
being capable “of similarly powerful data aggregation, but without the downside of centralized
control”. Decentralization, technically achieved by a distributed consensus mechanism across
a peer-to-peer network, is unanimously seen as principal characteristic and core value
proposition of a blockchain (Beck et al. 2017; Glaser 2017; Plansky et al. 2016; Seebacher and
Schüritz 2017; Werbach 2018). The distributed ledger technology, thus, is predestined to
fundamentally impact three closely interrelated concepts that serve as crucial enablers of
reputation portability: reputation systems, decentralized platforms, and digital identity
management.
Blockchain-based reputation systems
Previously, we discussed the role and importance of reputation systems in online transactions.
Blockchain technology could be instrumental in developing new designs for these systems, for
example, supporting digital platforms with the effective detection of fraudulent ratings (Cai
and Zhu 2016). While this helps to improve existing “centralized” reputation systems, Carboni
(2015) shows how a rating and feedback mechanism, like that of eBay, can be mimicked in a
decentralized and distributed way on the Bitcoin blockchain. Similarly, attempts at developing
blockchain-based reputation systems that are applicable in traditional e-commerce settings
(e.g., eBay) or peer-to-peer service platforms have been made (Dennis and Owenson 2016;
Schaub et al. 2016). These approaches have in common that reputation would be calculated
and stored in a decentralized manner, thus, diverging from the principle of central ownership,
which could support the factual establishment of portable reputation (Catalini 2017). DREP
(“Decentralized Reputation”), a cross-national decentralized reputation ecosystem, takes a
somewhat different approach. Specifically, they propose a reputation protocol that aims to
quantify and aggregate reputation across different centralized platforms (e.g., traditional e-
commerce, gaming, content, social networking) (DREP Foundation 2019). DREP plans to
connect reputation values from both traditional platforms and decentralized blockchains, but
in that approach also depends on the willingness of these players to participate in its solution.
Hence, the concept of a decentralized sharing economy platform may be more promising to
start with.
Decentralized platforms
There is broad scholarly consensus that blockchain technology bears the potential to disrupt
existing platform business models (Catalini and Gans 2016; De Filippi 2017; Glaser 2017;
Puschmann and Alt 2016; Werbach 2018). Especially multi-sided market platforms such as
Airbnb and Uber could be fundamentally challenged by decentralized blockchain technology,
poised to replace these providers altogether (Glaser 2017). Some suggest that most of a
platform’s mechanism design features (e.g., reputation systems, payment infrastructure,
resolution processes) “can be implemented in a distributed, verifiable way on top of a
blockchain” (Catalini and Gans 2016, p. 19).
Today, the most common technology and native blockchain for decentralized platforms to run
on is Ethereum. It was proposed by Buterin (2014) and allows the execution of so-called “smart
contracts” that function as software programs running on top of the distributed ledger. This
Chapter 2: Conceptualizing reputation portability
60
enables developers to script applications using Ethereum’s proprietary programming
language. These scripts can be executed using the built-in virtual machine, effectively turning
Ethereum into a distributed computing platform for all kinds of decentralized applications –
commonly referred to as DApps (Glaser 2017; Werbach 2018).
Based on the Bitcoin blockchain and Ethereum’s smart contract functionality, Buechler,
Eerabathini, Hockenbrocht, and Wan (2015) were among the first to design a decentralized
system to record transactions and calculate proprietary reputation scores for counterparties.
Following up, Bogner, Chanson, and Meeuw (2016) present a DApp for object sharing and Xu
et al. (2017) propose a privacy-preserving, blockchain-based sharing economy platform.
However, most of these proofs-of-concepts and prototypes, have remained rather theoretical,
lacking, for example, a visual user interface.
A range of Ethereum-based platforms are already in the prototyping phase. Swarm City, for
instance, identifies itself as a blockchain-based marketplace with a built-in reputation system
(Swarm City 2019). Users can create specific services (e.g., ridesharing or accommodation) and
build communities in which contextual reputation is earned. At doing so, Swarm City has been
described as “Ethereum-based blockchain infrastructure to empower sharing economy
applications” (Beck et al. 2018, p. 1023). Similarly, Arcade City, which the Swarm City team
forked off in 2015, started as blockchain-based ridesharing services, but recently pivoted to
“decentraliz[ing] and disintermediat[ing] the entire corporate sharing economy” (Arcade City
2019). Origin is another example of a DApp enabling the creation of peer-to-peer marketplaces
on Ethereum (Origin 2019). Notably, Hub introduces a Human Trust Protocol where one can
gain verifiable and portable reputation from interactions and transactions on a “professional
achievement network
21
” based on blockchain technology (Ly et al. 2019).
Common to all these models are the direct communication, interaction, and value exchange
between users – facilitated through smart contracts without intermediaries. For one, this
promises to be more cost-effective by eliminating provider fees
22
. What is more, these
ecosystems are built and maintained by decentralized peer-to-peer networks and their
participants retain full control over the data involved (Werbach 2018). This includes
transaction-based data as introduced in our conceptual model (Figure 9). Here, reputation is
publicly verifiable, fully resides with the individual user, and can be built across communities
in the decentralized environment where intermediaries’ (i.e., platforms’) claims over
ownership of reputational data are non-existent. Catalini and Gans (2016) envisioned this as
an open reputation platform with the “ability to port and use the resulting reputation scores
across different services and contexts” (p. 22). Since blockchains could play a substantial role
in achieving more interoperability in electronic markets (Glaser 2017), this reputation can
certainly be imagined to travel seamlessly across the boundaries of blockchain-based
platforms.
However, intermediaries today still serve a valuable role in matching demand and supply and
inserting trust in transactions between strangers. Near-term full disintermediation might thus
be inefficient and the sharing economy is unlikely to run entirely trust-free, that is, free of
“centralized trust” (Catalini and Gans 2016; Hawlitschek, Notheisen, et al. 2018). Moreover,
21
Note that the project’s CEO is Eric Ly, co-founder and former CTO at LinkedIn.
22
However, there might still be fees involved for validation of transactions on the network.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
61
the sociotechnical challenges of trusting and governing decentralized systems have been
outlined repeatedly (Beck et al. 2018; Glaser et al. 2019). Thus, given the early development
stage and limited user base of DApps so far, much experimentation, prototyping, and real-
world evidence will be required before reputation portability on decentralized platforms may
become a wide-spread reality. Until then, several (already observable and promising) attempts
at establishing portable reputation should be explored and assessed with regards to their
potential to co-exist and mutually benefit from blockchain-based approaches.
Digital identity management
For decentralized platforms as those discussed above, we explicitly contemplate
transaction-based data for reputation portability. Yet, a full digital identity management on
blockchain has the potential to encompass all three types of data we introduced in our
conceptual model in Figure 9. This includes decentral storage, user ownership and full control,
as well as proof-upon-consent for all types of personal data (e.g., citizenship, biometric
information, educational records, insurance and health/patient information, memberships)
(Garcia 2018; Sharples and Domingue 2016; Zyskind et al. 2015). Note that, for instance, one’s
personal health data stored on a blockchain could be leveraged to negotiate individually
tailored insurance contracts
23
. Similarly, blockchain technology could be used to share
validated educational records with potential employers or other entities. Likewise, profile &
activity data can be generated on distributed social networks where contributors gather
reputation and receive rewards
24
for upvoted content (De Filippi 2017). Current popular
examples include Steem.com, Minds.com, and HyperSpace.app (HyperSpace 2019; Minds
2019; Steem 2019).
Of course, digital identity is a much broader topic than online reputation. However, similar
principles apply. According to the ID2020 alliance
25
, the four P’s are core requirements for
digital identity to empower and protect individuals: private (i.e., control over shared data),
portable (i.e., accessible anywhere and through multiple methods), persistent (i.e., from life to
death), personal (i.e., unique) (ID2020 2019). Corporations, public organizations, and
governments are developing use cases for decentralized blockchain-based approaches to
verified digital identity that are owned and controlled entirely by individuals. For
comprehensive overviews of state-of-the-art solutions, we refer to Lim et al. (2018) as well as
Dunphy and Petitcolas (2018).
Building on Section 6.1., these solutions qualify as blockchain-based PIMS. They utilize the
decentralized technology infrastructure and cryptographically signed user permissions to
23
There are, however, significant legal challenges associated with storage and sharing of patient data
that we chose to not further dwell on here.
24
Usually paid out in the respective blockchain’s underlying cryptocurrency.
25
A consortium bringing together governments as well as public and private organizations such as
Accenture, Microsoft, the open source blockchain consortium Hyperledger (with Linux, IBM, SAP), and
the United Nations. Besides such joint approaches to establish decentral digital identity, several big
technology companies are eyeing opportunities to build decentral solutions for the management of
individual’s social, professional, and financial data. Among those are IBM’s Verify Credentials,
Microsoft’s planned Identity Overlay Network (ION) for decentralized identity (DID) built on top of the
Bitcoin blockchain, and Hyperledger’s shared infrastructure toolkit Aries for DLT-based digital identity
management.
Chapter 2: Conceptualizing reputation portability
62
share data items, previously provided by individuals, with third parties. This allows users to
retain full ownership and control of their data and is commonly referred to as “self-sovereign
identity” (Lim et al. 2018). Already we see PIMS following a similar path. Traity recently
introduced the blockchain-based “reputation network” as a decentralized ecosystem allowing
people to leverage their reputational data from a variety of sources for access to risk-based
services (Traity 2019c). PIMS seem to be jumping on the blockchain bandwagon, as offerings
are converging.
One challenge to overcome in the design of distributed systems for portable and aggregated
reputation may be blockchain’s transparency and immutability (Dunphy and Petitcolas 2018;
Werbach 2018). The GDPR framework grants users a right to be forgotten and obliges
controllers to erase user data upon request. Yet, data, once written on the ledger, is persistent,
irreversible, and often publicly available. Hence, providers must be careful to design
mechanisms which avoid scenarios of networked shame (Botsman 2017b). More importantly,
though, the irreversibility of blockchains and users’ right for deletion and correction of their
data seem fundamentally incompatible (FinTech Council 2019). Either GDPR needs to be
updated to also include distributed ledgers or the latter must be designed in a way that ensures
certain data may still be deleted after being written on the ledger. This, however, is in stark
contrast to the original paradigm of a public blockchain. Either way, the right to have one’s
data erased should not be falsely interpreted as an opportunity for cherry picking only the best
ratings to be displayed. As Dellarocas et al. (2009) suggested a decade ago – certainly
unbeknownst of distributed ledger technology – systems may be designed in a way that leaves
users with a discretional choice to disclose previous ratings. That is, either import all available
reviews for full transparency or chose to have none shown at all. For GDPR and reputation on
a blockchain there is no intuitive solution. In this regard, a first conceptual design and
high-level architecture for a GDPR-compliant blockchain-based personal data and identity
management system (BPDIMS) has been presented (Faber et al. 2019). Yet, further analysis
on how to map requirements of data protection law into decentral systems is imperative.
As with any business model, the success of new approaches to portable reputation and digital
identity will ultimately hinge on user experience and usability (Weinhardt 2017) – irrespective
of the technological basis. Time will tell whether PIMS and/or DLT can live up to the high
expectations of becoming viable alternatives to existing “centralized” reputation mechanisms.
And, in turn, whether they can solve the issue of reputation portability for individuals where
today’s platforms refrain from presenting satisfying solutions themselves.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
63
Outlook and conclusion
In this final section, we conclude by laying out archetypical scenarios for the future
development of portable reputation26 and highlighting implications for research and practice.
Scenarios for reputation portability
Based on our review and learnings, we briefly discuss potential future scenarios for reputation
portability. Figure 10 gives an overview on the four scenarios:
FIGURE 10: FUTURE SCENARIOS FOR REPUTATION PORTABILITY
Scenario 1: Platform silos—In this scenario, similar to today’s status quo, regulation around
portability of reputational data remains fuzzy (Martinelli 2019). Hence, major digital platforms
refrained from offering reputation import and/or export. Similarly, PIMS for reputation
aggregation and individual trust scores remain a niche offering (Teubner et al. 2019). At the
same time blockchain-based solutions were unable to gain traction in light of usability and data
privacy issues. Isolated approaches to portable reputation, however, are observed (Bonanza
2019a). In consequence, seamless cross-platform reputation portability remains a hypothetical
idea – as it has been for over two decades.
Scenario 2: Enforced portability—In this scenario, reputation portability is enforced by
regulation. Precisely, platforms are obliged to offer a) APIs for accessing reputational data from
outside with ease and b) import functions that allow users to incorporate reputation earned
elsewhere without hindrance. Especially between platforms from the same or similar
categories (i.e., multi-homing) users make ample use of this option. On a much smaller scale
and with some legal uncertainty, Bonanza and TrueGether have pioneered this practice by
allowing the import of eBay and Amazon ratings early on (Bonanza 2019a; TrueGether 2019).
Scenario 3: Universal PIMS—In this scenario, one or several PIMS – compliant with GDPR
and other data regulations – have emerged and serve as universal data custodians. Such
26 We thank the reviewer for the valuable idea of developing these scenarios.
Scenario 1: Platform silos Scenario 2: Enforced portability
Scenario 3: Universal PIMS Scenario 4: Distributed reputation
Chapter 2: Conceptualizing reputation portability
64
services are backed by platforms working with their preferred third-party service (i.e., PIMS)
and hence propelled towards consumer adoption (European Data Protection Supervisor 2016).
Eventually, a leading platform itself endorsed the development of a single intuitive and
wide-spread solution (think of Google’s or Facebook’s sign-in services). However, in the light
of discussions about monopolistic online platforms and increased competition in electronic
markets (European Commission 2017), this concentration of (reputational) data with a single
or only few providers faces harsh criticism by data protectionists who demand to break up large
data companies rather than to let them acquire additional market power (Forbes 2019b).
Scenario 4: Distributed reputation—In this scenario, decentral systems have replaced
today’s platform providers and are well-established as a de facto standard for facilitating
trusted online interactions (Glaser 2017). Ratings and reviews are not bound to a platform’s
reputation system but become transparent and verifiable in an entire ecosystem – specifically
across platform borders (Catalini and Gans 2016; Werbach 2018). However, crucial issues of
blockchain interoperability had to be addressed and solutions to “bridge” information across
several chains/ledgers (i.e., to create an “internet of blockchains”) were designed (EU
Blockchain Observatory & Forum 2019; Johnson et al. 2019).
Note that all depicted scenarios are to be understood as possibilities, deliberately leaning
towards the ends of the spectrum of conceivable future developments. Most likely, a
combination of all outlined scenarios will emerge, rather than a pure form of a single scenario.
In the future, today’s platforms, facilitators of reputation portability, and distributed
approaches may well co-exist and benefit from each other.
Conclusion
We conclude our conceptualization and review of reputation portability with a distillation of
challenges, opportunities, and associated implications for research and practice. Reputation
portability is found to be effective according to most of the few empirical studies (Section 5).
Naturally, this rather thin body of research can and should be extended in several ways:
• Research should assess cross-platform reputation’s tangible economic value – both for
users and platforms. Who profits from portable reputation in which way? When is it
beneficial to import reputation given different boundary conditions and signal types
and qualities? Answers to these questions would also extend signaling theory. To do so,
empirical and experimental approaches are needed. Existing research is limited in that
it is mostly based on (self-reported) survey data.
• More research on adequate import/aggregation mechanisms for reputational data is
required. Looking at Scenario 2, how should import mechanisms be designed? How
can reputation systems be (re-)designed to enable portable reputation? Specifically,
most existing studies are limited to one platform or source/target combination.
Multiple scenarios, platforms, and source/target fit should be considered.
As with any research, there are limitations to this paper. Intended to give a broad overview on
the topic, we acknowledge that our work relies on publicly available information in its
reasoning. Future research should – besides building on empirical analysis – use qualitative
methods to generate inside views. For instance, drawing on interviews with platform operators
and/or PIMS may help clarify why major platforms are currently not offering an import
function or other adequate means for reputation portability.
PART I: THE FOUNDATIONS OF REPUTATION PORTABILITY
65
Distributed ledger technology is here to stay (Scenario 4). Albeit in the conceptual and
prototyping phase, it is imperative for practitioners to explore and experiment with blockchain
technology’s potential – regardless of it being beneficial or disruptive to their own business
models. There are many open questions as to which role blockchains will play for portable
reputation and digital identity:
• DLT is said to promise giving individuals more control over their (reputational)
data. But are users willing to put their trust and data into technology rather than a
centralized actor? Can decentralized systems be designed to enable similar
usability, experiences, and (security) features as today’s platforms? What about
their business model?
• Using blockchain technology, value exchange, trust mechanisms, and organizations
may become decentralized. What are adequate governance frameworks for these
types of business models? On decentralized platforms with self-executing smart
contracts to facilitate the organization of economic activities, what type of dispute
and resolution mechanisms are in place? How will blockchains have to be designed
to comply with GDPR’s requirements? For a comprehensive agenda, we refer to
Beck et al. (2018)’s governance-related perspective as well as Risius and Spohrer
(2017) for an overarching blockchain research framework.
While research can help to underline the value of reputation portability with evidence-based
approaches, regulatory bodies, platforms, and third-party providers will have to develop
strategies to address increasing claims for portable and aggregated online reputation:
• In its current state, the GDPR does not explicitly cover reputation portability
(Section 4, Scenario 1). To let portable reputation between platforms become a reality,
policymakers will need to set out for an advancement (or, at minimum, a clarifying
interpretation) of the current legislation.
• The long list of (previously) unsuccessful startups and PIMS for reputation portability
and aggregation suggests this is not (thus far) a viable business model. How can PIMS
succeed in becoming mainstream solutions for reputation and identity management
(Scenario 3)? How do they ensure compliance with data privacy law?
With this work, we have introduced reputation portability as one aspect of a broader dialogue
on digital identity management. We are the first to provide a comprehensive conceptual model,
thus, touching upon the what, why, and how of portable reputation. Our model portrays and
defines the most important actors, mechanisms, data types, and external influencing factors.
By detailing these dimensions, we derive the need for clear regulatory guidance and identify
an obvious gap in empirical research. Where today’s leading platforms currently forgo
implementing adequate mechanisms for users, PIMS and blockchain technology might provide
the means to factually establish reputation portability and empower individuals with a digital
identity. However, for these approaches to appeal for wide-spread adoption, usable and
intuitive solutions will have to be designed and sociotechnical challenges need to be overcome.
To that end, we derived future scenarios, implications, and critical assessments for platforms,
PIMS, entrepreneurs, and governing bodies. We hope they will inform the ongoing debate
among researchers and practitioners and help with answering a (not so) simple question: Quo
vadis, reputation portability?
66
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
67
PART II: THE ECONOMICS OF
REPUTATION PORTABILITY
68
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
69
Chapter 3: User perceptions and a
market data perspective
After establishing theoretical and conceptual foundations, in this second part of the thesis, I
assess the effectiveness of reputation portability in digital markets. This chapter leverages
two approaches to develop respective answers: first, a survey asking potential platform
users to rate and discuss the importance of portable reputation as an additional trust cue,
and second, a case study analyzing actual platform data to understand the impact of applied
reputation portability on basic indicators of economic success.
Maik Hesse, Timm Teubner
27
Introduction
The management of reputation and digital identity has become imperative for any business
that attempts to sell products and/or services online (Tadelis 2016). On electronic commerce
platforms such as Amazon, eBay, Gumtree, or Bonanza, sellers build an online reputation
based on evaluations of prior customers. Importantly, sellers rely on this reputation since it
greatly determines their capability of 1) enforcing profitable prices and 2) attracting demand
(Dellarocas et al. 2009; Tadelis 2016).
It is quite common that sellers are active on multiple platforms (“multi-homing”) and, in this
case have to manage each reputation separately (Dakhlia et al. 2016; Dellarocas et al. 2009;
Teubner et al. 2019). Building up such a reputation as a signal of trustworthiness towards
prospective buyers requires effort and is costly (Dunham 2011). Especially when starting to sell
through a new channel, the lack of reputation may represent a crucial barrier for economic
success since, from the buyers’ perspective, such sellers represent dark horses with no credible
track record on the platform, a phenomenon which is also referred to as the “cold-start”
problem (Wessel et al. 2017).
This raises the question whether the reputation a seller acquires in one platform represents a
valuable signal of trustworthiness when imported on another platform. In other words – and
put as a concise research question:
RQ: Can sellers with little or no reputation on a platform draw on existing ratings from
another platform to their benefit?
With this exploratory work, we address buyers’ perceptions and sellers’ economic implications
of such cross-platform reputation portability. Until now, research has devoted broad attention
to either a) the roles of trust and reputation within the confines of a given platform
environment (Resnick and Zeckhauser 2002; Teubner et al. 2017) or b) the general concept of
27
This study was published in the proceedings of the Hawaii International Conference on System
Sciences with the title “Takeaway Trust: A Market Data Perspective on Reputation Portability in
Electronic Commerce” (Hesse and Teubner 2020b); see https://doi.org/10.24251/HICSS.2020.629.
Chapter 3: User perceptions and a market data perspective
70
cross-platform data portability (Drechsler 2018; Engels 2016). Overall, there is still an evident
gap of empirical work on the portability of reputation between platforms.
With this work, we seek to address this gap. We do so by considering 1) the perception of
imported reputation from the buyer perspective (Study 1; Survey) and 2) the effectiveness of
imported reputation based on actual market data from an e-commerce platform (Study 2;
Empirical Data). In a nutshell, our results show that e-commerce customers state to be
receptive for imported reputation in general. Specifically, they value the additional information
for increased levels of reliability and consistency in provider ratings—indicating their good (or
maybe not so good) online behavior across a range of sources. Looking at actual market data,
however, we observe a more intricate picture. While a large fraction of sellers makes in fact use
of reputation import and substantially increases rating scores by this, it does not seem to be
converted into tangible economic results. Quite to the contrary, we observe that reputation
imports and the resulting improved rating scores are associated with less demand and lower
prices. We discuss these peculiar findings in view of the rapid proliferation of the platform
economy and alternative approaches to reputation management (e.g., third-party services,
decentralized technology).
The remainder of this paper is organized as follows. We describe the general idea of cross-
platform reputation portability and review existing work on this subject in Section 2. Section 3
then reports the results of an online survey in which we roughly assess the potential of
reputation portability from the consumer perspective. We then analyze a real-world case of
reputation import based on market data from 4,506 sellers of an e-commerce platform in
Section 4. Section 5 provides a general discussion of our findings—linking them to current
developments in policy making and technology. Section 6 concludes.
Related Work
Trust and reputation on platforms
Today’s online platform landscape is ubiquitous, transforming almost every industry, and
pervading many levels of modern economic activity (Alt and Zimmermann 2014; De Reuver et
al. 2018). Digital platforms, such as Amazon, eBay, Taobao, and others mediate transactions
between sellers and buyers by providing them with means to exchange goods and services
online. For a platform business to thrive, the willingness of buyers and sellers to engage in
transactions is imperative. For this to happen, the most important antecedent is the
establishment and maintenance of trust between prospective transaction partners. Hence,
trust and reputation have been a popular research objects within and across platform
ecosystems over the past decades (Resnick and Zeckhauser 2002; Tadelis 2016).
In the context of digital platforms, a seller’s reputation is usually referred to as the accumulated
and documented evaluation by prior transaction partners based on their experiences with this
seller (Jarvenpaa et al. 2000). Reputation systems aggregate transaction-based feedback into
individual buyers’ rating scores to function, inter alia, as a signal of trustworthiness (Dellarocas
et al. 2009; Jøsang et al. 2007; Resnick et al. 2000). Since their introduction in electronic
commerce, reputation systems have been designed and developed to insert trust in online
transactions by providing “future buyers with a window into a seller’s past behavior with
previous buyers” (Tadelis 2016, p. 326). Besides the use of reputation systems to aggregate
third-party assessments, today, all major platforms deploy a variety of trust-enhancing
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
71
mechanisms that allow users (e.g., buyers, sellers, hosts, renters, etc.) to showcase a track-
record for reputable online conduct (Teubner and Dann 2018). Yet, feedback by others, that is,
third-party ratings (e.g., 1-5 stars, positive/negative) or written assessment (i.e., text reviews)
still represents the most frequently used signals to demonstrate trustworthiness in digital
environments (Mavlanova et al. 2012; Teubner et al. 2019).
There is a wide range of studies on the effects of these reputational elements within a given
platform environment—both regarding their impact on buyers’ perceptions of a seller’s
trustworthiness and, in turn, on the implied economics of seller reputation. First, ratings and
reviews have been shown to positively influence buyers’ trust in prospective sellers in a variety
of platform contexts. For instance, for previous buyers’ comments on eBay (Resnick and
Zeckhauser 2002) or, similarly, written feedback from prior transaction partners on sharing
economy platforms (Cheng et al. 2019; Zloteanu et al. 2018), as well as for star ratings and text
reviews on Airbnb (C. Chen and Chang 2018; Qiu et al. 2018),
Second, a plethora of research investigates the resulting economic effects of higher
trustworthiness, namely, the intuitive theoretical claim that it both drives demand/bids from
potential buyers and allows sellers to set and enforce higher prices. We refer to Bajari and
Hortaçsu (2004), Dellarocas et al. (2009), and Tadelis (2016) for more comprehensive
overviews of studies reporting empirical evidence for higher sales probability and price
premiums obtained by more reputable sellers.
Reputation portability
Reputation portability can be defined as “the effectiveness of a user’s reputation on a source
platform (e.g., a star rating score) in building trust on a different platform” (Teubner et al.
2019, p. 231). Considering the documented effects of reputation within a specific marketplace
and the increasing “platformization” of our world, it seems natural to ask questions regarding
the permeability of online reputation across platform borders—especially, since the idea of
portable reputation is anything but new.
In the 1990s, eBay and Amazon introduced early versions of reputation systems and—for some
time—Amazon allowed its sellers to directly import their ratings from eBay (Resnick et al.
2000). Back then, it seemed as though this might become a common practice, however,
apparently eBay was not appreciative of the fact. Upon claims of their ratings being proprietary
accompanied by legal threats, Amazon had to discontinue the service (Dellarocas et al. 2009).
There is a lively debate among legal scholars with regard to the ownership of transaction-based
feedback in the form of reviews—and hence regarding its ability to be ported between
platforms. In 2018, the EU introduced the General Data Protection Regulation (GDPR) in an
effort to foster and enable unprecedented competition in online markets by breaking the
monopoly of incumbents and large platforms (European Union 2016a). In fact, GPDR’s Article
20 mandates platform operators to grant “data subjects” the explicit right to data portability.
Note that any individual who provides data to the platform is subject to the regulation, that is,
both buyers and sellers. Legal scholars have welcomed the new legislation as a promising step
and strong tool to allow for increased competition between digital platforms and give
individuals more control over personal data (De Hert et al. 2018; Kathuria and Lai 2018).
However, given a lack of clear interpretations it needs to be clarified what type of data is
supposed to be ported from one platform to another (Engels 2016; Ursic 2018)—and if GDPR
Chapter 3: User perceptions and a market data perspective
72
in its current form encompasses reputational information. Reputation data (both in the form
of text reviews and ratings) is usually provided by others, that is, prior transaction partners, in
particular buyers. In the case of text reviews, Kathuria and Lai (2018) conclude that ownership
would most likely be with the author or the platform—but not with the individual seller.
Further, based on third-party (star) ratings, platforms calculate proprietary aggregated scores
via their reputation systems. The exact scope of data encompassed by GDPR is certainly
debatable, however, since sellers do not provide these rating themselves, Article 20 is unlikely
to cover them either (Graef et al. 2013; Vanberg and Ünver 2017). Also, obligatory (or better
yet: “forced”) interoperability to design systems so that data portability is supported was taken
out of the regulation—again following critical acclaims from the platforms themselves
(Drechsler 2018). In anticipation of a prisoner’s dilemma—even though platforms would win
over sellers by offering reputation import, they may eventually weaken their position due to
increased competition—Krämer (2018) suggests large players may silently agree to not offer
such functionality for direct competitors.
However unlikely that leading platforms will offer reputation import any time soon, the
economic benefits for sellers and entrant platforms are apparent: for instance, avoidance of
platform lock-in due to network effects, decreased switching cost, and mitigation of the “cold-
start” problem without prior on-site reputation.
Somewhat surprisingly—given the reluctance of long-established platforms such as eBay and
Amazon to offer reciprocal import functionality—it is still possible to have a seller’s ratings
ported from those platforms to newer, alternative e-commerce marketplaces. Bonanza and
TrueGether, for instance, do promote the direct import of both Amazon and eBay ratings into
sellers’ profiles on their platform (Bonanza 2019a; TrueGether 2019). On Bonanza, a seller’s
eBay rating (i.e., positive/neutral/negative) is translated into the equivalent value on a 5-star
rating scheme (i.e., 5, 3, or 1 stars), while on TrueGether both the external rating as well as the
corresponding feedback text are displayed. We analyze the dynamics of imported reputation
on these platforms in the next section.
These practical approaches come well-supported by calls for reputation portability to become
the gold standard. Resnick et al. (2000) have made this remark almost two decades ago stating
the difficulty stemming from “the lack of portability from system to system” (p. 48). More
recently, Puschmann and Alt (2016), in the context of sharing economy platforms, posed the
question how users (from both market sides) may connect their various profiles to
comprehensively manage their online identity. Following the same line of thought, a recent EU
report identifies “cross-platform reputation portability” as an important concept to address
issues of data ownership, prohibitive switching costs, lock-in effects, and platform competition
(European Commission 2017). Finally, Gans (2018) suggests a broader notion of “identity
portability” based on the artifacts used by platforms to generate reputation (i.e., ratings and
reviews).
As much advocacy there may historically be, the hypothesized value of portable reputation has
so far not been corroborated by empirical insight. There has been some quantitative work on
the potential of previously earned reputational data to predict trustworthiness and future
performance. However, this has either been based on social media data (as opposed to
transaction-based reputation; (Venkatadri et al. 2016)) or analysis was limited to different
categories within the same platform (Kokkodis and Ipeirotis 2016).
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
73
Only recently, first studies have set out to investigate the effectiveness of portable reputation
from the consumer’s perspective (Otto et al. 2018; Zloteanu et al. 2018). Otto et al. (2018) use
a controlled experiment to display imported ratings from Airbnb (a platform for peer-to-peer
accommodation rental) to potential passengers on BlaBlaCar (a ridesharing platform). Based
on self-reported scales, they find that imported signals positively impact trust in prospective
drivers on the platform. Similarly, Zloteanu et al. (2018) use a hypothetical scenario and
present study participants with different types of trust and reputation information (“TRI”)—
among which “online market reputation” denotes imported ratings from other platforms (in
this case Airbnb and Uber). They find that the availability of TRI increases the perceptions of
prospective hosts’ trustworthiness as well as participants willingness to book a stay. However,
it is neither possible to attribute this effect entirely to imported ratings nor to assess the latter’s
general importance from a consumer perspective, since the authors did not explicitly control
for it (or chose to not report it). Teubner et al. (2019) use survey data to emphasize today’s
platform multiplicity. They show that almost two out of three respondents are active on more
than one platform, underlining their claim for the relevance of “reputation transfer”. Following
up on this, Teubner et al. (2020) show that star ratings bear trust-building potential also across
platform boundaries for several platforms (i.e., Airbnb, BlaBlaCar, eBay, Uber) and that
thematic fit between source and target context is an important criterion for cross-platform
signaling to be effective. For a comprehensive literature review on reputation portability, we
refer to Hesse and Teubner (2020a).
Overall, the outlined studies unanimously find reputation portability to be effective, however
they all rely on survey data, that is, on self-reported scales and assessments made by the
participants. Given that such approaches are prone to intention-behavior gaps (Sheeran 2002),
we complement this view by additional data-based insights into the workings of portable
reputation. In the next section, we report the results from a consumer survey on perceptions
of portable reputation. We then explore the effects of reputation imports based on actual
market data (Bonanza.com). This allows us to empirically assess the impact of imported
reputation on sellers’ ability to a) attract demand and b) enforce higher prices—and come to a
surprising result. To the best of our knowledge, this paper is the first to do so.
A survey on portable reputation
Study design and data
In the first study, we conducted an online survey in which participants evaluated different
means of online reputation and trust
28
. Specifically, participants took the role of buyers and
evaluated the concept of sellers with imported ratings and/or reviews. We recruited 204
participants via Prolific.ac (Palan and Schitter 2018), ensuring respondents had sufficient
experience in online shopping on electronic commerce platforms (i.e., at least once a month)
and represented a diverse global audience (i.e., Northern America, EU, Middle East). A total
of 187 participants remained after attention and completeness checks. Using a
between-subjects design, participants were assigned to scenarios on five different types of
online platforms: accommodation (e.g., Airbnb, Homestay), P2P carsharing (e.g., easyCarClub,
Getaround), crowd work (e.g., TaskRabbit, Helpling), e-commerce (e.g., eBay, Bonanza), and
mobility (e.g., Uber, Lyft). Based on our previous work on the subject (Hesse and Teubner
28
See Appendix for further details.
Chapter 3: User perceptions and a market data perspective
74
2020a; Teubner et al. 2019, 2020), we introduced the concept of reputation portability as the
availability of additional reputation elements (i.e., star ratings and text reviews)—pertaining
to a prospective provider but stemming from a different platform. Participants then rated the
importance of these additional trust cues in deciding to buy from a prospective provider (single
item construct; 7-point Likert scales). We added an open-ended question asking why (if at all)
and in which cases subjects considered imported reputation to be helpful for their decision
process. These textual statements were reviewed to identify distinct aspects mentioned by
participants (e.g., importance of fit between platforms; helpfulness for new providers). The
survey was conducted in May 2019.
Results
Figure 11 summarizes the results of the survey and shows platform-specific differences. It
becomes evident that respondents generally value the concept of reputation portability for e-
commerce, scoring 5.6 on the importance Likert scale (1 to 7; corresponding to 76.67% when
standardized). Remarkably, other platform types (i.e., accommodation, crowd work, mobility)
exhibit higher scores for imported reputation. One potential explanation for these differences
may root in the fact that those platforms facilitate co-sharing and physical encounters are
prerequisite to service provision, that is, there occurs a generally higher degree of social
interaction (Hesse, Dann, et al. 2020). Hence, they may require higher levels of “inserted” trust
from external sources.
