ChapterPDF Available

Blockchain-Based Decentralized Business Models in the Sharing Economy: A Technology Adoption Perspective


Abstract and Figures

Recently, blockchain technology has increasingly been deemed to enable novel "decentralized" business models for the sharing economy and thereby potentially provide an alternative to extant "centralized" sharing economy business models. Using a technology adoption perspective, our article explores under which circumstances such blockchain-based decentralized sharing economy business models may be widely adopted. Building on extant research, we theorize on the factors that are relevant for adoption from the individual us-ers' perspective. We then derive eight potential adoption scenarios of blockchain-based decentralized sharing economy business models and explore adoption using an agent-based simulation for the short term vs. long term. Our analyses highlight the relatively high importance of individual attitudes toward decentralized business models vis-à-vis contextual influences, and show how adoption patterns vary depending on the time horizon for the different scenarios. We conclude our exploratory study by deriving research and practical implications for blockchain-based business models.
Content may be subject to copyright.
Cite this chapter as: Tumasjan, A. and Beutel, T. (Forthcoming). "Blockchain-based Decentralized Business
Models in the Sharing Economy: A Technology Adoption Perspective", in Treiblmaier, H. and Beck, R. "Busi-
ness Transformation through Blockchain", Volume 2, Cham, Switzerland: Palgrave Macmillan.
Andranik Tumasjan
Johannes Gutenberg University Mainz
Gutenberg School of Management and Economics
Chair of Management and Digital Transformation
Theodor Beutel
The University of Edinburgh
Business School
Recently, blockchain technology has increasingly been deemed to enable novel “decentralized” business
models for the sharing economy and thereby potentially provide an alternative to extant “centralized” sharing
economy business models. Using a technology adoption perspective, our article explores under which cir-
cumstances such blockchain-based decentralized sharing economy business models may be widely adopted.
Building on extant research, we theorize on the factors that are relevant for adoption from the individual us-
ers’ perspective. We then derive eight potential adoption scenarios of blockchain-based decentralized sharing
economy business models and explore adoption using an agent-based simulation for the short term vs. long
term. Our analyses highlight the relatively high importance of individual attitudes toward decentralized busi-
ness models vis-à-vis contextual influences, and show how adoption patterns vary depending on the time
horizon for the different scenarios. We conclude our exploratory study by deriving research and practical im-
plications for blockchain-based business models.
Keywords: sharing economy, business models, blockchain, technology adoption, agent-based modeling
Introduction: Blockchain technology and business models
Blockchain technology has recently gained strong attention in the media and in business practice (Gupta
2017). In a recent survey of US corporate executives, 61% claimed knowledge about blockchain technology
ranging from broad to expert, and among those with knowledge about blockchain, 42% believed that it will
disrupt their industry (Deloitte 2017). Moreover, in this group, 21% have already bought into blockchain
production, while 25% plan to do so in the next year (Deloitte 2017). In addition, beyond corporate block-
chain applications (that are mainly implementing private or federated blockchain solutions, also referred to
as “distributed ledger technologies”; BlockchainHub 2017; Tumasjan 2018), we currently witness an emerg-
ing landscape of startups using blockchain technology and creating entirely new business models. In this
vein, a recent study of blockchain startups found that worldwide more than 1,500 startups were building new
products, services, and business models based on blockchain technology, and that more than 1.5 billion US
dollars of venture capital were already invested in these startups (Friedlmaier et al. 2018).
Notably, based on Satoshi Nakamoto’s idea of adecentralized
digital currency (i.e., Bitcoin; Nakamoto
2008), many of these startups are explicitly building decentralized business models as an alternative to ex-
tant “centralized” business models. The idea is often formulated as “cutting out the middleman” and creating
a peer-to-peer network where a transaction is done directly between two parties as in the case of Bitcoin.
Transferring this idea to online platforms, several startups aim at creating an alternative to “sharing econ-
omy” platforms (e.g., Uber and AirBnB) instead of creating peer-to-peer business models where customers
and suppliers make transactions directly using blockchain technology. The goal in these cases is to build
“true” sharing economy business models in which users can make peer-to-peer transactions rather than
providing their data to and making transactions on centralized platforms entailing relatively high transaction
fees. These decentralized blockchain-based sharing economy business models
(BSEBM) are then presented
as an alternative to traditional, centralized sharing economy platforms (e.g., Uber, AirBnB, Upwork, and Fa-
cebook) with the goal of decentralized data storage and saving money on transaction fees (Tapscott and Tap-
scott 2016a). One example of such a BSEBM is OpenBazaar, a decentralized e-commerce marketplace based
on an open source software where people can buy and sell goods via a peer-to-peer network. Whereas its
centralized counterparts eBay or Amazon store data in a central database, charge fees, and use fiat cur-
rency payment, users on OpenBazaar distribute data across the network, do not charge fees, and use crypto-
Whereas Bitcoin was conceptualized as a decentralized digital currency by Nakamoto, we acknowledge that recent
research has found that Bitcoin and Ethereum are in fact much less decentralized than usually assumed (Gencer et al.,
For reasons of brevity, we refer to these as blockchain-based sharing economy business models (abbreviated as BSEBM)
in contrast to "traditional" sharing economy business models (abbreviated as SEBM).
currencies as a payment method (Patterson 2016). Similar concepts include, for example, La'Zooz (a decen-
tralized ridesharing solution), Filecoin (a decentralized data storage solution), and LaborX (a decentralized
labor hire marketplace)
While such BSEBM may be intriguing and worthwhile alternatives to their extant centralized sharing econ-
omy business models (SEBM), many challenges remain before BSEBM may reach mass-market adoption.
Among the main challenges are product/service usability and actual value added in the mass-market. In other
words, why should customers use decentralized solutions that may exhibit much lower levels of usability,
worse user experience, and possibly occasional technical problems, when they can likewise use extremely
convenient services such as Amazon, eBay or Upwork? Illustrating this point, a co-developer of OpenBazaar
stated that “[e]xisting centralized marketplaces like Amazon and eBay have had decades to build up an im-
pressive suite of features for their users. Our first release has the advantage of 0% fees and using Bitcoin, but
it will be a long time before we are as feature-rich as the big platforms (Patterson 2016). Moreover, among
other challenges, legal and regulatory challenges remain (e.g., settling money-back issues, dealing with
fraud, taxing issues etc.) as well as questions as to how the infrastructure is maintained, who it belongs to,
and which business paradigm changes are required to sustainably run such decentralized business models
(Beck et al. 2018; Seidel 2017; Voshmgir 2017).
Overall, our current knowledge of such BSEBM is scant. Aside from position papers, blogs, and popular ac-
counts (e.g., Tapscott and Tapscott 2016a), research has not yet systematically investigated under which con-
ditions such business models may be successful and which parameters influence the adoption of novel
BSEBM (for recent notable exceptions, see Beck et al. 2018). Therefore, our article addresses this gap in the
extant literature. We focus on customer adoption by building on a technology acceptance theoretical perspec-
tive (Karahanna et al. 1999; Venkatesh et al. 2003) to conceptualize relevant parameters of customer adop-
tion of BSEBM. Based on our conceptualization of relevant parameters, we then derive and theorize on dif-
ferent adoption scenarios that we subsequently explore in a simulation for the short vs. long term using an
agent-based modeling approach (Epstein 1999; Rand and Rust 2011; Treiblmaier 2017).
We make the following major contributions to the literature. First, we advance the current state of the litera-
ture by systematically theorizing on relevant parameters of user adoption of BSEBM. We do so by building
on extant research on technology acceptance (Dwivedi et al. 2017; Venkatesh et al. 2003), and conceptualiz-
ing relevant parameters which we use to derive possible adoption scenarios. Second, we contribute to the lit-
erature by exploring how these parameters interact in our adoption scenarios. Whereas prior research has of-
ten conceptualized adoption parameters in isolation or in a static fashion, our ABM approach allows us to
examine their joint influence in our adoption scenarios over time. Thereby, we explore user adoption in the
short vs. long term to identify potential differential temporal effects of adoption. Third, we contribute to the
In a comprehensive analysis, seventy-four startups building or operating blockchain-based decentralized sharing econ-
omy business models were identified as of 2017 (Schneck et al. 2018).
extant discourse by explicitly focusing on the users’ view. While prior research has often emphasized tech-
nological and governance issues of blockchain-based business models (Beck et al. 2018; Folkinshteyn et al.
2015; Voshmgir 2017), we advance the field by employing a user adoption perspective. Our approach is
based on the rationale that the success of such novel BSEBM will largely be decided by sustainable mass-
market adoption, and, hence, understanding the users’ perspective is crucial.
Theoretical conceptualization
In this exploratory study, we draw on several research streams that we will briefly review and integrate into
the following sections. Hence, the following paragraphs introduce digital platforms in the sharing economy,
followed by blockchain-based business models, and theories of technology adoption. Towards the end of this
section, we explain how we construct the core parameters of this study and present our research questions,
The sharing economy and concentration of power
Digital platforms for reallocating underutilized assets gave rise to SEBM that create value through such real-
location processes and charge a fraction of that value for their service of intermediation. SEBM shook in-
cumbents across industries from accommodation (e.g., AirBnB, HomeAway), transportation (e.g.,
BlaBlaCar, Uber), to household work and all kinds of freelance tasks (e.g., TaskRabbit, Upwork). These
business models go by many names, including the ‘sharing economy’ and ‘collaborative consumption’
(Botsman and Rogers 2011; Matzler et al. 2015). While there is an ongoing discussion whether and how
these and related terms differ, there are two common themes (Belk 2014). First, SEBM coordinate temporary
access to utilize goods and services. Second, to facilitate the former, SEBM rely on the Internet and mobile
devices (Belk 2014).
Some of these platforms reallocate assets (e.g., AirBnB for apartments), while others distribute labor (e.g.,
TaskRabbit for everyday services), and some a mix of both assets and labor (e.g., Uber for transportation ser-
vices). As a rule of thumb, startups may be seen part of the sharing economy if there is a true sense of shar-
ing and collaboration involved (Botsman 2015).
However, not only good comes along with these digital platforms of the sharing economy. Business models
of digital platforms are a prime example of winner-takes-all markets where economies of scale and profit
maximization foster market structures that are dominated by powerful platform owners (Brynjolfsson and
McAfee 2014). Ultimately, platform owners can use their power to establish or even enforce processes which
may disadvantage users and the public. Common concerns include privacy issues from irresponsible data us-
age, not transparent pricing mechanisms, and self-employment practices which unburden a platform owner
from paying employee benefits and employment tax (Edelman and Geradin 2016; Malhotra and van Alstyne
As these problems arise from a centralization of power, BSEBM may contribute to addressing such problem-
atic phenomena in the sharing economy (Abramova and Boehme 2016; Beck et al. 2018). In the following
section, we elaborate on the rationale underlying the argument that blockchain technology may be of value
for the sharing economy.
Blockchain-based business models in the sharing economy
Traditionally, online platform providers are a centralized entity that is, or rather needs to be, trusted by its
users. Through concepts such as platform cooperativism with peer-to-peer rather than centralized govern-
ance, blockchain technology could provide the infrastructure for decentralized trust (De Filippi 2017). How-
ever, different blockchain designs result in fundamentally different governance structures (e.g., hierarchy vs.
meritocracy) which require a thorough discourse among all involved stakeholders (Voshmgir 2017). To bal-
ance power in sharing economy settings, blockchain technology and the underlying protocols could essen-
tially replace a trusted platform operator in centralized structures with trust in the decentralized technology
itself (Jarvenpaa and Teigland 2017). The design of such decentralized technology and its implications for
trust (Seidel, 2017) is highly dependent on the particular blockchain technology in use.
In terms of terminology, the term ‘blockchain’ is criticized to be conceptually ambiguous in public discourse
(Jeffries 2018), as there is no such thing as ‘the blockchain’ but a variety of blockchains and blockchain-like
architectures (for review, see Yli-Huumo et al. 2016). Generally, blockchain technology may be defined as
a distributed database solution that maintains a continuously growing list of data records that are confirmed
by the nodes participating in it (Yli-Huumo et al. 2016, p. 1).
