Buying and selling knowledge-intensive tasks and services on the market has become a common practice for many organizations to gain on-demand access to a variety of expertise and knowledge [1
]. Ranging from small start-ups to those listed in Fortune 500, increasingly, companies are making use of crowdsourcing to access external knowledge and skills [3
]. These companies are task providers (TPs), having a knowledge demand and are buyers in market terms, whereas those who fulfill the task are crowdsourcing participants (CPs), e.g., knowledge providers being suppliers in market terms. In such crowdsourcing, TPs follow a tendering process in which they specify their needs and ask for offers from CPs on a market platform. CPs can provide their bids to the TP who can select the best offers and fulfill the e-commerce transaction. Today, an increasing number of knowledge-intensive tasks are crowdsourced, as companies need new expertise or capacity to address fast-changing technical and business environments. Knowledge-intensive crowdsourcing is regarded as one of the most promising areas for crowdsourcing, given its critical role in today’s knowledge-based economy [5
Crowdsourcing can be viewed as an online and decentralized problem-solving model [7
]. Typically, crowdsourcing starts with the need for a TP and the publishing of a request for proposals through an online market platform, with a description of the service or product needed, its expected duration, and possibly, a range of payment options. Potential CPs then bid on the task by submitting their proposals. If more than one proposal is received for a task, the proposals will be evaluated and compared, and a candidate will be selected to carry out the task. Once the tasks have been fulfilled, the electronic transaction can be settled, and the TP can decide to accept and pay for the work or refuse the deliverable if it does not fulfill their requirements. There are many discrete crowdsourcing platforms on the internet supporting such a tendering process, such as Amazon Mechanical Turk (AMT, www.mturk.com
), InnoCentive (www.innocentive.com
) and Upwork (www.upwork.com
). The majority of crowdsourcing platforms run their business on centralized servers with a revenue model that demands a commission ranging from 5% to 20% of the trade [8
While the demand for knowledge-intensive tasks is growing, there are three key challenges that hinder TPs from realizing the expected results. These challenges, include (1) fragmentation of CP expertise, (2) lack of trust between TPs and CPs, and (3) inability to learn from historical data due to a lack of openness of proprietary platforms.
Firstly, there is no single market, and many crowdsourcing platforms for conducting knowledge-intensive tasks exist. These many platforms result in scattering and fragmentation of demand and supply of expertise over multiple platforms. A market needs a minimal level of activities to be viable, whereas some platforms receive relatively few submissions. Furthermore, on some platforms, the quality of bids is an issue, as the majority of these bids do not meet the expectations of TPs [9
]. Secondly, TPs and CPs do not know each other. They might distrust each other, as they do not know if the CP is an imposter or has the right expertise and if the TPs will pay. Thirdly, most platforms do not provide the required openness to allow users to view past performance, experiences of others, similar requests, and fraud.
Central to overcoming the above challenges for knowledge-intensive crowdsourcing is the creation of a trusted and open crowdsourcing network connecting demand and supply [10
]. This requires storing, distributing, and updating the information on knowledge-intensive tasks and participants in a standardized way, and making user engagement transparent to avoid undesirable behavior. Blockchain
can create trust due to its decentralized storage of data records, which are hard to manipulate and using consensus mechanisms to ensure that all changes need to be confirmed, by the participating nodes [12
]. The information about every transaction completed in blockchain is recorded in a distributed ledger, which is shared among and available to all nodes [13
]. The distributed nature and consensus mechanism ensure that data cannot be altered by a single party. Blockchain is decentralized and can take over the roles of traditional platform intermediaries, as data cannot be easily tampered or manipulated [14
]. Blockchain provides a distributed data storage and verification solution, but it does not offer much functionality in data processing for implementing business logic. As a complementary technology, smart contracts on blockchains provide a general-purpose programmable infrastructure to deploy and run these programs [15
]. Using smart contracts, companies will be able to automate the terms of agreement [16
]. The automatic settlement of transactions will further lower coordination costs for companies and realize a disintermediated business model [17
]. Smart contracts can be used for arranging the agreements between TPs and CPs.