FIGURE 11: IMPORTANCE OF PORTABLE REPUTATION AND PREFERRED SIGNAL FOR IMPORT
Another observation concerns the preferred trust signal survey participants would like to see
being imported from other platforms. Here, participants appreciated written text reviews over
star ratings on all platforms except those for mobility services.
The reason for that can be found in the verbal statements. Participants were asked to comment
whether, why, and how (i.e., under which circumstances) imported reputation would be
helpful to them. As for their preference for text reviews, some participants mentioned star
rating scale skewness as a limiting factor (“people usually assign stars/points in extremes”). In
fact, positively skewed star rating distributions represent a common phenomenon for many
7.0
6.0
1.0
5.5
6.5
Crowd work Mobility
5.5
Accommodation Carsharing E-commerce
Ø 5.7
5.9
5.8
5.6
5.9
Importance (on 1-7 Likert scale)
Platform type
Preferred signal (text review vs. star rating)
0.0
0.5
1.0
1.5
2.0
2.5
0 5 10 15 20 25
Average Price (in $US)
Demand (=Additional transactions)
without
imported ratings
with
imported ratings
-.692
-.599
Ø 5.7
revenue
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
75
online platforms (Schoenmueller et al. 2018; Teubner and Dann 2018; Zervas et al. 2015). In
that sense, text reviews appealed as “more informative” and to provide a “clearer picture”.
Main reasons in favor of imported reputation were the general surplus of details (e.g., “the
more information the better”; 30 times), the ability to prove a provider’s good (or bad) conduct
on other platforms (26 times), or the opportunity to check on consistency/reliability of ratings
(e.g., could be “bought”, “fake”, fabricated”, or “written by oneself”; 20 times). In particular,
imported reputation was considered as potentially helpful as it allows novel providers with no
or little on-site reputation to mitigate the platform-inherent cold-start problem (18 times).
Specifically, one respondent stated that reputation portability could help in demonstrating an
“honest online footprint”.
Additionally, the issue of sellers getting rid of “a string of bad reviews” and “play[ing] the
system by using another platform” was repeatedly brought up (15 times). That is, consumers
would want the critical reviews from other platforms to show up. In this sense, it could prevent
bad actors from successfully continuing their poor game by starting over and over on new
platforms.
Last, the relevance of the fit between source and target platforms for porting reputation
received considerable attention (14 times). On the one hand, participants noted that the source
platform should be well-known, reputable, and trustworthy itself to instill confidence in the
ratings’ reliability and credibility. On the other hand, the service offering on source and target
platforms should be compatible and relevant to each other because, as one participant neatly
put it, “a fishmonger is not a good butcher” – an observation well in line with recent research
(Teubner et al. 2020).
Reputation import on Bonanza.com
Study design and data
In this second study, we follow up the first study’s findings and investigate how importing
reputation affects sellers’ market success, that is, their capability to 1) enforce price mark-ups
and 2) attract demand. In this sense, we now investigate the overall effect of the reputation
import strategy from an empirical perspective, that is, based on actual market data. Naturally,
this approach is limited to uncovering that certain effects occur but cannot speak to how or
why specifically they may appear. Thus, our approach is prone to issues of observational
studies such as omitted variables which we come back to it in the general discussion.
We collected data from Bonanza.com, an e-commerce platform that markets itself as an eBay
alternative. It targets business sellers, being a “marketplace that empowers entrepreneurs to
build a sustainable business based on repeat customers” (Bonanza 2019b). Founded in 2008,
the platform hosts about 50,000 registered sellers with an overall inventory of 35 million
items. Using a web-crawler to query the platform, we collected openly accessible data for a total
of 4,506 sellers and 124,067 items. The data was collected at two instances in April and May
2019.
For analysis, we consider individual sellers. For each seller, we consider data on i) overall rating
score (1-5 stars in steps of 0.1 stars), ii) average price of their items, iii) number of transactions
on the platform, and iv) whether they have imported ratings from eBay or not (binary variable:
Chapter 3: User perceptions and a market data perspective
76
yes/no). Moreover, we retrieved information on how many items a seller offers through the
platform as well as their total number of reviews as an approximation of business size.
The first pass of data collection in April 2019 was followed by a second pass four weeks later.
This longitudinal data structure allows us to account for sellers’ capability of generating
additional sales, that is, of attracting demand (i.e., delta between t=2 and t=1), as a result of
their rating score in t=1. Specifically, in contrast to “snapshot” data at only one single point of
time, this approach reduces endogeneity issues which make inferences on causal effects
notoriously difficult. The data can be characterized as follows:
i. The Score reflects a seller’s average star rating score (1-5 stars). Without import
(n=270): 4.52 stars; with import (n=4,236): 4.97 stars.
ii. The Price represents a seller’s average price across their items (without import: 19.82$;
with import: 19.22$).
iii. Demand describes a seller’s number of new transactions over the 4 weeks between data
collection 1 and 2 (without import: 1.81; with import: 1.12).
iv. Average number of Reviews
a. On Bonanza (without import: 21.85; with import: 9.85)
b. Imported from eBay (without import: 0; with import: 2,803)
Results
Overall, 94% of sellers import reputation. And for 9 out of 10 sellers who import reputation,
the imported ratings make up for more than 95% of their total number of ratings. Now, in order
to further understand the specific characteristics of and differences between sellers with and
without reputation imports, we conduct a series of OSL regressions on the above-mentioned
focus variables. Table 5 summarizes the regression models for the three dependent variables
Score (Model I), Price (Models II-IV), and Demand (Models V-VII) based on observed values
from 4,506 sellers on Bonanza.com.
i. Score
ii. Price
iii. Demand
Imported Reviews (y/n)
.450
***
-.314
.914
-.681
***
-.459
**
(.016)
(.783)
(.846)
(.144)
(.156)
Score (1-5)
-2.431
***
-2.727
***
-.641
***
-.492
***
(.664)
(.718)
(.123)
(.133)
#Reviews (10³)
.001
*
-.073
***
-.068
**
-.070
**
.019
***
.019
***
.020
***
(.000)
(.021)
(.021)
(.021)
(.004)
(.004)
(.004)
#Items (10³)
-.002
**
.124
***
.117
***
.118
***
.098
***
.097
***
.097
***
(.001)
(.033)
(.033)
(.033)
(.006)
(.006)
(.006)
Intercept
4.519
***
19.519
***
31.236
***
31.844
***
1.570
***
4.101
***
3.796
***
(.016)
(.761)
(3.286)
(3.334)
(.140)
(.607)
(.615)
#Observations
4,506
4,506
4,506
4,506
4,506
4,506
4,506
R2
.152
.005
.008
.008
.077
.078
.080
*** p<.001; ** p<.01; * p<.05; standard errors in parentheses
TABLE 5: OLS REGRESSION MODELS (SCORE, PRICE, DEMAND)
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
77
First, confirming the observation from above, average rating scores are significantly higher
for sellers with imported ratings by about half a star (b=.450, p<.001). Moreover, controlling
for a seller’s characteristics shows that the respective number of ratings and items have
significant (but minuscule) effects on rating score (b=.001, p<.05; b=-.002; p<.01; per
thousand ratings/items each).
Next, to assess the economic effect of importing/ not importing reputation, we consider
regression models for price and demand. The price models show that higher star rating scores
are negatively related to average prices (b=-2.727, p<.001) and that, beyond this, importing
reviews has no significant price effect (b=.914, n.s.). Next, also the demand regressions show
that higher star rating scores are negatively related to demand, that is, the number of sales a
seller had between the two points of observation (b=-.492, p<.001). Importantly, we observe
that beyond this rating score effect, there occurs an additional negative effect based on
importing itself (b=-.459, p<.01). Thus, even for identical rating scores, sellers who have
imported ratings attract less additional demand as compared to sellers who did not import.
Overall, these observations are puzzling, to say the least. The negative relation between higher
rating scores and prices/demand is counter-intuitive from a common-sense perspective and
inconsistent with theory and most of the literature. There exist, however, some studies that
report similar negative relations between reputation and economic indicators (Dellarocas et al.
2009) and we offer some approaches of explanation in the general discussion. Importantly,
when excluding the control variables from the models, none of the coefficients change in terms
of sign, magnitude, or significance.
Now, as a seller’s economic success can be well conflated by revenue (R), which emerges as the
interaction of price (p) and demand (d). Hence, revenue can be graphically interpreted as the
area of a rectangle with edges p and d, where R=p·d. Figure 12 shows average prices and
demand for sellers with versus without imported reputation (standard errors indicated by
error bars in both dimensions). We see that the overall revenue difference between importing/
not importing reputation is primarily driven by the lower additional demand rather than by
the difference of average prices (Figure 12).
FIGURE 12: REVENUE DIFFERENCE COMPOSITION
7.0
6.0
1.0
5.5
6.5
Crowd work Mobility
5.5
Accommodation Carsharing E-commerce
Ø 5.7
5.9
5.8
5.6
5.9
Importance (on 1-7 Likert scale)
Platform type
Preferred signal (text review vs. star rating)
0.0
0.5
1.0
1.5
2.0
2.5
0 5 10 15 20 25
Average Price (in $US)
Demand (=Additional transactions)
without
imported ratings
with
imported ratings
-.692
-.599
Ø 5.7
revenue
Chapter 3: User perceptions and a market data perspective
78
General discussion
Full transparency of previous online conduct across digital platforms—as appreciated by some
of our survey participants—would entail far-reaching implications. Having bad reviews
showing up everywhere we interact online would pose a considerable threat to Western world
liberal values. Along those lines, China is—as we speak—in fact implementing a national
reputation scheme (“Social Credit System”) that penalizes and rewards individuals and
businesses (Hatton 2015; Síthigh and Siems 2019). Importantly, the system explicitly includes
online behavior such as electronic shopping and social media conduct. By 2020, it intends to
assess every citizen nation-wide with a unique, standardized, and publicly ranked score.
Although most Western observers reject these plans outright (e.g., as dystopian or
“Orwellian”), others suggest that—with the ubiquity of digital platforms, predictive algorithms,
and scoring mechanisms—the Western world is not too far away from the Chinese system of
networked shame (Botsman 2017a). The EU will have to find its own position in dealing with
this development. The GDPR currently grants individuals the right to erasure (also labeled as
“right to be forgotten”; Article 17). This entitles them to have their personal data (stored by
platforms) deleted. Whether this also pertains to any reputational data—provided by others as
previously discussed—is still up to clarification.
However, cherry-picking only the best ratings to be ported from a given platform (or choosing
just those platforms where one has collected good average ratings) seems unreasonable, too.
To prevent this scenario of “takeaway trust” ad libitum, systems could be designed in a way
that leaves the seller with a choice: either import all available reviews (full transparency) or
none at all. In that sense, it would provide sellers with the discretional option to disclose their
previous ratings (Dellarocas et al. 2009). This could create an incentive for impeccable online
behavior and may be more in line with the foreseeable position of the EU between the US
(reputation capitalism) and China (reputation tyranny) as providing a third way: reputation
sovereignty.
We also learned that the source-target fit of a platform is of crucial importance when
considering portable reputation (e.g., a great Airbnb host could be a nightmarish driver).
Services might not be comparable across platforms and what is considered good in one
environment may not be such a desirable treat in another (Teubner et al. 2020). Furthermore,
ratings may hardly be comparable when, for instance, one platform exhibits skewness towards
highly positive scores whereas another might show a much more dispersed rating distribution
(Teubner and Dann 2018).
A puzzling market—As suggested by theory and common sense, there should be a positive
relation between reputation and buyers’ capability to attract demand as well as pricing power.
However, there have also been findings in the opposite direction suggesting either limited
effects of reputation or effects in the “wrong” direction (Dellarocas et al. 2009). For
Bonanza.com, we also observe a negative relation between rating scores and demand/prices.
Importantly, this also holds when considering only sellers without imported reputation. Given
these admittedly puzzling results, several (in parts rather speculative) approaches for
explanation come to mind. While survey participants expressed a preference for text reviews
as the to-be-imported signal, only star ratings are available which may impair their effect.
Import functions may hence be extended to textual information too. Also, as shown by the
cross-domain comparison (Figure 11), importing reputation is considered as relatively
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
79
unimportant for platforms such as eBay/Bonanza as compared to other domains with higher
degrees of social interaction. Thus, albeit being one of very few platforms on which reputation
imports can be studied, Bonanza.com may not be the best or most promising venue to do so.
In this sense, our findings may underestimate the potential benefits of reputation imports.
Also, a mechanism of reverse causality may cause “less expensive” sellers to be perceived as
having a better price-value, which is why they, in turn, receive better reviews. Hence, rather
than sellers’ reputation forcing/allowing them to set lower/higher prices, the prices themselves
may trigger biased ratings. This, however, is still at odds with most prior research on the
economic effects of reputation, where, very broadly, positive effects are reported. Specifically,
the latter rationale cannot explain why sellers with better reputation at a certain point in time
receive less demand subsequently. Another explanation could be that Bonanza.com represents
a secondary sales channel for many sellers through which they sell off their stock at discounts.
Following this logic, especially large (and successful) eBay sellers may treat their Bonanza
accounts with less care. To address this, a promising way forward could be to link the Bonanza
profiles back to their eBay counterparts. In addition, a more frequent sampling of larger sets
of Bonanza profiles could allow for a comparison of sellers before and after they have imported
ratings, enabling difference-in-differences estimation or regression discontinuity designs. In
view of our findings, one should critically ask whether importing large amounts (i.e.,
thousands) of ratings and reviews to a context in which typical sellers have only very few
ratings, somehow disturbs the effectiveness of market signaling altogether. Last, issues could
also root in the quite different group sizes of the “treatment/control” groups. Overall, given the
exploratory nature of this work, it is far too early to draw stark theoretical conclusions. Clearly,
in light of the quite puzzling results and the outlined limitations of this study, more research
based on larger data sets is needed to better understand the intricacies that govern reputation,
price, and demand for reputation portability in general, and for Bonanza.com in particular.
Overall, our findings also prompt the question whether the cross-platform employment of
reputation can and/or should be done differently. Should today’s platforms, for instance,
continue to forgo the implementation of user-friendly import mechanisms, Personal
Information Management Services (PIMS) and blockchain technology may provide means to
make reputation portability a more prevalent reality in the future.
Alternatives to direct import—Rather than drawing on direct “end-to-end” import, there
may exist viable alternatives to realize portable reputation. For instance, PIMS as third-party
services, aggregate and verify (reputational) data from disparate sources. Based on the
combined data (e.g., ratings/reviews, verified social/professional accounts, personal IDs)
these companies build individual trust-based data management tools, profiles, and
dashboards. PIMS can best be understood as a data intermediary with the primary aim of
putting people in control over their own reputation and digital identity (European Data
Protection Supervisor 2016). However, such services have thus far been mostly unsuccessful
with many failed attempts (e.g., trustcloud.com famust.com, peertru.st, tru.ly). Today, these
services still struggle to develop profitable business models and to gain market recognition
(Teubner et al. 2019).
Likewise, several endeavors based on blockchain (more generally: distributed ledger)
technology promise to empower individuals with more control over their data—irrespective of
intermediaries (Drescher 2017; Glaser 2017). Indeed, the technology’s potential for portable
Chapter 3: User perceptions and a market data perspective
80
reputation lies in its ability to decentralize control and ownership over (reputational) data
(Werbach 2018). In that way, it could break today’s siloed (i.e., platform-bound) approach to
reputation in favor of alternative business models for cross-platform portability (Catalini
2017). Especially, given the prevalence of “multi-homing” strategies with sellers/providers
offering their services on several platforms, a blockchain-based marketplace could become a
trusted, decentralized peer-to-peer network where participants retain full control over their
data (Werbach 2018). Reputation in these environments could seamlessly roam across
platform boundaries since it could be verified with ease, reside with individual sellers, and be
built across a range of markets. Catalini and Gans (2016) envision this as an open reputation
platform with the “ability to port and use the resulting reputation scores across different
services and contexts” (p. 22). Eventually, combining PIMS’ value proposition of self-sovereign
identity management with technology to build decentralized platforms may just be the right
approach in factually establishing portable reputation.
Conclusion
Digital platforms have become pervasive in basically all our daily lives. Specifically, two-sided
platforms have emerged as a quasi-standard within electronic commerce. In view of the ever-
increasing importance of this platform economy as well as its rules, mechanisms, and
paradigms, we expect that reputation and specifically its fluidity will gain further importance
for online businesses such as sellers on e-commerce marketplaces. With this first empirical
account on actual reputation imports, we hope to contribute to and spark an ongoing scientific
and public debate. Additionally, we expect the general topic of reputation portability to fuel a
vivid legal discourse in view of privacy regulation, data sovereignty, and competition among
platfirms.
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
81
Chapter 4: Trust measures and
import thresholds
Continuing the investigation of reputation portability’s economic effectiveness, in Chapter 4
I consider complementors’ ability to increase consumers’ trusting beliefs by importing
reputational data from another platform. The presented study introduces a stylized function
to measure trust and leverages data from an online experiment to estimate this function’s
parameters. As a result, I derive specific thresholds for when the import of star ratings from
an external platform is beneficial for complementors.
Maik Hesse, Timm Teubner, Marc T. P. Adam
29
Introduction
The recent years have seen a flourishing of multi-sided market platforms, for instance, in the
domains of accommodation sharing, mobility services, or crowd work. While reputation is
critical for complementors to attract demand, the reputation they have built on one platform
is, as of today, of limited use across platform boundaries. In this sense, reputation portability
refers to the idea of leveraging existing reputation from one platform and using it as a trust
signal on another (Hesse and Teubner 2020a; Teubner et al. 2019). For example, consider an
Airbnb host looking into additionally posting listings on a competing platform such as 9flats
or Homestay.com (i.e., planning to multi-home). Rather than having to re-build a reputation
from scratch on the entrant platform (e.g., Homestay.com), importing their reputational
history from the incumbent platform (e.g., Airbnb) would serve as an important credential to
attract demand from prospective guests.
Today, however, the concept of reputation portability is not a common practice. Two notable
exceptions are the e-commerce platforms TrueGether and Bonanza.com, which allow sellers to
import ratings and reviews from eBay and Amazon (Bonanza 2019a; TrueGether 2019). While
Bonanza translates eBay ratings into its own 5-star scheme, TrueGether also allows for the
import and display of written feedback.
30
Notwithstanding these practical implementations of
reputation portability in online marketplaces, a wider application across the entire platform
landscape is far from sight and it remains to be seen whether any of these applications will
prevail.
Yet, calls for research on reputation portability have been prevalent in academic literature for
almost two decades. In the early days of electronic commerce, Resnick et al. (2000) stated that
not being able to transfer one’s ratings from one platform to another is problematic as
29
This study was originally submitted to the journal Business & Information Systems Engineering and
carries the title “In stars we trust – A note on reputation portability in digital platform ecosystems”. It
is currently undergoing the third round of reviews. This version is based on the first revised manuscript.
30
Several other mechanisms to establish reputation portability are conceivable, including Personal
Information Management Systems (PIMS) and distributed ledger technology. These were discussed in
EU reports and previous literature (e.g., European Data Protection Supervisor 2016; see Hesse and
Teubner 2020a for an overview).
Chapter 4: Trust measures and import thresholds
82
“[l]imited distribution of feedback limits its effectiveness” (p. 48). In the mid-1990s, Amazon
allowed customers to import their existing ratings directly from eBay and the latter threatened
to press charges claiming these ratings to be proprietary content. About a decade later with
more platforms emerging and the advent of the sharing economy, Dellarocas et al. (2009)
made a case for portable reputation, in particular for multi-homing complementors. The
notion of managing online reputation by connecting profiles across platforms and enabling
“identity portability” on the basis of transferable ratings has since been picked up regularly
(Gans 2018; Puschmann and Alt 2016). These contributions unanimously point at the trust-
building potential of reputation portability and ample research demonstrates that reputation
represents an important driver of trust (i.e., consumers’ trusting beliefs towards
complementors; Resnick and Zeckhauser 2002; Tadelis 2016; Ter Huurne et al. 2017) and
economic success (as measured, for instance, in conversion rates; Lee and Hosanagar 2020).
However, only little scholarly work has investigated the effects of transferring reputation
across platform boundaries (Teubner et al. 2019). The EU has identified this lack of integration
as problematic and urges research on potential benefits and underlying mechanisms of
reputation portability (e.g., European Commission 2017, p. 131).
From a theoretical perspective, it is well conceivable that such supplementary reputation can
function as a signal of complementors’ trustworthiness. As building and maintaining a good
reputation is costly (Dunham 2011) and importing reputation would be at complementors’
discretion, signaling theory (Spence 1973) lends itself well as a theoretical grounding. Recently,
several studies have leveraged signaling as a theoretical underpinning for reputation
portability. For instance, Otto et al. (2018) found that consumers of a ride sharing service
exhibit higher willingness to pay and trust levels towards potential drivers (i.e.,
complementors) if their Airbnb ratings were available. Similarly, Teubner et al. (2020)
presented prospective consumers with complementor profiles across three common
application areas, namely accommodation sharing, ride/mobility services, and commodity
exchange. In these either the rating on the platform, an imported rating, or no rating at all was
available. They, too, found evidence for the trust-building potential of star ratings across
platform boundaries and, further, identified fit between source and target context as a driver
of this effect. However, it is important to note that the existing studies neither considered (1)
the availability of both on-site and imported rating simultaneously nor did they (2) account for
varying or sub-standard rating scores.
To address this gap, we consider the scenario outlined in the opening paragraph of this article:
Assuming reputation portability, under which circumstances would a complementor import
their ratings from an incumbent platform when starting on an entrant platform? Given that
several examples of platforms already offer this functionality, it is important to investigate
what drives a complementor’s decision to make use of reputation portability. With the goal of
maximizing trust from prospective consumers, we derive specific decision boundaries for
complementors to decide whether or not to import ratings from one platform to another.
Systematically varying both the on-site and imported rating scores, we estimate the resulting
“trust function’s” parameters by means of an online experiment (n=493). Based on the
overarching theoretical framing of (cross-platform) signaling and building on previous
empirical findings, we hypothesize that
H1: The effect of an imported rating for consumers’ trust in complementors positively depends
on the rating’s value.
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
83
H2: Whether or not importing a given external rating is beneficial for the complementor
depends on their on-site rating’s specific value.
We find support for both hypotheses. These are important results because, on one side, they
validate the assumption that reputation portability can increase trust in complementors – as
hypothesized and examined in recent empirical studies. On the other side, we also demonstrate
that there are boundaries to the positive effects of reputation import and that, depending on
the specific scores, different trust-maximizing strategies emerge. Thus, our work contributes
to both the trust literature as well as signaling theory in several ways. First, we extend the
general trust literature on online reputation by studying an additional (external) signal, which
can be imported from another platform. Second, we devise an experimental design with clear-
cut scenarios that consider all possible combinations of these two signals’ values. This provides
both a quantitative and a strategic perspective on online reputation management (Otto et al.
2018; Teubner et al. 2020). With our results at hand, we then lay out an agenda for future
studies on reputation portability, considering its practical relevance and discussing managerial
implications for platforms and complementors.
Reputation portability scenario
Consider a multi-homing complementor. Specifically, the complementor has an existing track
record on one platform (i.e., the incumbent platform i), and has additionally started to operate
on another platform (i.e., the entrant platform e). Reputation on both platforms is
conceptualized as a rating score ranging between 1.0 and 5.0 stars (the most common system).
The complementor has acquired a rating of 𝑟𝑖 on the incumbent platform and may (or may not)
have already acquired an on-site rating of 𝑟𝑒 on the entrant platform. We consider the
complementor’s decision whether to import their rating 𝑟𝑖 from the incumbent to the entrant
platform. The stylized function 𝑓(𝑟𝑒,𝑟𝑖) describes the relation between the complementor’s
(on-site and imported) rating scores and their trustworthiness in the eyes of prospective
consumers. Indeed, as previous research has shown, “good” reputation is a driving force for
trust (i.e., trusting beliefs) in online transactions (Dellarocas et al. 2009; Qiu et al. 2018;
Tadelis 2016). To be able to account for their individual and interactive effects, the two ratings
𝑟𝑒 and 𝑟𝑖 are not conflated into a single score but displayed independently. We thus distinguish
four conceptual cases: (1) no on-site rating exists and no rating is imported; (2) on-site rating
𝑟𝑒 exists and no rating is imported; (3) no on-site rating exists and a rating 𝑟𝑖 is imported; (4)
on-site rating 𝑟𝑒 exists and a rating 𝑟𝑖 is imported. This “trust function” 𝑓(𝑟𝑒,𝑟𝑖) can hence be
formalized as
𝑓𝑗(𝑟𝑒,𝑟𝑖) = 𝛼𝑗+𝛽𝑗𝑟𝑒+𝛾𝑗𝑟𝑖+𝛿𝑗𝑟𝑒𝑟𝑖=
{
𝛼1+𝛽1𝑟𝑒+𝛾1𝑟𝑖+𝛿1𝑟𝑒𝑟𝑖 (𝑗=1)
𝛼2+𝛽2𝑟𝑒+𝛾2𝑟𝑖+𝛿2𝑟𝑒𝑟𝑖 (𝑗=2)
𝛼3+𝛽3𝑟𝑒+𝛾3𝑟𝑖+𝛿3𝑟𝑒𝑟𝑖 (𝑗=3)
𝛼4+𝛽4𝑟𝑒+𝛾4𝑟𝑖+𝛿4𝑟𝑒𝑟𝑖 (𝑗=4) [a]
where the parameter vectors 𝛼,𝛽,𝛾,𝛿 capture the effects of the rating scores 𝑟𝑒 and 𝑟𝑖 on the
complementor’s trustworthiness across the four cases (𝑗=1,2,3,4)
31
. For the first case (1), a
single coefficient 𝛼1 suffices. For the cases in which either an on-site rating (2) or an imported
rating (3) exists, we assume a linear relationship between rating and trustworthiness. If both
31
Please note that 𝑓𝑗(𝑟𝑒,𝑟𝑖) represents a conceptual function here and not a regression. We hence do not
include residual/error terms in this representation.
Chapter 4: Trust measures and import thresholds
84
on-site and imported ratings exist (4), we allow for interaction between both ratings to occur
(𝛿4). A pair-wise comparison of all four cases results in several decision boundaries depending
on the availability of rating scores 𝑟𝑒 and 𝑟𝑖. First, to decide whether to import a rating of 𝑟𝑖 if
no on-site rating 𝑟𝑒 is available, the complementor compares the cases (1) and (3). Equating
these cases and solving for 𝑟𝑖 yields the import threshold
𝑟𝑖>𝑟∗= 𝛼1−𝛼3
𝛾3 [b].
Thus, the complementor will import a rating from the incumbent platform if it is greater than
𝑟∗. Second, if an on-site rating exists, the complementor compares cases (2) and (4), yielding
𝑟𝑖>𝑟∗∗(𝑟𝑒)= (𝛼2−𝛼4)+(𝛽2−𝛽4)𝑟𝑒
𝛾4+𝛿4𝑟𝑒 [c].
As opposed to 𝑟∗, note that 𝑟∗∗ is not a constant but a function of the on-site rating. We will
consider the pair-wise comparison of additional cases in the results section.
Experiment
Procedure and Sample—To estimate parameters for the outlined trust function, we conducted
an online experiment. The dependent variable was operationalized by prospective consumers’
trust in the complementor. Specifically, we recruited 500 participants via Prolific.co, which has
been praised for its comparably high data quality and subject diversity as well as the option to
pre-screen participants (Palan and Schitter 2018; Peer et al. 2017). In our case, we pre-filtered
participants to have at least some experience in online shopping and to represent a gender-
diverse and multi-national subject pool. Age ranged from 18 to 73 years (mean=35.0,
median=32) and 56% of participants were female. Respondents had to opt-in for their
participation, received 0.80£ upon successful completion, and took an average time of 74
seconds. Seven of 500 participants timed out and were excluded from further analysis.
Stimulus Material—Participants took the role of consumers and decided on booking a stay
with a complementor on the accommodation sharing platform Happystay.com (see Appendix
for details).
Treatment Design—Following a 2 × 2 between-subjects design, each participant experienced
only one of the four outlined scenarios. The complementor either had no on-site rating (cases
1 and 3) or a rating 𝑟𝑒 (randomly drawn from {1.0, 1.5, …, 5.0}; cases 2 and 4). Additionally, as
outlined above, the complementor had either imported a rating 𝑟𝑖 from the incumbent
platform (again, randomly drawn from {1.0, 1.5, …, 5.0}) or no rating was imported. All
possible combinations of rating scores were tested where the distribution was informed by the
skewed values observed on actual platforms (Schoenmueller et al. 2018). Thus, although
ratings above 3.0 stars occurred more frequently, the allocation of participants across cases (1)
to (4) was mainly driven by the number of possible rating combinations per condition and by
ensuring that at least a couple of observations were obtained for each possible combination.
Besides rating scores, we also varied the number of ratings the complementor had received
(ranging from lower single-digits to high double-digits) as a control variable. The allocation of
participants across the treatment conditions and specific rating scores is reported in the
Appendix.
Measurement—As a main dependent variable, we surveyed participants’ trust in the
complementor (i.e., their trust toward the prospective host; Gefen 2002). Moreover, as trust
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
85
represents an overarching theme and important prerequisite for the realization of transactions
on platforms (Tadelis 2016), we controlled for participants’ trust in the incumbent platform,
general trusting disposition (Gefen 2000), and, additionally, individual risk-affinity (Dohmen
et al. 2011). All variables were introduced as single-item 11-point Likert scales. Descriptive
statistics on the participant-specific controls and the measurement instrument are provided in
the Appendix.
Results
The trust-building effect of imported reputation
Moving beyond the conceptual representation of the trust function as specified in [a], we
conduct OLS regression analyses to estimate the parameter set for 𝛼,𝛽,𝛾, and 𝛿. Table 6
reports the results for all cases (1-4); the dependent variable (trust in complementor) was
standardized to the interval [0, 1]. Rating scores ranged from 1.0 to 5.0 (stars).
DV: Trust in complementor
Case (1)
(no rating)
Case (2)
(on-site only)
Case (3)
(imported only)
Case (4)
(both ratings)
Constant
.350
***
-.004
.016
.259
*
(.038)
(.075)
(.082)
(.116)
On-site Rating
.128
***
-.029
(.021)
(.032)
Imported Rating
.116
***
-.006
(.023)
(.033)
Interaction
.024
**
(.009)
Observations
24
81
83
305
Adjusted R2
—
.308
.237
.181
*** p<.001; ** p<.01; * p<.05; standard errors in parentheses
TABLE 6: PARAMETER ESTIMATES (TRUST)
These results are illustrated in Figure 13 (left: cases 1, 2 and 3; right: cases 2 and 4).
Chapter 4: Trust measures and import thresholds
86
FIGURE 13: CONSUMER TRUST IN COMPLEMENTOR BY RATING SCORE
Result 1: In the absence of an on-site rating, consumer trust in complementors is
driven by increasing values of the imported rating (𝑟𝑖; H1). However, importing a
rating will only be beneficial if its value is sufficiently high (i.e., exceeding 𝑟∗=2.9
stars). Otherwise, complementors should not import this rating
32
.
Second, when both on-site and imported ratings are available in case (4), these two positively
interact (𝛿4=.024, p<.01), that is, the effect of on-site ratings on trust in complementors
increases with better imported ratings – and vice versa. Figure 13 (right) shows this relation
for different values of imported star ratings.
The interplay of on-site and imported reputation
Moreover, if an on-site rating 𝑟𝑒 is available, the imported rating must exceed 𝑟∗∗ =−.263+.157𝑟𝑒
−.006+.024𝑟𝑒.
This function is displayed in Figure 14 (left). It illustrates the complementor’s decision
boundary for whether or not to import a rating (𝑟𝑖; y-axis) from the incumbent platform in case
they have to weigh it against an existing on-site rating (𝑟𝑒; x-axis). For instance, importing a
rating of 3.5 stars is beneficial for an on-site rating of 3.0 stars, while it is not for an on-site
rating of 4.0 stars. Likewise, when having already built up an on-site rating of 4.0 stars, a
complementor should import an incumbent rating of 4.5 stars while they should refrain from
importing an equally valued rating of 4.0 stars. Figure 14 (left) additionally depicts 𝑟∗.
Result 2: When an on-site rating is available, the decision threshold for importing a
rating depends on the on-site rating’s value (H2). This relation is positive and concave.
32
Please note that the value of exactly 2.9 stars (as well as other specific values utilized in this context),
has a representative and indicative character. Certainly, any import threshold will depend at least on 1)
the specific distribution (and skewness) of ratings on the source platform and 2) the applied rating scale.
For the sake of exemplary computations and to derive sufficiently generalizable implications, we assume
similar consumer perceptions of rating distributions across platforms and use a 5-star rating scheme.
(2)
(3)
(1)
Rating Type
Imported only (3)
On-site only (2)
None (1)
1.0 stars
Imported rating(s) (4)
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
87
Note that it is always beneficial to import an incumbent rating above 4.58 stars. Also, it is
always beneficial to import any rating if the on-site rating is below 1.91 stars.
FIGURE 14: IMPORT AND ACCOUNT ABANDONING THRESHOLDS
33
Account abandoning
Another way to look at the outlined scenario is via the strategy of account abandoning and
restart. Complementors who have received adverse reviews may decide to give up their entrant
platform account and create a new one. If no option for importing a rating from an incumbent
platform is available, the threshold to do this is equal to the above-mentioned intersection
between cases (1) and (2), that is, at 𝑟∗∗∗ =𝛼1−𝛼2
𝛽2=.350−.004
.128 ≈2.75 stars. Hence, if a
complementor’s rating drops below this value, it would be beneficial to start over (with no
rating at all). However, if importing a rating is an option, the threshold 𝑟∗∗∗∗ for resetting one’s
account on the entrant platform will depend on the value of the external rating (i.e., the rating
from the incumbent platform). Thus, a complementor will abandon their entrant platform
account if 𝑟𝑒 is lower than 𝑟∗∗∗∗, that is
𝑟𝑒<𝑟∗∗∗∗(𝑟𝑖)=(𝛼3−𝛼4)+(𝛾3−𝛾4)𝑟𝑖
𝛽4+𝛿4𝑟𝑖 [d],
which can be obtained by the comparison of cases (3) and (4). When substituting with the
estimated parameters from Table 6, we get 𝑟∗∗∗∗(𝑟𝑖)=−.243+.123𝑟𝑖
−.029+.024𝑟𝑖. Figure 14 (right) depicts this
relationship and indicates the complementor’s optimal strategy given the two ratings’ values.