Blockchains can be private or public (Morabito 2017). So-called federated or consortium blockchains are de-
signed for a set or pre-selected, trusted participants within a controlled network. These types of blockchains
such as Hyperledger Fabric or R3 Corda are most common in the financial industry (Valenta and Sandner
2017), and are often termed private blockchains. Private and public blockchains are subsumed under the um-
brella term “distributed ledger technology” (DLT).
If there are no limitations to who can read or write, a blockchain is often termed public (Morabito 2017).
Public blockchains allow anyone to participate and ensure that unknown (and potentially untrustworthy) in-
dividuals can coordinate among each other without requiring a central, trusted unit to coordinate because
these blockchains use game theory to disincentive malicious activity (Voshmgir 2017). Since BSEBM typi-
cally involve peer-to-peer interaction among individuals who do not necessarily know or trust each other,
public blockchains are well equipped to fulfill the requirements of SEBM.
Public blockchains (such as Ethereum and Bitcoin) could fundamentally change how a sharing economy is
implemented, once they unleash their potential for decentralized governance. In this vein, so-called decen-
tralized autonomous organizations (DAO) may promise a new type of organization that potentially requires
lower levels of transaction costs to organize itself while avoiding the pitfalls and dilemmas of centralized
governance (Voshmgir 2017). It is often argued that such DAOs may flatten or entirely remove hierarchical
structures in the digital economy by creating digital platforms that are governed and used on a peer-to-peer
basis by and for users (Beck et al. 2018). Effectively, BSEBM allow members to remain in control of the
platform, which could include rules of privacy and data protection as well as decisions on revenue generation
and profit distribution (Tapscott and Tapscott 2016a, 2016b).
Since current SEBM and digital platforms more broadly are highly centralized in nature and subject to a
range of shortcomings, could BSEBM give birth to a new, true sharing economy? To date, a variety of chal-
lenges and risks remain including technology maturity, regulatory uncertainties, and potentially still un-
known hurdles in designing functioning decentralized systems with effective incentive structures (Tapscott
and Tapscott 2016a). Especially such uncertainties in performance and characteristics of a system pose an
important challenge since there are no historical data to learn from when it comes to technology adoption and
the open question how users would decide when confronted with such a new BSEBM. To better understand
and investigate user adoption processes, in the next section, we elaborate on established frameworks for tech-
nology adoption on which we base our investigation.
Technology adoption
Established adoption theories encompass two major perspectives to explaining technology adoption. Macro
lens models describe and explain adoption at an inter-organizational or societal level, while micro lens mod-
els describe and explain intra-organizational or personal adoption (Frambach and Schillewaert 2002; Tarhini
et al. 2015).
Macro lens models such as the Bass diffusion model (Bass 1969) and Rogers (2003) diffusion of innovation
(DoI) theory describe the spread of a technology in a social system. Macro models, however, assume popula-
tions to be relatively homogenous in terms of preferences and connectedness among individuals. In the case
of DoI, heterogeneity is reflected to some extent as individual adopters are grouped into five categories in-
novators, early adopters, early majority, late majority, and laggards. These five groups of adopters behave
differently when confronted with a new technological innovation and shape the adoption rate and spread ac-
cording to DoI (Rogers 2003). While a macro view helps understanding emergent patterns in technology
adoption, for the purpose of modeling adoption and analyzing marketing activity for products and services,
micro-level models may be more appropriate. In particular, employing a micro lens is appropriate because
this perspective incorporates heterogeneous network structures on a local level, which allows creating mod-
els that are more realistic in terms of depicting social network structures (Bohlmann et al. 2010).
A variety of micro-level adoption models have been proposed to differentiate between heterogeneous indi-
viduals within a population. Exemplary technology adoption theories that model individuals’ behaviors at the
micro-level include the theory of planned behavior (TPB; Ajzen 1991), multiple versions of the technology
acceptance model (TAM; Davis et al. 1989; Venkatesh and Davis 2000), and two versions of the unified the-
ory of acceptance and use of technology (UTAUT; Venkatesh et al. 2003; Venkatesh et al. Xu 2012).
For our purpose of analyzing BSEBM, we posit that it is appropriate to employ a micro lens and focus on
individuals’ behavior. BSEBM are often conceptualized as two-sided marketplaces with peer-to-peer interac-
tions of individuals as consumers and suppliers at the individual level (Botsman and Rogers 2011). There-
fore, we argue that for our purpose, focusing on the predictors of individual adoption (rather than macro level
adoption) is more appropriate, and we therefore employ a micro lens to our research questions regarding user
In the following paragraphs, we therefore apply the logic of UTAUT because it is primarily designed for ex-
plaining individual adoption (Venkatesh et al. 2003) and has been widely used since it was proposed in the
early 2000s (for a review and meta-analysis, see Dwivedi et al. 2017). Thus, UTAUT serves as a conceptual
basis for our research by providing important elements for individual adoption modeling. However, we do
not attempt to replicate or apply the UTAUT model in its entirety for our research but rather use it as a basis
from which we derive some of our major parameters. We do so to preserve conceptual and computational
parsimony in our simulation model (see section Method and research design below), and to be able to inte-
grate further variables that are not part of the UTAUT model but that are nevertheless important for our mod-
eling of BSEBM. Thus, we build an integrative model that has UTAUT as a basis but also includes other pa-
rameters that are important for our purposes.
Hence, from the UTAUT model, we employ the construct performance expectancy (PE), defined as “the de-
gree to which an individual believes that using the system will help” them to benefit from its service (Ven-
katesh et al. 2003, 447). We also use the construct effort expectancy (EE), defined as „the degree of ease as-
sociated with the use of the system“ (Venkatesh et al. 2003, p. 450). However, we do not include the con-
struct of social influence because of mixed evidence of its utility and indications that social influence is only
significant among young women, older people or in mandatory settings (Hartwick and Barki 1994; Ven-
katesh et al. 2003). Moreover, facilitating conditions are excluded as they are predominantly conceptualized
for individuals in an organizational setting (Venkatesh et al. 2003).
Instead, and in line with TPB (Ajzen 1991), we include attitude (AT) which is “the user’s desirability of his
or her using the system” (Malhotra and Galletta 1999, p. 1) and is closely related to the intensity of an indi-
vidual’s value-based and ethical perception of the service (Dwivedi et al. 2017). AT is of central importance
in the pre-adoption phase when an individual decides to adopt a product (Karahanna et al. 1999) and is di-
rectly related to behavioral intention (Davis et al. 1989). Lastly, we also include pervasiveness (PV) as an
essential construct to reflect external influences from word-of-mouth and network effects which have been
deemed important for the adoption of two-sided digital platforms (Zutshi et al. 2014).
Conceptualization of adoption constructs
Drawing on the previous section where we introduce the theoretical constructs of technology adoption that
are relevant to our study, we will now elaborate on how we conceptualize these four main constructs for the
purpose of our investigation. We explain how the constructs relate to BSEBM and present eight scenarios
that logically emerge from our constructs.
Performance expectancy (PE). PE is related to constructs such as perceived usefulness (Davis et al. 1989)
and relative advantage (Moore and Benbasat 1991). Across multiple models, PE has been shown to be the
strongest predictor of behavioral intention to adopt a technology, which explicitly includes voluntary settings
such as our case (Venkatesh et al. 2003). In our context of BSEBM, PE describes how individuals would
make up their mind on whether they expect a BSEBM to meet their needs to a sufficiently high degree com-
pared to other SEBM. For simplicity, we model a direct effect of PE on use behavior. To illustrate, it is plau-
sible to assume that if individuals expect an BSEBM to serve their use case only to a small extent, the rela-
tive adoption of such a service is lower vis-à-vis other SEBM where the individuals expect to experience
greater value (e.g., receiving a quick transportation service from A to B, or staying at a local’s flat in a city
Effort expectancy (EE). Moreover, we include EE that, like PE, applies to voluntary usage contexts (Ven-
katesh et al. 2003). EE is similar to constructs from other theories describing the ease of use (Davis et al.
1989; Moore and Benbasat 1991) and complexity (Thompson et al. 1991). In their seminal work, Venkatesh
et al. (2003) find that EE is a strong mediator of facilitating conditions (e.g., the provision of support ser-
vices; Thompson et al. 1991) on behavioral use intention. Again, to keep the model reasonably simple, we
omit the construct of behavioral use intention and we model a direct effect on use behavior since intentions
have been shown to be the strongest predictor of actual behavior (Ajzen, 1991). In the present context, EE
becomes relevant in the process preceding BSEBM adoption, when individuals evaluate the effort they
would have to expend on installing the corresponding software for using the BSEBM and becoming familiar
with its user interface to perform the tasks that individuals want to accomplish (e.g., booking a ride or an ac-
commodation). If individuals expect that they would need to make a substantial effort before they could ex-
tract value from the BSEBM, a lower relative adoption would result as compared to other SEBM where the
expected effort is low.
Attitude (AT). The third construct of our model is AT which reflects “the user’s desirability of his or her us-
ing the system” (Malhotra and Galletta 1999, p. 1). In the adapted version of the UTAUT model (Dwivedi et
al. 2017), AT is influenced by social influence and by facilitating conditions, two constructs that are not part
of our model. Whereas AT also has an effect on intention, in their meta-analysis, Dwivedi et al. (2017) found
a comparatively stronger and direct effect of AT on use behavior. Including AT in our model is important for
our purpose of modeling BSEBM adoption. Since BSEBM can build upon the promises of distribution, im-
mutability, and trustlessness (contrary to conventional business models), a positive AT toward these features
will foster BSEBM adoption. Thus, if an individual values distribution (e.g., for increased resilience to avoid
the shutdown of services), immutability (e.g., for tamperproof records) or trustlessness (e.g., for substituting
trust in a third-party institution by trust in a distributed protocol), then this individual’s AT would be sub-
stantially stronger toward a BSEBM compared to other systems that do not provide these features.
Pervasiveness (PV). Our fourth construct, PV, is a measure of the distributedness of a technology or business
model within a population. For this study, we define PV as the result of external and internal forces that in-
fluence an individual’s technology adoption. For external forces, we refer to network effects and viral spread
(Gallagher and West 2009; Zutshi 2015). For internal forces, we include switching cost as the barriers an in-
dividual perceives when considering a change from one supplier to another (Farrell and Shapiro 1988). In the
context of digital business models, PV is fundamentally affected by network effects and viral spread (Zutshi
2015). In the context of digital platforms, network effects commonly describe a positive relationship between
the utility which a single consuming individual perceives and the amount of total consuming individuals of
the same product or service (Gallagher and West 2009). For BSEBM (and any SEBM) as two-sided market-
places, this means that these platforms typically face two-sided network effects where consumers perceive a
higher utility if the number of suppliers rises. Similarly, suppliers benefit from an increasing number of con-
sumers. The emergence of such network effects is facilitated by viral spread that occurs through word-of-
mouth among individuals within the reach of the radius of their personal network. At the same time, there is
a counteracting force in the form of habits and switching cost, which determines the minimum of additional
value an individual must expect before they decide to adopt a BSEBM (Farrell and Shapiro 1988). Likewise,
there is a threshold of negative user experience that, if exceeded, encourages an individual to abandon a
From these four parameters, we derive eight (two times four) possible scenarios describing the adoption of
BSEBM. In our first scenario analysis, which comprises four scenarios, the “perceived-benefit-matrix” (see
upper part of Table 1), we juxtapose PE and EE to model the value that users expect to extract from the
BSEBM’s features and usability. In an “ideal” adoption scenario, BSEBM are highly performant (high PE)
and easily usable (high EE), and, thus, competitive with other SEBM. In the opposite scenario of low PE and
low EE, user adoption would likely suffer. Intermediate combinations would lead to usability or performance
issues, with the well-performing parameter potentially offsetting the other low-performing parameter.
In our second scenario analysis, which comprises four scenarios, the internal/external trigger-matrix (see
lower part of Table 1), we juxtapose the combinations of high/low values of AT and high/low values of PV.