Blockchains could enable TPs to access the history data of CPs to assess their past performance, while smart contacts could allow TPs to define their own business rules in searching, selecting, and interacting with them. By exposing transactional data to all nodes for verification, blockchain-based applications are more transparent than crowdsourcing platforms in which executive transactions are centralized and run by a third party [18
]. The premise of blockchain is that no traditional intermediaries are needed for ensuring trust and conducting transactions [19
Despite several attempts to use blockchain to decentralize crowdsourcing in specific contexts e.g., [20
], it is not clear what such a distributed solution would look like. This raises the question of whether the premise of disintermediation using blockchain is right. Furthermore, the premise that no intermediaries are needed remains unproven. We will investigate this premise, in this paper, by evaluating whether a blockchain can make intermediaries redundant.
Although there have been many efforts and substantial investments, many blockchain projects fail, as they do not take into account the situation at hand or only provide limited benefits [23
]. In contrast, in this research, a reference architecture is developed to enable the distribution of tasks to a high number and variety of crowdsourcing participants without the involvement of a market intermediary. A reference architecture
consists of a set of principal design decisions, guidance for implementation, and a system structure and components ‘that are simultaneously applicable to multiple related systems, typically within an application domain’ [24
] (p. 58). This reference architecture presents the essential and high-level design of blockchain-based crowdsourcing solutions. This paper provides insight into the decentralized and disintermediated nature of blockchain-based solutions and how they might transform the current crowdsourcing paradigm. The knowledge-based view
(KBV) of the firm and the theory of search friction
, are used to theorize why a blockchain-based paradigm creates benefits for mediating electronic transactions. By this reference architecture we intend to solve the before mentioned three problems with the existing crowdsourcing platforms to improve knowledge-intensive crowdsourcing
The remainder of this paper is structured as follows. Section 2
explains the concept of KBV of the firm and the theory of search friction and introduces related works. In Section 3
the design science approach is presented. Whereas, Section 4
provides a step-by-step description of how the design science approach was implemented, including the design of the reference architecture, an illustrative case to demonstrate how the blockchain-based design could support core activities in knowledge-intensive crowdsourcing, and the evaluation of the architecture by checking if the requirements have been met by the designed reference architecture. Section 5
presents implications, and Section 6
Blockchain can be used for e-commerce transactions using smart contracts and, in this way, replace traditional intermediaries in crowdsourcing. Such a blockchain-based platform can overcome the challenges of fragmentation of expertise over multiple platforms, difficulty in preventing cheating, and the lack of access to historical data to learn from this. The emergence of blockchain technology offers the opportunity to transform the current crowdsourcing platforms into a blockchain-based e-commerce market, which is characterized by its distributed nature and consensus mechanism. The KBV of the firm and the theory of search friction predict that disintermediation allows for lower knowledge transfer costs in knowledge-intensive crowdsourcing. Blockchain technology offers the infrastructure for disintermediation in such a way that buyers and sellers can transact directly.
We followed a design science approach and developed a reference architecture for implementing blockchain-based crowdsourcing solutions. This architecture uses blockchain to store crowdsourcing records and distribute knowledge-intensive tasks. Smart contracts are used to coordinate and regulate the behavior of both TPs and CPs. Furthermore, business rule applications are in place to provide flexible data processing. This architecture can be used to assess, design, and implement real-world crowdsourcing systems that allow for disintermediated coordination and higher efficiency in managing knowledge-intensive crowdsourcing tasks. In comparison with conventional crowdsourcing platforms, blockchain-based crowdsourcing can have less search friction between TPs and CPs, enable a more powerful mechanism to avoid cheating, and have a greater openness to flexibly implement business logic. Traditional intermediaries providing a platform are not needed anymore, as the distributed blockchain application mediates the transactions, and smart contracts are automatically settled.
The blockchain solution for knowledge-intensive crowdsourcing is still nascent. The reference architecture is an abstract model of a system that has the potential to resolve similar problems in practice. We have only implemented, demonstrated and evaluated the basic processes of knowledge-intensive crowdsourcing to allow an evaluation of the architecture. In addition to the challenges mentioned above, further research into the business impact of using the blockchain-based platform is recommended.