In consequence, the decision threshold for account abandoning (based on the on-site rating)
depends on whether an incumbent rating exists and—if so—on this rating’s value. Hence,
enabling reputation import may result in a higher frequency of abandoned accounts as
complementors’ setbacks are smaller. This may, in turn, create new problems and necessitate
additional, preventive countermeasures on the part of platform operators. Ultimately, these
33
Both thresholds are based on consumers’ trust in complementors.
Without costs
∗∗∗
∗∗∗∗( )
Do not delete
account
Delete account on
entrant platform
NA
Rating on Incumbent Platform
On-site Rating
Without costs
∗
∗∗( )
Do not import
incumbent rating
Import
incumbent rating
NA
Chapter 4: Trust measures and import thresholds
88
measures come down to specific platform design choices, which we can touch upon only briefly
(please see the discussion section).
Trust-maximizing strategies
Next, we iterate on the idea of using the decision thresholds to identify a trust-maximizing
display strategy for platform complementors’ ratings. Consider a complementor who has both
an on-site and an incumbent rating and is making a decision about which of these rating(s) to
display. While, certainly, complementors would (and should) not have a choice to “drag &
drop” their reputation at will, we consider these hypothetical scenarios to investigate all
possible combinations and the resulting theoretical implications. Now, extending the pair-wise
comparison from above and analyzing each resulting decision threshold (i.e., the two constant
values 𝑟∗ and 𝑟∗∗∗; the two threshold functions 𝑟∗∗(𝑟𝑒) and 𝑟∗∗∗∗(𝑟𝑖); and two additional
boundaries) yields the strategy for obtaining the highest level of trust perceptions. Figure 15
plots all six resulting thresholds for the pair-wise case comparisons and highlights the
respective trust-maximizing strategies. Depending on the specific ratings, the complementor
may not display any rating, display either the imported or the on-site rating only, or both
ratings simultaneously. Irrelevant (sections of) boundaries are shown as dotted lines.
FIGURE 15: TRUST-MAXIMIZING RATING DISPLAY34
Control variable analysis and robustness checks
Extending scenario (4) in Table 6 by various (mean-centered) control variables shows that,
first, trust in the incumbent platform (𝛽=.014; 𝑝<.05), general disposition to trust (𝛽=
.029; 𝑝<.001), risk-affinity (𝛽=.031; 𝑝<.001) and Online Shopping Experience (𝛽=
−.007; 𝑝>.05) have significant effects, but that, second, the main coefficients are not
34 We assume the availability of both ratings.
Rating on Incumbent Platform
On-site Rating
Display incumbent rating only
Case (3)
Do not display any rating
Case (1)
Display on-site rating only
Case (2)
Display both ratings
Case (4)
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
89
substantially affected (the same holds for all other cases; see Appendix). Next, there did not
occur systematic differences with regard to the participants’ characteristics across treatments
(see Appendix).
Further, we conducted a robustness check for the interaction of star rating score and number
of reviews received, as the latter may enforce the validity of an already good/bad rating score
(positivity/negativity effects; Khare et al. 2011). The number of reviews had no systematic
effect on trust, nor did it affect the models’ main results (see Appendix).
As another supplementary analysis, we also investigated gender effects. In particular, we
analyzed how evaluations of trust were affected when, for instance, both host and potential
guest had the same gender or when a female guest is presented with a male host profile. None
of these variables had any significant effect (see Appendix).
As an additional check to account for the different sample sizes across the 81 treatment cells
in case (4) ({1.0, 1.5, … , 5.0}×{1.0, 1.5, … , 5.0}), we repeated the analysis using the data on an
aggregated level, also indicating robustness.
Discussion
This study provides a rationale for when to leverage reputation portability to increase trust
across platform ecosystems. Specifically, we take a complementor’s perspective and introduce
a model for the trust-building capabilities of importing an external rating and estimate the
model’s parameters. In line with nascent work on reputation portability (Otto et al. 2018;
Teubner et al. 2020), our results confirm that imported reputation can be leveraged to increase
consumer trust in complementors (H1). Extending previous research, we derive indicative
thresholds for complementors to decide on when to leverage external reputation to their
benefit – and when to refrain from doing so – which hinges on both on-site and to-be-imported
rating scores (H2). In doing so, this study is the first to identify trust-maximizing strategies
when both on-site as well as importable ratings are available.
Given today’s fragmented platform landscape, we acknowledge that the outlined scenario
remains somewhat hypothetical. As of 2020, only few platforms enable the import of ratings
(e.g., Bonanza.com, TrueGether.com). Hence, the concept can currently be described, at best,
as a niche application. However, since novel regulation such as the General Data Protection
Regulation (GDPR; European Union 2016a) fosters data subjects’ rights to data portability –
potentially including their reputational data – it is important to consider the strategic
consequences for incumbent and entrant platforms. After all, making decisions on the cross-
platform integration of reputation needs to take into account how complementors would
actually use this functionality. Especially, given the all-pervasiveness of platforms (Alt and
Zimmermann 2014), an abundancy of industries and services may be affected, including even
dark-web markets (De Reuver et al. 2018; Norbutas et al. 2020; Sundararajan 2016). Since
most users are active across multiple platforms (Teubner et al. 2019), we expect the concept of
reputation portability to draw increasing practical and academic attention.
Design of reputation portability—Directly importing ratings from one platform to
another has emerged as one of several potential means of drawing on one’s reputation across
platform boundaries. Researchers and practitioners may further explore both the design and
effectiveness of such tools – and define standards for interoperability and interfaces (Drechsler
2018; Vanberg and Ünver 2017). Specifically, they should implement measures to keep
Chapter 4: Trust measures and import thresholds
90
malicious actors from gaming such a system. Our results indicate that it is in the best interest
of the platform to limit the import option to a set of carefully curated source platforms and to
prevent exploitative strategies of frequent and/or abusive account creation and abandoning.
Similarly, mechanisms need to be established to avoid that someone else’s reputation is being
imported (e.g., single-sign-on or confirmation links).
Other approaches to reputation portability—Directly importing ratings is by no means
the only technological approach to facilitate reputation portability. Regulators and researchers
have discussed third-party services, so-called Personal Information Management Systems
(PIMS), to validate, certify, and aggregate reputational data from different sources into single
profiles, with users being in control of their personal data (European Data Protection
Supervisor 2016; Kathuria and Lai 2018; Martinelli 2019). Others envision distributed ledger
technology to enable seamless portability of digital reputation across platform ecosystems and
services (Catalini and Gans 2016; Werbach 2018). However, both these advances – just like
direct transfer between platforms – are in early stages and the list of unsuccessful PIMS
ventures is long (e.g., Deemly, erated.co, famust.com, peertru.st, tru.ly, trustcloud.com). Also,
decentralized applications (i.e., dApps) have thus far neither gained much traction nor a
critical user base (Hesse and Teubner 2020a).
Barriers to reputation portability—While the concept of reputation portability has been
touted since the early 2000s, it is striking that, so far, it has experienced only limited
proliferation (Resnick et al. 2000). As it stands, incumbent platforms (such as Airbnb, eBay,
and Uber) seek to preserve market power and invest in lock-in by (actively) inhibiting the
exploitation of their users’ reputation outside the platform (Krämer 2018). Further, the lack of
technical interoperability or, more precisely, the GDPR’s somewhat vague wording (data shall
be portable “where technically feasible”), may allow incumbents to shy away easily.
However, platform operators may not even have a say in whether “their” data is being
extracted, exported, or re-used. Consider the star ratings and/or text reviews on Airbnb. As on
most other platforms, these are publicly displayed for any Internet user to see (no login
required). Thus, if a complementor on Airbnb starts hosting on a competing platform and
wishes to benefit from their reputation on the incumbent platform, they can simply embed a
link to their Airbnb page within the profile on the entrant platform or use the HTML code
snippet Airbnb handily provides to users for such cases. They can also prove that they are
actually the rightful operator of the linked profile, for instance, by providing credentials or
confirmation through Airbnb’s communication system. Notably, this is precisely how Bonanza
and TrueGether allow complementors to import their existing reputation from eBay, with
limited options for the latter to prevent this practice. What is more, by tolerating such a
feature, incumbents may even enjoy free marketing if ratings from their platform are visibly
embedded on other platforms.
Moreover, especially users (i.e., consumers and/or complementors) could benefit in several
ways. The EU highlights explicitly that they would appreciate enhanced access to their data,
including ratings and reviews (European Commission 2018b). An empirical study by Hesse
and Teubner (2020b) confirms that consumers generally value the concept of reputation
portability – with an approval rating of around 77%. As of today, however, complementors face
a cold-start problem when beginning to operate on a new platform (Wessel et al. 2017). Hence,
while it can be debated how much of a pressing issue reputation portability is for
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
91
complementors, it may certainly be useful in multi-homing (e.g., Dellarocas et al. 2009).
Importantly, complementors may not want to rely on a single platform, especially since it may
cease to exist someday (European Commission 2018b). Further, most prior research suggests
that it is difficult to new complementors to collect “good” ratings, especially when they just
started out on a new platform and find themselves competing with other users that have
already gathered a positive track record (Gutt and Herrmann 2015; e.g., Resnick et al. 2006).
In this regard, note that 94% of all Bonanza sellers make use of the platform’s rating import
feature (Hesse and Teubner 2020b). Hence, it is safe to assume that they see value in doing
so.
Additional perspective on reputation portability—To augment the discourse, we
conducted a series of ad-hoc interviews with relevant platform practitioners. These included
current and former executives and researchers from Airbnb, Uber, Bonanza, the CEO of a
leading European job platform (InStaff), and the Director of the German Data Protection
Foundation (Stiftung Datenschutz). Interviewees unanimously acknowledged the potential
benefits of reputation portability for consumers and complementors – as well as for smaller
and entrant platforms. However, it was also hypothesized that especially leading platforms (as
of now) do not see a clear business value in making reputation portable – be it due to legal
uncertainty about the scope of portable data or because they simply refuse to share such
information in a competitive environment. There was consensus, though, that empirical
evidence (e.g., on either revenue increase or cost reduction resulting from ported reputation)
may be helpful in convincing them otherwise. It is somewhat surprising that major platform
operators seem to exhibit a lack of imagination with regard to the trust-building potential of
imported reputation and its benefits. Regarding the prevalence of reputation portability, it was
mentioned that a “consortium of interested platforms” may collaborate to explore the
opportunities of portable reputation among themselves. These platforms may do so by
developing and implementing the currently lacking “common design standard” for direct
import/export of reputation, thereby pushing its visibility and acceptance.
Policy making and reputation portability—Several recent EU reports (2016, 2017,
2018b) have called explicitly for reputation portability to be further explored and criticize the
current lack of interoperability and data access, in particular for its anticipated competitive
value in breaking up platform monopolies. In fact, incumbents such as Airbnb and Uber have
gained profound market power, which has created a lock-in that hinders complementors from
switching to a different platform. This, in turn, represents an entry barrier for new (i.e.,
entrant) platforms (Gans 2018; Van Alstyne et al. 2016). Hence, reputation portability could
change the strategic interplay between complementors, incumbents, and entrant platforms –
mitigating lock-in, facilitating multi-homing, and increasing platform competition (European
Commission 2016, 2017, 2018b). From a consumer protection perspective, stipulating the
right to reputation portability hence appears consequential.
While the GDPR’s Article 20 mandates portability of data provided by users (themselves),
ratings may not be subject to this right as they are provided by others. Some scholars have
voiced their concerns on this limiting scope (Graef et al. 2013; Inglese 2019; Kathuria and Lai
2018; Vanberg and Ünver 2017). However, in the ongoing legal debate, others argue that user
reputation is in fact subject to the right to data portability (Martinelli 2019; Van der
Auwermeulen 2017). It is also argued that “provided data” can be interpreted broader and that
the GDPR’s Article 15 grants users a general access to all data relating to their person (De Hert
Chapter 4: Trust measures and import thresholds
92
et al. 2018; Ursic 2018). Similarly, it was noted in the ad-hoc interviews that, since the
regulation is intended to foster data subjects’ rights and sovereignty, further updates will need
to reflect this notion and explicitly include reputational data. Hence, it is conceivable that
regulations such as the GDPR or California’s Consumer Privacy Act (CCPA) will encompass
reputation at some point.
Monetization of reputation portability—We have shown that imported reputation is a
driver of trust and may thus be used by complementors. This, in turn, also prompts the
question whether platforms should let complementors import or export reputation “for free”
or how they could, otherwise, monetize such a feature. While, initially, incumbents are likely
to prohibit the export of ratings to other (potentially competing) platforms (Dellarocas et al.
2009; as in the eBay vs. Amazon dispute; Resnick et al. 2000), they may yet discover
unexpected value in allowing for it. Indeed, they may explore how to capture the economic
value of portability, for instance, through a fee on importing/exporting reputation
35
.
Complementors’ decision-making under such regimes, associated platform strategies, and
competitive equilibria for optimal levels of data disclosure and portability promise interesting
further insights (see also Krämer and Stüdlein 2019).
Theoretical embedding of reputation portability—As a theoretical underpinning for
the notion of reputation portability, we have drawn on signaling theory (Mavlanova et al. 2012;
Spence 1973). Indeed, parallels can be drawn to the theory’s original example of job market
signaling. Similar to an educational track record, imported ratings are “observable
characteristics attached to the individual that are subject to manipulation by him [or her]”
(Spence 1973, p. 357). The complementor can make an effort to import an additional signal,
that is, by achieving a high star rating on a (reputable source) platform. However, such efforts
incur signaling costs – and these cost are higher for low-quality complementors as “costs of
signaling are negatively correlated with […] capability” (Spence 1973, p. 358). Similarly,
building and maintaining a “good” reputation is costly (Dunham 2011) – particularly when
doing so on several platforms.
This multiplicity of sources and trust cues warrants further investigation with regard to user
psychology. Visual hierarchies in information processing and the role of conflicting signals
(i.e., one exceedingly “good”, one “bad” rating) may play an important role for consumers’
decision-making and purchase intentions (Faraday 2000; Khare et al. 2011; Mudambi et al.
2014; Riasanow et al. 2015). We expect studies that track user attention and their foci on
ratings, reviews, and other trust cues to shed more light on such questions (Willems et al.
2019).
Concluding note
Recent regulation and the increasing market power of incumbent platforms have led to a
rediscovery of the idea of reputation portability. With this research note, we seek to contribute
to the ongoing debate on the nature and scope of data portability within the online platform
economy. Naturally, our study can only offer a very limited view into the complexity of actual
platform interactions. The chosen stylized experimental design is limited in providing a basic
stimulus of a transaction – stripped of potential other design elements and noise typically
35
However, the idea of having to pay for their own reputation may not necessarily resonate with
complementors.
PART II: THE ECONOMICS OF REPUTATION PORTABILITY
93
encountered in the field. While this setting provides a suitable testbed to study the anticipated
relation of (imported) reputation and trust, actual behavior will likely be influenced by
additional factors not accounted for in this research. Future work may embed and explore the
role of other, additional trust cues. Moreover, field experiments on actual platforms may shed
light on the (real-world) effectiveness of portable reputation.
Needless to say, obligatory or involuntary reputation portability is no desirable future. In this
regard, the multi-faceted debate around the Chinese Social Credit System comes to mind,
penalizing and/or rewarding individuals, inter alia, based on their online activity (Wong and
Dobson 2019). Note that full transparency of anyone’s online conduct would have far-reaching
implications and pose a considerable threat to the Western world’s liberal values. Eventually,
the EU may seek to find its own position in dealing with this complexity. The commencement
of the GDPR (“right to access”, “right to data portability”, “right to be forgotten”) can be seen
as a laudable, yet improvable step into this direction. Ultimately, the question may not be
whether users will have to manage their digital reputation across platforms but how
information systems should be designed to allow them to harness this data in an effective, fair,
and sustainable fashion.
94
PART III: THE DESIGN OF REPUTATION PORTABILITY
95
PART III: THE DESIGN OF
REPUTATION PORTABILITY
96
PART III: THE DESIGN OF REPUTATION PORTABILITY
97
Chapter 5: Signal discrepancy
and cognition
The third part of this thesis considers how reputation portability can be designed and
implemented across digital platform ecosystems. In Chapter 5, I set out to investigate the
phenomenon from a user psychology perspective. Leveraging experimental survey data, this
study examines the cognitive and economic effects of displaying reputation signals from two
different sources. Informing reputation systems’ design, the results show that conflicting
rating information confuses consumers and, ultimately, impairs purchase intentions.
Maik Hesse, Otto Hans-Martin Lutz, Marc T. P. Adam, Timm Teubner
36
Introduction
In today’s digitized world, consumers encounter reputation in various platform ecosystems (De
Reuver et al. 2018; i.e., the broader research context of this study; see Hein et al. 2020).
Starting off from the value of trust and reputation within platforms (e.g., Tadelis 2016), it has
been established that importing, transferring, and aggregating reputation between different
platforms can be utilized as a source of value too (Reputation Transfer; Teubner et al. 2019).
In that regard, reputation portability refers to the concept of leveraging existing reputation
from one platform and using it as a signal of service quality and trustworthiness to attract
demand on another platform. Examples include services such as Traity.com, aggregating user
ratings from different platforms and social networks into single trust scores or “reputation
profiles”. Moreover, end-to-end solutions for reputation import between platforms exist, for
instance, provided by the e-commerce sites Bonanza.com and TrueGether.com. These
platforms allow sellers to import star ratings from incumbent platforms (e.g., eBay or Amazon)
for inclusion in their profile.
While extant research investigated the effects of reputation in a single platform ecosystem –
including those of conflicting product or service reviews – the availability of ratings from
different sources (i.e., on-site reputation and imported reputation) prompts the question of
how users react to such information. Existing research on reputation portability shows that
imported ratings do in fact function as a driver of consumer trust and purchase intentions (Otto
et al. 2018; Teubner et al. 2020; Zloteanu et al. 2018). However, this research did not consider
whether and to what extent the effectiveness of reputation portability may be impaired by
discrepant rating information. When signals are sourced from more than one platform, they
may be complementary (e.g., a single aggregated rating score shown next to a profile image),
re-enforcing (e.g., two or more exceedingly good ratings being displayed together), or
conflicting (e.g., one excellent and one bad rating appearing simultaneously). Yet, research on
the impact of rating discrepancy on users’ purchase intentions and the role of attention and
36
This study was published in the proceedings of the International Conference on Information Systems
with the title “Gazing at the Stars: How Signal Discrepancy Affects Purchase Intentions and Cognition”
(Hesse, Lutz, et al. 2020) and won the award for best short paper in the Sharing Economy, Platforms
and Crowds track; see https://aisel.aisnet.org/icis2020/sharing_economy/sharing_economy/8/.
Chapter 5: Signal discrepancy and cognition
98
cognitive processing is scarce. To address this gap, we set out to answer the following research
question: How does the degree of discrepancy in ratings from different platforms affect
consumers’ cognitive processing and purchase intentions?
We apply a multi-method approach via two experimental studies. Thereby, we leverage
prospect theory (especially the assumption of risk averse consumers) and hypothesize on the
role of increased cognitive load and choice deferral for human information processing in the
presence of conflicting signals. In Study 1, we employ survey instruments in an online
experiment, assessing the effects of multiple ratings – as well as discrepancies in their values
– on purchase intentions. In Study 2, we propose a laboratory experiment with eye-tracking
measurements to determine the role of cognitive processes for human attention and behavior
(indicated by fixation durations and transitions) when either re-enforcing or conflicting
information is present. In a nutshell, our findings show that 1) larger rating discrepancies are
detrimental to consumer’s purchase intentions, however, 2) imported reputation does
represent a driver of purchase intentions, even in the presence of discrepant ratings. Further,
combining self-reported data from an online experiment (Study 1) and eye-tracking to explore
the mediating influence of gaze patterns in the proposed laboratory experiment (Study 2) will
allow us to take into account different perspectives on the phenomena of reputation portability
and conflicting information (Niehaves 2005; Vasseur et al. 2019).
Related work and hypotheses
Trust and reputation in platform ecosystems
Since the advent of e-commerce marketplaces such as Amazon and eBay in the mid-1990s,
trust-building and reputation mechanisms have played an important role in IS and Economics
research (Resnick et al. 2000; Tadelis 2016). With the proliferation of digital platform
ecosystems, today’s major players – from Airbnb to Uber – employ a wide range of
mechanisms to build and maintain trust as a prerequisite for platform activity (Hesse, Dann,
et al. 2020). The most frequently used signal in this regard is on-site reputation – often in the
form of third-party assessments such as (star) rating scores or text reviews (Mavlanova et al.
2012; Teubner et al. 2019). Complementors (i.e., service providers) on these platforms rely on
their reputation as it provides potential customers with an opportunity to assess past behavior
(Tadelis 2016). However, acquiring a reputation as a sign of trustworthiness is costly (Dunham
2011). Still, complementors benefit economically from accumulated reputation. Ample
research demonstrates that “good” on-site reputation drives perceptions of trustworthiness
and purchase intentions (C. Chen and Chang 2018; Jarvenpaa et al. 2000; Resnick et al. 2006;
Ter Huurne et al. 2017). That, in turn, increases complementors’ ability to attract demand
and/or enforce price markups (e.g., Tadelis 2016).
Reputation portability—People use an increasing number of different platforms and, yet,
their valuable online reputation is restricted to one platform at a time. This has led to repeated
calls for reputation portability (Dellarocas et al. 2009; Puschmann and Alt 2016; Resnick et
al. 2000). Only recently, several contributions have picked up this concept. While Teubner et
al. (2019) delineate the topic from the general dynamics of trust and reputation, Hesse and
Teubner (2020a) present a conceptual model and identify a need for more empirical research.
Some groundwork, however, has been done on the economic potential of portable reputation.
Several studies find support for the impact of imported signals – usually, star ratings – on trust
or purchase intentions (Otto et al. 2018; Teubner et al. 2020; Zloteanu et al. 2018).
PART III: THE DESIGN OF REPUTATION PORTABILITY
99
Discrepant ratings and economic prospects
Importantly, when introducing multiple sources of reputation, one must consider how users
process this information and deal with re-enforcing or conflicting signals (i.e., discrepancies).
Note that prior studies on reputation portability did neither simultaneously display two rating
scores (only on-site or imported rating visible) nor were ratings’ values varied (only 4 or 5 stars
used), that is, no discrepancy was apparent (Otto et al. 2018; Teubner et al. 2020). Thus,
existing research lacks insights into (1) how an interplay of specific on-site and imported
ratings affects users’ economic decision making and (2) the role of discrepancy in displayed
information and how it affects consumers’ information processing and attention.
To embed consumers’ decision making and purchase intention in the presence of discrepant
information signals in well-established economic literature, we draw upon Kahneman and
Tversky (1979)’s prospect theory. One of the theory’s main tenets is the assumption of risk
aversion. Specifically, when given a choice, the theory posits that people generally prefer a
certain outcome (say, $50) over any lottery with the same expected value (say, a 50:50 gamble
on $0 or $100). This notion goes back to the concept of expected utility (rather than expected
value) with decreasing marginal utility gains.
37
Applying this frame of thought, we can
understand a platform-mediated transaction as an economic prospect where available rating
scores serve as signals of potential outcomes. In this vein, a prospective host, driver, or seller
with (at least) two such scores may be evaluated similarly to a lottery where the available rating
scores define the space of economic outcomes. Applying the concept of risk aversion and given
equal expected values, people will choose the prospect with lower variance. Since variance is
directly related to the discrepancy between ratings, higher discrepancy will be associated with
a lower likelihood of the prospect being chosen, that is, the apartment or ride being booked, or
the product being purchased (again, assuming identical average values).
Example: A host on Airbnb has two ratings, an on-site and an imported rating, both of which
are 4 stars. Another host has two ratings of 3 and 5 stars, respectively. Hence, both hosts have
an average score of 4 stars. However, risk averse consumers are likely to have more confidence
in the host with less variance (i.e., with no discrepancy). Even though consumers may have
higher prospects if the actual service quality was 5 stars as in one of the options, they will be
aware of the fact that the “bad” rating (i.e., 3 stars) indicates lower potential prospects and,
hence, will tend to the safer bet, that is, two equal rating scores. We expect this notion of
consumer decision making to be reflected in their willingness to purchase and thus suggest:
H1: A higher discrepancy between the on-site and the imported rating has a negative
impact on consumers’ purchase intentions.
37
The theory’s applicability to human decision making has proven wildly successful and Kahneman was
awarded the Nobel prize, “having integrated insights from psychological research into economic science
[…] concerning human judgment and decision-making under uncertainty” (NobelPrize 2002). The
authors’ seminal work reached over 50.000 citations in the academic literature.
Chapter 5: Signal discrepancy and cognition
100
Study 1: Online experiment on rating discrepancy and user
intention
Addressing H1, Study 1 assesses how simultaneously displayed on-site and imported rating
scores affect consumers’ willingness to purchase. Using an online experiment, participants
take the role of potential guests on the accommodation sharing platform happystay.com and
state their purchase intentions (i.e., the willingness to book a stay with the displayed
prospective host). For that purpose, consumers were presented with a host profile that
displayed both an on-site rating as well as imported reputation from another accommodation
sharing platform (Airbnb).
Treatment Structure—Using a full-factorial between-subjects design of on-site and
imported rating score combinations, the prospective host had an on-site rating (randomly
drawn from {1.0, 1.5, …, 5.0}) and, additionally, had imported their rating from another
platform (randomly drawn from the same set). Both ratings were displayed simultaneously
with some distance (Figure 16; left). The rating values’ distribution across treatments reflected
the skewness towards positive scores commonly observed on many platforms (Schoenmueller
et al. 2018). That is, even though all possible combinations of rating scores were tested, those
above 3.0 stars occurred more frequently. At the same time, we ensured that a sufficient
number of discrepancy cases could be observed per absolute value of the difference in ratings
(Figure 16; right).
FIGURE 16: EXPERIMENTAL SETUP AND TREATMENT DISTRIBUTION
Stimulus Material—Figure 16 shows exemplary stimulus material used for this experiment.
Host names and (blurred) profile images were randomly drawn from a selection of common
first names and typical images. Each participant was presented with one host’s profile only.
Names and images had no significant effects on the dependent variable. The (random)
numbers of reviews were displayed to create a more realistic setting as participants are used to
this kind of representation of star ratings on platforms. Their effect was negligible in
magnitude and significant for on-site reviews only.
Measures—Consumers’ purchase intentions were measured on an 11-point Likert scale
(Gefen and Straub 2003). To delineate participant-specific effects, we control for trust in the
incumbent platform, general disposition to trust, as well as individuals’ risk-affinity. We
On-site rating (stars)
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
∑
Imported rating (stars)
1.0
2 2 2 2 2 4 4 4 4
26
1.5
2 2 2 2 2 4 4 4 4
26
2.0
2 2 2 2 2 3 4 3 4
24
2.5
2 2 2 2 2 4 4 4 4
26
3.0
2 2 2 2 2 3 4 4 4
25
3.5
3 3 4 4 4 7 6 6 7
44
4.0
3 3 4 4 4 7 6 6 7
44
4.5
2 3 4 3 4 7 7 7 7
44
5.0
3 3 4 4 4 7 7 7 7
46
∑
21
22
26
25
26
46
46
45
48
305
Rating discrepancy
Difference
0
0.5
1
1.5
2
2.5
3
3.5
4∑
Cases
37
63
55
45
34
29
22
13
7
305
PART III: THE DESIGN OF REPUTATION PORTABILITY
101
measured rating discrepancy as the absolute difference between the values of on-site and
imported ratings.
Sample—We recruited 309 participants with prior experience in online shopping using
Prolific.ac (Palan and Schitter 2018). Participants received 0.80£ for successfully completing
the experiment. On average, participants answered the survey in 76 seconds (median: 68,
shortest: 22) and responses taking longer than 5 minutes were excluded from further analysis
(4 cases). The final sample includes 305 participants (163 female, 142 male) from diverse
Western countries and regions (e.g., Americas, Australia EU, CH and UK) and age ranged from
18 to 70 years (mean: 35.9, median: 34). The sample size was adequate for detecting a 5%
deviation in purchase intentions with a statistical power of 97%.
Results
Table 7 reports the OLS regression results where the dependent variable (consumers’ purchase
intention) was standardized to the interval [0,1] and rating scores ranged from 1.0 to 5.0
(stars). First, Model I shows that rating discrepancy indeed has a significant negative impact
(b=-.064, p<.001; Cohen’s f²=.034, small effect size) on purchase intentions, supporting
hypothesis H1. As expected, the average score (Model II) has a positive impact on purchase
intentions (b=.165, p<.001; f²=.315, medium) with a comparably high fraction of variance
explained (R2=.251). Importantly, controlling for both rating discrepancy and average scores
the results hold (Contr. III), meaning a difference in displayed star ratings is detrimental to
complementors’ economic prospects – whether comparing 1 and 3 stars vs. twice 2 stars or 3
and 5 stars vs. two times 4 stars. As per Controls IV and V, both on-site and imported
reputation have positive and significant effects on consumers’ purchase intentions, supporting
the findings in previous literature. Remarkably, the imported rating exhibits a higher
coefficient (b=.087, p<.001; f²=.176, medium) than the on-site rating (b=.066, p<.001;
f²=.103, small) and explains more variance (R2=.185 vs. R2=.123), which may partly result from
participants’ exalted levels of trust in incumbent (7.04 on the 11-point Likert scale; ~60.4%)
compared to the unknown (i.e., fictitious) target platform happystay.com.
As a robustness check, we repeated the previous analyses including all measured control
variables, that is, participant-specific controls (gender, trust in incumbent, general disposition
to trust, and risk-affinity) as well as numbers of reviews (both on-site and imported).
Extending all models by these variables shows that, first, participant-specific controls have
expected effects, but that, second, the main models’ coefficients were not substantially affected
in terms of sign, magnitude, or significance. Further, no systematic differences occurred for
evaluations of purchase intentions and the effect of discrepancies when controlling for
different geographical regions of participants or excluding the quickest 5% of responses (<34
seconds).
Chapter 5: Signal discrepancy and cognition
102
One aspect to consider is an alternative explanation to the findings with regard to rating
discrepancy and purchase intentions. It can be argued that, rather than discrepancy itself, the
lower of the two ratings (i.e., the minimum score) mainly determines consumers’ willingness
to purchase. We concur with this notion as it was shown, for instance, for vendor reviews, that
consumers draw particularly on negative information (Steward et al. 2018). Also, anecdotal
evidence suggests that negative star ratings or reviews are more influential as consumers
attribute a higher weight to them (negativity effect; Khare et al. 2011; Sen and Lerman 2007).
Hence, as a control analysis, we tested the fraction of variance explained by different
aggregations of on-site and imported rating scores when only this single variable was used.
Comparing the ratings’ minimum (adjusted R2=.246), maximum (.127), average (.251), on-site
value only (.088), and imported value only (.169), shows the high explanatory power of the
minimum score. Note, however, that it is impossible to control for the minimum rating while
holding both difference and average rating score constant since two of these variables will
always directly determine the third one.
Study 2: An eye-tracking study on attention to conflicting
signals
In this second study, addressing H1 and additionally H2, we employ eye-tracking to add an
additional explanatory layer of observational data (Niehaves 2005), potentially uncovering
mediating variables such as fixation duration, distribution and gaze patterns. This will allow
us to understand the role of cognitive processes in assessing different rating combinations in
the presence of reputation portability and how these ultimately influence users’ behavioral
intention. Recently, eye-tracking was used to show how discrepancies in online ratings (of non-
profit organizations) attract users’ visual attention and that individuals rather focus on two
unequal star ratings (i.e., a visual discrepancy) than on two identical ones in the presence of
other (non-reputational) information (Willems et al. 2019). Importantly, visual attention to
certain cues can be interpreted as central cognitive processing related to decision making based
on these information (Rötting 2001; Schulte-Mecklenbeck et al. 2017). Following the eye-mind
assumption (i.e., objects hit by the visual axis are subject to central cognitive processing in that
Dependent variable: Consumers’ Purchase Intentions
Model I
H1
Model II
Avg. Score
Contr. III
H1 (Avg. Score)
Contr. IV
H1 (on-site)
Contr. V
H1 (imported)
Contr. VI
Interaction
Rating Discrepancy
-.064
***
-.031
*
-.054
***
-.039
**
.026
(.016)
(.014)
(.015)
(.015)
(.069)
Average Score
.165
***
.156
***
.171
***
(.016)
(.017)
(.025)
On-site Rating
.066
***
(.013)
Imported Rating
.087
***
(.012)
Discrepancy x Average
-.018
(.022)
Constant
.494
***
-.150
**
-.076
.257
***
.168
**
-.126
(.028)
(.056)
(.066)
(.053)
(.052)
(.089)
Observations
305
305
305
305
305
305
Adjusted R2
.050
.251
.260
.123
.185
.260
*** p<.001; ** p<.01; * p<.05; standard errors in parentheses
TABLE 7: OLS REGRESSION MODELS (PURCHASE INTENTIONS)
PART III: THE DESIGN OF REPUTATION PORTABILITY
103
given moment) and the immediacy assumption (i.e., fixation duration corresponds to the
amount of central cognitive processing; Rötting 2001, p. 76), eye-tracking data helps to draw
inferences on cognitive processes. More specifically, (Bera et al. 2019) have shown that eye-
tracking can be used to measure cognitive integration processes of information from two or
more sources. We follow Willems et al. (2019)’s call to consider more variety in displayed
ratings and to investigate their impact on decision making. Additionally, our study answers to
several other calls for research on reputation and its transferability. First, we follow the EU’s
urge to investigate potential benefits and underlying mechanisms of reputation portability to
break up platform monopolies and avoid user lock-in (European Commission 2016, 2017).