In an ideal scenario, again, we would expect fast-paced and sustainable adoption rates because users would
not only initially become aware of the BSEBM from a general hype (high PV) but would also stick to the
BSEBM due to their favorable AT toward the BSEBM. A low/low combination of AT and PV, however,
would negatively impact the viral spread and reduce customers’ motivation to use the BSEBM. Similarly, a
local or niche adoption describes a situation where there is low virality but a smaller community of people
who are sufficiently motivated by their AT toward the BSEBM. On the other hand, a temporary buzz may
also occur resulting from a constellation of high interest initially but no intention for individuals to perma-
nently adopt the BSEBM due to negative AT toward the value of BSEBM. Table 1 summarizes the scenarios
emerging from the combinations of PE and EE, and AT and PV, respectively.
------------------------------ Insert Table 1 about here ------------------------------
Late bloomer
BSEBM struggles because
users find it inconvenient
and insufficient.
Users expect BSEBM to
perform well but struggle
with barriers to use it.
Users expect BSEBM to be
good enough in usability
but experience performance
BSEBM is widely adopted
because users perceive little
barriers to adoption and
sufficient performance.
At risk of
Users are not keen to use
the BSEBM. Neither do
they hear from others nor
they talk themselves about
the service.
Local or niche
Users find the BSEBM ap-
pealing, but word of mouth
has limited reach.
BSEBM lands a viral hit
but fails to convince for
permanent usage.
mass adoption
Users are intrinsically moti-
vated by the BSEBM’s dis-
tinct features and willingly
spread the word.
Table 1: Scenarios of perceived-benefit-matrix and internal/external trigger-matrix
------------------------------ Insert Table 1 about here ------------------------------
Research questions
Above, we have argued that BSEBM could have relative advantages over centralized platforms, and, thus,
may potentially become widely adopted. To explore the eight adoption scenarios that we have derived above
in more detail, we model the underlying parameters using agent-based modeling. We do so by addressing the
following three research questions.
First, we have derived four parameters from the extant literatures on SEBM, blockchain technology, and user
adoption of technology, and study the influence of the parameters as reflected in the following research ques-
RQ1: How are the adoption parameters performance expectancy, effort expectancy, attitude, and pervasive-
ness related to the adoption of BSEBM?
Second, we derived eight scenarios based on scenario analyses of our four parameters, respectively, resulting
in the following research question:
RQ2: Which adoption patterns result for the scenarios and how do the patterns differ across the eight sce-
Third, to investigate sustainable adoption in the long term, we explore how these eight scenarios evolve in
the short vs. the long term as reflected in the following research question:
RQ3: How does the model predict adoption within the same scenarios in the short term vs. the long term?
In the next section, we operationalize our theoretical approach and introduce our methodology.
Methods and research design
We conduct a simulation study to explore the complex interdependencies in our variables. As discussed in
the previous section, traditional frameworks investigate adoption patterns typically through macro and micro
lenses. Both lenses can provide useful methods to gather and analyze data on the adoption of technology. On
the one hand, macro-level frameworks such as Rogers(2003) diffusion of innovation model help under-
standing market saturation. On the other hand, micro-level frameworks such as UTAUT (Venkatesh and Da-
vis 2000; Venkatesh et al. 2003) provide sophisticated ways to investigate the motivation of individuals and
their intentions. We intend to contribute to bridging this gap between micro and macro lenses by agent-based
modeling (ABM). Various studies have shown that ABM can significantly enhance theories for technology
adoption which are often inherently static, such as UTAUT (Bohlmann et al. 2010; Treiblmaier 2017).
Agent-based modeling
ABM is a “category of computational models invoking the dynamic actions, reactions and intercommunica-
tion protocols among the agents in a shared environment, in order to evaluate their design and performance
and derive insights on their emerging behavior and properties” (Abar et al. 2017, p. 14). In the past twenty
years, ABM has evolved into a widely accepted methodology in the social sciences, and in research on inno-
vation diffusion in particular (Garcia and Jager 2011), yet it remains arguably underused (Axelrod 1997; Ep-
stein 2012).
By modeling interdependencies between individuals rather than aggregating independently conceptualized
intentions of individuals, we take into account contingent behavior. Thereby, we consider how each individ-
ual’s behavior develops independently over time and model the macro lens accordingly (Bruch and Atwell
2015). Especially in an environment potentially coined by peer-to-peer interaction rather than centralized
governance, as indicated in the previous section, the result of reciprocal interaction among individuals is of
considerable importance. Such a context indicates that using ABM is appropriate since the methodology is
considered beneficial if research aims at investigating phenomena emerging from interactive entities (Rand
and Rust 2011). Doing so is particularly relevant in our case modeling diffusion and adoption of technology
as these entities are autonomous and heterogeneous in nature.
As of today, the blockchain ecosystem is still in its fluid phase of technological experimentation and has yet
to shift from experimentation to optimization (Iansiti and Lakhani 2017). Moreover, peers are naturally di-
verse in this context (i.e., potential adopters of this new technology across all groups of a society). Due to
these heterogeneous characteristics not only among agents but also among blockchain technologies, it is
challenging to model the adoption of BSEBM. ABM generally allows for taking such heterogeneity into ac-
count ( Bohlmann et al. 2010).
Rand and Rust (2011) propose six key indicators for evaluating the appropriateness of an ABM approach:
medium numbers, local and potentially complex interactions, heterogeneity, rich environments, temporal as-
pects, and adaptive agents.
First, medium numbers indicate that ABM is neither ideal for a very small number of agents where game the-
ory is superior, nor a very large number where statistical regression is typically more efficient. A medium-
sized population is typically facing a few, yet important interactions among individual agents (Casti 1996).
This criterion applies to our case, which indicates the appropriateness of ABM. Second, local and complex
interaction patterns are common to consumer adoption (Rand and Rust 2011) and characterize the case at
hand particularly well as the sharing economy relies heavily on local network effects (Malhotra and van
Alstyne 2014). Third, the characteristics of agents are indicative of using ABM if they are highly heterogene-
ous. In our model, we describe local networks with a high degree of individuality due to randomly allocated
preferences and continuous word-of-mouth among agents. Therefore, heterogeneity applies, too. Fourth, rich
or dynamic environments describe a complex interaction between an agent and its surroundings, which,
again, is warranted by assuming continuous word-of-mouth among agents. Fifth, temporality is the only nec-
essary factor and describes the need for investigating processes over time rather than at a static point in time.
With technology adoption, time and rate of adoption play a major role, for instance, when crossing the
chasm (Moore 2014). Sixth, the indicator of adaptive agents describes individuals who may change their
behavior dynamically based on new information. In the case at hand, we assume agents to act upon influence
from their network and change their behavior once they adopt a product, and are therefore adaptive. Thus,
overall, ABM can be considered appropriate for our study. In the following paragraphs, we describe the de-
sign of our model.
Design of the model
ABM consist of three components: agents, interactions (or relationships), and an environment (Epstein
2012; Rand and Rust 2011). By running the model for a certain amount of time units (also referred to as
steps or ticks), agents execute their behavior and engage in interactions with every time unit. Thus,
ABM is based on discrete events from activities over a certain amount of time units in which the agents in-
teract with each other in a given environment (Macal and North 2010). In the present case, we model the
adoption of a BSEBM over a time of 100 time units vs. 1,000 time units, and compare the impact of multiple
We constructed our model using NetLogo (Wilensky 1999) an ABM environment with sophisticated capa-
bilities to implement and run agent-based simulations and one of the most widely used tools for ABM (Rails-
back et al. 2006). NetLogo is considered suitable due to the software’s proven track record in the social sci-
ences and beyond, its comprehensive documentation and online support, and its feature-rich yet simple pro-
gramming language (Railsback et al. 2006). The models were constructed and run in NetLogo version 6.0.2.
To explore our research questions, we employ an existing NetLogo model, DYNAMOD
(Zutshi 2015), as a
basis. DYNAMOD is specifically designed to model the adoption of technologies and can be readily modi-
fied and extended (Zutshi 2015; Zutshi et al. 2013, 2014). We find DYNAMOD particularly useful as it fo-
cuses on digital business models such as multi-sided marketplaces and includes functionalities for modeling
network effects and viral spread. Moreover, DYNAMOD provides a useful construct of influence and satis-
faction values among to model adoption and rejection of a product.
Our model includes two groups of agents, blockchainers” and non-blockchainers”, describing their status
as a user of a BSEBM. Following the original DYNAMOD model, we set the total population to 2,400
We sincerely thank Aneesh Zutshi for providing the DYNAMOD model code for our research.
agents (max-agents). Out of the total population, 1% are assumed to be initial-blockchainer (i.e., users) from
the beginning (i.e., early adopters). In each group, a variable percentage of agents (moving-agents) are as-
sumed to be extraordinarily well connected to the whole population. We model these individuals by moving
their agent to another, randomly-determined location with every time unit within the reach of their network
Together, moving-agents and radius reflect the viral spread and word-of-mouth component of PV, a core pa-
rameter as described in the previous section. As a second component, minimum-satisfaction and minimum-
influence model the responsiveness of an agent toward the formerly described word-of-mouth component.
All agents that are non-blockchainers are subject to influence by their peers and by general network effects.
If an agent’s minimum-influence threshold is surpassed, then adoption would take place. The agent then be-
comes a blockchainer and subsequently receives a satisfaction score. If that score is lower than the minimum-
satisfaction threshold, the agent abandons the BSEBM and adoption is reversed.
Varying the parameter moving-agents and radius therefore affect the environment, whereas the parameters
minimum-satisfaction and minimum-influence determine the extent to which agents act on such influence.
Modeling performance expectancy and effort expectancy. As described in the previous section, we model
agents’ PE and EE regarding BSEBM. Comparing these two parameters allows revealing how favorably po-
tential users perceive the service. An agent may expect a BSEBM to perform well but high usability barriers
may keep the agent from adopting it. Likewise, a BSEBM with good usability would ultimately fail in deliv-
ering value if the agents expect low-class performance.
Modeling pervasiveness and attitude. Moreover, we juxtapose the two constructs of AT and PV. A direct
comparison of AT and PV reveals the impact of both on emerging patterns in user adoption. The reasoning is
to differentiate between external and internal drive toward adoption. Whereas PV reflects an external influ-
ence on an agent, AT reflects the agent’s own beliefs.
All our parameters PE, EE, AT and PV are modeled with values ranging from 0 to 100, except radius which
is scaled in whole numbers ranging from 0 and 20 (Zutshi 2015). For PE, EE, AT and PV, a higher value im-
plies a stronger impact on adoption. It is important to note that a higher parameter value of EE, therefore, re-
flects low (i.e., feasible) expected effort.
Model construction
For the previously described construct of PV, we utilize DYNAMOD’s (Zutshi 2015) capabilities of simulat-
ing network effects. A network-effect-coefficient is introduced which varies between -0.5 (network-effect-
lower) and 0.5 (network-effect-higher) depending on the number of adopters within the radius of an agent.
The network-effect-coefficient is multiplied by a network-effect-constant and added to the influence that is
continuously assigned to each agent at every time unit.
Adapted from DYNAMOD, we construct influence from an influence-constant and five other factors (pe-
influence, ee-influence, at-influence, nb-influence, and gl-influence). These five factors are dynamically
changed with every time unit and are the product of their corresponding parameter with a k-value to weigh
their impact (Zutshi 2015).
The last two factors, gl-influence and nb-influence, describe PV through a viral spread within the global con-
text (gl) and within an agent’s neighborhood (nb). Their corresponding parameter for gl-influence is the aver-
age influence of all other agents; for nb-influence, it is the average influence of other agents within the ra-
The parameters of the former three factors performance-expectancy for pe-influence, effort-expectancy for
ee-influence, and attitude for at-influence are part of the scenario analyses, as described earlier.
The k-values for PV (nb, gl) originate from DYNAMOD (Zutshi 2015). The remaining k-values for PE, EE
and AT stem from a UTAUT meta-analysis (Dwivedi et al. 2017) and underwent standardization. By multi-
plying Dwivedi et al.’s (2017) five-point Likert-scale values with 25 and subtracting 25, we transformed the
Likert-scale to values between 0 and 1. This scale, ranging from 0 to 1 with three decimals, is already availa-
ble in Zutshi (2015) and used in our model too.
The parameter avg-satisfaction is a key element of the DYNAMOD model as it plays an important role in
the decision-making process of an individual agent to be or not to be a user, depending on their satisfaction
and the minimum-satisfaction threshold.
The initial value of satisfaction has a major impact on the course of a model run and therefore requires rigor-
ous testing. Adapted from DYNAMOD, the average satisfaction avg-satisfaction ranges from -1 to 1 with a
standard deviation sd-satisfaction of 0.2. Its absolute value is closely connected to the absolute value of mini-
mum-satisfaction, as the difference between these two parameters is equivalent to the amount to which each
individual’s satisfaction level needs to rise for that particular individual to adopt the technology. Since vary-
ing the parameter minimum-satisfaction is part of a scenario-analysis, the parameter avg-satisfaction under-
goes a robustness test. Using NetLogo’s BehaviorSpace, the parameter avg-satisfaction is tested for the val-
ues -0.2 (default), -0.1 (higher) and -0.3 (lower). At the end of the next section, we elaborate on robustness
Overall, our model aims at exploring various potential scenarios. The context in question business models
for the sharing economy that are based on blockchain technology is currently emerging and therefore does
not yet provide robust empirical data to build on. Therefore, we base our parameters on previous related em-
pirical research on technology adoption and digital business models (Dwivedi et al. 2017; Venkatesh et al.
2003; Zutshi 2015) with the limitation that we thereby can only approximate our parameter values. There-
fore, we employ an explicitly exploratory and preliminary approach in our analyses.
To address our three research questions, we simulated the eight scenarios of our matrices (Figure 1). The first
matrix (perceived-benefit-matrix) reflects the combination of PE and EE. The second matrix (internal/exter-
nal trigger-matrix) relates AT to PV.
------------------------------ Insert Figure 1 about here ------------------------------
Figure 1: Perceived-benefit-matrix (left) and internal/external trigger-matrix (right)
------------------------------ Insert Figure 1 about here ------------------------------
In the following two sections, we describe the results for both matrices in detail.
The perceived benefit-matrix
When deciding to adopt the BSEBM, agents evaluate their expected performance and effort. The relationship
between the two parameters is summarized in Figure 1. In an ideal case of high expectations on both per-
formance and effort, the BSEBM may be competitive against traditional SEBM which are mostly of good
usability and performance. However, problems in usability and/or performance could diminish the relative
adoption and even result in failure.
Figures 2 and 3 show how the parameter values of PE, EE, AT, and PV, as described in our four scenarios
(Figure 1), are related to adoption patterns of BSEBM.
------------------------------ Insert Figure 2 about here ------------------------------
Figure 2: Short-term scenarios of perceived-benefit-matrix (100 time units)
------------------------------ Insert Figure 2 about here ------------------------------
For each scenario, we repeated the simulation for the long term. While Figure 2 presents the short term (100
time units), Figure 3 illustrates the long term (1,000 time units).
------------------------------ Insert Figure 3 about here ------------------------------
Figure 3: Long-term scenarios of perceived-benefit-matrix (1,000 time units)
------------------------------ Insert Figure 3 about here ------------------------------
Table 2 displays relative adoption in the short term (100 time units) and in the long term (1,000 time units) at
a glance. Each of the four cells presents a distinct percentage value describing the relative adoption of
BSEBM in each scenario in the short term and in the long term respectively. The remaining percentage val-
ues to 100% describe the total adoption of other SEBM for each scenario and duration. The numerical results
of each scenario in Table 2 are mean values from 50 repetitions and, hence, may slightly diverge from the
exemplary curves in Figures 2 and 3 which display a single run out of 50 runs.
------------------------------ Insert Table 2 about here ------------------------------
Effort Expectancy
Short term: 16.8%
Long term: 84.4%
Short term: 23.2%
Long term: 92.3%
Short term: 7.2%
Long term: 51.2%
Short term: 15.0%
Long term: 80.3%
Table 2: Perceived-benefit-matrix by market share of BSEBM (in %)
------------------------------ Insert Table 2 about here ------------------------------
In the following sections, we elaborate on these adoption patterns in each scenario in the short term and in
the long term.
Usability issues. If the expected performance is low but the expected effort is ideal, as shown in cell 1a in
Figure 2, the BSEBM fails to reach a majority of users. After 100 time units, only 16.8% of the agents
choose to use a BSEBM. Therefore, this constellation of parameters has a rather small impact on short-term
relative adoption. Initially, agents seem optimistic about the BSEBM and create notable growth in adoption
(Figure 2). However, the initial growth is soon to be corrected and is subsequently transformed into a tiny,
yet positive growth rate, indicating performance issues.
However, with 84.4% of the agents being a user after 1.000 time units, the model indicates that this constel-
lation of parameters may lead to substantial adoption in the long term (see cell 1b in Figure 3), and surpasses
the “late bloomer” scenario.
Late bloomer. The next scenario, that we coin “late bloomer”, is illustrated in cell 2a in Figure 2. After 100
time units, 7.2% of the agents choose to use a BSEBM. Therefore, this constellation reveals a lack of sub-
stantial traction among the agents to adopt the BSEBM in the short term. In the present case of low levels of
both PE and EE, a picture of stagnation emerges. An initial rise which is not an uncontrolled peak may seem
to indicate a rise toward sustainability at first but quickly flattens out and leads to near-zero growth until the
100th time unit. Expecting neither a feasible effort nor a valuable performance, the agents see no reason for
adopting the BSEBM.
However, the model indicates that this constellation of parameters may change fundamentally in the long
term and may eventually lead to substantial adoption among a majority of the agents (see cell 2b in Figure
3). After 1,000 time units, 51.2% of the agents have adopted the solution.
Performance issues. With a higher PE but poor EE, as shown in cell 3a in Figure 2, relative adoption remains
poor. After 100 time units, 15.0% of the agents choose to use a BSEBM. Therefore, this constellation of pa-
rameters has a similarly small impact on short-term relative adoption as the previously introduced set of pa-
rameter values. Like in the previous scenario, an initial rise in relative adoption is not maintained and subse-
quently flattens to a minor, yet non-zero growth rate, as Figure 2 shows. The result of usability issues, there-
fore, has a significant effect on relative adoption but still allows for growth.
Yet again, like before, with 80.3% of the agents being users after 1.000 time units, the model indicates con-
tinuous growth which eventually flattens out but only due to potential market saturation, as indicated by cell
3b in Figure 3. With the present set of assumptions, neither performance issues nor usability issues seem to
persist in the long term.
Competitive. The combination of high degrees of both PE and EE results in a striking degree of adoption
with 23.2% of the agents having adopted a BSEBM after only 100 time units. A minor initial peak and im-
mediate plunge in adoption do not stop subsequent growth in relative adoption (Figure 2, cell 4a). With a
solid number of users after already 100 time units, the BSEBM is considered competitive if it satisfies the
agents with both convenient usability and effective performance.
Moreover, with 92.3% of the agents being a user after 1,000 time units, the model indicates that these param-
eters continue the path of the short-term prediction and eventually lead to market saturation (Figure 3, cell
The internal/external trigger-matrix
Our second approach to modeling relative adoption of BSEBM entails contrasting internal (AT) and external
(PV) forces. On the one hand, agents are influenced by PV through viral spread and network effects. On the
other hand, their inner strength of AT also impacts their adoption behavior. Only a combination of both high
levels of AT and high levels of PV may be capable of reaching sustainable mass adoption. Merely having
high levels of PV may result in lacking sustainability, whereas low levels of PV may merely lead to niche
adoption or even extinction.
------------------------------ Insert Figure 4 about here ------------------------------
Figure 4: Short-term scenarios of internal/external trigger-matrix (100 time units)
------------------------------ Insert Figure 4 about here ------------------------------
As can be seen in Figures 4 and 5, relative adoption varies considerably both across the four scenarios and
also between the short term (100 time units) and the long term (1,000 time units).
------------------------------ Insert Figure 5 about here ------------------------------
Figure 5: Long-term scenarios of internal/external trigger-matrix (1,000 time units)
------------------------------ Insert Figure 5 about here ------------------------------
For an overview, Table 3 summarizes these results in a numerical format reflecting the relative adoption.
Like before in Table 2, all percentage values show how popular a BSEBM is in terms of adoption. The re-
maining percentage values to 100% reflect the total adoption including other SEBM. Like with Table 2, the
following numerical results in Table 3 show the mean value of 50 repetitions and may represent a slightly
different pattern than the exemplary model runs shown in Figures 4 and 5.
------------------------------ Insert Table 3 about here ------------------------------
Strength of
Short term: 8.3%
Long term: 39.5%
Short term: 16.9%
Long term: 95.0%
Short term: 3,2%
Long term: 6.4%
Short term: 6.0%
Long term: 6.3%
Table 3: Internal/external trigger-matrix by market share of BSEBM
------------------------------ Insert Table 3 about here ------------------------------
In the following sections, we elaborate on the results of each scenario.
Local or niche adoption. High AT but low PV, as shown in cell 5a in Figure 4, leads to an adoption of 8.3%
after 100 time units. This constellation of parameters, therefore, has a minor impact on short-term relative
adoption that also does not reach a mainstream of agents. As illustrated in cell 5a Figure 4, a pattern of local
or niche adoption emerges. The growth in adoption is tiny, yet almost continuously positive. This means that
there is a group of interested users who see value in the BSEBM but widespread adoption is not materializ-
ing due to low overall PV (e.g., minor viral spread).
Intriguingly, however, the long-term analysis after 1,000 time units with 39.5% of the agents reveals that this
set of parameters may eventually lead to substantial adoption (Figure 5, cell 5b).
At risk of extinction. As shown in Figure 4 in cell 6a, users are not at all motivated to adopt the BSEBM. Af-
ter 100 time units, only 3.2% of the agents choose to use a BSEBM, representing a low adoption over 100
time units. Thus, as expected, in the case of minor strength of AT and low overall PV, the BSEBM is not
adopted by a significant number of users. Potential users are neither motivated due to their own attitudes or
values nor are they triggered by their environment. The graph clearly indicates that adoption is stagnant and
does not peak at any given moment (Figure 4, cell 6a).
Even more intriguing, with 6.4% of the agents being users after 1,000 time units, the model continues to
grow at a near-zero rate. Notably, as shown in Figure 5 in cell 6b, the model does not result in extinction af-
ter 1,000 time units, yet the relative adoption is even farther away from total adoption.
Temporary buzz. With low AT and high PV (Figure 4, cell 7a), the resulting relative adoption is also ne-
glectable. After 100 time units, 6.0% of the agents choose to use a BSEBM.
As shown in Figure 4 the initial peak remains to have almost no long-term effect as the growth rate of users
after the peak remains almost perfectly steady at a value near zero. Initially, a large share of agents is prone
to adopt the BSEBM due to external influence from word-of-mouth and viral spread so that a large propor-
tion of agents chooses to try the service. However, due to their low internal strength of AT, the time as a user
is extremely short-lived for many agents who abandon the BSEBM immediately and never return.
With 6.3% of the agents being users after 1,000 time units (Figure 5, cell 7b), the model stagnates over the
remaining 900 time units and does not show any sign of growth or long-term potential.
Sustainable mass adoption. The fourth scenario, as expected, describes sustainable mass adoption. As illus-
trated in cell 8a in Figure 4 after 100 time units, 16.9% of the agents choose to use a BSEBM. Therefore, this
constellation of parameters has a positive yet medium impact on short-term adoption. With high AT as well
as high PV, a distinct pattern results that appears to illustrate organic growth. Even though, at first, the initial
peak reminds of a hype as it is followed by a strong decline in user count immediately after the peak. After
the drop, however, the BSEBM keeps growing continuously and reaches a remarkably steady user base.