Second, we take up Tadelis’ comment to “combine economic thinking and analyses with
human-computer interaction” in platform ecosystems (2016, p. 338).
Scenario—The basic scenario is identical to that of Study 1. Again, participants will take the
role of guests on a sharing platform, evaluate prospective providers, and state purchase
intentions based on the simultaneous availability of signals from two sources (i.e., the on-site
and the imported rating score).
Treatments—Similar to Study 1, we systematically vary on-site and imported rating scores.
We thereby employ a 4 by 3 full-factorial within-subjects design (on-site rating = {NA, low,
medium, high}, imported rating = {low, medium, high}). Every participant experiences all
4×3=12 treatment conditions in random sequence. By introducing the “NA” condition as a
control, we include the possibility of complementors not having obtained any on-site rating
yet. Each of the 12 booking scenarios is displayed on one web page.
Eye-Tracking—Gaze data will be recorded using a SensuMotoric Instruments RED-m
binocular eye-tracker at a 120Hz frame rate (Mele and Federici 2012). Stimuli are presented
on a 21-inch external screen. Participants are able to move freely but will be asked to keep their
head movements to a minimum. For each participant, we perform a 5-point calibration and
successive validation procedure at the experiment’s start. As glasses are prone to cause major
issues with eye-tracking stability and data quality, recruiting is limited to persons not wearing
glasses and participants are asked not to wear mascara or make-up in the eye region as this
can significantly impair pupil detection (Holmqvist et al. 2011, p. 119ff). Areas of interest
(AOIs) for gaze data analysis are defined for each experimental page: 1) on-site rating score, 2)
imported rating score, 3) the prospective provider’s face, 4) platform logo, and 5) the Likert
scale on which participants state and submit booking intentions. The on-site and imported
rating elements are designed in a very similar manner to ensure internal validity (Orquin and
Holmqvist 2018) and have comparably high visual saliency to stand out in visual hierarchy
(Faraday 2000). The corresponding star rating AOIs have identical dimensions.
Procedure—After arriving at the lab, participants sign a consent and data protection form.
Following a brief presentation of the eye-tracking setup, the experiment starts with an
introductory page explaining its sequence, scenario, and task. Subsequently, experimental
treatment pages are presented in random order. We estimate the overall procedure to take 30-
40 minutes (i.e., arrival and briefing, eye-tracking setup and calibration, questionnaire,
debriefing, and payout). Participants are rewarded by a cash payout (e.g., 7€). The
experimental setup complies with ethical standards (approval obtained; available upon
request).
Chapter 5: Signal discrepancy and cognition
104
Measurements—On each page, participants state purchase intentions as in Study 1. The
scale appears at the page’s bottom after a delay of six seconds. After completing all 12 pages,
we survey participants on demographic and common controls. They may add general thoughts
or comments on factors that influenced their decision process. The main focus variables are
individual and accumulated fixation durations on the two rating AOIs, the number of
transitions between them (AOI run count) and normalization of fixations (for subject
comparability; Bera et al. 2019). Besides analyzing the impact of rating discrepancy on
purchase intentions (as in Study 1; H1), these gaze patterns will be used to analyze mediating
effects on participants’ purchase intention and, hence, to shed light on subcomponents of
hypothesis H2 (i.e., attention to ratings, cognitive conflict, decision length). We will apply
multivariate regressions to conduct the analyses. Order effects (e.g., fatigue or learning) are
counterbalanced by randomizing and controlling for page sequence.
Outlook—Due to the COVID-19 lockdown of our institution, the experiment could not be
conducted as originally planned. Still, the experimental setup was pre-tested, successfully
demonstrating functionality. Initial test data for Study 2 uncovered promising results with
regard to hypothesis H2. Power analysis suggests a sample of n~40 for this experiment as
adequate to detect a 20% difference in effect size with a statistical power of .95 (ntotal~492).
Participants will be recruited from the local subject pool at TU Berlin using ORSEE (Greiner
2015). We plan to proceed with the experiment as soon as the university re-opens and allows
on-premises lab research.
Discussion and conclusion
The multiplicity of platforms in today’s digital economy has led to repeated calls for reputation
portability across platform boundaries (Dellarocas et al. 2009; Puschmann and Alt 2016;
Resnick et al. 2000). While previous research has shown that complementors could,
theoretically, benefit from reputation portability, the notion of the interplay between specific
rating scores – including the possibility of conflicting information – has not been considered
thus far. We set out to address this gap by means of a multi-method approach. We devised (1)
an online experiment, assessing consumers’ purchase intentions and (2) a laboratory
experiment using eye-tracking to explore consumers’ cognitive processing of ratings with
different levels of discrepancy.
Key findings—Answering one part of the research question, our study’s main result is that
higher rating discrepancy is in fact detrimental to purchase intentions (H1). We offer a first
explanation for this finding based on general human risk aversion and the modeling of
platform-based transactions as economic “prospects”. Naturally, other explanations are
conceivable and were discussed (e.g., minimum rating). Our findings prompt a subsequent
question, namely which cognitive processes – in addition to subjects’ stated perceptions – may
be at play (H2). To investigate this line of thought further, we propose and sketch out a
follow-up experiment to gather an additional layer of data that may help explain these nuances
in human decision making based on eye-tracking. With regard to the economic potential of
imported reputation, we find a positive relationship – which is well in line with previous
literature (Otto et al. 2018; Teubner et al. 2020). We address these studies’ limitations of
considering only one single rating and/or value at a time (e.g., on-site or imported; 5.0 stars)
by investigating the concurrency of these ratings as well as a wide array of values for both scores
(1.0 to 5.0 stars).
PART III: THE DESIGN OF REPUTATION PORTABILITY
105
Considering managerial implications, our paper makes a contribution regarding design
features of online reputation systems (Dellarocas et al. 2009; Tadelis 2016). In a scenario
where reputational data is likely to be free-floating across platforms, platform operators must
decide how to best represent users’ accumulated reputation from across different sources while
avoiding overburdening consumers with too many pieces of information. Limitations in
human information processing will likely have implications for designing adequate
mechanisms. For instance, the level of interactivity of elements (e.g., several ratings or both
ratings and reviews) may affect users’ working memory. Further, the modality effect may
impact user judgment in the presence of different types of reputation signals, for instance,
when written (e.g., text reviews) and visual information (e.g., star ratings) must be integrated
by requiring consumers’ split attention (see also Sweller et al. 1998). In particular Study 2 may
shed light on the question whether or not individual scores should be retained, or information
be aggregated into one single score. There may emerge a tradeoff between information
interpretability and the potential harms of conflicting cues. In this sense, the optimal
reputation display may even depend on the structure of scores (e.g., aggregating conflicting
scores but retaining consistent ones). Finally, by pursuing a multi-method approach,
combining survey instruments with eye-tracking data and qualitative insights, we are positive
to contribute to the field of online platforms, trust, and reputation systems – both from a
theoretical and methodological perspective.
106
PART III: THE DESIGN OF REPUTATION PORTABILITY
107
Chapter 6: Designing reputation
portability
In this penultimate chapter of my dissertation, I provide concrete practical guidelines for the
design of reputation portability in digital platform ecosystems. Using a hybrid approach of
deductive and inductive reasoning, this study explores portable reputation’s impact as a
strategic lever, identifies design characteristics for direct reputation transport between
platforms, and highlights potential ways to implement solutions. Chapter 6 complements the
previous positivist approaches of investigating reputation portability by adding qualitative
data, which was collected in a series of interviews with relevant platform stakeholders, as
an additional explanatory layer.
Maik Hesse
38
Introduction
In recent years, digital platforms have shaped how products and service are offered and
consumed by both businesses and individuals (De Reuver et al. 2018; Hein et al. 2020). The
rise of early electronic markets such as Amazon and eBay as well as the advent of sharing
economy platforms such as Airbnb, Uber, and TaskRabbit have provided Internet users with a
multitude of opportunities to sell and buy online (Alt and Zimmermann 2014; Sundararajan
2016). An important and inherent part of orchestrating the supply and demand in these
platform ecosystems is the establishment of trust between participants. Complementors (i.e.,
service providers; sellers, hosts, drivers) can, for instance, achieve this by showcasing to
potential consumers (e.g., buyers, guests, passengers) their solid reputation in an online
marketplace (e.g., Resnick et al. 2000; Tadelis 2016). Hence, platforms today support users
with a variety of trust-building mechanisms, allowing them to build, maintain, and display
their reputation (Hesse, Dann, et al. 2020).
What is more, complementors may use hard-earned reputation to demonstrate their
trustworthiness and attract demand in an environment other than the one it originated from
(Reputation Portability; Hesse and Teubner 2020a; Teubner et al. 2019). In fact, some smaller
platforms already allow complementors to import (star) ratings and/or (text) reviews
39
. While
extant research has considered design features of reputation systems across a range of existing
platforms (Dellarocas et al. 2009; Jøsang et al. 2007; Tadelis 2016), little is known about how
to display multiple reputation signals on a single platform. In other words, assuming
reputation data were “free-floating across platforms, platform operators must decide how to
best represent users’ accumulated reputation from across different sources” (Hesse, Lutz, et al.
2020, p. 7). With this study we address the gap by answering the following overarching
38
This study was first published as a working paper on ResearchGate with the title “Exploring the design
of reputation portability in digital platform ecosystems – A hybrid approach” (Hesse 2020); see
https://doi.org/10.13140/RG.2.2.20583.47520.
39
For instance, the e-commerce platforms and self-labeled eBay alternatives Bonanza.com and
TrueGether.com.
Chapter 6: Designing reputation portability
108
research question: “How should reputation portability be designed and implemented in
digital platform ecosystems?” In doing so, we first and foremost consider a direct import
scenario between operators, in which one platform allows complementors to leverage their
reputation from another platform.
Our exploratory study uses a hybrid approach of deductive and inductive reasoning to inform
the design of reputation portability between platforms (Gregory and Muntermann 2011). To
begin with, we review the emerging body of literature on reputation portability as well as prior
work on (the design of) platforms’ reputation systems (deduction). This allows us to 1) develop
the basic analysis dimensions to explore a design blueprint of reputation portability in digital
platform ecosystems, and 2) identify suitable stakeholder groups to be involved in deriving
concrete design guidelines for portable reputation. Next, building on the theoretical
groundwork, we conduct a series of exploratory semi-structured interviews (induction) with
representatives from the identified stakeholder groups (direct stakeholders: platform
operators/complementors/consumers, indirect stakeholders: policymakers, legal experts,
members of academia). Following the steps outlined by Braun and Clarke (2006), we use the
qualitative data thus generated to carry out a thematic analysis. Thereby, we uncover both
overarching themes as well as sub-themes that are instrumental in the design of reputation
portability. Combining the elements of our deductive and inductive research, we present a
comprehensive thematic framework to address the research question. We discuss findings
along the five overarching themes and develop design guidelines for practitioners and
researchers. In short, our analysis comprises the why, what, and how of designing reputation
portability as well as the most relevant contributing factors.
This paper proceeds as follows. In Section 2, we review previous research on reputation
portability and reputation system design in online platform markets. In Section 3, we
introduce the method(s) applied in our research. Then, in Section 4, we present the result of
our hybrid approach by introducing the thematic framework and discussing its components.
We thereby develop five concrete and actionable design guidelines. Subsequently, Section 5
discusses findings, highlights our study’s limitations, and proposes avenues for future
research. Finally, Section 6 concludes.
Background and related work
Reputation portability in digital platform ecosystems
Reputation portability is a nascent albeit not entirely new phenomenon. The idea of leveraging
existing reputation built in one platform and using it as a signal of reliable conduct in another
– or, more precisely, the issue of not being able to do so – has previously been noted (e.g.,
Dellarocas et al. 2009; Gans 2018; Puschmann and Alt 2016; Resnick et al. 2000). Teubner et
al. (2019) formalized the concept as reputation transfer and delimited it from the overall
process of trust and reputation building as well as trust transfer between platforms and their
users. Hesse and Teubner (2020a) subsequently developed a conceptual overview to delineate
relevant terms, actors, and mechanisms; thus “touching upon the what, why, and how of
reputation portability” (p. 332). We chose this triad as a starting point for the development of
our (deductive) analysis dimensions in Section 3.
Recent work used online experiments and surveys to investigate the impact of imported (star)
ratings on trust in complementors and consumers’ purchase intentions (Hesse, Teubner, et al.
PART III: THE DESIGN OF REPUTATION PORTABILITY
109
2020; Otto et al. 2018; Teubner et al. 2020; Zloteanu et al. 2018). While prior research had
only hypothesized on the associated benefits, these studies were able to empirically confirm
the positive relationships. Others have dealt with the legal intricacies of porting reputation data
(i.e., ratings and reviews) across platform borders, for instance by pointing to respective
sections of the EU’s General Data Protection Regulation (GDPR; European Union 2016a).
While some legal scholars voiced their concern that the limiting scope of current regulation
may hinder portability (Inglese 2019; Kathuria and Lai 2018; Vanberg and Ünver 2017), others
argue that user reputation is in fact subject to the GDPR’s right to data portability (Martinelli
2019; Van der Auwermeulen 2017). The European Union itself, has urged the investigation of
potential benefits and underlying mechanisms of reputation portability, anticipating its
usefulness in breaking up platform monopolies and averting user lock-in (European
Commission 2016, 2017). Finally, Hesse, Adam, et al. (2020) implemented a stylized scenario
of reputation portability to analyze the impact of rating discrepancy on purchase intentions.
The authors deduce the need for more work on designing adequate portability mechanisms as
well as the optimal display and arrangement of reputation information from different sources.
Designing reputation systems
The design of reputation and feedback mechanisms has been subject to scholarly investigation
since the advent of digitally mediated commerce. Resnick et al. (2000) were among the first to
discuss design features, which platforms like Amazon and eBay exhibited in their early days.
For instance, the authors highlighted the nuances in leaving a (written) feedback, displaying a
running total of positive/neutral/negative feedback points, or using a score that is averaged on
a rating scale of 1 to 5 (stars). Jøsang et al. (2007) first define trust and reputation classes as
well as semantics before providing a detailed account of reputation network architectures and
computation engines. In the context of reputation system design, the authors highlight the
importance of security features such as robustness against attacks and manipulation of
reputation scores. Designing feedback mechanism is also the central focus of Dellarocas et al.
(2009)’s work which derives concrete practical conclusions with regard to reputation scales,
metrics, and the amount of information that should be visible to users. Remarkably, the
authors note that allowing users to export their reputation and apply it across different
markets, will improve reputation systems – by stimulating more platform use (i.e., multi-
homing) and rendering the overall platform landscape more effective. Similarly, Tadelis (2016)
discusses ideas to improve the practical design of reputation and feedback mechanisms.
Importantly, he suggests two areas for future research: First, users’ experience in interacting
with displayed reputation information and whether they respond differently depending on the
reputation format (e.g., distribution of ratings vs. average scores). Second, the aggregation of
an individual’s reputation across different platforms to create a meta-profile, allowing these
users to jump-start their initial trust on novel platforms. Lastly, Hesse, Dann, et al. (2020)
systematically compare features and functionality of trust-building mechanisms and
reputation management tools across a set of eleven platforms. The authors use an online survey
asking participants (i.e., platform consumers) to rate the overall importance of different design
artifacts. They find that, on average, respondents deemed rating scores slightly more important
than text reviews. Though some of the above studies explicitly refer to the idea of reputation
portability, neither of them makes a suggestion as to what should be displayed and integrated
on the target platform neither do they hint at how reputation portability schemes ought to be
operationalized by platforms.
Chapter 6: Designing reputation portability
110
Hybrid approaches in developing design guidelines
There exists a rich body of literature using qualitative approaches, specifically, thematic
analyses, to explore the design of technology and human computer interactions (e.g., Brown
and Stockman 2013; Hodge et al. 2018; McCurdie et al. 2012; Tanaka et al. 2012). We complete
our review of related work by highlighting two such studies leveraging a hybrid approach in
deriving design considerations based on multiple stakeholder perspectives (Aljaroodi et al.
2020; Noorbergen et al. 2019). Noorbergen et al. (2019) explore the design of mHealth systems
using mobile biosensors for health behavior change. In their exploratory study, the authors
develop a theoretical framework and six concrete design guidelines based on semi-structured
interviews with health stakeholders. Similarly, Aljaroodi et al. (2020) investigate how avatars
can be designed in the context of Arabian culture. The authors first develop an integrative
theoretical framework based on a literature review and follow up with a thematic analysis to
establish their design guidelines. We leverage these studies’ proposal of combining deductive
with inductive reasoning to conduct a series of qualitative interviews with platform ecosystem
stakeholders and a thematic analysis based on Braun and Clarke (2006)’s approach.
Research methods
Despite the ubiquitous use of digital platforms and research on both reputation system design
and portability, no prior work has been conducted to derive practical design guidelines for
direct portability of reputation between platforms. To address this gap and answer the research
question, our exploratory study leverages a hybrid approach comprising deductive and
inductive reasoning (Gregory and Muntermann 2011). Since deduction is often referred to as
concept-driven or theory-driven, it requires the ex-ante (or top-down) establishment of a
meaningful structure for the content of the data. On the contrary, induction describes a
data-driven approach, in which patterns form directly (or bottom-up) based on iterative
analysis of the data set (Kuckartz and Rädiker 2019). That is, deduction is based on concepts,
while inductive knowledge forms empirically (Aljaroodi et al. 2020). Applying a hybrid
approach allows us to build on the established literature on reputation portability and
reputation system design, while at the same time incorporating the view of multiple platform
ecosystem stakeholders in exploring design guidelines.
Deduction: analysis dimensions and stakeholders
Based on our theoretical groundwork in Section 2, we first developed analysis dimensions as
well as an overview of relevant stakeholders in the context of reputation portability (Kaiser
2014). Figure 17 introduces these five basic dimensions (left) and the stakeholder overview
(right).
PART III: THE DESIGN OF REPUTATION PORTABILITY
111
FIGURE 17: ANALYSIS DIMENSIONS AND STAKEHOLDER GROUPS
The five analysis dimensions were derived based on the critical reading of academic literature
(i.e., Section 2) and previous conceptual work on the subject (e.g., Hesse and Teubner 2020a).
Particularly, in this case the triad of why, what, and how serves as the first step in
‘operationalizing’ the overarching research question (Kaiser 2014). To derive general design
guidelines, we hence plan to explore i) why reputation portability should be introduced (that
is, answering the question which specific purpose reputation portability serves), ii) what
should be displayed and made visible to user, and iii) how reputation portability may be
implemented on and between platforms. Further, since previous research discussed the
potential for fraud and manipulation in reputation system design, we added risks & challenges
as an important analytic dimension to our initial framework. Finally, based on our review of
data portability laws in digital markets and the ongoing legal debate on the scope of these
legislations, we identified policy & regulation as an important dimension.
As stakeholders (i.e., Figure 17 right), we identified those actors most frequently involved with
or subject to instances of portable reputation. We follow previous research that provided
conceptual overviews of all relevant actors (Hesse and Teubner 2020a) and leverage a simple
classification of direct and indirect stakeholders (Noorbergen et al. 2019). Direct stakeholders
are those actors immediately involved with the utilization of portable reputation: platform
operators (i.e., enabling reputation portability features in a platform ecosystems), platform
complementors (i.e., leveraging said reputation portability features on their profiles), and
platform consumers (i.e., interacting with the complementor profiles using reputation
portability). Indirect stakeholders are those influencing and/or leading the public discourse
about data and reputation portability. This group comprises of policymakers (i.e., those actors
shaping policymaking from nonprofit organizations, foundations, or think tanks) as well as
legal experts and members of academia. Next, we will leverage the analysis dimensions and
stakeholder groups to inform our inductive theorizing.
Induction: qualitative interviews and thematic analysis
In the inductive part of our study, we built on the analysis dimensions and stakeholder groups
to conduct a series of semi-structured interviews as input for a thematic analysis and the
development of tangible design guidelines. Throughout this process, we closely followed Kaiser
(2014)’s guidelines for qualitative (expert) interviews. Specifically, we operationalized the five
analysis dimensions by developing a comprehensive interview guide consisting of open-ended
questions along each dimension (see Appendix for details). The interview guide was pre-tested
to ensure clarity in wording and common understanding.
Overarching thematic map / framework:
A first outline of the main thematic frame has been developed
4© Maik Hesse
Analysis FrameworkExploring the Design of Reputation Portability
…should
reputation
portability be
designed –for
which purposes?
…should be visible
to users?
…should it be
implemented in
and between
platforms?
POLICY & REGULATIONRISKS & CHALLENGES
WHY? HOW?WHAT?
REPUTATION PORTABILITY
Stake-
holders
Platform
operators
Platform
complementors
Platform
consumers
Indirect stakeholders
(e.g., policymakers,
lawyers, academia)
Overarching thematic map / framework:
A first outline of the main thematic frame has been developed
4© Maik Hesse
Analysis FrameworkExploring the Design of Reputation Portability
…should
reputation
portability be
designed –for
which purposes?
…should be visible
to users?
…should it be
implemented in
and between
platforms?
POLICY & REGULATIONRISKS & CHALLENGES
WHY? HOW?WHAT?
REPUTATION PORTABILITY
Stake-
holders
Platform
operators
Platform
complementors
Platform
consumers
Indirect stakeholders
(e.g., policymakers,
lawyers, academia)
Chapter 6: Designing reputation portability
112
Interview sample—For the interviews, we recruited representatives from each of the
identified stakeholder groups. Besides our own private and professional networks, we
leveraged Internet searches on Google (Scholar) and LinkedIn to find suitable candidates.
Thus, we ensured that participants possessed the required experience – that is, operational,
contextual, and interpretive knowledge (Kaiser 2014; see Appendix for an overview of
participants’ background). A total of ten interviews were conducted, distributed equally across
the four groups (see Table 8). It is worth noting that each participant additionally qualified as
a platform consumer (i.e., every individual had frequent personal experience using platforms).
The sample size was sufficient for redundancy (or theoretical saturation) to occur early in the
interview process (e.g., after five or six data sources; Guest et al. 2006; Kuzel 1999). Thus, given
the novelty of the phenomenon, in which our aim was to uncover common design
considerations for reputation portability but not to achieve maximum information variety, we
limited the number of interviews accordingly.
Stakeholder group
Number of interviews
Interview length
Platform operators (OP)
3 (0 female, 3 male)
43-54 min
Platform complementors (COMP)
2 (0 male, 2 male)
25-42 min
Platform consumers (CONS)
2*) (1 female, 1 male)
29-32 min
Indirect stakeholders (IN)
3 (2 female, 1 male)
48-65 min
Total
10 (3 female, 7 male)
25-65 min (>7 hours total)
*) counting exclusive experience as a consumer only; please note that all 10 interviewees qualify as frequent platform consumers
TABLE 8: INTERVIEW SUMMARY TABLE
Interview procedure—Having agreed to be interviewed, all participants were contacted via
mail and provided with a briefing document containing background information on reputation
portability, the goal of the interview (i.e., design of reputation portability), and formalities on
the process. Interviewees were asked for their informed consent to audio recording and
(anonymized) data collection/analysis for academic purposes, and were given the option to
receive the interview guide in preparation of the actual interview (Kaiser 2014). Participation
was voluntary and no reward or compensation was involved. Interviews were either conducted
in person (3) or online via Zoom video conferencing (7) and took 44 minutes on average. Before
each interview began, the interviewer confirmed that participants familiarized themselves with
the provided background information (i.e., the previously sent briefing document) and/or
repeated the most important parts thereof. Throughout, interviewees were encouraged to ask
questions and clarify potential misunderstandings. The interviews were of semi-structured
nature and allowed for deviation from the five analysis dimensions whenever probing for
additional information was necessary. Depending on participants’ represented stakeholder
group and expertise, each interviewee received a slightly different set of questions (see
Appendix). The focal part of each interview, however, dealt with general design considerations
for direct import of reputation data between platforms. Data was collected between June and
November 2020. Concurrently, interviews were transcribed verbatim using Otter.ai software;
each transcript was then thoroughly proof-read and double-checked against audio recordings
by the author of this study.
Thematic analysis—Driven by the research question and leveraging our pre-existing
theoretical knowledge as outlined in Section 2, we conducted a rigorous reflexive thematic
analysis to find patterns of shared meaning across the data set (Braun and Clarke 2006, 2019).
PART III: THE DESIGN OF REPUTATION PORTABILITY
113
We chose thematic analysis for two reasons: 1) since it is considered a highly flexible research
method in its own rights that allows for providing a rich and detailed account of the data within
pre-defined analysis dimensions (Nowell et al. 2017), and 2) since it has been affirmed as a
robust approach when exploring a novel phenomenon such as (the design of) reputation
portability (Vaismoradi et al. 2013). Figure 18 provides a detailed illustration on the thematic
analysis process.
FIGURE 18: ADAPTED THEMATIC ANALYSIS APPROACH BASED ON BRAUN AND CLARKE (2006)
To start with (step 1), we reviewed all interview transcripts to familiarize with the data –
meanwhile taking notes, creating memos, and ideating potential codes40 . This step also
included re-contacting several interviewees when follow-up questions arose on the transcript
or to clarify parts of the recording. Next, we started a systematic coding process (i.e., open/axial
coding; Corbin and Strauss 2008), in which we highlighted interesting features in the data
(step 2). Some of the initial codes thus generated were informed by the interview guide or
overall structure of the analysis dimensions (i.e., by deduction), however, we made sure to
approach this phase openly and inductively to capture all semantic nuances in the data (see
Fereday and Muir-Cochrane 2006). Hence, we assigned a substantial number of initial codes
(127) in the early phases of data analysis. Leveraging those, we entered a process of creative
coding in MAXQDA, in which we organized and sorted the initial codes, and kept condensing
information to eventually form broader sub-themes and early thematic maps (step 3).
Importantly, sub-themes were built inductively based on codes and without trying to fit them
into pre-existing categories (Nowell et al. 2017). Next, we reviewed the code system (e.g.,
combine, separate, discard), which resulted in a refined list of 19 sub-themes (step 4). All codes
were then checked for fit against the assigned sub-theme and overall fit of sub-themes against
each other was also validated. At this point, we went full circle with the hybrid approach by
revisiting our previously introduced analysis dimensions. Namely, to develop a full thematic
framework, inductively derived sub-themes were mapped against deductive analysis
dimensions. Thus, we wed the concept-driven analysis dimensions – to serve as overarching
themes – with the data-driven sub-themes (Gregory and Muntermann 2011). We conducted a
workshop and facilitated several discussions with fellow researchers (i.e., peer debriefing;
40 We used MAXQDA software to analyze the qualitative data set (Kuckartz and Rädiker 2019).
We use a reflexive thematic analysis approach to identify patterned
meaning –or themes –across the qualitative dataset
2© Maik Hesse
Method
Exploring the Design of Reputation Portability
Sources: Braun, V., and Clarke, V. 2006. “Using Thematic Analysis in Psychology,” Qualitative Research in Psychology (3:2); Braun, V., and Clarke, V. 2019. “Reflecting on Reflexive
Thematic Analysis,” Qualitative Research in Sport, Exercise and Health (11:4); Kaiser, R. 2014. Qualitative Experteninterviews, Wiesbaden: Springer Fachmedien Wiesbaden;
Kuckartz, U., and Rädiker, S. 2019. Analyzing Qualitative Data with MAXQDA, Cham: Springer International Publishing.
Figure 1. Adapted thematic analysis approach based on Braun & Clarke (2006) and Becker et al. (2017)
Hybrid (deductive &
inductive) approach
Creative &
open coding
Workshops & 1:1
discussions
REPORT
PRODUCTION
6
DATA
FAMILIARIZATION
1
CODE
GENERATION
2
THEME
SEARCH
3
THEME
REVIEW
4
THEME
DEFINITION
5
INPUT
PROCESS
OUTPUT
10 transcribed interviews
POLICY & REGULATIONRISKS & CHALLENGES
WHY? HOW?WHAT?
5analysis dimensions Prior literature
127 unique initial codes 19 sub-themes 3key & 2 contributing themes
WHY HOWWHAT
…
..
…
…
…
…
RISKS & CHALLENGES
……
POLICY & REGULATION
……
Thematic framework & 5 design guidelines
Chapter 6: Designing reputation portability
114
Nowell et al. 2017) to refine specifics of the coding system, sub-themes, and the broader
thematic framework (step 5). This ran parallel to the development of an overall narrative for
the data and the definition of the (overarching) themes. At this stage, we classified three key
themes (corresponding to the triad of why, what, and how) and two contributing themes.
Finally, we selected extracts from the data to emphasize the essence of a given (sub-)theme,
developed five practical design guidelines (i.e., one per each overarching theme), and wrote up
the report (step 6).
Findings
Figure 19 shows the full thematic framework anchoring both our deductive and inductive
reasoning. It comprises five overarching themes, which, by and large, root in the previous
analysis dimensions but have been further substantiated. Particularly, three themes were
identified as key themes to answer the research question on reputation portability design: the
impact of portable reputation (i.e., the “why”), the features and functionality (i.e., the “what”),
and the requirements for operationalization (i.e., the “how”). Our analysis also showed that
risks and challenges as well as policy and regulation classify as contributing themes playing
an important role in whether and how reputation portability can be established. These
contributing themes are interrelated with the key themes; however, they do not directly inform
the design considerations for platform operators. Yet, disregarding them would not allow us to
sketch out a coherent picture of reputation portability in digital platform ecosystems.
FIGURE 19: EXPANDED THEMATIC FRAMEWORK
41
41
This full thematic framework includes both the five overarching themes and nineteen sub-themes.
2nd order sub-themes:
Overarching themes break down into (sub-)themes
6© Maik Hesse
Thematic FrameworkExploring the Design of Reputation Portability
RQ: “How should reputation portability be designed and
implemented in digital platform ecosystems?”
Figure 3. Expanded thematic framework including overarching themes and sub-themes
Impact of portable
reputation
WHY
Requirements for
operationalization
HOW
Features and
functionality
WHAT
≙Interrelation
Data sovereignty (u) Imported signal type
Visualization and display
User experience (UX)
Security requirements
Uniform standards
Value proposition
Platform collaboration
SUB-THEMES
Trustworthy supply (b)
User transparency (b)
KEY
THEMES
CONTRIBUTING
THEMES
Competitive head start (b)
u = users; p = platform operators; b = both
RISKS & CHALLENGES
Stakeholder opposition Governmental involvement Obligatory portability
POLICY & REGULATION
GDPR and other regulations Legal barriersRegulatory enforcement
Fit and contextualizationNew platform advantage (p)
PART III: THE DESIGN OF REPUTATION PORTABILITY
115
Figure 19 also provides an overview of the inductively identified sub-themes within the
overarching thematic framework
42
. In what follows, we are going to tell the story of each
overarching theme and its respective sub-themes. We start with a definition, highlighting the
essence of what each theme is about, before analyzing the qualitative data along sub-themes.
Interview quotes from participants will be used to underscore analyses (Braun and Clarke
2006; Nowell et al. 2017). Finally, we conclude each theme’s discussion with a practical design
guideline that addresses relevant stakeholders, thus also summarizing key takeaways for both
researchers and practitioners.
Why: impact of portable reputation
It is generally assumed that one of the most critical properties of any new technology, product,
or innovation is to succeed in creating value or anticipating user needs (e.g., Mahmoud et al.
2018; Thomke and Von Hippel 2002). Reputation portability is no different. To design a direct
import functionality between platforms, one first needs to understand why it will provide
utility. Our interviewees identified several areas where reputation portability could benefit
direct stakeholders, that is, platform complementors and/or consumers (i.e., users; De Reuver
et al. 2018), platform operators, or both. The upper left column of Figure 19 indicates whether
these effects occur predominantly towards a particular stakeholder group (i.e., u = users;
p = platform operators; b = both). Hence, this first key theme provides a rationale as to why
reputation portability should be utilized. It encompasses the benefits interviewees associated
with the introduction of portable reputation and differentiates them by stakeholder group. In
doing so, it gives designing reputation portability its purpose.
Data sovereignty was discussed as the single most important benefit pertaining to users
exclusively. Our interlocutors characterized this sub-theme as user empowerment and control
over their online data or as the option to “realize informational self-determination” [IN1]. In
this sense, it refers to users’ (and especially complementors’) ability to manage their digital
identity and reputation across platform ecosystems or, as IN3 said: “Users gain mobility, users
gain power on data, and users gain more choice concerning platforms. And that's only a good
thing for them.” However, in this context some interviewees did caution that any reputation
portability solution be designed in a way that upholds the data sovereignty of all involved
actors including those that originally provided the reputation (see Section 4.5; “policy”).
Within the “why” theme, the remaining four sub-themes overlap. Uniting them is the notion
that additional reputation brings along advantages in acting on, using, or operating a platform.
First, the creation of trustworthy supply is based on the idea that reputation is a driver of trust
in digital platform ecosystems (e.g., Tadelis 2016). Interviewees highlighted that (positive)
ratings and reviews transport a lot of trust and hence porting these across platforms may
benefit platforms and users alike. For platforms, “not the demand is the problem but the
supply, and actually generating supply that is trustworthy has a really big impact” [CONS1].
That observation is in line with Dellarocas et al. (2009)’s argument that registering new “good
sellers” via reputation portability may improve the overall quality of trade on platforms.
42
Due to the extent of data variety, sub-themes come in different forms: as fully realized patterns of
shared meaning united around a core concept (Braun and Clarke 2019) or – for instance, in the case of
less consensus – as domain patterns capturing the diversity of meaning with regard to a particular focus
area (Boyatzis 1998). Our research question and the novelty of the investigated topic justify this nuanced
representation of (sub-)themes.