Moreover, as shown in cell 8bin Figure 5 , with 95.0% of the agents being users after 1,000 time units, the
model indicates that this constellation of parameters may lead to a potential saturation of the market in the
long term.
In the next section, we explore the robustness of our model and the previously presented results.
Robustness of the model
Using NetLogo’s BehaviorSpace feature, we conducted thorough robustness checks of our model (following
Wilensky and Rand 2015, see also Alden et al. 2014 and Groenhuijzen and Verhagen 2016). BehaviorSpace
is a batch experiment tool in NetLogo to run a model multiple times with different settings (Wilensky and
Rand 2015), allowing for repeated model runs with identical parameters and further repetitions with varied
parameters. In the following sections, we refer to our original model with default values of both k-values and
avg-satisfaction as the “standard model”.
Random seed. NetLogo simulations include an integer random-seed which by default is generated pseudo-
randomly from the current date and time, thus is generated deterministically (Wilensky and Rand 2015). The
exact date and time will always lead to the same random seed. However, depending on the random seed, re-
sults can vary dramatically. This intentionally enriches the data to reflect on unknown unknowns that are
exogenous to this model. To preserve reproducibility of this model, an average was taken of each simulation
after running the model for 50 times with 50 different random seeds.
By taking the average result across 50 model runs with 50 different deterministically generated random seed
integers, our model, thus, ensures reproducibility while preserving randomness.
Robustness of k-values. We also evaluated how robust our model is to variations of the k-value correspond-
ing with each parameter by testing each scenario for lower (50%) and higher (100%) k-values.
In the perceived-benefit-matrix with 100 time units, reducing the k-values to 50% of the standard model re-
sults in an adoption between 48% and 93% compared to the standard model. The adoption rates in cells 1, 3
and 4 are similarly low (54%, 48% and 62% respectively), indicating a change that is roughly proportionate
to the standard model. Cell 2, however, stands out with an adoption rate of 93% compared to the standard
model, indicating that the combination of low/low PE and EE is barely affected by changed k-values and
does not respond proportionally, whereas the three other cells with low/high or high/high combinations of PE
and EE do vary.
Raising k-values to 200% compared to the standard model equally preserves their overall shape and leads to
a more pronounced initial peak. All cells 1, 2, 3 and 4 are similarly affected (188%, 179%, 172% and 181%
respectively) compared to the adoption rates in the standard model, indicating proportional change across the
In the internal/external trigger-matrix, after 100 time units, with reduced k-values of 50% of the standard
model, adoption rates range between 56% and 69% compared to the standard model. Thus, the cells 5, 6, 7
and 8 are all similarly affected by reduced k-values.
In the case of increased k-values of 200% compared to the standard model, the adoption rates after 100 time
units are between 96% and 180%. Cells 5, 6 and 8 are all affected at the upper end of the range (180%,
176%, 161%), indicating that any combination of low/high or high/high of AT and PV values changes the
adoption rate roughly proportionate to the increase of the k-values. The adoption rate in cell 7, however,
amounts to 96% compared to the standard model, showing that the combination of low AT and high PV val-
ues is barely affected by higher k-values and does not change proportionally.
From these analyses, we conclude that variations of the k-values keep the overall form of the curves in place
and reveal an impact on adoption rates that is mostly proportional to the change of the k-value. Hence, we
consider the model robust to changed k-values.
Robustness of satisfaction. As with the k-values, we tested our model for robustness to changed values of the
parameter avg-satisfaction. The following paragraphs present the results of changing avg-satisfaction from -
0.2 in the standard model to a lower (-0.3) and a higher (-0.1). These values are closely related to minimum-
In the perceived-benefit-matrix after 100 time units, a lower avg-satisfaction of -0.3 results in a graph of sim-
ilar shape with an equally significant initial peak. The adoption rate after 100 time units is between 29% and
45% compared to the standard model. Cell 2 with a low/low combination of the parameters PE and EE is
most affected with 45%. Cells 1, 3 and 4 are affected with 29%, 31%, and 36% respectively, showing that
any combination of the parameters PE and EE other than low/low is less affected by lower satisfaction lev-
Moreover, a higher avg-satisfaction of -0.1 leads to adoption rates between 226% and 377% compared to the
standard model, showing how sensitive the model responds to the avg-satisfaction parameter. However, as
before, the curves are similarly shaped. Cell 2 is most affected by 377%, indicating that low/low combina-
tions of PE and EE are highly affected by higher satisfaction levels. Compared to cell 2, the adoption rates in
cells 1, 3 and 4 are less affected (246%, 226% and 229% respectively). These results show that low/high or
high/high combinations of PE and EE are highly affected by adoption rates, but not as much as the low/low
In the internal/external trigger-matrix, again after 100 time units, both reducing and increasing avg-satisfac-
tion results in a similarly shaped graph with a comparably considerable initial peak in adoption rates. Tests
with the lower value (-0.3) of avg-satisfaction show that cells 5 and 6 result in adoption rates of 46% and
47% respectively, compared to the standard model. However, cells 7 and 8 with higher levels of PV result in
adoption rates of 19% and 29% of the standard model, respectively, indicating that scenarios with lower lev-
els of PV are less affected by lower satisfaction levels than scenarios with higher levels of PV.
Furthermore, increasing avg-satisfaction to -0.1 results in adoption rates between 226% and 394% in com-
parison to the standard model. Cells 6 and 7 are most affected (329% and 394% respectively), indicating that
the low/low and the low/high combinations of AT and PV values are more affected by higher satisfaction
levels than cells 5 and 8. The latter are less affected than cells 6 and 7 but with adoption rates of 237% and
226%, respectively, the cells 5 and 8 are still highly affected in comparison to the standard model.
As expected, we conclude that our robustness tests show that all eight scenarios are sensitive to varying val-
ues of the parameter avg-satisfaction. However, the overall form of all eight curves remains intact. Moreo-
ver, varying values of avg-satisfaction are not always resulting in uniform changes among adoption rates
(e.g., for a higher value of avg-satisfaction with the internal/external trigger-matrix, the adoption rates range
relatively widely from 226% and 394% as compared to the standard model).
Summary of results
We set out to examine the emerging phenomenon of blockchain, a technology that may give rise to novel de-
centralized business models in the sharing economy (Beck et al 2018; Voshmgir 2017; Tapscott and Tapscott
2016a, 2016b). In this vein, the objective of our article was to explore the adoption of such decentralized
blockchain-based sharing economy business models (BSEBM). More specifically, we investigated how
BSEBM adoption is impacted by the parameters PE, EE, AT, and PV that we derived from prior research
(Dwivedi et al. 2017, Zutshi 2015). We demonstrated that the adoption of BSEBM is dependent on users’
expectations on BSEBM performance, expected effort and usability, a user’s attitude and value-based per-
ception of BSEBM, and the overall pervasiveness of a BSEBM (e.g., in terms of network effects and switch-
ing barriers). Notably, the personal attitude of users toward BSEBM has a profound effect on whether adop-
tion is sustainable or remains temporary (e.g., in the form of a trend or hype). In other words, we found that
even strong pervasiveness is not sufficient for BSEBM to be permanently adopted if a user’s value-based
perception is not favorable to a particular BSEBM.
Moreover, our results show that time (i.e., short term vs. long term) has a profound impact on some but not
all scenarios. For instance, whereas a combination of low PE and low EE leads to negligible and stagnant
relative adoption in the short term, the same combination does eventually reach significant adoption in the
long term. On the other hand, low AT prevents adoption even when PV is high and even in the long term.
Implications for research
Our article demonstrates the importance of taking a user adoption perspective to BSEBM. While the extant
discourse often centers on technological and governance issues, we posit that for BSEBM the user perspec-
tive needs to be taken into account to a much larger extent. Our perspective is consistent with current ap-
proaches (e.g., UTAUT; Venkatesh et al. 2003) and earlier accounts of user adoption (e.g., technology diffu-
sion theory; Rogers, 2003) in highlighting the pivotal role of users’ attitudes and behaviors as a prerequisite
for potential mass-market technology adoption. Whereas prior research has conceptually examined to what
extent blockchain technology already meets central attributes of innovation diffusion (Friedlmaier et al.
2018), we extend this research stream by theoretically deriving possible adoption scenarios and simulating
these scenarios in the short vs. long term. While our results are congruent with prior studies in pointing to the
importance of including time (i.e., short vs. long term) as a crucial dimension in adoption models, our re-
search makes an exploratory effort, and we need much more fine-grained research to address temporal ques-
tions of user adoption (Zutshi 2015; Zutshi et al. 2013, 2014). Moreover, research is needed that includes
moderating variables, such as the moderating variables present in the UTAUT model (e.g., gender, age, or
experience; Venkatesh et al. 2003). Such research could reveal further adoption patterns among these varia-
bles that may otherwise be “averaged out. Moreover, including the UTAUT parameter social influence (i.e.,
“the degree to which an individual perceives that important others believe he or she should use the new sys-
tem”; Venkatesh et al. 2003, 451) would allow conclusions regarding society-wide values and norms. It
would be interesting to test whether social influence (external and value-driven) significantly differs from
AT (internal and value-driven) and PV (external and non-value-driven). Finally, we encourage further stud-
ies using variables from UTAUT2 (Venkatesh et al. 2012) and including its three novel constructs of hedonic
motivation, price value, and habit to theorize on their influence in BSEBM.
Our article also advances the current debate on BSEBM by highlighting the importance of usability variables
vis-à-vis users’ positive AT toward BSEBM. We maintain that for mass-market adoption, positive user atti-
tudes toward BSEBM will not be sufficient to motivate a considerable number of users to convert to BSEBM
but that rather usability and accessibility issues are of high importance for considerable adoption beyond a
niche. Users will not necessarily appreciate blockchain technology or the concept of decentralization per se
unless it is more than or at least as convenient as traditional online platforms of the sharing economy. In
fact, similar to most current Internet users being unaware of how the Internet actually works, users of
BSEBM will not necessarily understand the technical details of blockchain or use it for merely “cognitive”
or “ideological” reasons (e.g., decentralization). By pointing to these issues, our article lays the groundwork
for more research to investigate how customers make trade-off decisions between using decentralized
BSEBM vs. centralized SEBM. Since decentralized BSEBM have only recently been made possible by novel
technology, we need more theory and empirical research on when (i.e., under which circumstances and in
which contexts) decentralization vs. centralization is appropriate and actually desirable. A promising path-
way to evaluating decentralized vs. centralized business model designs could be to apply our model to the
first decentralized blockchain applications receiving widespread attention across a major population beyond
early adopters and technology enthusiasts. Understanding early successful applications that prove the value
of blockchain technology for novel business models (i.e., so called killer applications) may be an oppor-
tunity to conceptualize its value proposition.
Implications for practice
Our findings also have implications for practice both at an infrastructure level and at an application level. In
the light of some blockchains such as Bitcoin and Ethereum not being as decentralized as expected (Gencer
2018), users may conclude that a BSEBM may not sufficiently fulfill their attitudes and corresponding val-
ues and beliefs. As we showed in this study, the level of AT can have a profound effect on adoption. Moreo-
ver, users place a high value on convenience and accessibility when they anticipate the effort and perfor-
mance of a BSEBM. Therefore, blockchain startups need to ensure superior or at least sufficient usability
and user experience among BSEBM when they want to compete against their centralized counterparts.
To act on such suggestions, there need to be sound, well-founded educational resources and initiatives.
Building blockchain technology is fundamentally different from traditional software development in some
regards, for instance, in terms of testing environments, bug fixing procedures or system performance (Porru
et al. 2017). Likewise, developing a business model in the blockchain space including a clear idea of how a
blockchain-based business model creates value brings up completely new challenges. Thus, we need to build
competencies by employing extensive and early on education on blockchain for developers, designers, and
business developers alike.
Importantly, new blockchain startups need to clearly communicate where they see the benefit of blockchain
technology and explain the concrete value proposition over and above extant solutions in a precise and spe-
cific way. There is no point in building blockchain for the sake of blockchain, and decentralizing for the sake
of decentralization (Tumasjan 2018; Welpe et al. 2015). Instead, doing so requires careful investigation in-
cluding potential downsides. With blockchain still receiving an inflated degree of expectations according to
the Gartner Hype Cycle (Walker 2017), it is important to assess for which use cases blockchain is valuable,
and for which uses cases it does not make sense. Clearly, blockchain is no panacea, and startups should
therefore clearly point out how they build better solutions than existing ones.