Chapter 6: Designing reputation portability
116
Similarly, more trustworthy complementors may translate into more reliable choices for
platform consumers. For emerging platforms direct import could thus prove as an effective
tool to tackle the chicken-and-egg problem of attracting both enough complementors and
consumers to offer a valid value proposition (Hein et al. 2020). Second, interviewees
associated reputation portability with increased levels of user transparency saying it would
reduce the amount of fraud or scam accounts and that having more reputation in one place
may induce an “added layer of perceived security” [OP2]; a construct, which has proven
instrumental in driving trust in platforms (Mao et al. 2020; Ter Huurne et al. 2017). Third, it
was expected that complementors were to gain a competitive head start leveraging their prior
reputation. That is because it allows them to overcome the “cold-start problem”, which
complementors face not having conducted any transactions yet on the new platform (Kokkodis
and Ipeirotis 2016; Wessel et al. 2017). Thus, for these new users without prior reputation
history, it will be “more difficult to signal trust to other users. Being able to import reputation
collected on other platforms may kickstart reputation scores” [IN2]. At the same time,
platforms may equally benefit from enabling the import – as OP1 confirmed:
We have a rating and review system after the first job but for us, [reputation portability]
would probably be particularly interesting before the first job. Because that's still a
problem we face that we cannot reliably tell if someone is actually going to work and
behave well on the platform.
Fourth, some of our interviewees noted that the positive impact of reputation portability
(especially of allowing direct import) may be more pronounced for niche, smaller, or entrant
platforms. In this sense, offering reputation portability may lead to a new platform advantage
“because [these platforms] could expect that users are more easily willing to leave the big
incumbent platforms when it's easier to port their reputation” [IN3]. This idea is in line with
the EU’s thinking that portability may facilitate more platform competition and break
monopolies (European Commission 2016, 2017). However, a new platform may even wish to
import reputation from an incumbent platform that is active in a different market domain
because “it might be very beneficial to have some sort of history from other sites to create
trustworthy service providers” [COMP2]. Moreover, incumbent platforms themselves (such as
Airbnb, eBay, Uber) may benefit from portable reputation. Specifically, OP3 pointed to the
potential for operational savings: more trustworthy supply and user transparency could
generate “a reduction in the number of transactions that do not go as planned,” which will
ultimately result in “a cost reduction on customer burden, defined as customer complaints.”
Finally, promotional effects for established platforms might be at play if reputational data
would be (visibly) integrated on other platforms: “it's marketing. […] That is basically the most
prolific reason for bigger platforms” [OP1]. We will discuss next, what exactly will have to be
displayed on platforms for reputation portability to work effectively.
Practical design guideline #1 (to operators): Design reputation portability with the end
users in mind; while you may equally benefit from increased levels of trust and transparency
on the platform, the central design directive should be individuals’ data sovereignty in online
ecosystems.
What: features and functionality
The second key theme explores the concrete features and functionality of direct reputation
import between platforms. It addresses the question what type of reputation should be made
PART III: THE DESIGN OF REPUTATION PORTABILITY
117
available, its visualization and aspects regarding end user’s interaction with the solution (i.e.,
user experience; UX).
Third-party assessments, such as (star) ratings or (text) reviews, represent the most commonly
used reputation signal to engender trust in digital platform ecosystems (Mavlanova et al. 2012;
Teubner et al. 2019). Unsurprisingly, discussing the portability of reputation across platforms,
the imported signal type emerged as a relevant sub-theme. Interestingly, the preferences for
importing either numerical ratings or text reviews were roughly split between interviewees.
While CONS1 made an argument in favor of ratings as being “the most common [and] really
learned from online shopping behavior,” platform operators underlined the informational
value of written feedback: “there's no question in my mind that [text] reviews are the most
powerful indicator where we can mine more data” [OP3]. However, there were also doubts
about ownership of these reviews (see Section 4.5; “policy”). It was also suggested by some that
platforms should include both ratings and reviews because they fulfill different functions and
only together provide a full picture. Some interviewees went beyond to argue that other types
of reputation should be made portable by platforms. These included professional certificates
users acquired on gig or freelance platforms and badges like the super host status on Airbnb
43
for import or embedding on other platforms or personal sites.
The second central design (sub-)theme revolves around the visualization and display of
reputation when imported onto another platform. Importantly, the display of numerical
ratings received more attention as this aspect proved more intuitive for interviewees to
envision (compared to imported reviews). Now, most participants preferred the imported
reputation to be distinctively displayed, for instance, as a separate widget or dashboard. “I
would say, you can display two or three of these widgets from other platforms” [OP1]. This
suggests that platforms provide some sort of exportable format for their reputation that can
then be showcased on other platforms. Notably, this is contrary to what the few
implementations of reputation portability (e.g., Bonanza) are currently offering as they are
integrating and merging (in this case eBay) ratings into their own rating scheme. Some
interviewees were skeptical of this practice saying that “a lot of information may be lost through
uncontrolled aggregations” [IN2]. Still, the idea of score aggregation was picked up, albeit for
different reasons. OP3 argued that importing, for instance, a mediocre but unrelated score
(e.g., from Airbnb to Lyft) may “hurt [a complementor] in the eyes of consumers and reduce
the chances of entering into interactions. […] And so, disaggregation of scores may create more
problems for the platform. I wouldn't disaggregate, I would just give one score that is
compounded somehow under the hood.” Others seconded this thought and hypothesized on
the introduction of an overarching score, which may be further broken down by platform
domain types such as accommodation, mobility, and so forth. Tadelis (2016) described this as
a meta-reputation score requiring diligently trustworthy behavior across platforms. We will
discuss implications and risks of such a universal rating later (see Section 4.4; “risks”).
Irrespective of whether reputation is shown distinctively or separate, interviewees suggested
that information be expandable as one clicks on or hovers over it. This would allow “looking
separately at reviews from different platforms” [CONS1], “breaking down the distribution of
stars” [OP3], and potentially even “link[ing] directly to the other platform” [IN2]. However, as
43
As a matter of fact, Airbnb readily provides such a badge to platform users in the form of an exportable
HTML-code.
Chapter 6: Designing reputation portability
118
we know, not all platforms use the common 1 to 5 stars system for reputation scores but several
other metrics can be observed on digital platforms (e.g., 0-10 points, 1-100%, etc.; Hesse,
Dann, et al. 2020). Thus, several interviewees highlighted the importance of rating
comparison: “you need to have some way of translating one reputation into another” [IN1].
Especially for distinctively displayed ratings, a scale indicating the imported rating value’s
translation into the target platform’s rating score standard was considered essential. As one
platform operator hypothesized: “in the ideal world, there will be some sort of an agreement
where everybody's using the same scale. Because every time you translate, you're introducing
distortions” [OP2]. Discussing the (source) platform’s visibility, OP1 stated: “obviously brand
name has to be clear and visible. The ratings would have to be from famous or well-known
platforms,” while COMP1 added: “I would like to have reputation portability from renowned
platforms”; simply because imported ratings from smaller platforms may not be so valuable.
However, it was also understood that for competitive reasons, platforms may not want to
‘advertise’ other platforms by displaying a foreign brand logo on their own website. For those
platforms, an aggregated or collated rating display may be more attractive. Other ideas were
discussed, for instance, timestamping imported reputation (i.e., when a rating/review was
received) or the option to ‘pin’ a favorable review, rating, or endorsement from other platforms
to one’s profile page (akin to the practice of cross-site linking on social or professional
networks; Y. Chen et al. 2014; Silvestri et al. 2015).
As Figure 19 indicates, another prominently discussed sub-theme was user experience for
reputation import. First, relating back to the notion of data sovereignty, interviewees suggested
that users retain full control and ownership over data including the right to opt-out of
reputation portability at any step of the process – even retrospectively. A viable way to do that
could be to prompt users, for instance with a pop-up, and give them the option to import
reputation from – or, vice versa, export existing on-site reputation to – other platforms. By
doing so, platforms should consider laying bare their intentions and being fully transparent:
“This data is used for x, y, and z, are you okay with this? – and it doesn't have to be a one
time in your life decision. Instead that can be revisited” [OP3]. Second, almost all interviewees
expected reputation import to be simple, quick, intuitive, and easy to use. CONS2 referred to
“seamless design” and likened the desired process to that of setting up a new Apple device from
an existing backup. Similarly, IN2 stressed the importance of simplicity in reputation
portability design, given various (hypothetical) paths for customization: “there's already a lot
of complexity when it comes to trust cues on platforms. If you’re adding additional trust cues,
such as imported scores, and then also add written text reviews, I would see that in many ways
too complex.” Third, our respondents suggested the design of a reputation portability import
functionality be appealing (i.e., sleek, beautiful). Taken together, these results are little
surprising since both ease-of-use and aesthetical appeal are generally understood as important
factors for how users perceive website usability and evaluate design (Nielsen and Loranger
2006; K. E. Schmidt et al. 2009). IN3 summarized in saying:
I think users expect the most convenient design you can think of. If there's a one-click
solution, users would use it. If they would have to go through a cascade of menus and
decision points, they might jump off the process. I think from the users’ view it’s really
important to make the process super easy to go through.
Importantly, platforms need to ensure inclusivity and create a sense of empathy and emotional
care for consumers by applying design principles to “reduce biases and create technology that
PART III: THE DESIGN OF REPUTATION PORTABILITY
119
is truly available for everyone” [OP3]. This perspective resonates with recent paradigms of
emotional web design (Bollini 2017).
Practical design guideline #2 (to operators): Ensure comparability of imported
reputation from a set of curated source platforms. Disclose full intentions, allow users to retain
data/process control, and design appealing as well as easy-to-navigate interfaces.
How: requirements for operationalization
The third key theme considers the broader picture of how reputation portability ought to be
designed and implemented. That is, it enunciates security requirements and uniform
standards as well as an overarching value proposition towards stakeholders. It also investigates
how platforms may group together in establishing solutions for reputation exchange.
Our interviews deemed the fulfillment of adequate security requirements indispensable when
establishing reputation portability in practice. For instance, by preventing manipulation and
fraud at any point of the process (i.e., ex ante or post hoc import). To do so, platforms will have
to eliminate opportunities for malicious actors to tamper with data, which is subject to porting.
Moreover they must minimize the risk of importing fake reviews, which have become a
prominent issue in platform ecosystems (Gutt et al. 2019; Wu et al. 2020). CONS1 suggested:
“It would be a risk if it's necessary to download your reputation and upload it somewhere else.
But if reputation is automatically transferred without you having access to the data, it would
help. Of course, this never avoids 100% the risk of fake reviews.” One way to minimize said risk
is for platforms themselves to curate a set of trustworthy sources for reputation import. As a
platform operator, “you need to make sure that where you import data from is a credible source
and you can trust it. Because if it's not credible, it can mess around your with business”
[COMP2]. This could imply for operators to “define the 10 most relevant platforms where we
think the information is most valid for us” [OP3]. Additionally, platform operators will have to
find ways to validate users’ claim to the reputation data on the source platform, for instance,
by confirming that account details, email addresses, or personal ID documents match – similar
to KYC (know your customer) procedures in banking. Doing so also guarantees that reputation
from a particular source platform can only be imported to one account at the time. This is
crucial to avoid the occurrence of secondary reputation markets, in which rating scores (for
import) are illegally offered for sale (just like in the case of fake reviews).
Interlocutors expressed their desire for uniform standards in designing and implementing
reputation portability across platforms: “The most ambitious solution would be a general
standard of reputation, which all relevant platforms could consent to, for example, five stars.
This would be an ideal environment to import and export. But this is perhaps a very distant
vision” [IN3]. It was also hypothesized that the EU may have an interest in promoting such a
standardization of reputation scores to avoid that major platforms push their own agenda in
setting de facto standards: “I think it would be nice to have some sort of standard rating from
the EU that has a meaning, which everyone understands. Cross-national and cross-platform. I
think achieving something like this, may be the first step” [COMP1]. To warrant a level and
unbiased playing field, customization options for importing reputation may have to be limited.
That is, from the platforms complementors are given an option to import from, they may either
import all of their reputation or no reputation at all (i.e., cherry-picking of ratings or reviews
should not be supported): “What you want to avoid is somebody just porting their positive
reputation data, ideally” [IN1].
Chapter 6: Designing reputation portability
120
For operators, a viable way of approaching reputation portability is by exploring opportunities
for platform collaboration. If platforms were to partner up in “a collaboration with another
platform and they use APIs and [log-in] credentials to move the data, it would make me trust
these ratings more” [COMP2]. Certainly, for multi-homing complementors entrusting their
account details to participating platforms may require a leap-of-faith (Greiner et al. 2018).
However, this barrier could be substantially lowered if an interested group of platforms can
form together to create and advertise a ‘trusted reputation portability consortium’. This route
may prove particularly interesting for smaller or niche platforms spearheading the
development of a potential solution. It would also provide users with the prospect of real-time
reputation portability and automatic rating update via APIs: “I would expect something like
that to lazy load while just accessing the page. We wouldn't actually import the data on our
platform, we would actually just ping the API of eBay or Uber and show what it gives us” [OP1].
Two additional sub-themes are notable requirements in operationalizing direct reputation
import between platforms. First, reputation portability needs to offer a value proposition to
all involved stakeholders but, most importantly, to users. To properly engage them, effective
design must highlight reputation portability’s utility and purpose (Garett et al. 2016). This can
be achieved by explaining and demonstrating the benefits of portable reputation (see Section
4.1; “why”). “If you're trying to design it, you have to start with the user need. You have to create
something that is valuable [and] where a real problem is solved” [CONS2]. For instance,
platform operators could explain in simple terms how reputation portability works, and which
benefits it promises to complementors (e.g., increase consumers’ trust and purchase
intentions; see Section 2.1). Second, most if not all participants considered platform fit and
contextualization to be an important issue. Some argued that imported ratings would have
limited explanatory power for service provision on different platform domains and that it was
at least debatable if importing reputation is helpful in those cases. “I think that a general fit is
nice. If I buy, for example, some screws on eBay and I get a positive review that has definitely
not the same value as a two month stay on Airbnb” [COMP1]. Indeed, Teubner, Adam, et al.
(2020) show that reputation portability effectiveness crucially depends on the fit between
source and target platform. In this sense, whether ratings should be converted or retained in
their original form (see Section 4.2; “what”) could also be tied to the degree of platform fit: “if
you choose to have one rating, then you really should make sure that the ratings are specific to
the same job – or the same industry at least. I think you can still show ratings from different
industries, but separately” [COMP2].
Practical design guideline #3 (to operators): Explore opportunities to jointly develop
prototypes with uniform standards. Solutions must demonstrate a clear value proposition to
platform users and require evaluation regarding their security and reliability across platform
ecosystems.
Risks and challenges
This theme draws attention to obstacles identified in the establishment of portable reputation
such as stakeholders’ opposition, governmental involvement, and the consequences of non-
voluntary reputation portability schemes. Thus, the theme contributes to developing an
understanding of the issues that must yet be overcome for reputation portability to be effective,
successful, and factually established.
PART III: THE DESIGN OF REPUTATION PORTABILITY
121
Our analysis thus far foreshadows some of the challenges that could be encountered in
developing solutions for reputation portability, more specifically, direct import between
platforms. Stakeholder opposition, for instance, is likely to be expected, especially from
established incumbent platforms (such as Airbnb, eBay, Uber). In the absence of physical
assets (e.g., properties, stores, vehicles), the biggest asset for these players are their user base
and associated data (Sundararajan 2016). Leading platforms invest heavily in user lock-in and
data monopolies, thus, their voluntary participation in reputation portability and data sharing
is unlikely. “Nowadays reputation is platforms’ currency, they see a lot of value in their
community” [COMP1]. This leads many of our interviewees to believe that incumbent
platforms do not see monetizable value in reputation portability. Said OP3:
The fundamental challenge is that I don't see platforms transferring data to other
platforms. There's no business model … they don't see the benefit. […] The people at
the executive level, the CEOs, or the investors – they're not thinking that way yet.
Maybe the industry will mature to a level where they will start to see the benefit of
having a reputation that crosses platforms [but] we are not close, not even remotely.
We have previously presented ways in which platforms may benefit from reputation portability
(i.e., trustworthy supply, competitive head start; see Section 4.1). While for these reasons
platforms would probably not oppose importing data, they will have to weigh anticipated
benefits of gaining more trustworthy complementors against very real threats of increased
competition. This may lead to a prisoner’s dilemma, in which incumbent platforms chose to
forgo any action that could weaken their own position (Krämer 2018). Hence, our interviewees
argued that demand for reputation portability must come from users, for instance, encouraged
by smaller or new platforms offering direct import: “larger platforms could expect a drain of
users if there are smaller competitors with more attractive services” [IN3]. By making
reputation portability attractive and desirable for users “you're going to create pressure
through the users. It's the only way how it's going to work because if they lose users, they will
react to that threat” [CONS2]. While the idea is plausible and we discussed ways to create a
need (i.e., value proposition; see “how”), users, too, will first have to be convinced of reputation
portability’s value. Currently, platform users are not aware of direct import functionalities on
platforms, simply because not many of them exist. Also, CONS1 cautioned that demand for
reputation portability may be overestimated: “maybe people are actually not working on many
platforms; you rather have people who are very active on one platform than being active on
five at the same time.” Yet, there is reason to believe that complementors would leverage
reputation import if it were available. In fact, while import on Bonanza is not obligatory, 94%
of all sellers still make use of this feature and chose to import their reputation from eBay (Hesse
and Teubner 2020b). Similarly, almost two thirds of users are active on more than one
platform making multi-platform use the rule rather than the exception (Teubner et al. 2019).
Restraints with regard to data sharing may also impede user interest in reputation portability,
especially when governmental involvement is expected. Interviewees unanimously pointed
out that only private entities should be mandated with handling users’ portable (reputation)
data and offering services on that basis. As IN3 accentuated: “here in Europe, we wouldn't like
the possibility to have an exchange of data between the private and the public system”, and IN1
specified this further in saying: “no civil rights should be dependent on mandatory reputation
portability." Discussions around this sub-theme were motivated by the ongoing development
of state-sponsored reputation schemes in China (Social Credit System; Liang et al. 2018).
Chapter 6: Designing reputation portability
122
These data surveillance tools are designed to reward/penalize citizens based on credit scores
and reputable conduct to determine their access to education, markets, or tax deductions.
Notably, leading platforms and technology companies work with the government in testing and
establishing the system’s technical infrastructure (e.g., Alibaba, Baidu, WeChat; Botsman
2017a; Liang et al. 2018). Participants resented this notion saying “as soon as the state comes
into play with tracking your reviews and your behavior, it's getting highly risky” [CONS1].
Direct fluidity of reputation between platforms – as discussed in this study – is certainly far off
authoritarian control systems, which are outside Western cultural and political tradition.
However, it is worth noting that with the proliferation of online markets, Westernized societies
have been embracing a rating culture and the continuous collection of user data and behavior
has become the norm (Wong and Dobson 2019). Thus, since Western platforms’ reputation
systems and the Chinese Social Credit System are not entirely incommensurable, “an
ecosystem will need to be built such that it doesn't become big brother, so that we don't follow
the route of other places. I think it has to be constantly in the public discourse” [OP3]. Looking
at opportunities to learn from the systems in place, CONS2 added: “we see a lot of things in
China and we should think about how to adapt them while keeping democracy in the right
space.” This approach concurs with Síthigh and Siems (2019)’ suggestion of closely studying
the Chinese model to assess implications for Western societies and to inform regulatory
intervention and law-making.
In a similar vein, interviewees dreaded obligatory portability. Specifically, if failure to fully
disclose gathered reputation would bring about disadvantages for platform users. “I see the
danger that this becomes the new standard at some point. And if you don't have that rating, if
you don't have the visibility or transparency, then you are discriminated against” [COMP1].
Study participants likened this scenario to recent discussions on having to demonstrate the use
of a COVID-19 app or issuing immunity passports to control entrance to shops and access to
venues or gatherings in times of a globally surging pandemic. “There could arise pressure from
the common usage of platforms when users are forced sublimely to port reputation data. […]
This introduces the problem of an indirect duty to reputation management. So then [retail]
shops might say: four-star customers only” [IN3]. In this sense, reputation portability between
platforms may already forebode a more extensive use of reputation scores across online and
offline services.
Practical design guideline #4 (to all stakeholders): In the long run, seek alignment on
the necessity of reputation portability, for instance, by increasing user awareness and creating
a need for portability services. Avoid reputation data sharing between platforms and
governmental branches and prevent users from being obliged to involuntary reputation
disclosure.
Policy and regulation
This theme comprises a regulatory view of reputation portability; including potential
enforcement by means of regulation (e.g., GDPR) as well as anticipated restraints of current
legislation on reputation data transfer, privacy, and ownership. Thus, the theme contributes to
the overall research question by developing an understanding of current and potential future
regulations and how they might affect the design and implementation of portable reputation.
As things stand, it is hard to imagine that leading platforms would be willing to work together
on devising uniform standards and interoperability, especially given different design choices
PART III: THE DESIGN OF REPUTATION PORTABILITY
123
in their reputation systems and the resulting lack of compatibility (Dellarocas et al. 2009;
Tadelis 2016). However, “portability without interoperability is only half of the way” [IN3].
Hence, assuming that (at least major) platforms are reluctant to enable direct import, our
interviewees wondered whether regulatory enforcement of standards and interoperability
could be a viable approach: “A lot of these companies do not like to share data with other
people. I don't think incentives would be a catalyst towards them for cooperation. You know
the carrot and the stick model? In this situation, unfortunately, the stick might have to be used”
[OP2]. However, especially policymakers cast doubt on whether it will be necessary for
legislation to specify standards: “To set the conditions is not the task of legislators, there could
be self-regulation within the market. This is, of course, a question of time. Will the market
players align their structures? Will the market solve this problem to the benefit of users? If
nothing happens, perhaps we need to have a discussion on compulsory interoperability” [IN3].
Interestingly, enforced interoperability (i.e., requiring platforms to adapt reputation systems
to support data portability) did not become part of the GDPR, although originally foreseen;
platforms criticized the passage and it was later taken out of the regulation’s draft (Drechsler
2018). Still, our interviewees argue that there is at least “a good case for making reputation
export mandatory” [IN1]. In this context, many people point to GDPR and other regulations,
however, as we have discussed in Section 2, it is up to clarification whether GDPR’s right to
data portability (i.e., Article 20) even applies to reputation data. According to IN3 this may not
be the case:
The European Data Protection Board has set forth what should be portable are those
data where users’ concrete actions or use of connected devices are observed. I think
most of the time reputation is given, those are actions of other users and some
algorithm is involved. And so, other users’ recommendations or assessments are not
data provided on the basis of consent to the platform. It's not comprised today under
Article 20.
The position is corroborated by several legal scholars (Inglese 2019; Kathuria and Lai 2018;
Vanberg and Ünver 2017). However, if GDPR is to be understood as a tool for individuals’ data
sovereignty and an instrument for user empowerment, “then Article 20 should be expanded
and also include reputation data [and] this idea should be promoted in further possible
modifications of GDPR” [IN3]. Van der Auwermeulen (2017) takes a similar stand in arguing
that reputation scores should fall under the scope of data portability as personal data.
However, interviewees also discussed alternative regulatory paths, for instance, leveraging the
EU’s soon-to-be adopted Digital Service Act, which, inter alia, mandates leading platforms to
share collected data with competitors: “Let's put this into a separate regulation, for example,
under the Digital Services Act, so that we get proper real-time data portability. It's much clearer
if we have data portability to promote competition, than personal data portability to realize
informational self-determination” [IN1].
As the final sub-theme, we will consider legal barriers to designing reputation portability.
While regulation may help with promoting portability, these hurdles must be overcome or, at
least, require further clarification. “There are a lot of legal barriers, stopping companies from
transferring data or creating reputation across platforms. I think that the new privacy regimes
are making that even more of a challenge” [OP3]. Indeed, when differentiating between types
of reputation (i.e., imported signal type; see Section 4.2; “what”), it appears unlikely that the
copyright for a text review (i.e. written feedback) remains with the user who demands for the
Chapter 6: Designing reputation portability
124
reputation to be ported (Kathuria and Lai 2018). Interviewees with a legal background
confirmed this notion: “in the case of [written] reviews, copyright might be an issue; also
written lines may include names of others. This would have to be anonymized otherwise […]
data protection might be touched” [IN3]. One option, though, for platforms to include user
reviews in reputation portability’s scope, could be to demand a waiver of rights, so that the
feedback becomes a public property to be used at the platform’s will (Hesse and Teubner
2020a). As IN1 stated: “you could give users the option to consent. The person writing the
review could wave their IP, like ‘default on/off’. If they decide their host can port the
reputation, then why not?” Similarly, ownership of (numerical) ratings will have to be clarified
before wide-spread solutions can be deployed. Taking the form of a numerical value, such
ratings (i.e., stars, points, percentages) are, in most cases, anonymized out of their nature.
Although this appears to make that signal type less prone to copyright issues, this assumption
should not be taken for granted. As IN2 remarked:
Thinking about data and ownership as a uni-dimensional construct is too simple. […]
If everyone has to agree to the export or portability of reputation – the person who
received a review, the person who left a review, and the platform – it is very unlikely
that all stakeholders will always agree to that. And so, as a default, you would end up
with a ported score that is flawed.
Thus, it is not entirely clear whether (aggregated) ratings can be used without the consent of
either the score provider (i.e., the person assigning the rating) and/or the processing platform,
both of which may ‘own’ the data. A look at practical examples tells an ambiguous story: on the
one hand, back in the mid-1990s, Amazon allowed its users to import their reputation scores
from eBay. However, after the latter claimed ratings as proprietary and threatened with legal
actions, Amazon discontinued the service (Dellarocas et al. 2009; Resnick et al. 2000). On the
other hand, the few platforms, which have been offering reputation import for some time (e.g.,
Bonanza and Truegether), enable their users to directly import ratings and reviews from both
eBay and Amazon. We can only hypothesize whether the incumbents tolerate this practice
because they do not see a potential competitive threat in these smaller players. Presumably,
this outlines a viable path for smaller and nascent platforms to jump on the bandwagon.
Eventually, for reputation portability to come to fruition, it is desirable for all involved
stakeholders to receive more legal guidance “on what reputation data can be used and
demanded for” [IN1].
Practical design guideline #5 (to regulators): Enhance data portability policies and
regulation by making the export of online reputation (especially numerical ratings) mandatory
for platforms, thus empowering users and encouraging competition in digital markets.
Discussion
General discussion
While an emerging literature stream explores reputation portability, recent studies mostly
focused on conceptualizations and investigating the impact on consumers’ evaluation of trust
and purchase intentions. So far, however, no research has considered the concrete design of
reputation portability (especially, in the form of direct reputation import) in digital platform
ecosystems. Contributing to the nascent body of knowledge, we use a hybrid approach of
deductive (Section 3.1) and inductive (Section 3.2) reasoning to develop a thematic framework
PART III: THE DESIGN OF REPUTATION PORTABILITY
125
of reputation portability and analyze its main components. In doing so, we conducted a series
of interviews with direct and indirect platform stakeholders, applied thematic analysis
methods on the qualitative data, and derived five practical guidelines to inform the design of
reputation portability. Both the resulting thematic framework and the design guidelines can
serve as a starting point for researchers and practitioners to develop reputation portability
mechanisms.
While our study’s findings, especially the guidelines, outline specific steps for stakeholders to
design, enable, and establish wide-spread reputation portability, many issues still require
attention. Importantly, the short-term viability of reputation portability is largely dependent
on platforms’ willingness to participate in proposed solutions. Despite the positive impact
portable reputation may entail (see Section 4.1; “why”), our analysis indicates that leading
platforms are unlikely to trigger the development of direct import solutions themselves. While
incumbents currently tolerate (or are unaware of) the few offerings that do exist on smaller
platforms, they may not remain patient in the longer run and claim ownership of their
reputation (as they did in the past; see Section 4.5; “policy”). Our interviewees hinted at the
possibility that reputation portability may well be obstructed by leading platforms on purpose,
labeling this as a “generational issue” [CONS2]. We see at least three potential ways to
overcome this stalemate. First, a consortium of interested platforms (e.g., smaller, or emerging
operators) could form to explore and test reputation portability features. Creating a trusted
group of partner platforms may allow these players to pilot the proposed design elements (e.g.,
clear value proposition, uniform security/layout standards, waiver of rights for portable
reputation, etc.) and increase user awareness and acceptance. Second, requesting and pushing
for (reputation) portability must not be limited to interested platforms but extend to users,
who should be empowered to demand respective services from operators. As OP3 stated:
“making consumers willing to participate is the cornerstone.” Third, discussions of users’
digital sovereignty and identity management must be extended to include reputation
portability. For online workers, digital reputation is their most important currency and in a
world of increasing digitization and platformization, regulation must aim to enforce and
strengthen users’ mobility and their rights vis-à-vis platform operators. This is particularly
important since operators consider complementors as contractors without access to the same
rights that normal employees of the platform enjoy (e.g., minimum wage, unemployment
benefits, health insurance; The Guardian 2020). Eventually, a combination of these three
aspects may pave the way forward to more reputation portability in digital platform
ecosystems.
Limitations and recommendations for future research
Despite its contribution, our study naturally faces limitations. For one, we develop guidelines
for the design of reputation portability, however, we do not build, or evaluate an actual
solution. We do believe, however, that our findings serve as an important foundation for
practitioners and researchers to explore the concrete design of an IS artifact. Future research
should thus leverage a design science approach to develop models, methods, and instantiations
of direct reputation import for platforms (Hevner et al. 2004). Our thematic analysis and
practical design guidelines can be used as a starting point (e.g., as meta-requirements or core
components) to develop respective reputation portability products and theories (Gregory and
Muntermann 2011). As part of this process, prototypes could be tested and evaluated by
interested platforms (see Sections 4.3, “how”, and 5.1).
Chapter 6: Designing reputation portability
126
Further, the direct import of reputation between platforms is by no means the only mechanism
to establish reputation portability. Most notably, Personal Information Management Systems
(PIMS) and blockchain technology have been discussed as viable alternatives (Hesse and
Teubner 2020a). In fact, some interviewees suggested that third-party solutions, such as PIMS,
may provide more data sovereignty and customization options to users because they act
independent of platforms. Similarly, several of our interviewees believed that a decentral
solution may be best suited to solve the presented challenges of reputation data ownership,
security, and verification. Recent research has pointed to distributed ledger technology’s
potential for disrupting existing platform business models and enabling reputation portability
(Bellini et al. 2020; Glaser 2017; Werbach 2018). Though it is certainly conceivable that these
technological advances will play a role in establishing the portability of personal reputation in
the long run, we deliberately chose to focus our analysis on the more straightforward solution,
that is, direct import between platforms. Hence, while these alternative mechanisms are
beyond the scope of this study, further research is needed to understand whether, and if so
how, our practical design guidelines extend to these contexts.
From a methodological perspective, we contribute to the growing body of literature on
reputation portability by conducting the first qualitative study on the concept. By conflating
the essence of relevant stakeholders’ views, this work informs platform design theory “tak[ing]
into account the iterative shaping and redefinition of what is the platform by multiple
distributed actors with divergent goals.” (De Reuver et al. 2018, p. 130). However, to further
corroborate (or even contrast) our findings and achieve maximum information variety, more
interviews may be conducted. In particular, the viability of desired features and functionality
would benefit from achieving a higher gender/regional diversity as well as including (visually)
impaired participants to realize inclusive design principles (see Section 4.2; “what”).
Throughout this research, we made sure to address the trustworthiness criteria (e.g.,
credibility, transferability, confirmability) put forth by Nowell et al. (2017). However, future
research may wish to selectively include additional criteria, such as triangulation of different
data collection modes and additional feedback loops with interviewees to confirm the findings’
validity and fit (i.e., member checking
44
).
Finally, future research should aim to provide empirical evidence for reputation portability’s
effectiveness beyond trust and purchase intentions (i.e., demonstrable value for stakeholders;
see Sections 2.1 and 4.3; “how”). For instance, this could be achieved by testing hypothesized
business case drivers to assess implications on platforms’ financial performance (e.g.,
operational savings; see Section 4.1; “why”). Similarly, effects should be differentiated by
imported signal type, since extant research so far only investigated the positive effects of (star)
ratings, however, existing studies have not yet considered whether imported text reviews have
a comparable impact.
44
Note, however, that (beyond re-contacting interviewees for clarification) we chose to omit this step
since it is not desirable practice according to Braun and Clarke (2006)’s reflexive approach to thematic
analysis. It promotes the active, interpretive role of the researcher and, hence, does not necessarily seek
to provide an account of the data that is recognizable to study participants.
PART III: THE DESIGN OF REPUTATION PORTABILITY
127
Concluding remarks
Reputation portability has been established as a viable approach to increase trust and stimulate
purchase intentions in online transactions. Direct reputation import, as enabled by a few
existing platforms, may provide a straightforward and intuitive mechanism to facilitate
seamless portable reputation between platform ecosystems. To answer how platforms should
design and implement such solutions, we leveraged a hybrid approach, conducting a thematic
analysis based on a literature review and a series of interviews with diverse stakeholder groups.
We summarized our findings by deriving five practical design guidelines. These provide
researchers and practitioners with hands-on advice on how to develop mechanisms for
reputation portability. In a nutshell, our study answers why reputation portability should be
employed by platforms in general, what should be displayed in potential solutions, and how
these may be implemented by operators across platform ecosystems.
Eventually, the ideal of seamless reputation transfer should be independent of specific
mechanisms. As the saying goes, “all roads lead to Rome”. In this case, reputation portability,
and specifically direct import, is only one means of promoting mobility in online markets.
While the broader notion of self-determined digital identity management is beyond our study’s
scope, it is important to ask how users can be supported in leveraging their online reputation
across contexts and what this data can ultimately be used for. In the words of COMP1: “In
general, human trust is something that exists across platforms and maybe also across services.
Perhaps, there's a good system to get something out of this [reputation] information.” Thus, in
a digitized world where online and offline environments merge, it becomes imperative to
design adequate tools that protect, empower, and promote individuals’ data sovereignty and
choice.
128
PART IV: OUTLOOK
129
PART IV: OUTLOOK
130
PART IV: OUTLOOK
131
Chapter 7: Quo vadis, reputation
portability?
In the chapters preceding this outlook, I introduced reputation portability in digital platform
ecosystems and addressed the nascent research topic from a diverse range of angles. To start
with, I provided an overview of the general dynamics of trust-building and reputation and
formally conceptualized cross-platform reputation portability. Thereafter, I investigated the
effectiveness of porting reputation, using both survey-based and experimental approaches,
by analyzing the impact on basic indicators of economic success such as prices, demand,
trusting beliefs, and purchase intentions. Finally, I explored reputation portability design
with relevant platform stakeholders and presented practical guidelines to implement
adequate solutions. This last chapter summarizes the results of this cumulative dissertation,
and circles back to the starting point by answering the research questions, which were
introduced in Chapter 1. To conclude, I discuss limitations of my work, provide avenues for
future research, and consider the broader context of digital identity management.
Re-visiting the research questions
In this dissertation, I investigate trust-building and reputation portability on digital platforms
as the central research objective. In doing so, five overarching research questions – introduced
in Chapter 1 – are addressed. In what follows, I will briefly summarize the main findings and
provide an answer to each research question RQ1 to RQ5.
RQ1: How do platforms build trust and how are these trust-building mechanisms perceived
by platforms users?
Platforms employ a variety of trust and reputation mechanisms to support users in curating
and maintaining their online presence. Chapter 1 introduces a taxonomy for trust-building
cues and finds that basically all platforms utilize a combination of transaction-based
mechanisms, expressive user profiles, identity verification, and implicit information to build
user trust (see Table 2). We find that, by and large, transaction-based signals (i.e., numeric
ratings or text reviews provided by prior transaction partners) are perceived as most important
in the eyes of potential consumers (Table 3). Notably, both the frequency of specific trust cues
provided by platforms and how their users perceive these tools are linked to a platform’s
degree of social interaction (DoSI). Hence, depending on the platform type, DoSI seems to
provide a reliable indicator of platforms’ choice for adequate trust-building mechanisms.
Moreover, as Chapters 1 and 3 indicate, platform users value the idea of imported reputation
as a potential means to build trust – especially in high-DoSI environments (Figure 11).
RQ2: How can reputation portability be conceptualized along the existing literature and
current practical implementations?
Referring to our model developed in Chapter 2 (Figure 9), reputation portability can be
conceptualized along its most relevant dimensions. These include digital platforms and
complementors acting thereon, feedback providers (i.e., consumers), and data transport
between otherwise confined ecosystems (e.g., through linking and embedding of profiles or by
Chapter 7: Quo vadis, reputation portability?
132
importing reputation signals). Moreover, external influencing factors play an important role,
most notably novel regulation of digital platforms and services as well as the emergence of
technological solutions such as PIMS and decentralized technology.
RQ3: What is the economic effectiveness of reputation portability?
As a platform complementor, leveraging reputation portability to increase the chances of
selling a service to consumers can work. Specifically, Chapters 4 and 5 show that reputation
portability, especially rating import, is a driver of trustworthiness and purchase intentions
(i.e., two proxies of complementors’ economic success). Yet, there are limits to these positive
effects on platforms. As was demonstrated in Chapter 4, the decision boundary for whether it
is beneficial to import reputation from another platform depends critically on the on-site and
imported ratings’ specific values (Figure 14). In other words: reputation portability’s economic
effectiveness is determined by both the existing and the imported signals’ strength.
However, as shown in Chapter 3 when analyzing a real-world example of reputation portability
on an e-commerce platform, counterintuitive effects, (e.g., on prices and demand) may occur
(Figure 12). We can attribute these outcomes in part to the context-specific implementation of
rating import on this platform. Hence, we provide suggestions for the adequate design of
reputation portability in later parts of this thesis (e.g., Part III; Chapter 6).
RQ4: How does reputation signal discrepancy affect cognition and purchase intentions?
As stated in the previous answer to RQ3, there are limits to reputation import’s effectiveness
depending on the specific ratings’ values. A closer look at survey data in Chapter 5 reveals that
a higher rating discrepancy is indeed detrimental to consumers’ purchase intentions (Table 7).
Similarly, cognitive psychology predicts that conflicting information and the arrangement and
display of two or more ratings, will have implications for consumers’ processing of these
signals as well as their decision making. Therefore, these considerations are an integral part of
the exploration of reputation portability’s design in Chapter 6 of this thesis.
RQ5: How should reputation portability be designed and implemented in digital platform
ecosystems?
Establishing reputation portability requires buy-in from all involved stakeholders. To explore
the design of adequate solutions, a diverse selection of representatives from the identified
stakeholder groups are interviewed. Importantly, I provide five practical design guidelines for
direct reputation import between platforms: 1) design to achieve data sovereignty for end
users, 2) ensure comparability across platforms and disclose full intentions in easy-to-use
interfaces, 3) explore joint (platform) opportunities to develop secure and reliable prototypes
with uniform standards, 4) create a user need for reputation portability services but prevent
governmental data access and involuntary schemes, and 5) expand current regulation and
make reputation export mandatory for platforms.
These guidelines exemplify the minimum set of requirements that will have to be addressed to
successfully design reputation portability. However, within the scope of this dissertation, the
presented studies’ goal was to provide first insights into the conceptualization, effectiveness,
and design of portable reputation in digital platform ecosystems. Naturally, further research
efforts are required to fully understand each dimension of reputation portability.
PART IV: OUTLOOK
133
Conclusion, limitations, and future research
My aim with this cumulative dissertation was to provide first comprehensive groundwork,
laying a basis for further studies on the concept of reputation portability in digital platform
ecosystems. The presented results therefore have important implications for both theory and
practice. To begin with, the research program makes several noteworthy contributions to the
continuously growing body of platform literature in IS and adjacent fields:
i) A delineation of relevant terms, actors, mechanisms, and influencing factors in a
first-of-its-kind conceptualization of reputation portability,
ii) empirical evidence for reputation portability’s trust-building potential and its
ability to positively impact consumers’ purchase intentions in electronic markets,
iii) the exploration of design requirements as an input (e.g., meta-requirements or core
components) to develop respective reputation portability products and theory.
My research also has direct implications for the practical establishment of adequate reputation
portability mechanisms. Practitioners in the domain of IS (i.e., platform operators but also
entrepreneurs considering a market entry) can benefit from the findings presented in this
thesis. First, market entrants and smaller platforms may wish to establish reputation
portability as a competitive tool against incumbent platforms. Second, the practical design
guidelines, proposed in Chapter 6, may serve as a viable blueprint to design direct import
solutions for and between platforms. Moreover, this work provides a summary of current
discussions around policies and regulation governing online (reputation) data (e.g., GDPR)
and identifies several existing legal barriers to portability. Policymakers and legislators may
interpret these findings as a call for action to further strengthen consumer rights in digital
markets.
Notwithstanding these theoretical and practical contributions, some important limitations
should be addressed by future research. Most importantly, Chapter 1 introduced different
types of transaction-based trust cues on platforms, that is, numerical rating scores (e.g., stars)
and written assessments (e.g., text reviews). In Part II of this thesis, I investigate the
effectiveness of portable ratings for trust-building. However, there is reason to believe that text
reviews may as well be ported and leveraged across platform boundaries. In fact, participants
across our studies frequently pointed to the extraordinary informational value of written
assessments (e.g., in Chapters 3 and 6). Previous research confirms this notion and
demonstrated a strong positive impact of (good) text reviews for both economic and social
value expectations on platforms when compared to other trust-building artifacts (Abramova
et al. 2017; Dann et al. 2020). However, as Fresenda and Gefen (2020) show, the helpfulness
of a review depends on different aspects of the textual content (i.e., its word entropy). Now,
written feedback is highly contextualized depending on the type of service and/or platform.
Hence, future research should investigate which parts of the informational content are
particularly helpful for promoting trust, how this information can best be extracted from
reviews, and whether it is transferable across different service categories and platforms.
What is more, although I elaborate on the design of reputation portability in this thesis, an
actual IS artifact was not built. As stated in Chapter 6, design science researchers may leverage
the provided guidelines to build, implement, and evaluate solutions for reputation portability.
In this context, a resulting artifact could be investigated in several ways. For instance, to
Chapter 7: Quo vadis, reputation portability?
134
analyze the external validity of an instantiation of reputation portability, a field experiment
could be conducted (Edelman et al. 2017; Hesse, Teubner, et al. 2020). Also, the reputation
portability scenario, which we leveraged for the experiments in Chapters 4 and 5, was stripped
of most other contextual platform design elements to avoid introducing too many confounding
factors. This setup should be adapted for further studies to create more realistic scenarios and
to compare the effectiveness of reputation portability against other typically employed trust
and reputation cues on platforms (e.g., Zloteanu et al. 2018). Such a more nuanced website
design, integrating reputation portability (e.g., via direct rating import), could then be tested
in a laboratory experiment with users, for instance, via eye tracking (see Chapter 5; Hesse,
Lutz, et al. 2020).
Beyond that, reputation portability has the potential to affect individuals more broadly than
just in the context of digital platforms. I will briefly discuss some implications for technology,
policymaking, and society in the closing section of this research program.
Outlook
Since I started my work on this dissertation, the importance of digital identity and reputation
management, both as a research and societal issue, has lost none of its appeal. The digitization
and platformization of our world are progressing fast – and this development is only
accelerated by a global pandemic that forces us to rely more increasingly on digital technology
and business models. Accordingly, managing online reputation is of vital importance,
especially when “no reputation means ‘you don’t exist’ in the eyes of consumers” (Gandini
2020, p. 20). As was indicated in Chapter 3 (Figure 11), the concept of portable reputation is
considered a strong tool especially for crowd work type platforms (incl. mobility services). And
indeed, there is a compelling case to be made for gig workers (i.e., complementors) providing
their services on such platforms. On a global scale, the United States are the leading market
for the provision of gig economy services mediated through platforms and the industry has
been growing consistently during the COVID-19 crisis. That is why California’s Proposition 22,
which received a majority vote in the 2020 US presidential election and was funded in large
parts by Uber and Lyft, is regarded as an alarming turning point in online labor law (Oxford
Internet Institute 2020). In effect, for these types of platforms, the new proposition overrules
a previously adopted law that would have ultimately granted platform workers more rights as
regular employees. Thus, in several aspects the new proposition arguably deteriorates
complementors’ situation vis-à-vis platforms and has foreboding implications as a precedent
for similar regulations across the globe. Without the protection and benefits of employment,
crowd workers remain at the mercy of platforms’ goodwill. This is precisely why reputation
portability can make a difference for them and strengthen their rights.
As of late 2020, portability of reputation receives ongoing research interest and sojourns in
the public discourse (Basili and Rossi 2020; Eckhardt 2020; Möhlmann and Teubner 2020).
However, the maturity of different mechanisms seems to differ vastly. First, at the time of
writing, both Deemly.co and Traity.com are inactive or have gone out of business
45
, adding to
the long list of failed PIMS and startups working on reputation portability (see Chapter 2).
Second, blockchain-based platforms and decentralized reputation systems are still heavily
45
Therefore, their websites, as refenced in pre-published parts of this dissertation and listed in the
Reference section of this work, are inaccessible at the current point in time.
PART IV: OUTLOOK
135
researched study objects with several existing prototypes and use cases (Almasoud et al. 2020;
Bellini et al. 2020; De Filippi et al. 2020; Glaser et al. 2019). Nevertheless, as Gartner in its
technology hype cycle estimates, such blockchain platforms and the decentralized Internet are
5 to 10 years (if not more than a decade) away from transformational impact (Gartner 2020).
In consequence, direct reputation transfer between platforms persists as an intuitive
short-term option for more wide-spread adoption. In this regard, the European Union’s efforts
to investigate portable reputation have recently picked up steam (European Commission 2016,
2017, 2018b). Contrary to the US, European countries are historically known for strong labor
law and, compared to developments across the Atlantic, the EU now seems to be going down
a distinctively different route in strengthening complementors’ and consumers’ rights. Most
importantly, the German government, representing the continent’s largest economy, just
recently announced to take on the issue of reputation portability in their latest effort to
regulate platform work and thereby outspokenly distanced itself from practices in the US
(BMAS 2020). In further comments, the current German Ministers for Justice and Labor
position themselves clearly with regard to the portability of online reputation:
“[P]latform workers must be able to take their ratings to a different platform each time.
So far, this has not been possible because each platform uses its own system. This
makes these service providers dependent and means that they often, nolens volens,
stay on one platform to avoid having to start over somewhere else. Consequently, this
strengthens the monopoly position of individual platforms and simultaneously leads
to less competition to the detriment of consumers and complementors. […] We
therefore want to ensure that platform workers can transfer their data to other
platforms at any time via neutral interfaces. This strengthens competition and freedom
of choice and, thus, also benefits consumers.” (Lambrecht and Heil 2020; own
translation)
It remains to be seen how Germany’s planned regulation and strong commitment to reputation
portability will be embedded and reflected in the broader EU’s Digital Services Act and Digital
Markets Act. These policy packages themselves, scheduled to be released in late 2020, aim at
holding liable the big technology platforms and limiting their extend of power by fostering
innovation and competitiveness in European online environments. Importantly, the
regulations are expected to introduce mandated interoperability and improve data portability
for consumers seeking to switch platforms (Broadbent 2020). With Germany currently holding
the presidency of the EU Council, considerable overlap of the policies should be expected.
Finally, the investigated phenomenon of reputation portability will have to be considered in
the broader context of digital identity and reputation management. As Kokkodis and Ipeirotis
(2016) demonstrate, category-specific performance of online workers is predictive of their
success in other, comparable tasks. One naturally wonders whether this transferability of
reputation and skills persists when considering different scenarios, including the transfer of
online reputation to offline contexts. As already noted by Publilius Syrus in Ancient Rome: “A
good reputation is more valuable than money.” These days, who can tell what our accumulated
digital reputation will eventually be worth and what it may “buy” in the future?
136
Appendix
137
APPENDIX
Appendix
138
Supplementary Material Part I
Chapters 1 and 3: Survey design
FIGURE 20: ONLINE SURVEY
46
46
This survey was administered online via Prolific.co and Google Forms. It was conducted across five
different platform types (i.e., accommodation, carsharing, mobility, e-commerce, crowd work). Here, an
excerpt from the survey on “Mobility Platforms” is shown exemplary. Wording, logos, and introductory
texts were adapted slightly between surveys to fit the specific platform type. The overall study content
and questions remained identical across contexts. Please refer to Chapters 1 and 3 for further
information on the study’s methodology, specific items to be rated by respondents, and sample.
Appendix
139
Chapter 2: Literature search on Google Scholar
47
47
Notes:
1) As of June 2019; patents were excluded from the list of results.
2) Besides these terms, various combinations of terms such as reputation, signal, rating, review,
platform, aggregation, portability, transfer were searched (using Boolean operators AND/OR).
3) To specify the search in an additional step, we used combinations with platform names.
Results since 2010
Search term
Term by itself
Combined with
AND (Airbnb OR eBay OR Uber)
"Reputation portability"
21
12
Reputation AND portability
16,600
1900
"Portable reputation"
32
21
"Portability of reputation"
7
4
Portability AND ratings
16,000
889
Portability AND reviews
27,200
1,780
“Portable reviews”
25
1
"Review portability"
19
0
“Reputation transfer”
220
28
"Reputation transference"
10
0
“Transferable reputation”
17
4
"Reputational data"
237
44
"Reputational data" AND portability
41
18
Reputation AND platform
129,000
17,400
"Cross-platform signal”
8
0
"Cross-platform signaling”
0
0
“Cross-platform” AND signal
18,500
850
“Cross-platform” AND signaling
7,960
133
“Cross-platform reputation”
10
8
“Cross-platform” AND reputation
5,200
593
“Cross-platform portability”
694
14
“Cross-platform” AND portability
16,900
712
“Cross-platform” portability AND (ratings OR reviews)
5,170
335
“Cross-platform” AND “review portability”
1
0
“Cross-platform” AND “portable reviews”
0
0
“Cross-platform” AND (ratings OR reviews)
17,300
1,050
“Reputation signal”
216
28
“Reputation signaling”
254
32
Reputation AND signal
136,000
11,700
Reputation AND signaling
23,500
3,960
Reputation AND (signal OR signaling) AND portability
6,680
1,870
“Reputation aggregation”
754
229
“Aggregation of reputation”
89
37
“Aggregated reputation”
254
97
“Digital Identity Management” AND reputation
570
108
“Digital Identity Management” AND ratings
353
154
“Digital Identity Management" AND reviews
1,470
128
"Digital Identity Management" AND platforms
1,580
131
"Reputation passport"
7
6
"Reputation dashboard"
17
4
"Reputation economy"
1,010
242
“Personal Information Management System” AND reputation
56
5
TABLE 9: STRUCTURED LITERATURE SEARCH
Appendix
140
Supplementary Material Part II
Chapter 4: Stimulus material
Imagine that you would like to book a stay via the (hypothetical) Internet platform happystay.com.
On this platform, private hosts rent out their apartments and rooms (very similar to Airbnb). You have
already identified a potential host and, in the following, consider the host's profile page. Note that you
will see a very stylized form of this page (i.e., with a blurred profile photo).
Based on the host’s (star) ratings, we will ask you to state how likely you are to book and how trustworthy
you consider the host to be.
• 1: Note, however, that the host has not collected any ratings yet.
• 2: Note that the host has already collected ratings from previous guests on happystay.com.
• 3: Note that the host has not collected any ratings on happystay.com yet but has imported
previous ratings from her/his Airbnb profile.
• 4: Note that in addition to her/his ratings on happystay.com, the host has imported previous
ratings from her/his Airbnb profile.
FIGURE 21: LANDING PAGE FOR ONLINE EXPERIMENT PARTICIPANTS
Names: (equally distributed; randomly drawn and allocated to blurred female or male image)
Alex, Emma, Laura, Lisa, Lucas, Michael, Paul, Philipp, Pia, Sarah, Sophia, Tim
Images:
FIGURE 22: BLURRED PROFILE IMAGES OF COMPLEMENTORS ON PLATFORM
Appendix
141
Chapter 4: Treatment design
Table 10 shows the experiment’s treatment matrix. As can be seen, when no rating is available in the 2x2
between subjects design, the condition has been coded as “NA”. The participants’ allocation across cases
(1) to (4) is indicated by the color-coding, namely case (1) (no rating; 24 observations; yellow area), case
(2) (on site only; 81 observations; blue area), case (3) (imported only, 83 observations; red area), and
case (4) (both ratings; imported and/or on-site rating ranges from 1.0 to 5.0; 305 observations; grey
area).
Imported rating
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
NA
∑
On-site rating
1.0
2
2
2
2
2
3
3
2
3
7
28
1.5
2
2
2
2
2
3
3
3
3
7
29
2.0
2
2
2
2
2
4
4
4
4
6
32
2.5
2
2
2
2
2
4
4
3
4
7
32
3.0
2
2
2
2
2
4
4
4
4
7
33
3.5
4
4
3
4
3
7
7
7
7
11
57
4.0
4
4
4
4
4
6
6
7
7
13
59
4.5
4
4
3
4
4
6
6
7
7
11
56
5.0
4
4
4
4
4
7
7
7
7
12
60
NA
6
6
7
6
7
13
12
13
13
24
107
∑
32
32
31
32
32
57
56
57
59
105
493
TABLE 10: TREATMENT MATRIX
Table 11 reports descriptive statistics (i.e., mean/std. dev.) of participant-specific control variables. All
controls were evenly distributed, and no systematic differences existed for participants’ key
characteristics across treatments (i.e., results were not driven by confounding factors).
Case
Control
Scale
1
(no rating)
2
(on-site only)
3
(imported only)
4
(both ratings)
All cases
Disposition to Trust
1-11
6.1 (2.5)
6.3 (2.6)
5.8 (2.5)
5.9 (2.6)
6.0 (2.6)
Trust in Airbnb
1-11
7.4 (2.2)
6.9 (2.5)
7.1 (2.4)
7.0 (2.5)
7.1 (2.5)
Risk-Affinity
1-11
3.8 (2.3)
4.4 (2.7)
4.4 (2.6)
4.5 (2.7)
4.4 (2.7)
Online Shopping Experience
1-5
3.0 (1.2)
2.5 (1.3)
2.9 (1.3)
2.9 (1.3)
2.9 (1.3)
Age
18-70
33.1 (9.5)
32.8 (12.0)
34.7 (11.9)
35.9 (12.2)
35.0 (12.0)
Gender Female
{0, 1}
66%
59%
57%
53%
56%
Note: Mean values reported (standard errors in parentheses)
TABLE 11: BALANCE OF PARTICIPANT-SPECIFIC CONTROLS ACROSS CASES
Table 12 shows that no significant differences occurred for the number of reviews that were displayed in
the different groups. Please also refer to the section “Chapter 4: Supplemental analysis” in this
Appendix, in which we additionally investigate the interaction of rating score and the number of reviews.
Number of ratings
Type of rating
n
Minimum
Maximum
Mean
St. Dev.
Median
On-site
386
1
89
22.4
28.1
9
Imported
388
1
89
24.4
29.9
9
TABLE 12: NUMBER OF REVIEWS ACROSS TREATMENT VARIATIONS
Appendix
142
Chapter 4: Questionnaire and measurement instruments
Control
Question
Source
Trust in Complementor
I would consider this host as trustworthy.
(adapted from Gefen 2002)
Risk-Affinity
In general, I am very willing to take risks.
(adapted from Dohmen et al. 2011)
Disposition to Trust
I generally trust other people.
(adapted from Gefen 2000)
Trust in Airbnb
I think that Airbnb is a trustworthy
company.
(adapted from Gefen 2002)
TABLE 13: INSTRUMENT VARIABLES
48
Chapter 4: Supplemental analysis
Full control variable analysis—Table 14 reports the full control variable analysis of all used
variables across the four cases. For each case 1 to 3 we show the original model results next to an
extended model comprising all relevant controls per case (i.e., participant-specific controls and number
of reviews; Models 1.1/2, 2.1/2, 3.1/2). Additionally, we repeated the analyses for case 4 including the
original model (i.e., Model 4.1), including all controls (Model 4.2), and a model without interaction
(Model 4.3; to check for significance when controlling for both ratings). Please note that for these
analyses we introduced mean-centered control variables to account for the fact that the original
coefficients are determined by these controls and can only be interpreted in a meaningful manner when
applied to “the average population” (which is achieved by mean-centering). In particular, it can be seen
that in case 1, the coefficient 𝛼1, calculated as .350 in our model, is only slightly affected when including
all (mean-centered) controls (i.e., equating to .376), despite the statistical significance of some of the
control variables. The same holds for all other relevant coefficients from the “trust function”, which can
be checked in the control models 2.2, 3.2, and 4.2 in the Table 4.1 – they do not differ substantially in
their sign, significance, or magnitude.
Interaction of rating score and number of reviews—Table 15 reports the results of a robustness
check for the interaction of rating score and number of reviews. For that purpose, we calculated a series
of multivariate OLS regression models to analyze the interaction effect across the three cases where at
least one rating was visible (i.e., cases 2, 3, and 4). Neither in case 2 nor 3 (onsite or imported rating
only; Models 2a/b/c, 3a/b/c) the number of reviews has any significant impact. Both when controlling
only for the number of reviews and when additionally controlling for the interaction with the respective
rating. Similarly, in case 4 (both ratings; Models 4a/b/c) the number of reviews does have a significant
impact (Model 4b; b=.001, p<.05), albeit negligibly small – it would take 1,000 additional reviews to
increase the evaluation of trust by one unit on the 1-7 Likert scale. Additionally, this significant effect
vanishes when also controlling for the interaction(s) between rating and number of reviews (Model 4c).
In conclusion, the controls barely have any influence on coefficients – neither in terms of magnitude nor
significance.
48
Measured on 11-point Likert scales
Appendix
143
143
DV: Trust in Complementor; OLS regressions
Model 1.1
(Case 1)
Model 1.2
(Case 1)
Model 2.1
(Case 2)
Model 2.2
(Case 2)
Model 3.1
(Case 3)
Model 3.2
(Case 3)
Model 4.1
(Case 4)
Model 4.2
(Case 4)
Model 4.3
(Case 4)
Constant
0.350
***
0.376
***
-0.004
0.081
0.016
0.021
0.259
*
0.290
**
-0.010
(0.038)
(0.037)
(0.075)
(0.062)
(0.082)
(0.071)
(0.116)
(0.099)
(0.057)
On-site Rating
0.128
***
0.106
***
-0.029
-0.017
0.050
***
(0.021)
(0.017)
(0.032)
(0.027)
(0.012)
Imported Rating
0.116
***
0.116
***
-0.006
-0.005
0.076
***
(0.023)
(0.020)
(0.033)
(0.028)
(0.012)
On-site × Imported
0.024
**
0.017
*
(Rating)
(0.009)
(0.008)
Disposition
to Trust
0.018
0.005
0.039
**
0.029
***
(0.014)
(0.009)
(0.012)
(0.005)
Risk-Affinity
0.043
*
0.028
**
0.023
*
0.031
***
(0.017)
(0.009)
(0.010)
(0.005)
Trust in Airbnb
0.001
0.021
*
0.013
0.014
*
(0.017)
(0.010)
(0.012)
(0.006)
Online Shopping
Experience
-0.004
0.058
**
-0.032
-0.007
(0.029)
(0.017)
(0.020)
(0.009)
Age
-0.0004
-0.004
*
0.001
0.002
(0.004)
(0.002)
(0.002)
(0.001)
Participant Female
-0.013
-0.049
0.062
-0.023
(0.081)
(0.044)
(0.052)
(0.025)
#Reviews (ons.)
-0.0003
0.001
(0.001)
(0.0004)
#Reviews (imp.)
0.001
-0.001
(0.001)
(0.0004)
Observations
24
24
81
81
83
81
●)
305
300
●)
305
R2
-
0.476
0.317
0.625
0.246
0.534
0.189
0.472
0.170
Adjusted R2
-
0.292
0.308
0.583
0.237
0.482
0.181
0.451
0.165
Res. Std. Error
0.184
(df = 23)
0.155
(df = 17)
0.244
(df = 79)
0.189
(df = 72)
0.257
(df = 81)
0.214
(df = 72)
0.255
(df = 301)
0.209
(df = 288)
0.258
(df = 302)
F Statistic
2.579
(df = 6; 17)
36.646***
(df = 1; 79)
14.988***
(df = 8; 72)
26.471***
(df = 1; 81)
10.309***
(df = 8; 72)
23.389***
(df = 3; 301)
23.364***
(df = 11; 288)
30.978***
(df = 2; 302)
●) Age information missing for some participants; observations excluded from analysis
*** p<.001; ** p<.01; * p<.05; standard errors in parentheses
TABLE 14: EXTENSION BY (MEAN-CENTERED) CONTROL VARIABLES
Appendix
144
144
DV: Trust in Complementor; OLS regressions
Model 2a
(Case 2)
Model 2b
(Case 2)
Model 2c
(Case 2)
Model 3a
(Case 3)
Model 3b
(Case 3)
Model 3c
(Case 3)
Model 4a
(Case 4)
Model 4b
(Case 4)
Model 4c
(Case 4)
Constant
-0.004
0.002
0.050
0.016
0.010
-0.044
0.259
*
0.217
0.214
(0.075)
(0.078)
(0.095)
(0.082)
(0.085)
(0.109)
(0.116)
(0.119)
(0.129)
On-site Rating
0.128
***
0.128
***
0.114
***
-0.029
-0.022
-0.024
(.012)
(0.021)
(0.027)
(0.032)
(0.032)
(0.035)
Imported Rating
0.116
***
0.116
***
0.131
***
-0.006
0.001
0.003
(0.023)
(0.023)
(0.030)
(0.033)
(0.033)
(0.035)
On-site × Imported
0.024
**
0.022
*
0.022
*
(Rating)
(0.009)
(0.009)
(0.009)
#Reviews (ons.)
-0.0003
-0.002
0.001
*
0.001
(0.001)
(0.003)
(0.001)
(0.001)
Rating × #Reviews
(on-site)
0.001
0.0001
(0.001)
(0.0004)
#Reviews (imp.)
0.0003
0.002
-0.0004
0.00002
(0.001)
(0.003)
(0.0005)
(0.001)
Rating × #Reviews
(imported)
-0.001
-0.0001
(0.001)
(0.0004)
Observations
81
81
81
83
83
83
305
305
305
R2
0.317
0.318
0.324
0.246
0.247
0.253
0.189
0.204
0.204
Adjusted R
2
0.308
0.300
0.298
0.237
0.228
0.224
0.181
0.191
0.186
Res. Std. Error
0.244
(df = 79)
0.245
(df = 78)
0.246
(df = 77)
0.257
(df = 81)
0.258
(df = 80)
0.259
(df = 79)
0.255
(df = 301)
0.254
(df = 299)
0.255
(df = 297)
F Statistic
36.646***
(df = 1; 79)
18.152***
(df = 2; 78)
12.323***
(df = 3; 77)
26.471***
(df = 1; 81)
13.136***
(df = 2; 80)
8.910***
(df = 3; 79)
23.389*** (df =
3; 301)
15.329***
(df = 5; 299)
10.896***
(df = 7; 297)
*** p<.001; ** p<.01; * p<.05; standard errors in parentheses
TABLE 15: INTERACTION OF RATING SCORE AND NUMBER OF REVIEWS
Appendix
145
Gender effects—We first analyzed mean trust values (on a scale of 0 to 1) to check for a general
tendency towards differences in the assigned value of trust scores among all possible host/guest
combinations. Table 16 shows these mean values for all tested combinations of male/female hosts (i.e.,
profile pages) as well as male/female guests (i.e., survey participants). As we can see, evaluations of trust
are generally higher for female hosts than they are for males and, further, the combination female
guest/participant and male host exhibits the lowest overall trust average.
Case
Observations
Trust Score
Male host, male guest
122
.393
Male host, female guest
131
.379
Female host, male guest
97
.431
Female host, female guest
143
.436
All combinations
493
.410
TABLE 16: NUMBER OF OBSERVATIONS AND AVERAGE TRUST SCORES49
To better understand these results, we introduced binary dummy variables for Female Host and Female
Guest (i.e., participant) to check if above “effects” can be validated statistically. As we see in Table 17, a
female host has a positive (albeit small) impact on participants’ evaluations of trust (Model 1) –
irrespective of the prospective guest’s gender. However, the effect itself – as any of the investigated
gender control variables – is not statistically significant. Similarly, female guests do not exhibit a
significantly different evaluation of trust in hosts (Models 2/3; also, when controlling for host gender.
Further, we checked the interaction term for the combination female hosts/ female guests (Model 4). As
expected, this validates our results on the trust scores as reported in Table 16, however, the gender
effects have no significant impact on our model. As a final check, we also controlled for the influence of
the same gender on trust (i.e., a female guest evaluating a female host or vice versa with males; Model
5). Again, the effect is negligible in size and not significant (b=.015, p=.55). Table 17 summarizes the
results, and we conclude that there is no systematic gender bias in our data.
DV: Trust in Complementor; OLS regressions
Model (1)
(all cases)
Model (2)
(all cases)
Model (3)
(all cases)
Model (4)
(all cases)
Model (5)
(all cases)
Constant
0.386
***
0.410
***
0.389
***
0.393
***
0.401
***
(0.018)
(0.019)
(0.022)
(0.025)
(0.019)
Female Host
0.048
0.048
0.037
(0.025)
(0.025)
(0.038)
Female Guest
-0.001
-0.005
-0.014
(0.026)
(0.026)
(0.035)
Interaction
0.019
(0.051)
Same Gender
0.015
(0.025)
Observations
493
493
493
493
493
Adjusted R2
0.005
-0.002
0.003
0.002
-0.001
*** p<.001; ** p<.01; * p<.05; standard errors in parentheses
TABLE 17: EFFECTS OF GENDER (HOST/PARTICIPANT) ON ASSIGNED TRUST SCORES
49
Per host/guest combination
Appendix
146
Last, we also checked for the effect of host names and profile pictures on assigned trust scores. Figure 23
reports the mean assigned trust values (incl. their standard deviation) per either host name or blurred
profile image as previously introduced in “Chapter 4: Stimulus material” in this Appendix. The
analysis indicates the tendency that, on average, some female host names (e.g., Emma, Laura) exhibit
higher trust scores than some of their male counterparts (e.g., Philipp, Paul, Michael). Similarly, blurred
profile images with a female contour exhibit higher trust scores than male blurred profile pictures.
However, neither of the observed effects is statistically significant at the 5%-level (p=.068 for names;
p=.406 for images).
FIGURE 23: MEAN TRUST SCORES PER HOST NAME AND PROFILE IMAGE
Online Appendix
Appendix
147
Supplementary Material Part III
Chapter 6: Interview Guide
Opening the interview: The interviewer introduced themself to the interviewees and thanked
them for their participation. Then, background information of the research and definitions of
all terms and jargons used in the interview guide were clearly explained to the interviewees
before the interview. Interviewees were asked if they had any questions with regard to the
previously shared material prior to the interview. Interviewees were further briefed regarding
their informed consent.
Opening questions:
Benefits
1. In your opinion, what are the general benefits for users and/or platforms if
complementors are able to port their reputation between platforms? [ALL]
Mechanisms
2. Considering different mechanisms of porting reputation (e.g., direct import between
platforms, 3rd party providers like PIMS), which of these concepts promises to be most
successful? Why do you think so? [ALL]
3. Do you consider other mechanisms – besides those mentioned – as suitable to enable
seamless portability of online reputation across services/contexts? [OP, IN]
Main part: In the following, we will focus on direct import functions for reputation.
Design
4. In general, what do you think is important when designing a direct import/export
functionality for portable reputation between platforms? [ALL]
5. If external reputation could be "imported" onto a platform, what type of information
should be portable in the first place? Why? (star ratings, text reviews, both) [ALL]
6. Considering the import of ratings (e.g., a 1 to 5-star rating, a 1-100% rating, or 1 to 10
points), how should these be displayed on the target platform? [ALL]
i. Should these ratings' value(s) be converted into the standard on the target
platform (e.g., into a star rating)?
ii. Should these ratings be integrated into the on‑site/target ratings or displayed
distinctively from the on-site rating as an additional information?
iii. If displayed distinctively, should they be retained in their original value?
iv. Is there another way we have not discussed that these ratings should be
displayed?
7. How frequently, if it all, should these ratings be updated? That is, should it be a one-
time option to pull the data from the source platform or should this be designed in a
way that the ratings will be continuously updated? At what frequency? (weekly,
monthly, continuously) [OP, COMP]
8. In your opinion, should platforms (enabling the import of reputational data) define
themselves from which platforms their users are allowed to import the data? Why? Why
not? [OP, COMP]
Appendix
148
9. Similarly, if complementors are allowed to import reputation how much "choice" shall
they be given on the data to be pulled? That is, should they have the chance to pick the
ratings they would like to import from a particular platform, or should this be designed
as "all-or-nothing"? Why? [OP, COMP]
10. How feasible do you think is a "standardized" format for reputational data that is easy
to handle between platforms? In your opinion, what may such a standard look like?