Our study also has several limitations. First, ABM as a methodology is different from “traditional” empirical
research. Simulation-based models like ours are inherently limited due to their simplified design and, thus,
prone to predicting biased scenarios due to incompleteness (Seidl 2014). Moreover, like other agent-based
models, our model faces a trade-off between the number of agents and computational resources (Wilensky
and Rand 2015). Since increasing the number of agents requires significantly more computing power and
time, we restricted the population to a size of 2,400 agents.
Second, our own model has various inherent limitations. We only focus on four parameters to analyze our
adoption scenarios. We evaluated theories of technology adoption and provided a rationale on why we se-
lected the four parameters of PE, EE, AT, and PV. However, the literature on technology adoption provides
many other concepts that may likewise contribute to explaining the adoption of BSEBM, and which are not
part of our early exploratory study.
Third, our model illustrates adoption patterns in terms of their relative and roughly approximated magnitude
(rather than precise values) by comparing high and low values of each parameter using scenario analyses.
Moreover, our model computes heterogeneous agent parameters by normally distributing each parameter
through an inbuilt function in NetLogo. Therefore, the emerging patterns only represent an exploratory start-
ing point and, of course, require further validation.
Future research
Building on the results of our study, we encourage future research to validate and extend our model. First,
empirical research is needed to refine our parameters describing the driving forces in BSEBM adoption. Our
four parameters need to be replicated and refined through survey data on actual user data once such signifi-
cant amounts of empirical data on BSEBM adoption are available. Moreover other parameters should be
considered to enhance our scenarios and modeling of user behavior.
Second, decentralized business models overall require much more research. Many open questions remain and
include the following: What are the advantages and disadvantages of BSEBM? What are decisive criteria for
users to adopt such decentralized services? When, and in what way are BSEBM superior to centralized busi-
ness models? What are common trade-offs that users make between decentralized and centralized business
models? With our study, we lay the groundwork for future research reassessing and extending our findings
both conceptually and empirically.
Third, the sharing economy, in particular, provides many opportunities for further research on the specifics
of the potentially dawning decentralized paradigm of blockchain. What are conditions for decentralized busi-
ness models in the sharing economy to materialize? How do BSEBM compete against SEBM from incum-
bents? Which factors or stakeholders are decisive in fostering change?
Fourth, we encourage scholars from a variety of disciplines beyond computer science and management re-
search such as design research, law, economics, sociology, psychology, philosophy to investigate
BSEBM to explore their potential implications for economies and societies.
Fifth, we see great potential in ABM applications in the social sciences and follow Epstein’s (2012) call for
further experimentation and applications. To improve our model, we would initially propose to include more
parameters and data of actual BSEBM use. Moreover, we propose to extend our scenario analysis through
temporal analysis in order to explore new patterns and replicate existing results.
The emergence of blockchain as a peer-to-peer technology poses the question of how blockchain technology
and BSEBM may be adopted depending on individuals’ attitudes and their environment. We analyzed adop-
tion patterns across eight scenarios in the short term and in the long term. We found that users’ expectations
and attitudes play a crucial role for the adoption of BSEBM, and that adoption patterns vary considerably for
the short term vs. the long term. We hope that future research will build on our results to further advance our
knowledge on the increasingly important topic of blockchain-based business models.
Abar, S., G.K. Theodoropoulos, P. Lemarinier, and G.M.P. O’Hare. “Agent based modelling and simulation
tools: a review of the state-of-art software.” Computer Science Review 24 (2017): 1333.
Abramova, S., and R. Boehme. “Perceived benefit and risk as multidimensional determinants of bitcoin use:
a quantitative exploratory study.” Proceedings of the 37th International Conference on Information Sys-
tems (2016): 120.
Ajzen, I. “The theory of planned behavior.” Organizational Behavior and Human Decision Processes 50,
no. 2 (1991): 179211.
Alden, K., J. Timmis, and M. Coles. "Easing parameter sensitivity analysis of NetLogo simulations using
SPARTAN." Proceedings of the 14th International Conference on the Synthesis and Simulation of Living
Systems (2014): 622-628.
Axelrod, R. “Advancing the art of simulation in the social sciences.” In Simulating Social Phenomena.
vol. 456. Edited by G. Fandel et al., 2140. Lecture notes in economics and mathematical systems. Berlin,
Heidelberg: Springer Berlin Heidelberg, 1997.
Bass, F.M. “A new product growth for model consumer durables.” Management Science 15, no. 5 (1969):
Beck, R., C. Müller-Bloch, and J.L. King. “Governance in the blockchain economy - a framework and re-
search agenda.” Journal of the Association for Information Systems 19, no. 3 (2018): Forthcoming.
Belk, R. “You are what you can access: Sharing and collaborative consumption online.” Journal of Business
Research 67, no. 8 (2014): 15951600.
BlockchainHub. “Blockchains & distributed ledger technologies.”
and-distributed-ledger-technologies-in-general/ (accessed March 22, 2018).
Bohlmann, J.D., R.J. Calantone, and M. Zhao. “The effects of market network heterogeneity on innovation
diffusion: An agent-based modeling approach.” Journal of Product Innovation Management 27, no. 5
(2010): 741760.
Botsman, R. “Defining the sharing economy: What is collaborative consumption and what isn’t?”.
tion-and-what-isnt (accessed March 22, 2018).
Botsman, R., and R. Rogers. Whats mine is yours: How collaborative consumption is changing the way we
live. London: Collins, 2011.
Bruch, E., and J. Atwell. “Agent-based models in empirical social research.” Sociological Methods & Re-
search 44, no. 2 (2015): 186221.
Brynjolfsson, E., and A. McAfee. The second machine age: Work, progress, and prosperity in a time of bril-
liant technologies. New York, London: W.W. Norton & Company, 2014.
Casti, J.L. “Seeing the light at El Farol: A look at the most important problem in complex systems theory.”
Complexity 1, no. 5 (1996): 710.
Davis, F.D., R.P. Bagozzi, and P.R. Warshaw. “User acceptance of computer technology: A comparison of
two theoretical models.” Management Science 35, no. 8 (1989): 9821003.
De Filippi, P. “What blockchain means for the sharing economy.” Harvard Business Review (2017): 25.
Deloitte. “Deloitte survey: Blockchain reaches beyond financial services with some industries moving
faster”. (accessed March 27,
Dwivedi, Y.K., N.P. Rana, A. Jeyaraj, M. Clement, and M.D. Williams. “Re-examining the Unified Theory
of Acceptance and Use of Technology (UTAUT): Towards a revised theoretical model.” Information Sys-
tems Frontiers 110, no. 3 (2017): 392407.
Edelman, B.G., and D. Geradin. “Efficiencies and regulatory shortcuts: How should we regulate companies
like Airbnb and Uber?” Stanford Technology Law Review 19, no. 2 (2016): 293328.
Epstein, J.M. “Agent-based computational models and generative social science.” Complexity 4, no. 5
(1999): 4160.
——— . Generative social science: Studies in agent-based computational modeling. Princeton studies in
complexity 11. Princeton: Princeton University Press, 2012.
Farrell, J., and C. Shapiro. "Dynamic competition with switching costs." The RAND Journal of Economics
(1988): 123-137.
Folkinshteyn, D., N. J. Glassboro, M. Lennon, P. A. Altoona, and T. Reilly. “A tale of twin tech: Bitcoin and
the WWW.” Journal of Strategic and International Studies X, no. 2 (2015). 82-90.
Frambach, R.T., and N. Schillewaert. “Organizational innovation adoption: A multi-level framework of de-
terminants and opportunities for future research.” Journal of Business Research 55, no. 2 (2002): 163
Friedlmaier, M., A. Tumasjan, and I.M. Welpe. “Disrupting industries with blockchain: The industry, ven-
ture capital funding, and regional distribution of blockchain ventures.” Proceedings of the 51st Hawaii
International Conference on System Sciences (2018): 35173526.
Gallagher, S., and J. West. “Reconceptualizing and expanding the positive feedback network effects model:
A case study.” Journal of Engineering and Technology Management 26, no. 3 (2009): 131147.
Garcia, R., and W. Jager. “From the special issue editors: Agent-based modeling of innovation diffusion.”
Journal of Product Innovation Management 28, no. 2 (2011): 148151.
Gencer, A.E., S. Basu, I. Eyal, R. van Renesse, and E.G. Sirer. "Decentralization in Bitcoin and Ethereum
Networks." arXiv preprint arXiv:1801.03998 (2018): 118.
Groenhuijzen, M.R., and P. Verhagen. "Testing the robustness of local network metrics in research on arche-
ological local transport networks." Frontiers in Digital Humanities 3 (2016): 6.
Gupta, V. “A brief history of blockchain.” Harvard Business Review (2017): 24.
Hartwick, J., and H. Barki. “Explaining the role of user participation in information system use.” Manage-
ment Science 40, no. 4 (1994): 440465.
Iansiti, M., and K.R. Lakhani. “The truth about blockchain.” Harvard Business Review 95, no. 1 (2017):
Jarvenpaa, S., and R. Teigland. “Trust in digital environments: From the sharing economy to decentralized
autonomous organizations.” Proceedings of the 50th Hawaii International Conference on System Sci-
ences; HICSS (2017): 58125816.
Jeffries, A. “’Blockchain is meaningless.” blockchain-
bitcoin-ethereum-cryptocurrency-meaning (accessed March 27, 2018).
Karahanna, E., D.W. Straub, and N.L. Chervany. “Information technology adoption across time: A cross-
sectional comparison of pre-adoption and post-adoption beliefs.” MIS Quarterly 23, no. 2 (1999): 183
Macal, C.M., and M.J. North. “Tutorial on agent-based modelling and simulation.” Journal of Simulation 4,
no. 3 (2010): 151162.
Malhotra, A., and M. van Alstyne. “The dark side of the sharing economy … and how to lighten it.” Commu-
nications of the ACM 57, no. 11 (2014): 2427.
Malhotra, Y., and D.F. Galletta. "Extending the technology acceptance model to account for social influence:
theoretical bases and empirical validation." Proceedings of the 32nd Hawaii International Conference on
System Sciences; HICSS (1999).
Matzler, K., V. Veider, and W. Kathan. “Adapting to the sharing economy.” MIT Sloan Management Review
56, no. 2 (2015): 7177.
Moore, G.C., and I. Benbasat. “Development of an instrument to measure the perceptions of adopting an in-
formation technology innovation.” Information Systems Research 2, no. 3 (1991): 192222.
Moore, G.A. Crossing the chasm: Marketing and selling disruptive products to mainstream customers. 3rd.
ed. New York: Harper Collins, 2014.
Morabito, V. Business innovation through blockchain: The B³ perspective. Management/business for profes-
sionals. Cham: Springer International Publishing, 2017.
Nakamoto, S. “Bitcoin: A peer-to-peer electronic cash system.” (accessed No-
vember 30, 2017).
Patterson, S. “Openbazaar in depth: Interview with COO Sam Patterson.”
in-depth-interview-with-coo-sam-patterson/ (accessed March 25, 2018).
Porru, S., A. Pinna, M. Marchesi, and R. Tonelli. “Blockchain-oriented software engineering: challenges and
new directions.” Proceedings of the 39th International Conference on Software Engineering Companion
(2017): 169171.
Railsback, S.F., S.L. Lytinen, and S.K. Jackson. “Agent-based simulation platforms: Review and develop-
ment recommendations.” Simulation 82, no. 9 (2006): 609623.
Rand, W., and R.T. Rust. “Agent-based modeling in marketing: Guidelines for rigor.” International Journal
of Research in Marketing 28, no. 3 (2011): 181193.
Rogers, E.M. Diffusion of Innovation. 5th. New York: The Free Press, 2003.