[OP, COMP, IN]
11. What is your opinion on the role of “fit” between platforms (i.e., porting ratings between
Uber and Lyft seems like a more natural fit than porting ratings between Uber and
Airbnb or eBay)? Also, the gig economy faces these (cross-sector) challenges while on
e-commerce platforms, users’ experience with “selling stuff” is more homogenous [OP,
COMP, CONS]
12. How do you think platform operators could be encouraged to offer an import/export
function for reputational data? [OP, COMP, IN]
13. Similarly, how could users be incentivized to use such a feature to import reputational
data from another platform? [ALL]
Policy & Regulation
14. The General Data Protection Regulation (GDPR) mandates data to be portable between
data controllers (i.e., subjects). Do you think this regulation should extend to
reputational data as well? That is, should general reputation portability under GDPR
Art. 20 be enforced by GDPR? [COMP, CONS, IN]
15. Do you see other implications with regard to the GDPR and its applicability to
reputation portability – in particular for direct import functionality – under Art. 20?
[IN]
16. Do you have any other suggestions with regard to policy making and regulation for
reputation portability? [COMP, IN]
Risks & Challenges
17. Do you think porting reputation could raise the risk of post hoc manipulation of data
or is there another way in which such a system of rating import could be "played" or
misused? If so, how? How could that be avoided? [ALL]
18. Do you see a risk that very specific/niche platforms and their communities will lose
their (unique) identity with imported ratings? [ALL]
19. When thinking, for instance, of the Chinese Social Credit System, what is your opinion
on the trade-off between the right of self-determination and (online) data privacy on
one side and the risk of "transparent" citizens on the other side (i.e., ratings obligatory
showing up everywhere)? [ALL]
20. Do you see a potential conflict of interest between different stakeholder groups (i.e.,
operators, complementors, consumers) with regard to reputation portability? [ALL]
Closing Questions:
21. Besides portability between platforms/services, what other types of transfer of online
reputation to other contexts, for instance in the offline world, could you envision?
[ALL]
22. In your opinion, why is reputation portability rarely observed in today's practice? [ALL]
23. Are there any other design considerations that need to be taken into account? [ALL]
Appendix
149
Chapter 6: Sample
Stakeholder group
ID
Background
Gender
Platform operators
(OP)
OP1
CEO at a leading European temporary job platform
Male
OP2
Former Country Head and Operations Manager at Uber
Male
OP3
Senior researcher and trust scientist at leading sharing platforms
Male
Platform consumers
(CONS)
CONS1
Frequent platform user; Marketing Manager at an online store
Female
CONS2
Frequent platform user; CEO & Co-Founder at a car management startup
Male
Platform complementors
(COMP)
COMP1
Host on Airbnb (accommodation sharing)
Male
COMP2
Senior UX Designer on Fiverr (freelance services)
Male
Indirect stakeholders
(IN)
IN1
Project Director for data economy at a think tank; Economist
Female
IN2
Professor of Information Systems Management; Trust expert
Female
IN3
Director of a European Data Protection Foundation; Lawyer
Male
TABLE 18: BACKGROUND OF INTERVIEW PARTICIPANTS
150
References
151
References
Abramova, O., Krasnova, H., & Tan, C.-W. (2017). How much will you pay? Understanding the
value of information cues in the sharing economy. In ECIS 2017 Proceedings (pp. 1–18).
Guimaraes, Portugal.
Aljaroodi, H. M., Chiong, R., & Adam, M. T. P. (2020). Exploring the design of avatars for users
from Arabian culture through a hybrid approach of deductive and inductive reasoning.
Computers in Human Behavior, 106, 106246. https://doi.org/10.1016/j.chb.2020.106246
Almasoud, A. S., Hussain, F. K., & Hussain, O. K. (2020). Smart contracts for blockchain-based
reputation systems: A systematic literature review. Journal of Network and Computer
Applications, 170, 102814. https://doi.org/10.1016/j.jnca.2020.102814
Alt, R., & Zimmermann, H.-D. (2014). Editorial 24/3: Electronic Markets and general research.
Electronic Markets, 24(3), 161–164. https://doi.org/10.1007/s12525-014-0163-9
Arcade City. (2019). Pivoting Upward - Announcements - Arcade City.
https://forum.arcade.city/t/pivoting-upward/37. Accessed 12 April 2019
Ba, S., & Pavlou, P. A. (2002). Evidence of the Effect of Trust Building Technology in Electronic
Markets: Price Premiums and Buyer Behavior. MIS Quarterly, 26(3), 243–268.
https://doi.org/10.2307/4132332
Bajari, P., & Hortaçsu, A. (2004). Economic Insights from Internet Auctions. Journal of
Economic Literature, 42(2), 457–486. https://doi.org/10.1257/0022051041409075
Barnes, S. J., & Mattsson, J. (2017). Understanding collaborative consumption: Test of a
theoretical model. Technological Forecasting and Social Change, 118, 281–292.
https://doi.org/10.1016/j.techfore.2017.02.029
Barrett, C. (2019). Are the EU GDPR and the California CCPA becoming the global de facto
standard for data privacy and protection? Scitech Lawyer, 15(3), 24–29.
https://search.proquest.com/docview/2199825726/abstract/C384F029DD21402FPQ/1?
accountid=14264
Basili, M., & Rossi, M. A. (2020). Platform-mediated reputation systems in the sharing
economy and incentives to provide service quality: The case of ridesharing services.
Electronic Commerce Research and Applications, 39, 100835.
https://doi.org/10.1016/j.elerap.2019.100835
Beck, R., Avital, M., Rossi, M., & Thatcher, J. B. (2017). Blockchain Technology in Business and
Information Systems Research. Business & Information Systems Engineering, 59(6),
381–384. https://doi.org/10.1007/s12599-017-0505-1
Beck, R., Müller-Bloch, C., & King, J. L. (2018). Governance in the Blockchain Economy: A
Framework and Research Agenda. Journal of the Association for Information Systems,
19(10), 1020–1034. https://doi.org/10.17705/1jais.00518
Bellini, E., Iraqi, Y., & Damiani, E. (2020). Blockchain-Based Distributed Trust and Reputation
Management Systems: A Survey. IEEE Access, 8, 21127–21151.
https://doi.org/10.1109/ACCESS.2020.2969820
Bellotti, V., Ambard, A., Turner, D., Gossmann, C., Demkova, K., & Carroll, J. M. (2015). A
Muddle of Models of Motivation for Using Peer-to-Peer Economy Systems. In CHI 2015
Proceedings (pp. 1085–1094). New York, USA. https://doi.org/10.1145/2702123.2702272
Bera, P., Soffer, P., & Parsons, J. (2019). Using eye tracking to expose cognitive processes in
understanding conceptual models. MIS Quarterly, 43(4), 1105–1126.
References
152
https://doi.org/10.25300/MISQ/2019/14163
BMAS. (2020). So schaffen wir die soziale digitale Marktwirtschaft. German Federal Ministry
of Labour and Social Affairs.
https://www.bmas.de/DE/Service/Presse/Reden/Hubertus-Heil/2020/2020-11-19-
namensartikel-plattformoekonomie.html. Accessed 1 February 2021
Bogner, A., Chanson, M., & Meeuw, A. (2016). A Decentralised Sharing App running a Smart
Contract on the Ethereum Blockchain. In IoT 2016 Proceedings (pp. 177–178). New York,
USA. https://doi.org/10.1145/2991561.2998465
Bohn, S., Friederici, N., & Gümüsay, A. A. (2020). Too big to fail us? Platforms as systemically
relevant. Internet Policy Review. https://policyreview.info/articles/news/too-big-fail-us-
platforms-systemically-relevant/1489. Accessed 21 November 2020
Bollini, L. (2017). Beautiful interfaces. From user experience to user interface design. The
Design Journal, 20(sup1), 89–101. https://doi.org/10.1080/14606925.2017.1352649
Bolton, G., Greiner, B., & Ockenfels, A. (2013). Engineering Trust: Reciprocity in the
Production of Reputation Information. Management Science, 59(2), 265–285.
https://doi.org/10.1287/mnsc.1120.1609
Bonanza. (2019a). All About the eBay Feedback Importer.
https://www.bonanza.com/site_help/general_selling/feedback_importer. Accessed 12
December 2019
Bonanza. (2019b). About Us – Bonanza. https://www.bonanza.com/about_us. Accessed 20
April 2020
Boston Consulting Group. (2017). Hopping Aboard the Sharing Economy. BCG Henderson
Institute. https://www.bcg.com/publications/2017/strategy-accelerating-growth-
consumer-products-hopping-aboard-sharing-economy. Accessed 15 December 2020
Botsman, R. (2012). Welcome to the new reputation economy. Wired Magazine.
https://www.wired.co.uk/article/welcome-to-the-new-reputation-economy. Accessed 12
February 2019
Botsman, R. (2017a). Big data meets Big Brother as China moves to rate its citizens. Wired
Magazine. https://www.wired.co.uk/article/chinese-government-social-credit-score-
privacy-invasion. Accessed 4 November 2020
Botsman, R. (2017b). Who Can You Trust?: How Technology Brought Us Together and Why
It Might Drive Us Apart. New York, USA: PublicAffairs.
Botsman, R., & Rogers, R. (2010). What’s Mine Is Yours: The Rise of Collaborative
Consumption. UK: Harper Business.
Boyatzis, R. E. (1998). Transforming qualitative information: Thematic analysis and code
development. Thousand Oaks, USA: SAGE Publications, Inc.
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in
Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa
Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research
in Sport, Exercise and Health, 11(4), 589–597.
https://doi.org/10.1080/2159676X.2019.1628806
Broadbent, M. (2020). The Digital Services Act, the Digital Markets Act, and the New
Competition Tool. Center for Strategic & International Studies.
https://www.csis.org/analysis/digital-services-act-digital-markets-act-and-new-
competition-tool. Accessed 5 December 2020
References
153
Brown, N., & Stockman, T. (2013). Examining the Use of Thematic Analysis as a Tool for
Informing Design of New Family Communication Technologies. In HCI 2013 Proceedings
(pp. 1–6). London, UK. https://doi.org/10.14236/ewic/HCI2013.30
Bucher, E., Fieseler, C., & Lutz, C. (2016). What’s mine is yours (for a nominal fee) – Exploring
the spectrum of utilitarian to altruistic motives for Internet-mediated sharing. Computers
in Human Behavior, 62, 316–326. https://doi.org/10.1016/j.chb.2016.04.002
Buechler, M., Eerabathini, M., Hockenbrocht, C., & Wan, D. (2015). Decentralized Reputation
System for Transaction Networks. Dept. of CIS - Senior Design. Philadelphia, USA:
University of Pennsylvania.
http://www.seas.upenn.edu/~cse400/CSE400_2014_2015/reports/07_report.pdf
Burkhardt, D., Werling, M., & Lasi, H. (2018). Distributed Ledger. In 2018 IEEE International
Conference on Engineering, Technology and Innovation (ICE/ITMC) (pp. 1–9).
Stuttgart: IEEE. https://doi.org/10.1109/ICE.2018.8436299
Business Daily Africa. (2018). Uber driver loans from Sidian, Stanbic hit 500 - Business Daily.
https://www.businessdailyafrica.com/corporate/companies/Uber-driver-loans-from-
Sidian--Stanbic-hit-500/4003102-4740554-n8p00k/index.html. Accessed 3 April 2019
Buterin, V. (2014). A next-generation smart contract and decentralized application platform.
Ethereum White Paper, 1–36. https://github.com/ethereum/wiki/wiki/White-Paper
Cai, Y., & Zhu, D. (2016). Fraud detections for online businesses: a perspective from blockchain
technology. Financial Innovation, 2(20), 1–10. https://doi.org/10.1186/s40854-016-
0039-4
Campbell, C. (2019). How China is using big data to create a social credit score. Time
Magazine. http://time.com/collection/davos-2019/5502592/china-social-credit-score/.
Accessed 4 April 2019
Carboni, D. (2015). Feedback based Reputation on top of the Bitcoin Blockchain.
http://arxiv.org/abs/1502.01504
Catalini, C. (2017). Blockchain, explained | MIT Sloan. https://mitsloan.mit.edu/ideas-made-
to-matter/blockchain-explained. Accessed 4 June 2019
Catalini, C., & Gans, J. S. (2016). Some Simple Economics of the Blockchain. SSRN Electronic
Journal. https://doi.org/10.2139/ssrn.2874598
CBInsights. (2019). Blockchain Trends To Watch In 2019 - CB Insights Research.
https://www.cbinsights.com/research/blockchain-trends-2019/. Accessed 6 February
2019
Chen, C., & Chang, Y. (2018). What drives purchase intention on Airbnb? Perspectives of
consumer reviews, information quality, and media richness. Telematics & Informatics,
35(5), 1512–1523. https://doi.org/10.1016/j.tele.2018.03.019
Chen, X., Huang, Q., & Davison, R. M. (2017). Economic and social satisfaction of buyers on
consumer-to-consumer platforms: The role of relational capital. International Journal of
Electronic Commerce, 21(2), 219–248.
Chen, X., Huang, Q., Davison, R. M., & Hua, Z. (2015). What drives trust transfer? The
moderating roles of seller-specific and general institutional mechanisms. International
Journal of Electronic Commerce, 20(2), 261–289.
Chen, Y., Zhuang, C., Cao, Q., & Hui, P. (2014). Understanding Cross-site Linking in Online
Social Networks. In Social Network Mining and Analysis 2014 Proceedings (pp. 1–9).
New York, USA. https://doi.org/10.1145/2659480.2659498
References
154
Cheng, X., Fu, S., Sun, J., Bilgihan, A., & Okumus, F. (2019). An investigation on online reviews
in sharing economy driven hospitality platforms: A viewpoint of trust. Tourism
Management, 71, 366–377. https://doi.org/10.1016/j.tourman.2018.10.020
Corbin, J., & Strauss, A. (2008). Basics of Qualitative Research: Techniques and Procedures
for Developing Grounded Theory (2nd ed.). Thousand Oaks, USA: SAGE Publications.
https://doi.org/10.4135/9781452230153
Dakhlia, S., Davila, A., & Cumbie, B. (2016). Trust, but Verify: The Role of ICTs in the Sharing
Economy. In Lecture Notes in Information Systems and Organisation (pp. 303–311).
https://doi.org/10.1007/978-3-319-28907-6_20
Dann, D., Duong, M. H., Hoeven, E., Teubner, T., & Weinhardt, C. (2018). How Airbnb conveys
social and economic value through user representation. In Proceedings of the Workshop
on Designing User Assistance in Interactive Intelligent Systems (pp. 32–43).
Dann, D., Teubner, T., Adam, M. T. P., & Weinhardt, C. (2020). Where the host is part of the
deal: Social and economic value in the platform economy. Electronic Commerce Research
and Applications, 40, 100923. https://doi.org/10.1016/j.elerap.2019.100923
Daugherty, T., Eastin, M. S., & Bright, L. (2013). Exploring Consumer Motivations for Creating
User-Generated Content. Journal of Interactive Advertising, 8(2), 16–25.
https://doi.org/10.1080/15252019.2008.10722139
Davis, F. D. (1989). Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology. MIS Quarterly, 13(3), 319. https://doi.org/10.2307/249008
De Filippi, P. (2017). What Blockchain Means for the Sharing Economy. Harvard Business
Review Digital Articles. https://hbr.org/2017/03/what-blockchain-means-for-the-
sharing-economy. Accessed 3 April 2019
De Filippi, P., Shimony, O., & Tenorio-Fornés, A. (2020). Glossary of distributed technologies:
Reputation. Internet Policy Review. https://policyreview.info/open-abstracts/reputation.
Accessed 5 December 2020
De Hert, P., Papakonstantinou, V., Malgieri, G., Beslay, L., & Sanchez, I. (2018). The right to
data portability in the GDPR: Towards user-centric interoperability of digital services.
Computer Law and Security Review, 34(2), 193–203.
https://doi.org/10.1016/j.clsr.2017.10.003
De Reuver, M., Sørensen, C., & Basole, R. C. (2018). The digital platform: A research agenda.
Journal of Information Technology, 33(2), 124–135. https://doi.org/10.1057/s41265-
016-0033-3
De Winter, J. C. F., Zadpoor, A. A., & Dodou, D. (2014). The expansion of Google Scholar
versus Web of Science: a longitudinal study. Scientometrics, 98(2), 1547–1565.
https://doi.org/10.1007/s11192-013-1089-2
Deemly. (2019a). Build your trusted community. https://deemly.co/. Accessed 20 April 2020
Deemly. (2019b). What The GDPR Means For The Sharing Economy - deemly.
https://deemly.co/blog/what-the-gdpr-means-for-you-your-data/. Accessed 20 April
2020
Dellarocas, C., Dini, F., & Spagnolo, G. (2009). Designing reputation mechanisms. In N.
Dimitri, G. Piga, & G. Spagnolo (Eds.), Handbook of Procurement (pp. 446–482).
Cambridge University Press. https://doi.org/10.1017/cbo9780511492556.019
Dennis, R., & Owenson, G. (2016). Rep on the Roll: A Peer to Peer Reputation System Based on
a Rolling Blockchain. International Journal for Digital Society, 7(1), 1123–1134.
References
155
https://doi.org/10.20533/ijds.2040.2570.2016.0137
Dohmen, T., Falk, A., Huffman, D., Sunde, U., Schupp, J., & Wagner, G. G. (2011). Individual
risk attitudes: Measurement, determinants, and behavioral consequences. Journal of the
European Economic Association, 9(3), 522–550. https://doi.org/10.1111/j.1542-
4774.2011.01015.x
Dolnicar, S. (2017). Peer-to-Peer Accommodation Networks. Peer-to-Peer Accommodation
Networks. Goodfellow Publishers. https://doi.org/10.23912/9781911396512-3454
Drechsler, L. (2018). Practical challenges to the right of data portability in the collaborative
economy. In Proceedings of the 14th International Conference on Internet, Law &
Politics. Barcelona, Spain.
DREP Foundation. (2019). DREP - A Blockchain-Based Decentralized Reputation System.
https://www.drep.org/. Accessed 11 April 2019
Drescher, D. (2017). Blockchain Basics: A Non-Technical Introduction in 25 Steps. Berkeley,
CA: Apress. https://doi.org/10.1007/978-1-4842-2604-9
Dunham, B. (2011). The Role for Signaling Theory and Receiver Psychology in Marketing. In G.
Saad (Ed.), Evolutionary Psychology in the Business Sciences (pp. 225–256). Berlin,
Heidelberg: Springer. https://doi.org/10.1007/978-3-540-92784-6_9
Dunphy, P., & Petitcolas, F. A. P. (2018). A First Look at Identity Management Schemes on the
Blockchain. IEEE Security & Privacy, 16(4), 20–29.
https://doi.org/10.1109/MSP.2018.3111247
Eckhardt, G. M. (2020). Playing the Trust Game Successfully in the Reputation Economy. NIM
Marketing Intelligence Review, 12(2), 10–17. https://doi.org/10.2478/nimmir-2020-
0011
Economist. (2013). The rise of the sharing economy. The Economist.
https://www.economist.com/leaders/2013/03/09/the-rise-of-the-sharing-economy.
Accessed 14 May 2019
Edelman, B., Luca, M., & Svirsky, D. (2017). Racial discrimination in the sharing economy:
Evidence from a field experiment. American Economic Journal: Applied Economics, 9(2),
1–22. https://doi.org/10.1257/app.20160213
Engels, B. (2016). Data portability among online platforms. Internet Policy Review, 5(2), 1–17.
https://doi.org/10.14763/2016.2.408
Ert, E., & Fleischer, A. (2019). The evolution of trust in Airbnb: A case of home rental. Annals
of Tourism Research, 75, 279–287. https://doi.org/10.1016/j.annals.2019.01.004
Ert, E., Fleischer, A., & Magen, N. (2016). Trust and reputation in the sharing economy: The
role of personal photos in Airbnb. Tourism Management, 55, 62–73.
https://doi.org/10.1016/j.tourman.2016.01.013
EU Blockchain Observatory & Forum. (2019). Scalability Interoperability and Sustainability
of Blockchains. www.eublockchainforum.eu/reports.
European Commission. (2016). The future of the EU collaborative economy – Using scenarios
to explore future implications for employment. Available at:
https://ec.europa.eu/jrc/en/publication/eur-scientific-and-technical-research-
reports/future-eu-collaborative-economy-using-scenarios-explore-future-implications-
employment. https://doi.org/10.2760/354417
European Commission. (2017). Exploratory study of consumer issues in online peer-to-peer
platform markets. Available at: https://ec.europa.eu/newsroom/just/item-
References
156
detail.cfm?item_id=77704.
European Commission. (2018a). Study to Monitor the Economic Development of the
Collaborative Economy at sector level in the 28 EU Member States.
https://doi.org/10.2873/83555
European Commission. (2018b). Study on data in platform-to-business relations. Available at:
https://ec.europa.eu/digital-single-market/en/news/study-data-platform-business-
relations. https://doi.org/10.2873/602765
European Data Protection Supervisor. (2016). EDPS Opinion on Personal Information
Management Systems. Towards more user empowerment in managing and processing
personal data. Opinion 9 / 2016. Available at: https://edps.europa.eu/data-
protection/our-work/publications/opinions/personal-information-management-
systems_en. https://edps.europa.eu/data-protection/our-
work/publications/opinions/personal-information-management-systems_en
European Union. (2016a). Regulation 2016/679: “General Data Protection Regulation.”
Official Journal of the European Communities. Available at: https://eur-
lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32016R0679&from=EN.
European Union. (2016b). Recital 68 - Right of data portability | General Data Protection
Regulation (GDPR). https://gdpr-info.eu/recitals/no-68/. Accessed 19 March 2019
Faber, B., Michelet, G., Weidmann, N., Mukkamala, R. R., & Vatrapu, R. (2019). BPDIMS: A
Blockchain-based Personal Data and Identity Management System. In HICSS 2019
Proceedings (Vol. 6, pp. 6855–6864). https://doi.org/http://hdl.handle.net/10125/60121
Fagerstrøm, A., Pawar, S., Sigurdsson, V., Foxall, G. R., & Yani-de-Soriano, M. (2017). That
personal profile image might jeopardize your rental opportunity! On the relative impact of
the seller’s facial expressions upon buying behavior on Airbnb. Computers in Human
Behavior, 72, 123–131.
Fang, Y., Qureshi, I., Sun, H., McCole, P., Ramsey, E., & Lim, K. H. (2014). Trust, satisfaction,
and online repurchase intention: The moderating role of perceived effectiveness of e-
commerce institutional mechanisms. MIS Quarterly, 38(2), 407–427.
Faraday, P. (2000). Visually Critiquing Web Pages. In 6th Conference on Human Factors & the
Web (pp. 155–166). Austin, TX. https://doi.org/10.1007/978-3-7091-6771-7_17
Fereday, J., & Muir-Cochrane, E. (2006). Demonstrating Rigor Using Thematic Analysis: A
Hybrid Approach of Inductive and Deductive Coding and Theme Development.
International Journal of Qualitative Methods, 5(1), 80–92.
https://doi.org/10.1177/160940690600500107
Fertik, M., & Thompson, D. C. (2015). The reputation economy: how to optimize your digital
footprint in a world where your reputation is your most valuable asset. Crown Business.
https://www.penguinrandomhouse.com/books/223918/the-reputation-economy-by-
michael-fertik-and-david-thompson/9780385347600/
FinTech Council. (2019). Statement concerning the blockchain strategy of the German
government in the context of public consultations.
https://www.bundesfinanzministerium.de/Content/EN/Downloads/2019-04-17-
statement-fintech-council.pdf?__blob=publicationFile&v=5. Accessed 26 June 2019
Forbes. (2018). Global Digital Commerce Sales To Near $6 Trillion By 2022. Forbes.com.
https://www.forbes.com/sites/jordanmckee/2018/09/11/global-digital-commerce-sales-
to-near-6-trillion-by-2022/. Accessed 25 February 2019
Forbes. (2019a). When Is “Social Credit” Orwellian? Forbes.com.
References
157
https://www.forbes.com/sites/rhockett/2019/01/03/when-is-social-credit-orwellian/.
Accessed 4 April 2019
Forbes. (2019b). Warren Wants To Break Up Amazon, Facebook, Google. Forbes.com.
https://www.forbes.com/sites/kennethcorbin/2019/03/08/warren-wants-to-break-up-
amazon-facebook-google/. Accessed 10 February 2020
Fresneda, J. E., & Gefen, D. (2020). Gazing at the stars is not enough, look at the specific word
entropy, too! Information & Management, 57(8), 103388.
https://doi.org/10.1016/j.im.2020.103388
Gal-Oz, N., Grinshpoun, T., & Gudes, E. (2010). Sharing Reputation Across Virtual
Communities. Journal of Theoretical and Applied Electronic Commerce Research, 5(2),
1–25. https://doi.org/10.4067/S0718-18762010000200002
Gandini, A. (2020). Reaching for the Stars: The Role and Value of Digital Reputation. NIM
Marketing Intelligence Review, 12(2), 18–21. https://doi.org/10.2478/nimmir-2020-
0012
Gans, J. (2018). Enhancing Competition with Data and Identity Portability. The Hamilton
Project, (June 2018), 1–28.
http://www.hamiltonproject.org/assets/files/Gans_20180611.pdf
Garcia, P. (2018). Biometrics on the blockchain. Biometric Technology Today, 2018(5), 5–7.
https://doi.org/10.1016/S0969-4765(18)30067-5
Garett, R., Chiu, J., Zhang, L., & Young, S. D. (2016). A Literature Review: Website Design and
User Engagement. Online Journal of Communication and Media Technologies, 6(3).
https://doi.org/10.29333/ojcmt/2556
Gartner. (2020). Hype Cycle for Blockchain Technologies, 2020. Gartner Research.
https://www.gartner.com/en/documents/3987450/hype-cycle-for-blockchain-
technologies-2020. Accessed 5 December 2020
Gefen, D. (2000). E-commerce: The role of familiarity and trust. Omega, 28(6), 725–737.
https://doi.org/10.1016/S0305-0483(00)00021-9
Gefen, D. (2002). Reflections on the Dimensions of Trust and Trustworthiness among Online
Consumers. Data Base for Advances in Information Systems, 33(3), 38–53.
https://doi.org/10.1145/569905.569910
Gefen, D., Karahanna, E., & Straub, D. (2003). Trust and TAM in Online Shopping: An
Integrated Model. MIS Quarterly, 27(1), 51–90. https://doi.org/10.2307/30036519
Gefen, D., & Straub, D. (2003). Managing User Trust in B2C e-Services. e-Service Journal,
2(2), 7–23. https://doi.org/10.2979/esj.2003.2.2.7
Glaser, F. (2017). Pervasive Decentralisation of Digital Infrastructures: A Framework for
Blockchain enabled System and Use Case Analysis. In HICSS 2017 Proceedings (pp.
1543–1552). https://doi.org/10.24251/HICSS.2017.186
Glaser, F., Hawlitschek, F., & Notheisen, B. (2019). Blockchain as a Platform. In H. Treiblmaier
& R. Beck (Eds.), Business Transformation through Blockchain (pp. 121–143). Cham:
Springer. https://doi.org/10.1007/978-3-319-98911-2_4
Graef, I., Verschakelen, J., & Valcke, P. (2013). Putting the right to data portability into a
competition law perspective. The Journal of the Higher School of Economics, Annual
Review, 53–63. https://ssrn.com/abstract=2416537
Granovetter, M. (1985). Economic Action and Social Structure: The Problem of Embeddedness.
American Journal of Sociology, 91(3), 481–510. https://doi.org/10.1086/228311
References
158
Granovetter, M. (2005). The Impact of Social Structure on Economic Outcomes. Journal of
Economic Perspectives, 19(1), 33–50. https://doi.org/10.1257/0895330053147958
Gregory, R. W., & Muntermann, J. (2011). Theorizing in design science research: Inductive
versus deductive approaches. In ICIS 2011 Proceedings (pp. 888–904). Shanghai, China.
Greiner, B. (2015). Subject pool recruitment procedures: organizing experiments with ORSEE.
Journal of the Economic Science Association, 1(1), 114–125.
https://doi.org/10.1007/s40881-015-0004-4
Greiner, B., Teubner, T., & Weinhardt, C. (2018). Grundfragen der Plattformökonomie – Wie
man Vertrauen designt. In U. Blaurock, M. Schmidt-Kessel, & K. Erler (Eds.), Plattformen
(pp. 59–76). Nomos Verlagsgesellschaft mbH & Co. KG.
https://doi.org/10.5771/9783845292298-59
Grinshpoun, T., Gal-Oz, N., Meisels, A., & Gudes, E. (2009). CCR: A Model for Sharing
Reputation Knowledge Across Virtual Communities. In 2009 IEEE/WIC/ACM
International Joint Conference on Web Intelligence and Intelligent Agent Technology
(pp. 34–41). https://doi.org/10.1109/WI-IAT.2009.13
Guest, G., Bunce, A., & Johnson, L. (2006). How Many Interviews Are Enough?: An
Experiment with Data Saturation and Variability. Field Methods, 18(1), 59–82.
https://doi.org/10.1177/1525822X05279903
Gusenbauer, M. (2019). Google Scholar to overshadow them all? Comparing the sizes of 12
academic search engines and bibliographic databases. Scientometrics, 118(1), 177–214.
https://doi.org/10.1007/s11192-018-2958-5
Gutt, D., & Herrmann, P. (2015). Sharing Means Caring? Hosts’ Price Reaction to Rating
Visibility. In ECIS 2015 Proceedings (pp. 1–13). Münster, Germany.
http://aisel.aisnet.org/ecis2015_riphttp://aisel.aisnet.org/ecis2015_rip/54
Gutt, D., Neumann, J., Zimmermann, S., Kundisch, D., & Chen, J. (2019). Design of review
systems – A strategic instrument to shape online reviewing behavior and economic
outcomes. The Journal of Strategic Information Systems, 28(2), 104–117.
https://doi.org/10.1016/j.jsis.2019.01.004
Guttentag, D. (2015). Airbnb: disruptive innovation and the rise of an informal tourism
accommodation sector. Current Issues in Tourism, 18(12), 1192–1217.
https://doi.org/10.1080/13683500.2013.827159
Harris, L., & Rae, A. (2011). Building a personal brand through social networking. Journal of
Business Strategy, 32(5), 14–21. https://doi.org/10.1108/02756661111165435
Hatton, C. (2015). China “Social Credit”: China Sets up Huge System. BBC News.
https://www.bbc.com/news/world-asia-china-34592186. Accessed 4 April 2019
Hawlitschek, F., Notheisen, B., & Teubner, T. (2018). The limits of trust-free systems: A
literature review on blockchain technology and trust in the sharing economy. Electronic
Commerce Research and Applications, 29, 50–63.
https://doi.org/10.1016/j.elerap.2018.03.005
Hawlitschek, F., Teubner, T., & Gimpel, H. (2018). Consumer motives for peer-to-peer sharing.
Journal of Cleaner Production, 204, 144–157.
https://doi.org/10.1016/j.jclepro.2018.08.326
Hawlitschek, F., Teubner, T., & Weinhardt, C. (2016). Trust in the Sharing Economy. Die
Unternehmung, 70(1), 26–44. https://doi.org/10.5771/0042-059x-2016-1-26
Heiland, H. (2018). Review-Artikel: Zum aktuellen Stand des Plattformkapitalismus.
References
159
Industrielle Beziehungen. Zeitschrift für Arbeit, Organisation und Management, 25(1),
128–139. https://doi.org/10.3224/indbez.v25i1.06
Hein, A., Schreieck, M., Riasanow, T., Setzke, D. S., Wiesche, M., Böhm, M., & Krcmar, H.
(2020). Digital platform ecosystems. Electronic Markets, 30(1), 87–98.
https://doi.org/10.1007/s12525-019-00377-4
Hesse, M. (2020). Exploring the design of reputation portability in digital platform ecosystems
– A hybrid approach. ResearchGate, 1–25. https://doi.org/10.13140/RG.2.2.20583.47520
Hesse, M., Dann, D., Braesemann, F., & Teubner, T. (2020). Understanding the Platform
Economy: Signals, Trust, and Social Interaction. In HICSS 2020 Proceedings (pp. 5139–
5148).
Hesse, M., Lutz, O. H.-M., Adam, M. T. P., & Teubner, T. (2020). Gazing at the Stars: How
Signal Discrepancy Affects Purchase Intentions and Cognition. In ICIS 2020 Proceedings
(pp. 1–9). Hyderabad, India.
Hesse, M., & Teubner, T. (2020a). Reputation portability – quo vadis? Electronic Markets,
30(2), 331–349. https://doi.org/10.1007/s12525-019-00367-6
Hesse, M., & Teubner, T. (2020b). Takeaway Trust: A market data perspective on reputation
portability in electronic commerce. In HICSS 2020 Proceedings (pp. 5119–5128).
Hesse, M., Teubner, T., & Adam, M. T. P. (2020). Bring your own stars - the economics of
reputation portability. In ECIS 2020 Proceedings (pp. 1–10). Marrakesh, Morocco.
Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems
research. MIS Quarterly, 28(1), 75–105. https://doi.org/10.2307/25148625
Hodge, J., Balaam, M., Hastings, S., & Morrissey, K. (2018). Exploring the design of tailored
virtual reality experiences for people with dementia. In CHI 2018 Proceedings (pp. 1–13).
New York, USA. https://doi.org/10.1145/3173574.3174088
Holmqvist, K., Nystrom, M., Andersson, R., Dewhurst, R., Jarodzka, H., & van de Weijer, J.
(2011). Eye Tracking: A Comprehensive Guide to Methods and Measures. Oxford
University Press.
Hovorka, D. S., & Lee, A. S. (2010). Reframing interpretivism and positivism as understanding
and explanation: Consequences for information systems research. ICIS 2010 Proceedings,
1–13.