Schneck, P., A. Tumasjan, and I.M. Welpe. Disrupting the disruptors? An in-depth analysis of the implica-
tions of blockchain technology for today’s sharing economy.” Working Paper, 2018.
Seidel, M.D.L. “Questioning centralized organizations in a time of distributed trust.” Journal of Management
Inquiry 19, no. 1 (2017): 1-5.
Seidl, R. “Social scientists, qualitative data, and agent-based modeling.” Proceedings ot the Social Simula-
tion Conference (2014): 14.
Tapscott, D., and A. Tapscott. Blockchain revolution: How the technology behind Bitcoin is changing
money, business and the world. London: Portfolio Penguin, 2016a.
——— . “The impact of the blockchain goes beyond financial services.” Harvard Business Review (2016b):
Tarhini, A., N.A.G. Arachchilage, R. Masadeh, and M.S. Abbasi. “A critical review of theories and models
of technology adoption and acceptance in information system research.” International Journal of Tech-
nology Diffusion 6, no. 4 (2015): 5877.
Thompson, R.L., C.A. Higgins, and J.M. Howell. “Personal computing: Toward a conceptual model of utili-
zation.” MIS Quarterly 15, no. 1 (1991): 125143.
Treiblmaier, H. “The world isn’t static, so why are we? How agent based modeling helps to create and test
dynamic IS models and theories.” Proceedings of the International Conference on Information Systems
(2017): 116.
Tumasjan, A. “Blockchain-Technologie und das Internet of Things: Kurzfristiger Hype oder eine Symbiose
für neue IoT-Geschäftsmodelle?” Industrie 4.0 Management 24, no. 2 (2018): 2932.
Valenta, M., and P. Sandner. “Comparison of Ethereum, Hyperledger Fabric and Corda.” FSBC Working Pa-
per (2017).
Venkatesh, V., M.G. Morris, and F.D. Davis. “User acceptance of information technology: Toward a unified
view.” MIS Quarterly 27, no. 3 (2003): 425478.
Venkatesh, V., and F.D. Davis. “A theoretical extension of the technology acceptance model: Four longitudi-
nal field studies.” Management Science 46, no. 2 (2000): 186204.
Venkatesh, V., J.Y.L. Thong, and X. Xu. “Consumer acceptance and use of information technology: Extend-
ing the Unified Theory of Acceptance and Use of Technology.” MIS Quarterly 36, no. 1 (2012): 157
Voshmgir, S. “Disrupting governance with blockchains and smart contracts.” Strategic Change 26, no. 5
(2017): 499509.
Walker, M.J. “Hype cycle for emerging technologies 2017.”
cycle-emerging-technologies (accessed February 12, 2018).
Welpe, I.M., A. Tumasjan, and C. Theurer. “Der Blick der Managementforschung“. In Das demokratische
Unternehmen: Neue Arbeits- und Führungskulturen im Zeitalter digitaler Wirtschaft. Edited by T. Sattel-
berger et al., 89103. Freiburg: Haufe, 2015.
Wilensky, U. "NetLogo”. Center for connected learning and computer-based modeling, Northwestern Uni-
versity, 1999.
Wilensky, U., and W. Rand. Introduction to agent-based modeling: Modeling natural, social, and engi-
neered complex systems with NetLogo. Cambridge, Massachusetts: The MIT Press, 2015.
Yli-Huumo, J., D. Ko, S. Choi, S. Park, and K. Smolander. "Where is current research on blockchain tech-
nology? A systematic review." PloS one 11, no. 10 (2016): 127.
Zutshi, A. “DYNAMOD – a dynamic agent based modelling framework for digital businesses.” PhD diss.,
Universidade Nova de Lisboa, 2015.
Zutshi, A., A. Grilo, and R. Jardim-Gonçalves. "DYNAMOD: A modelling framework for digital businesses
based on agent based modeling." Proceedings of the IEEE International Conference on Industrial Engi-
neering and Engineering Management; IEEM (2013): 13721376.
——— . “A dynamic agent-based modeling framework for digital business models: Applications to Face-
book and a popular Portuguese online classifieds website.” In Digital enterprise design & management.
vol. 261. Edited by P.J. Benghozi et al., 10517. Advances in intelligent systems and computing.
Cham: Springer International Publishing, 2014.
... Organizational alignment refers to the tension resulting from blockchain's intrinsic value proposition to reduce transaction costs through disintermediation and the use of the technology for proprietary purposes to create economic value (Schweizer et al., 2017;Tumasjan and Beutel, 2019). The latter is often accompanied by the centralization of infrastructures and insights into data, value capture in the ecosystem as a platform provider, and users' dependence on the technological infrastructure (Bogusz and Morisse, 2018;Schweizer et al., 2017). ...
... Companies that want to configure and implement a blockchain should, thus, carefully weigh the vulnerabilities of tensions that are accompanied by (partial) centralization and decentralization of blockchain's architecture (Renwick and Gleasure, 2021;Schweizer et al., 2017). Tensions in the area of environmental sensing might lead to urging companies to consider the loss of synergistic value through cooperation versus a loss of control over data and actors in the blockchain ecosystem (Schweizer et al., 2017;Tumasjan and Beutel, 2019). This involves design decisions that are reflected in the area of organizational alignment (Bogusz and Morisse, 2018;Renwick and Gleasure, 2021). ...
... This involves design decisions that are reflected in the area of organizational alignment (Bogusz and Morisse, 2018;Renwick and Gleasure, 2021). Companies should carefully consider whether a loss of operational efficiency due to the decentralization of data and infrastructure and the associated idea of social value outweigh the loss of democratization of data and thus, potentially, distributed innovation and technological mashups being beneficial for all participants in the ecosystem (Schweizer et al., 2017;Tumasjan and Beutel, 2019). (Cousins et al., 2019;Scholl et al., 2020;Yin et al., 2019) Data ...
Conference Paper
Full-text available
Given its decentralized nature, blockchain technology can lead to the emergence of various organizational or technological tensions, especially when used as an infrastructure for inter- and intra-corporate information exchange. These tensions are potentially a barrier to the adoption of the technology and can lead to vulnerabilities of the organizations and its long-term viability, so it is important that researchers have an overview, as well as an in-depth understanding, of the consequences of these tensions. We use a paradox lens and a literature review to synthesize the tensions emerging through blockchain adoption and implementation already identified in the IS literature. We develop an initial framework that classifies and provides an overview of different tensions, including governance-, regulatory-, and process-related issues. This resulting framework provides a holistic overview on different types of tensions that have so far only been considered case-specifically or separately in the IS literature.
... Using their monopoly-like status, the companies behind PBM are criticized for insufficiently distributing the produced value among those who mainly contributed to the value production, namely the buyer and the seller (De Filippi, 2017; Wright & De Filippi, 2015). Addressing this matter, new forms of platform-based business models, so-called decentralized platforms (DP) (Tumasjan & Beutel, 2019), emerged, building upon technological innovation, the blockchain technology. DP are considered to have a disruptive potential (Brennan et al., 2019), thereby becoming the harbinger of a new economic era, the 'true' platform economy (Belk, 2014), where monopolistic companies, in their role as the platform owner, become obsolete (Trabucchi et al., 2019;Voshmgir, 2019) while the value produced is being directly distributed among the platform users (Wright & De Filippi, 2015). ...
... Today, we observe that DP such as OpenBazaar (decentralized marketplace), OPUS (decentralized music sharing platform), and Steemit (decentralized social media platform) are starting to take their first steps into market segments being dominated by monopolistic centralized platforms (De Filippi, 2017; Perscheid et al., 2020; Tumasjan & Beutel, 2019). While DP are still far from representing a severe threat to those monopolistic centralized platforms today, the platform economy's future remains uncertain. ...
... In addition to the unique functionalities, DP offer its users, the usability of the services plays an essential role when choosing to use a DP (Tumasjan & Beutel, 2019). To this end, Sam Patterson, the COO of OB1 (the company that developed OB1), acknowledged in an interview that "Lots of Bitcoin applications have the ability to do technically challenging (and valuable) 2016). ...
Conference Paper
Full-text available
Digital platforms connect producers and consumers worldwide. The platform itself acts as a trusted intermediary matchingmarket participants and assisting the conclusion of contracts. Today, most platforms have become quasi-monopolies, exploiting their market power by charging monopoly fees and monetizing customer data.With the advent of blockchain technology, a new form of platform-based business model, so-called decentralized platforms (DPs), emerged. Using self-executing digital contracts (smart contracts) and blockchain technology, DPs automate the matching process and contract execution without the need for a centralized intermediary. By doing so, DPs drastically reduce transaction fees and return data sovereignty to the platform user.Consequently, DPs are considered to potentially disrupt the platform economy. To analyze DPsdisruptive potential, the characteristics distinguishing DPs from their centralized counterparts were identified and analyzed followingChristensen's theory of disruptive innovation. The findings attest that DPs have the potential to disrupt the platform economy in the sense of a low-end or even new-market disruption. In particular, the increasing popularity of blockchain and cryptographic tokens accelerates the growing prevalence of DPs, paving their way to become a severethreat to centralized platforms.
... Addressing this matter, new forms of decentralized platform-based business models (DPBMs) have emerged, building upon blockchain technology (Tumasjan and Beutel, 2019). These business models are considered to be the harbinger of a new economic era, i.e., the 'true' platform economy (Belk, 2014). ...
... While DPBMs are technically feasible, it is not yet clear whether or not business models without a governing entity are reasonably practicable, particularly because it is unclear to what extent and how decentralized governance can be realized (Voshmgir, 2017;Tumasjan and Beutel, 2019). Researchers' efforts to answer these questions have led to various conceptualizations of decentralized or blockchain governance (e.g., Atzori, 2015;Beck et al., 2018), which should be delimitated from centralized forms of decision making, accountability, and incentives (Beck et al., 2018). ...
... Since transparency and user participation are related to platform control and commercialization (Hein et al., 2016), centralized platforms do not provide the platform user with information on why governance processes take place (Staykova and Damsgaard, 2015;Tumasjan and Beutel, 2019). Although certain units within the company may be involved and, therefore, be given access to information, this does not account for all units and participants of a platform (Staykova and Damsgaard, 2014). ...
Conference Paper
Full-text available
Driven by blockchain technology, decentralized platform-based business models emerged that challenge current conceptualizations of platform governance. While researchers developed first conceptualizations of decentralized platform governance, hitherto, there does not exist a comprehensive framework capable of determining decentralized forms of governance. To develop such a framework, we ask, "What determines platform governance, and how can we capture gradients of decentralization?". To answer this question, we use a structured literature review to develop a framework encompassing three governance types: centralized, decentralized, and autonomous platform governance. Subsequently, we work out five components based on which the respective governance types are defined, thereby allowing a fine granular analysis of platform governance. We assert that the developed platform governance framework will assist researchers to determine divergent types of platform governance. Thereby, this paper delivers a deeper understanding and a basis for discussion on how decentralized governance may, or should look like.
... Blockchain and the Internet of things (IoT) represent two recent technological breakthroughs that can transform the way firms cooperate. These technologies have given rise to blockchain-based shared economy business models (BSEBMs), which have a highly decentralized nature compared to conventional SEBMs in digital platform ecosystems (Tumasjan & Beutel, 2019). In Section 3.1, I focus on role that these two technologies play in business model innovation and inter-firm cooperation. ...
... Even though businesses have transformed from traditional models to digital business models through IT-based business processes, BSEBMs may help an ecosystem to further evolve with seemless interfacing of decentralized business processes across organizations. Thus, BSEBMs pose different challenges from the more prevalent, centralized sharing economy platforms such as Uber and Airbnb (Chong, Lim, Hua, Zheng, & Tan, 2019;Tumasjan & Beutel, 2019). ...