Huang, Q., Chen, X., Ou, C. X., Davison, R. M., & Hua, Z. (2017). Understanding buyers’ loyalty
to a C2C platform: The roles of social capital, satisfaction and perceived effectiveness of e-
commerce institutional mechanisms. Information Systems Journal, 27(1), 91–119.
Human Rights Watch. (2017). China’s Chilling ‘Social Credit’ Blacklist.
https://www.hrw.org/news/2017/12/12/chinas-chilling-social-credit-blacklist. Accessed
4 April 2019
HyperSpace. (2019). HyperSpace - Your Content. New Audiences. Real Income.
https://hyperspace.app/. Accessed 12 April 2019
ID2020. (2019). ID2020 | Digital Identity. https://id2020.org/digital-identity. Accessed 14
April 2019
Inglese, M. (2019). Regulating the Collaborative Economy in the European Union Digital
Single Market. Cham: Springer. https://doi.org/10.1007/978-3-030-30040-1
Jarvenpaa, S. L., Tractinsky, N., & Vitale, M. (2000). Consumer trust in an Internet store.
Information Technology and Management, 1(1/2), 45–71.
https://doi.org/https://doi.org/10.1023/A:1019104520776
References
160
Johnson, S., Robinson, P., & Brainard, J. (2019). Sidechains and interoperability.
http://arxiv.org/abs/1903.04077
Jøsang, A., Ismail, R., & Boyd, C. (2007). A survey of trust and reputation systems for online
service provision. Decision Support Systems, 43(2), 618–644.
https://doi.org/10.1016/j.dss.2005.05.019
Jung, J., & Lee, K.-P. (2017). Curiosity or Certainty? A Qualitative, Comparative Analysis of
Couchsurfing and Airbnb User Behaviors. In CHI 2017 Proceedings (pp. 1740–1747).
https://doi.org/10.1145/3027063.3053162
Kahneman, D., & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk.
Econometrica, 47(2), 263–291. https://doi.org/10.2307/1914185
Kaiser, R. (2014). Qualitative Experteninterviews. Wiesbaden: Springer Fachmedien
Wiesbaden. https://doi.org/10.1007/978-3-658-02479-6
Kathuria, V., & Lai, J. C. (2018). User review portability: Why and how? Computer Law and
Security Review, 34(6), 1291–1299. https://doi.org/10.1016/j.clsr.2018.05.018
Kenney, M., & Zysman, J. (2016). The rise of the platform economy. Issues in Science and
Technology, 32(3). https://issues.org/the-rise-of-the-platform-economy/
Khare, A., Labrecque, L. I., & Asare, A. K. (2011). The Assimilative and Contrastive Effects of
Word-of-Mouth Volume: An Experimental Examination of Online Consumer Ratings.
Journal of Retailing, 87(1), 111–126. https://doi.org/10.1016/j.jretai.2011.01.005
Kobie, N. (2019). The complicated truth about China’s social credit system. Wired Magazine.
https://www.wired.co.uk/article/china-social-credit-system-explained. Accessed 5 April
2019
Kokkodis, M., & Ipeirotis, P. G. (2016). Reputation Transferability in Online Labor Markets.
Management Science, 62(6), 1687–1706. https://doi.org/10.1287/mnsc.2015.2217
Krämer, J. (2018). Datenschutz 2.0 – ökonomische Auswirkungen von Datenportabilität im
Zeitalter des Datenkapitalismus. In T. Kretschmer et al. (Ed.), Datenkapitalismus – eine
ökonomische Betrachtung (Vol. 98, pp. 459–480). Wirtschaftsdienst.
https://doi.org/10.1007/s10273-018-2318-3
Krämer, J., & Stüdlein, N. (2019). Data portability, data disclosure and data-induced switching
costs: Some unintended consequences of the General Data Protection Regulation.
Economics Letters, 181, 99–103. https://doi.org/10.1016/j.econlet.2019.05.015
Kuckartz, U., & Rädiker, S. (2019). Analyzing Qualitative Data with MAXQDA. Cham:
Springer. https://doi.org/10.1007/978-3-030-15671-8
Kuzel, A. J. (1999). Sampling in Qualitative Inquiry. In B. F. Crabtree & W. L. Miller (Eds.),
Doing Qualitative Research (pp. 33–45). Thousand Oaks: SAGE Publications, Inc.
Kwon, K. H., Halavais, A., & Havener, S. (2015). Tweeting Badges: User Motivations for
Displaying Achievement in Publicly Networked Environments. Cyberpsychology,
Behavior, and Social Networking, 18(2), 93–100.
https://doi.org/10.1089/cyber.2014.0438
Lambrecht, C., & Heil, H. (2020). Digitalpolitik: Sollten Digitalplattformen stärker reguliert
werden? WirtschaftsWoche. https://www.wiwo.de/politik/deutschland/digitalpolitik-so-
schaffen-wir-die-soziale-digitale-marktwirtschaft/26637328.html. Accessed 5 December
2020
Lee, D., & Hosanagar, K. (2020). How Do Product Attributes and Reviews Moderate the
Impact of Recommender Systems Through Purchase Stages? Management Science, in
References
161
press. https://doi.org/10.1287/mnsc.2019.3546
Legris, P., Ingham, J., & Collerette, P. (2003). Why do people use information technology? A
critical review of the technology acceptance model. Information & Management, 40(3),
191–204. https://doi.org/10.1016/S0378-7206(01)00143-4
Liang, F., Das, V., Kostyuk, N., & Hussain, M. M. (2018). Constructing a Data-Driven Society:
China’s Social Credit System as a State Surveillance Infrastructure. Policy & Internet,
10(4), 415–453. https://doi.org/10.1002/poi3.183
Lim, S. Y., Tankam Fotsing, P., Almasri, A., Musa, O., Mat Kiah, M. L., Ang, T. F., & Ismail, R.
(2018). Blockchain Technology the Identity Management and Authentication Service
Disruptor: A Survey. International Journal on Advanced Science, Engineering and
Information Technology, 8(4–2), 1735–1745. https://doi.org/10.18517/ijaseit.8.4-2.6838
Ly, E., Miller, R., & Matsumura, M. (2019). Human Trust Protocol.
https://www.hubtoken.org/images/hub-white-paper.pdf. Accessed 11 April 2019
Ma, X., Hancock, J. T., Mingjie, K. L., & Naaman, M. (2017). Self-disclosure and perceived
trustworthiness of Airbnb host profiles. In CSCW 2017 Proceedings (pp. 1–13).
Mahmoud, M. A., Hinson, R. E., & Anim, P. A. (2018). Service innovation and customer
satisfaction: the role of customer value creation. European Journal of Innovation
Management, 21(3), 402–422. https://doi.org/10.1108/EJIM-09-2017-0117
Mao, Z. (Eddie), Jones, M. F., Li, M., Wei, W., & Lyu, J. (2020). Sleeping in a stranger’s home:
A trust formation model for Airbnb. Journal of Hospitality and Tourism Management,
42, 67–76. https://doi.org/10.1016/j.jhtm.2019.11.012
Martín-Martín, A., Orduna-Malea, E., Thelwall, M., & Delgado López-Cózar, E. (2018). Google
Scholar, Web of Science, and Scopus: A systematic comparison of citations in 252 subject
categories. Journal of Informetrics, 12(4), 1160–1177.
https://doi.org/10.1016/j.joi.2018.09.002
Martinelli, S. (2019). Sharing Data and Privacy in the Platform Economy: The Right to Data
Portability and “Porting Rights.” In L. Reins (Ed.), Regulating New Technologies in
Uncertain Times (pp. 133–152). T.M.C. Asser Press. https://doi.org/10.1007/978-94-
6265-279-8_8
Mauri, A. G., Minazzi, R., Nieto-García, M., & Viglia, G. (2018). Humanize your business. The
role of personal reputation in the sharing economy. International Journal of Hospitality
Management, 73, 36–43. https://doi.org/10.1016/j.ijhm.2018.01.017
Mavlanova, T., Benbunan-Fich, R., & Koufaris, M. (2012). Signaling theory and information
asymmetry in online commerce. Information and Management, 49(5), 240–247.
https://doi.org/10.1016/j.im.2012.05.004
Mazzella, F., Sundararajan, A., Butt d’Espous, V., & Möhlmann, M. (2016). How Digital Trust
Powers the Sharing Economy: The Digitization of Trust. IESE Insight, (30), 24–31.
https://doi.org/10.15581/002.ART-2887
McCurdie, T., Taneva, S., Casselman, M., Yeung, M., McDaniel, C., Ho, W., & Cafazzo, J.
(2012). mHealth Consumer Apps: The Case for User-Centered Design. Biomedical
Instrumentation & Technology, 46(s2), 49–56. https://doi.org/10.2345/0899-8205-
46.s2.49
Mele, M. L., & Federici, S. (2012). Gaze and eye-tracking solutions for psychological research.
Cognitive Processing, 13, 261–265. https://doi.org/10.1007/s10339-012-0499-z
Microsoft. (2018a). Cognitive Services | Face. https://azure.microsoft.com/en-
References
162
us/services/cognitive-services/face/. Accessed 14 May 2019
Microsoft. (2018b). How to: Sentiment analysis and Opinion Mining.
https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/how-tos/text-
analytics-how-to-sentiment-analysis. Accessed 14 May 2019
Minds. (2019). Home | Minds. https://www.minds.com/. Accessed 12 April 2019
Mingers, J. (2001). Combining IS Research Methods: Towards a Pluralist Methodology.
Information Systems Research, 12(3), 240–259.
https://doi.org/10.1287/isre.12.3.240.9709
Möhlmann, M. (2015). Collaborative consumption: determinants of satisfaction and the
likelihood of using a sharing economy option again. Journal of Consumer Behaviour,
14(3), 193–207. https://doi.org/10.1002/cb.1512
Möhlmann, M. (2016). Digital Trust and Peer-to-Peer Collaborative Consumption Platforms: A
Mediation Analysis. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2813367
Möhlmann, M., & Geissinger, A. (2018). Trust in the Sharing Economy. In N. Davidson, M.
Finck, & J. Infranca (Eds.), The Cambridge Handbook of the Law of the Sharing
Economy (pp. 27–37). Cambridge University Press.
https://doi.org/10.1017/9781108255882.003
Möhlmann, M., & Teubner, T. (2020). Navigating by the Stars: Current Challenges for
Ensuring Trust in the Sharing Economy. NIM Marketing Intelligence Review, 12(2), 22–
27. https://doi.org/10.2478/nimmir-2020-0013
Mudambi, S. M., Schuff, D., & Zhewei Zhang. (2014). Why Aren’t the Stars Aligned? An
Analysis of Online Review Content and Star Ratings. In HICSS 2014 Proceedings (pp.
3139–3147). https://doi.org/10.1109/HICSS.2014.389
Nakamoto, S. (2008). Bitcoin: A Peer-to-Peer Electronic Cash System. www.bitcoin.org.
https://bitcoin.org/bitcoin.pdf
Narayanan, A., Bonneau, J., Felten, E., Miller, A., & Goldfeder, S. (2016). Bitcoin and
Cryptocurrency Technologies: A Comprehensive Introduction. Princeton University
Press. http://bitcoinbook.cs.princeton.edu/
Neumann, J., & Gutt, D. (2017). A homeowner’s guide to airbnb: Theory and empirical
evidence for optimal pricing conditional on online ratings. In ECIS 2017 Proceedings (pp.
997–1010). Guimaraes, Portugal.
Niehaves, B. (2005). Epistemological perspectives on multi-method information systems
research. In ECIS 2005 Proceedings (pp. 1–12). Münster, Germany.
Nielsen, J., & Loranger, H. (2006). Prioritizing Web Usability. Thousand Oaks, USA: New
Riders Publishing.
NobelPrize. (2002). The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred
Nobel 2002. https://www.nobelprize.org/prizes/economic-sciences/2002/summary/.
Accessed 14 April 2020
Noorbergen, T. J., Adam, M. T. P., Attia, J. R., Cornforth, D. J., & Minichiello, M. (2019).
Exploring the design of mhealth systems for health behavior change using mobile
biosensors. Communications of the Association for Information Systems, 44(1), 944–981.
https://doi.org/10.17705/1CAIS.04444
Norbutas, L., Ruiter, S., & Corten, R. (2020). Reputation transferability across contexts:
Maintaining cooperation among anonymous cryptomarket actors when moving between
markets. International Journal of Drug Policy, 76, 102635.
References
163
https://doi.org/10.1016/j.drugpo.2019.102635
Norcie, G., De Cristofaro, E., & Bellotti, V. (2013). Bootstrapping Trust in Online Dating: Social
Verification of Online Dating Profiles. In A. Adams, M. Brenner, & M. Smith (Eds.),
Financial Cryptography and Data Security. FC 2013. Lecture Notes in Computer
Science, vol 7862 (pp. 149–163). Berlin, Heidelberg: Springer.
https://doi.org/10.1007/978-3-642-41320-9_10
Nowell, L. S., Norris, J. M., White, D. E., & Moules, N. J. (2017). Thematic Analysis: Striving to
Meet the Trustworthiness Criteria. International Journal of Qualitative Methods, 16(1),
1–13. https://doi.org/10.1177/1609406917733847
Origin. (2019). Origin Protocol. https://www.originprotocol.com/. Accessed 12 April 2019
Orquin, J. L., & Holmqvist, K. (2018). Threats to the validity of eye-movement research in
psychology. Behavior Research Methods, 50(4), 1645–1656.
https://doi.org/10.3758/s13428-017-0998-z
Otto, L., Angerer, P., & Zimmermann, S. (2018). Incorporating external trust signals on service
sharing platforms. In ECIS 2018 Proceedings (pp. 1–17). Portsmouth, UK.
https://aisel.aisnet.org/ecis2018_rp/78
Oxford Internet Institute. (2020). California’s Proposition 22 Presents an Alarming Turning
Point in Labor Law. Fairwork. https://fair.work/en/fw/blog/proposition-22/#continue.
Accessed 5 December 2020
Oyedele, A., & Simpson, P. (2018). Emerging adulthood, sharing utilities and intention to use
sharing services. Journal of Services Marketing, 32(2), 161–174.
https://doi.org/10.1108/JSM-09-2016-0344
Palan, S., & Schitter, C. (2018). Prolific.ac—A subject pool for online experiments. Journal of
Behavioral and Experimental Finance, 17, 22–27.
https://doi.org/10.1016/j.jbef.2017.12.004
Peer, E., Brandimarte, L., Samat, S., & Acquisti, A. (2017). Beyond the Turk: Alternative
platforms for crowdsourcing behavioral research. Journal of Experimental Social
Psychology, 70, 153–163. https://doi.org/10.1016/j.jesp.2017.01.006
Petter, S. C., & Gallivan, M. J. (2004). Toward a framework for classifying and guiding mixed
method research in information systems. HICSS 2004 Proceedings, 4061–4070.
https://doi.org/10.1109/hicss.2004.1265614
Pingel, F., & Steinbrecher, S. (2008). Multilateral Secure Cross-Community Reputation
Systems for Internet Communities. In S. Furnell, S. Katsikas, & A. Lioy (Eds.), Trust,
Privacy and Security in Digital Business. TrustBus 2008. Lecture Notes in Computer
Science, vol 5185 (pp. 69–78). Berlin, Heidelberg: Springer. https://doi.org/10.1007/978-
3-540-85735-8_8
Plansky, J., O’Donnell, T., & Richards, K. (2016). A Strategist’s Guide to Blockchain.
strategy+business. https://www.strategy-business.com/article/A-Strategists-Guide-to-
Blockchain. Accessed 9 April 2019
Puschmann, T., & Alt, R. (2016). Sharing Economy. Business & Information Systems
Engineering, 58(1), 93–99. https://doi.org/10.1007/s12599-015-0420-2
PwC. (2015). The Sharing Economy. Consumer Intelligence Series.
https://www.pwc.com/us/en/industry/entertainment-media/publications/consumer-
intelligence-series/assets/pwc-cis-sharing-economy.pdf. Accessed 9 February 2019
Qiu, W., Parigi, P., & Abrahao, B. (2018). More Stars or More Reviews? Differential Effects of
References
164
Reputation on Trust in the Sharing Economy. In CHI 2018 Proceedings (pp. 1–11).
Montréal, Canada. https://doi.org/10.1145/3173574.3173727
Resnick, P., & Zeckhauser, R. (2002). Trust among strangers in internet transactions:
Empirical analysis of eBay’s reputation system. In M. R. Baye (Ed.), The Economics of the
Internet and E-commerce (Advances in Applied Microeconomics) (Vol. 11, pp. 127–157).
Bingley, UK: Emerald Group Publishing Limited. https://doi.org/10.1016/S0278-
0984(02)11030-3
Resnick, P., Zeckhauser, R., Friedman, E., & Kuwabara, K. (2000). Reputation systems.
Communications of the ACM, 43(12), 45–48. https://doi.org/10.1145/355112.355122
Resnick, P., Zeckhauser, R., Swanson, J., & Lockwood, K. (2006). The value of reputation on
eBay: A controlled experiment. Experimental Economics, 9(2), 79–101.
https://doi.org/10.1007/s10683-006-4309-2
Riasanow, T., Ye, H. J., & Goswami, S. (2015). Generating Trust in Online Consumer Reviews
through Signaling: An Experimental Study. In HICSS 2015 Proceedings (pp. 3307–3316).
https://doi.org/10.1109/HICSS.2015.399
Risius, M., & Spohrer, K. (2017). A Blockchain Research Framework: What We (don’t) Know,
Where We Go from Here, and How We Will Get There. Business & Information Systems
Engineering, 59(6), 385–409. https://doi.org/10.1007/s12599-017-0506-0
Rötting, M. (2001). Parametersystematik der Augen- und Blickbewegungen fr
arbeitswissenschaftliche Untersuchungen. In Schriftenreihe Rationalisierung und
Humanisierung; 34 (p. 212).
Schaffner, D., Georgi, D., & Federspiel, E. (2017). Comparing Customer Experiences and Usage
Intentions Regarding Peer-to-Peer Sharing Platforms with Conventional Online Booking
Websites: The Role of Social, Hedonic, and Functional Values. In P. Rossi (Ed.), (pp.
1049–1056). Cham: Springer. https://doi.org/10.1007/978-3-319-47331-4_208
Schaub, A., Bazin, R., Hasan, O., & Brunie, L. (2016). A Trustless Privacy-Preserving
Reputation System. In IFIP Advances in Information and Communication Technology
(pp. 398–411). https://doi.org/10.1007/978-3-319-33630-5_27
Scheiber, N. (2017). How Uber Uses Psychological Tricks to Push Its Drivers’ Buttons. The New
York Times. https://www.nytimes.com/interactive/2017/04/02/technology/uber-
drivers-psychological-tricks.html?emc=eta1. Accessed 14 May 2019
Schmidt, F. A. (2017). Digital Labour Markets in the Platform Economy. In Friedrich Ebert
Stiftung Bibliothek. http://library.fes.de/opus4/frontdoor/index/index/docId/43760
Schmidt, K. E., Liu, Y., & Sridharan, S. (2009). Webpage aesthetics, performance and usability:
Design variables and their effects. Ergonomics, 52(6), 631–643.
https://doi.org/10.1080/00140130802558995
Schoenmueller, V., Netzer, O., & Stahl, F. (2018). The Extreme Distribution of Online Reviews:
Prevalence, Drivers and Implications. SSRN Electronic Journal.
https://doi.org/10.2139/ssrn.3100217
Schulte-Mecklenbeck, M., Johnson, J. G., Böckenholt, U., Goldstein, D. G., Russo, J. E.,
Sullivan, N. J., & Willemsen, M. C. (2017). Process-Tracing Methods in Decision Making:
On Growing Up in the 70s. Current Directions in Psychological Science, 26(5), 442–450.
https://doi.org/10.1177/0963721417708229
Seebacher, S., & Schüritz, R. (2017). Blockchain Technology as an Enabler of Service Systems:
A Structured Literature Review. In S. Za, M. Drăgoicea, & M. Cavallari (Eds.), Lecture
Notes in Business Information Processing, vol 279 (pp. 12–23). Cham: Springer.
References
165
https://doi.org/10.1007/978-3-319-56925-3_2
Sen, S., & Lerman, D. (2007). Why are you telling me this? An examination into negative
consumer reviews on the web. Journal of Interactive Marketing, 21(4), 76–94.
https://doi.org/10.1002/dir.20090
Sharples, M., & Domingue, J. (2016). The Blockchain and Kudos: A Distributed System for
Educational Record, Reputation and Reward. In K. Verbert, M. Sharples, & T. Klobučar
(Eds.), Adaptive and Adaptable Learning. EC-TEL 2016. Lecture Notes in Computer
Science, vol 9891 (pp. 490–496). Cham: Springer. https://doi.org/10.1007/978-3-319-
45153-4_48
Sheeran, P. (2002). Intention—Behavior Relations: A Conceptual and Empirical Review.
European Review of Social Psychology, 12(1), 1–36.
https://doi.org/10.1080/14792772143000003
Silvestri, G., Yang, J., Bozzon, A., & Tagarelli, A. (2015). Linking accounts across social
networks: The case of StackOverflow, Github and Twitter. In Proceedings of the 1st
International Workshop on Knowledge Discovery on the WEB (pp. 41–52). Cagliari,
Italy.
Síthigh, D. Mac, & Siems, M. (2019). The Chinese social credit system: A model for other
countries? EUI Department of Law Research Paper No. 2019/01.
Snyder, H. (2019). Literature review as a research methodology: An overview and guidelines.
Journal of Business Research, 104, 333–339.
https://doi.org/10.1016/j.jbusres.2019.07.039
Spence, M. (1973). Job Market Signaling. The Quarterly Journal of Economics, 87(3), 355–
374. https://doi.org/10.2307/1882010
Stanbic Bank. (2017). Stanbic Bank and Uber in Kenya launch an innovative vehicle financing
solution. https://www.stanbicbank.co.ke/kenya/About-Us/News/Stanbic-Bank-and-
Uber-in-Kenya-launch-an-innovative-vehicle-financing-solution. Accessed 3 April 2019
Steem. (2019). Powering Communities and Opportunities - Steem. https://steem.com/.
Accessed 12 April 2019
Steward, M. D., Narus, J. A., & Roehm, M. L. (2018). An exploratory study of business-to-
business online customer reviews: external online professional communities and internal
vendor scorecards. Journal of the Academy of Marketing Science, 46(2), 173–189.
https://doi.org/10.1007/s11747-017-0556-3
Stol, K.-J., Ralph, P., & Fitzgerald, B. (2016). Grounded theory in software engineering
research. In Proceedings of the 38th International Conference on Software Engineering
(pp. 120–131). New York, USA. https://doi.org/10.1145/2884781.2884833
Stummer, C., Kundisch, D., & Decker, R. (2018). Platform launch strategies. Business &
Information Systems Engineering, 60(2), 167–173.
Sundararajan, A. (2014). Peer-to-Peer Businesses and the Sharing (Collaborative) Economy:
Overview, Economic Effects and Regulatory Issues.
https://docs.house.gov/meetings/SM/SM00/20140115/101613/HHRG-113-SM00-
20140115-SD003-U1.pdf
Sundararajan, A. (2016). The Sharing Economy: The End of Employment and the Rise of
Crowd-Based Capitalism. The MIT Press.
Swarm City. (2019). Swarm City. https://thisis.swarm.city/. Accessed 11 April 2019
Sweller, J., Van Merrienboer, J. J. G., & Paas, F. G. W. C. (1998). Cognitive Architecture and
References
166
Instructional Design. Educational Psychology Review, 10(3), 251–296.
https://doi.org/10.1023/A:1022193728205
Tadelis, S. (2016). Reputation and Feedback Systems in Online Platform Markets. Annual
Review of Economics, 8(1), 321–340. https://doi.org/10.1146/annurev-economics-
080315-015325
Tanaka, A., Parkinson, A., Settel, Z., & Tahiroglu, K. (2012). A Survey and Thematic Analysis
Approach as Input to the Design of Mobile Music GUIs. In Proceedings of the
International Conference on New Interfaces for Musical Expression (pp. 200–203). Ann
Arbor, USA.
Ter Huurne, M., Ronteltap, A., Corten, R., & Buskens, V. (2017). Antecedents of trust in the
sharing economy: A systematic review. Journal of Consumer Behaviour, 16(6), 485–498.
https://doi.org/10.1002/cb.1667
Teubner, T. (2018). The web of host–guest connections on Airbnb: a network perspective.
Journal of Systems and Information Technology, 20(3), 262–277.
https://doi.org/10.1108/JSIT-11-2017-0104
Teubner, T., Adam, M. T. P., Camacho, S., & Hassanein, K. (2014). Understanding resource
sharing in C2C platforms: The role of picture humanization. In ACIS 2014 Proceedings
(pp. 1–10).
Teubner, T., Adam, M. T. P., & Hawlitschek, F. (2020). Unlocking Online Reputation. Business
& Information Systems Engineering, 62(6), 501–513. https://doi.org/10.1007/s12599-
019-00620-4
Teubner, T., & Dann, D. (2018). How Platforms Build Trust. SSRN Electronic Journal.
https://doi.org/10.2139/ssrn.3266472
Teubner, T., & Hawlitschek, F. (2016). The Economics of Peer-to-Peer Online Sharing. In P.
Albinsson (Ed.), Karlsruher Institut für Technologie (KIT) (pp. 1–28). ABC-CLIO.
https://publikationen.bibliothek.kit.edu/1000078546
Teubner, T., Hawlitschek, F., & Adam, M. T. P. (2019). Reputation Transfer. Business &
Information Systems Engineering, 61(2), 229–235. https://doi.org/10.1007/s12599-018-
00574-z
Teubner, T., Hawlitschek, F., & Dann, D. (2017). Price Determinants on Airbnb: How
Reputation Pays Off in the Sharing Economy. Journal of Self-Governance and
Management Economics, 5(4), 53. https://doi.org/10.22381/JSME5420173
The Guardian. (2020). Prop 22: why Uber’s victory in California could harm gig workers
nationwide. https://www.theguardian.com/us-news/2020/nov/11/california-
proposition-22-uber-lyft-doordash-labor-laws. Accessed 14 November 2020
Thomke, S., & Von Hippel, E. (2002). Customers as innovators: A new way to create value.
Harvard Business Review. https://hbr.org/2002/04/customers-as-innovators-a-new-
way-to-create-value. Accessed 13 November 2020
Traity. (2019a). Traity | Helping people trust one another. https://traity.com/. Accessed 19
March 2019
Traity. (2019b). Traity | Alternative Credit Scoring. https://traity.com/scoring. Accessed 3
April 2019
Traity. (2019c). Reputation Network | Blockchain Privacy Protocol.
https://reputation.network/#home-usecases. Accessed 14 April 2019
TrueGether. (2019). TrueGether: About Us. https://www.truegether.com/aboutus.html.
References
167
Accessed 20 April 2020
Trustbond. (2019). Learn more about Trustbond. https://trustbond.com/about. Accessed 8
April 2019
Tussyadiah, I. P. (2015). An Exploratory Study on Drivers and Deterrents of Collaborative
Consumption in Travel. In I. P. Tussyadiah & A. Inversini (Eds.), Information and
Communication Technologies in Tourism 2015 (pp. 817–830). Cham: Springer.
https://doi.org/10.1007/978-3-319-14343-9_59
Tussyadiah, I. P. (2016a). Factors of satisfaction and intention to use peer-to-peer
accommodation. International Journal of Hospitality Management, 55, 70–80.
https://doi.org/10.1016/j.ijhm.2016.03.005
Tussyadiah, I. P. (2016b). Strategic Self-presentation in the Sharing Economy: Implications for
Host Branding. In A. Inversini & R. Schegg (Eds.), Information and Communication
Technologies in Tourism 2016 (pp. 695–708). Cham: Springer.
https://doi.org/10.1007/978-3-319-28231-2_50
Ursic, H. (2018). Unfolding the New-Born Right to Data Portability: Four Gateways to Data
Subject Control. SCRIPT-ed, 15(1), 42–69. https://doi.org/10.2966/scrip.150118.42
Vaismoradi, M., Turunen, H., & Bondas, T. (2013). Content analysis and thematic analysis:
Implications for conducting a qualitative descriptive study. Nursing & Health Sciences,
15(3), 398–405. https://doi.org/10.1111/nhs.12048
Van Alstyne, M. W., Parker, G. G., & Choudary, S. P. (2016). Pipelines, platforms, and the new
rules of strategy. Harvard Business Review, 4, 54–63.
Van der Auwermeulen, B. (2017). How to attribute the right to data portability in Europe: A
comparative analysis of legislations. Computer Law and Security Review, 33(1), 57–72.
https://doi.org/10.1016/j.clsr.2016.11.012
Van Dijck, J. (2009). Users like you? Theorizing agency in user-generated content. Media,
Culture and Society, 31(1), 41–58. https://doi.org/10.1177/0163443708098245
Vanberg, A. D., & Ünver, M. B. (2017). The right to data portability in the GDPR and EU
competition law: odd couple or dynamic duo? European Journal of Law and Technology,
8(1), 1–22.
Vasseur, A., Léger, P.-M., & Sénécal, S. (2019). Eye-tracking for IS Research: A Literature
Review. In 2019 Pre-ICIS Workshop (pp. 13–18). Munich, Germany.
Venkatadri, G., Goga, O., Zhong, C., Viswanath, B., Gummadi, K. P., & Sastry, N. (2016).
Strengthening Weak Identities Through Inter-Domain Trust Transfer. In Proceedings of
the 25th International Conference on World Wide Web (pp. 1249–1259). New York, USA:
ACM Press. https://doi.org/10.1145/2872427.2883015
Venkatesh, V., Thong, J., & Xu, X. (2016). Unified Theory of Acceptance and Use of
Technology: A Synthesis and the Road Ahead. Journal of the Association for Information
Systems, 17(5), 328–376. https://doi.org/10.17705/1jais.00428
Walsham, G. (1995). The Emergence of Interpretivism in IS Research. Information Systems
Research, 6(4), 376–394. https://doi.org/10.1287/isre.6.4.376
Wang, D., & Nicolau, J. L. (2017). Price determinants of sharing economy based
accommodation rental: A study of listings from 33 cities on Airbnb.com. International
Journal of Hospitality Management, 62, 120–131.
https://doi.org/10.1016/j.ijhm.2016.12.007
Watt, M., & Wu, H. (2018). Trust mechanisms and online platforms: A regulatory response. M-
References
168
RCBG Associate Working Paper No. 97, 1–83.
https://www.hks.harvard.edu/centers/mrcbg/publications/awp/awp97
Weber, R. (2004). The Rhetoric of Positivism versus Interpretivism: A Personal View. MIS
Quarterly, 28(1), iii–xii. https://doi.org/10.2307/25148621
Weinhardt, C. (2017). Platforms – Business Models and Contracts. In T. Pertot & M. D’Onofrio
(Eds.), European Consumer and Market Law (Vol. 6, p. 171). München: Verlag C.H. Beck
oHG.
Werbach, K. (2018). The Blockchain and the New Architecture of Trust. (S. Braman, Ed.). The
MIT Press. https://mitpress.mit.edu/books/blockchain-and-new-architecture-trust
Wessel, M., Thies, F., & Benlian, A. (2017). Competitive Positioning of Complementors on
Digital Platforms: Evidence from the Sharing Economy. In ICIS 2017 Proceedings (pp. 1–
18). Seoul, South Korea.
Willems, J., Waldner, C. J., & Ronquillo, J. C. (2019). Reputation Star Society: Are star ratings
consulted as substitute or complementary information? Decision Support Systems, 124,
113080. https://doi.org/10.1016/j.dss.2019.113080
Wong, K. L. X., & Dobson, A. S. (2019). We’re just data: Exploring China’s social credit system
in relation to digital platform ratings cultures in Westernised democracies. Global Media
and China, 4(2), 220–232. https://doi.org/10.1177/2059436419856090
World Economic Forum. (2018). Digital Identity – Why It Matters and Why It’s Important We
Get It Right. World Economic Forum. https://www.weforum.org/press/2018/01/digital-
identity-why-it-matters-and-why-it-s-important-we-get-it-right/. Accessed 12 March
2019
Wu, Y., Ngai, E. W. T., Wu, P., & Wu, C. (2020). Fake online reviews: Literature review,
synthesis, and directions for future research. Decision Support Systems, 132, 113280.
https://doi.org/10.1016/j.dss.2020.113280
Xie, K., & Mao, Z. (2017). The impacts of quality and quantity attributes of Airbnb hosts on
listing performance. International Journal of Contemporary Hospitality Management,
29(9), 2240–2260. https://doi.org/10.1108/IJCHM-07-2016-0345
Xu, L., Shah, N., Chen, L., Diallo, N., Gao, Z., Lu, Y., & Shi, W. (2017). Enabling the Sharing
Economy: Privacy Respecting Contract based on Public Blockchain. In Proceedings of the
ACM Workshop on Blockchain, Cryptocurrencies and Contracts (pp. 15–21). New York,
USA: ACM Press. https://doi.org/10.1145/3055518.3055527
Zarifis, A., Cheng, X., & Kroenung, J. (2019). Collaborative consumption for low and high trust
requiring business models: from fare sharing to supporting the elderly and people with
disability. International Journal of Electronic Business, 15(1).
https://doi.org/10.1504/IJEB.2019.099059
Zervas, G., Proserpio, D., & Byers, J. (2015). A First Look at Online Reputation on Airbnb,
Where Every Stay is Above Average. SSRN Electronic Journal, 1–22.
https://doi.org/10.2139/ssrn.2554500
Zloteanu, M., Harvey, N., Tuckett, D., & Livan, G. (2018). Digital Identity: The effect of trust
and reputation information on user judgement in the Sharing Economy. PLoS ONE,
13(12), 1–18. https://doi.org/10.1371/journal.pone.0209071
Zyskind, G., Nathan, O., & Pentland, A. S. (2015). Decentralizing Privacy: Using Blockchain to
Protect Personal Data. In 2015 IEEE Security and Privacy Workshops (pp. 180–184). San
Jose, USA: IEEE. https://doi.org/10.1109/SPW.2015.27