Full-text available
In this exploratory research, I develop new knowledge on trust in inter-firm cooperation that leverages recent technologies such as blockchain and the Internet of things in a digital platform ecosystem. In a digital network, advanced algorithms govern and shape inter-firm business processes. While such algorithms introduce efficiency in inter-firm business processes, their limitations, especially their apparent lack of transparency, may affect the key trust dimensions (i.e., reliability, fairness, and goodwill) in the relationships among the participating firms. I introduce algorithmic relationship, a label that embeds the concepts of smart contracts in inter-firm cooperation. Algorithmic relationships involve autonomous and semi-autonomous implementations of smart contracts in all lifecycle stages of inter-firm cooperation. By analyzing extant literature on trust, inter-firm cooperation, business model innovation, and digital platforms, I demonstrate how various factors influence whether firms adopt smart contracts: perceptions about other participants’ trustworthiness, participants’ own propensity to trust, participants’ shared goals and resource embeddedness in the network, perceived risks in inter-firm interactions, and complexity and time criticality of inter-firm interactions. Taking a temporal perspective, I also recognize the present lacunae with smart contracts from various perspectives (algorithm development, algorithm implementation, algorithm governance, and the availability of appropriate legal resources in the event that disputes occur) and demonstrate how these drawbacks impede shared value creation. Keywords: Algorithmic Relationships, Blockchain, Internet of Things, Inter-firm Cooperation, Smart Contracts, Trust.
... By looking at the earlier discussion in this section, we can look for new dimensions for Software Process Improvement as in this era of technological advancement where everything is being upgraded to the best possible solutions; we can turn the software process improvement towards a new direction by introducing blockchain in it as blockchain gives an effective method to tackle issues using distributed, protected, collective and authorized ledgers [26]. The Authors [27] claim that blockchain-based applications are rapidly evolving and most of the startups are developing their services by using blockchain technology as they think that blockchain technology will be the dominant technology in near future and the majority of the business models would be transformed by using it. ...
Full-text available
Progressively, Software development organizations are investing their resources, time and, money on Software Process Improvement (SPI) since it is beneficial in the enhancement of product quality, reduction in development time, and cost of software projects. However, the existing methodologies and approaches are time-consuming and costly and their major focus is on the SPI of Large Scale Enterprises (LSEs) therefore, we are introducing blockchain in SPI to overcome its major issues such as reliance on a central body of standardization for certification, knowledge management, high cost, resource management and change in organizational culture, etc. We have performed an exploratory case study to identify the different barriers of traditional SPI approaches. To overcome the identified issues, we have proposed and implemented a new approach by performing two case studies. The first case study was performed to identify the barriers in traditional SPI approaches and the second case study was performed to validate our proposed approach. We have performed our experiments on 55 representatives of 50 organizations. According to the results of proposed approach 56.4% of the population agreed that the SPI cost will decrease, 61.8% agreed that time of SPI will decrease and 60.3% of the population agreed that BBSPI will decrease resource utilization. Moreover, 69.1% of the population agreed with the fact that proposed BBSPI will make effective knowledge management and 83.3% of the population said that an organization can mature its processes equaliant to the central certification (CMMI, ISO) body by employing proposed BBSPI. Our results affirm that the BBSPI can reduce the time, cost, resources and helps to manage knowledge used to perform SPI. Moreover, results also depict that the BBSPI can be an efficient substitute of central bodies that could help small and medium-sized organizations to conform to common process improvement models by spending less money, time, and resources with effective knowledge management.
Technology-Driven Innovation in Gulf Cooperation Council (GCC) Countries: Emerging Research and Opportunities is an essential reference source that showcases state-of-the-art approaches to and successful applications of technology-driven innovation and outlines how new, advanced, international business strategies can expand the economic- and social-sustainability frontiers in the GCC region. The chapters included in this volume will communicate and disseminate recent GCC region-oriented academic, computer engineering, and international business research and success stories and will feature topics such as business leadership, electronic marketplace, and international business. Students, educators, international business experts, computer engineers, scientists, and researchers interested in innovation, entrepreneurship, and international business will benefit from the content examined within this publication.
Full-text available
Üblicherweise basieren plattformbasierte Geschäftsmodelle auf zentral gesteuerten digitalen Plattformen. Dazu zählen Geschäftsmodelle von Unternehmen wie Amazon, Alibaba, Apple, Uber etc. Deren Plattformen fungieren als Vermittler („Gatekeeper“), die die beteiligten Nutzerseiten koordinieren und so den Austausch von Waren und Dienstleistungen ermöglichen. Die Plattformen besitzen die angebotenen Produkte und Dienstleistungen i.d.R. nicht selbst. Aufgrund der dadurch entfallenden Produktions- und Lagerkosten profitieren plattformbasierte Geschäftsmodelle, verglichen mit herkömmlichen Geschäftsmodellen, von einer hohen Skalierbarkeit, ohne zwingend größere Investitionen dafür vornehmen zu müssen. Einige Plattformen haben sich dadurch in kurzer Zeit zu Quasi-Monopolen, inklusive der entsprechenden Preisgestaltung, entwickelt. Aufgrund dessen werden diese Unternehmen oftmals dafür kritisiert, den produzierten Wert nicht ausreichend mit denjenigen zu teilen, die erheblich zur Wertschöpfung beigetragen haben – nämlich die Käufer und Verkäufer.
Full-text available
Banking and Information Technology (BIT) Jg. 22(2), S. 69-78. Die Nutzung digitaler Plattformen ist zu einem selbstverständlichen Teil unseres Lebens geworden. Statt Geschäfte und Filialen aufzusuchen, führt uns der Weg auf Vergleichsplattformen, auf Social-Media-Plattformen, zu Suchmaschinen oder direkt auf Handelsplattformen. Üblicherweise basieren die Geschäftsmodelle die-ser Plattformen auf einer zentralen Steuerung. Aufgrund hoher Skalierbarkeit des Geschäfts haben sich einige dieser Plattfirmen in kurzer Zeit zu Quasi-Monopolen entwickelt. Aufgrund dessen werden diese Unternehmen oftmals dafür kritisiert, den produzierten Wert nicht ausreichend mit denjenigen zu teilen, die erheblich zur Wertschöpfung beigetragen haben – nämlich den Käufern und Verkäufern. Als Alternative zu den zentral gesteuerten Plattformen werden dezentrale Plattformen gesehen. Der Charme dieser Plattformen liegt darin, dass sie Anbieter und Kunden direkt miteinander verbinden und es somit ermöglichen, den Intermediär sowie die von ihm verlangten Margen zu umgehen. Dazu werden die Blockchain-Technologie und selbstausführende Verträge genutzt. In diesem Beitrag skizzieren die Autoren das Konzept, die Charakteristika und die Entwicklung dezentraler Plattformen. Weitere Aspekte sind die Rolle der Blockchain-Technologie in solchen Netzwerken und die Klassifikation von dezentralen Plattformen. Außerdem befassen sich die Autoren mit den Perspektiven sowie der Bedeutung dieser Plattformen für Banken.
How do venture capitalists (VCs) incorporate weak and strong signals in the valuation of technology-based startups? Based on a sociocognitive perspective of signaling theory, we introduce Twitter sentiment as a novel and weak signal, which we juxtapose with patents as a traditional, strong signal. While we find a positive association between both signals and VCs' venture valuations, our results reveal that Twitter sentiment does not correlate with actual long-term investment success, whereas patents do. Additionally, we identify and test novelty and experience characteristics (i.e., startup age and VC firm experience) as boundary conditions for our proposed signal-valuation relationships.
Blockchain technology refers to a digital, immutable, distributed ledger that registers completed transactions in a well-ordered manner and near real time. Blockchain security creates a decentralized environment that bars any third-party organization from controlling the cryptographically validated transactions and data. Blockchain technology fosters business innovation by creating a peer-to-peer networking that prevents one central server from accessing as well as processing data belonging to all companies in the network. Cryptocurrency can be defined as a digital asset built to facilitate completed transactions using cryptography. It helps in providing protection to the completed transactions and controlling the creation of additional units of the currency. In the recent years, the application of blockchain technology has been associated with governance. Blockchain governance has been applied in different fields; for example, it can be used to create permanent laws that cannot be violated by any third party.
Conference Paper
Full-text available
The Information Systems discipline builds upon a well-established set of theories and models. A frequently used methodological procedure is to build a model based on a theory and to apply a multivariate method such as Structural Equation Modeling to test the fit of the model and the significance of its parameters. This approach does not take into account the important role of temporal dynamics. Taken by itself, the Technology Acceptance Model cannot explain the Diffusion of Innovations model, and vice versa, in spite of the fact that both deal with adoption patterns. In this paper we call for a methodological paradigm shift and show how Agent Based Modeling in combination with Structural Equation Modeling can help to create models which combine the strengths of dynamic modeling approaches with rigorous multivariate statistics. This will lead to more meaningful models that incorporate temporal dynamics and enable new ways to test validity.
Full-text available
Blockchain technology is often referred to as a groundbreaking innovation and the harbinger of a new economic era. Blockchains may be capable of engendering a new type of economic system: the blockchain economy. In the blockchain economy, agreed-upon transactions would be enforced autonomously, following rules defined by smart contracts. The blockchain economy would manifest itself in a new form of organizational design—decentralized autonomous organizations (DAO)—which are organizations with governance rules specified in the blockchain. We discuss the blockchain economy along dimensions defined in the IT governance literature: decision rights, accountability, and incentives. Our case study of a DAO illustrates that governance in the blockchain economy may depart radically from established notions of governance. Using the three governance dimensions, we propose a novel IT governance framework and a research agenda for governance in the blockchain economy. We challenge common assumptions in the blockchain discourse, and propose promising information systems research related to these assumptions.
Full-text available
Blockchain-based cryptocurrencies have demonstrated how to securely implement traditionally centralized systems, such as currencies, in a decentralized fashion. However, there have been few measurement studies on the level of decentralization they achieve in practice. We present a measurement study on various decentralization metrics of two of the leading cryptocurrencies with the largest market capitalization and user base, Bitcoin and Ethereum. We investigate the extent of decentralization by measuring the network resources of nodes and the interconnection among them, the protocol requirements affecting the operation of nodes, and the robustness of the two systems against attacks. In particular, we adapted existing internet measurement techniques and used the Falcon Relay Network as a novel measurement tool to obtain our data. We discovered that neither Bitcoin nor Ethereum has strictly better properties than the other. We also provide concrete suggestions for improving both systems.
Full-text available
Imagine meeting a stranger and entering into a trusted economic exchange without needing a third party to vouch for you. What changes in your theoretical perspective in such a world? That model of interaction is what distributed trust technologies such as blockchain bring. I introduce the basic concept of distributed trust, describe some early instances, and highlight how organizational theories need to be updated to no longer rely upon fundamental assumptions about trust which are becoming outdated. Distributed trust fundamentally transforms boundaries of organizations, and challenges assumptions about internalizing organizational functions to overcome market trust coordination issues. Implicit assumptions about the legitimacy and power of central network positions no longer ring true. This is very fertile ground for organizations research as the core tenet of the field – what roles and functions should group together within an organization - is being called into question at the most fundamental level.
Full-text available
Based on a critical review of the Unified Theory of Acceptance and Use of Technology (UTAUT), this study first formalized an alternative theoretical model for explaining the acceptance and use of information system (IS) and information technology (IT) innovations. The revised theoretical model was then empirically examined using a combination of meta-analysis and structural equation modelling (MASEM) techniques. The meta-analysis was based on 1600 observations on 21 relationships coded from 162 prior studies on IS/IT acceptance and use. The SEM analysis showed that attitude: was central to behavioural intentions and usage behaviours, partially mediated the effects of exogenous constructs on behavioural intentions, and had a direct influence on usage behaviours. A number of implications for theory and practice are derived based on the findings.
Blockchain as an engine for auto-enforceable smart contracts could disrupt traditional governance structures by reducing bureaucracy through lower transaction costs, solving principal–agent issues, and subsequent moral hazard. While machine consensus can radically reduce transaction costs and disrupt traditional governance structures, there is a gap between initial conceptualizations of blockchains and their first instantiations. First use cases show that as circumstances change, protocols can become inappropriate for the new environment and require modification. Modification of blockchain code happens through majority consensus, but reaching consensus in a distributed multi-stakeholder network with sometimes unaligned interests is complex, potentially introducing new agency issues.