Next Article in Journal
Can Relocation Influence Human Acceptance of Connected and Automated Vehicles?
Previous Article in Journal
Soft Systems Methodology in Standardizing the Method for Applying Dolphin-Assisted Therapies in Neurodivergent Patients: Case Study of Delfiniti Mexico
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Systems
Volume 12
Issue 8
10.3390/systems12080295
systems-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

How Can Scientific Crowdsourcing Realize Value Co-Creation? A Knowledge Flow-Based Perspective

by
Ran Qiu
1,
Guohao Wang
1,
Liying Yu
1,
Yuanzhi Xing
1 and
Hui Yang
2,*
1
School of Management, Shanghai University, Shanghai 200444, China
2
Policy Research Office, Shanghai University, Shanghai 200444, China
*
Author to whom correspondence should be addressed.
Systems 2024, 12(8), 295; https://doi.org/10.3390/systems12080295
Submission received: 5 June 2024 / Revised: 7 August 2024 / Accepted: 9 August 2024 / Published: 11 August 2024
Browse Figures
Versions Notes

Abstract

:
Presently, the practice of scientific crowdsourcing still suffers from user loss, platform operational inefficiency, and many other dilemmas, mainly because the process mechanism of realizing value co-creation through interaction between users and platforms has not yet been elaborated. To fill this gap, this study takes Kaggle as the research object and explores the realization process and internal mechanism of scientific crowdsourcing value co-creation from the perspective of knowledge flow. The results show that the operation process of Kaggle-based scientific crowdsourcing can be decomposed into five progressive evolutionary stages, including knowledge sharing, knowledge innovation, knowledge dissemination, knowledge application, and knowledge advantage formation. The knowledge flow activates a series of value co-creation activities of scientific crowdsourcing, forming a dynamic evolution and continuous optimization of the value co-creation process that includes the value proposition, value communication, value consensus, and all-win value. Institutional logic plays a key role as a catalyst in the value co-creation of scientific crowdsourcing, effectively facilitating the realization of value co-creation by controlling and guiding the flow of knowledge. The study unlocks the “gray box” from knowledge flow to value co-creation, providing new theoretical support and guidance for further enhancing the value co-creation capacity and accelerating the practice of scientific crowdsourcing.

1. Introduction

Scientific crowdsourcing, as an outcome of the big data and internet era, consolidates distributed innovation resources, thereby surpassing the conventional linear and chain-based cooperation logic. It is characterized by non-linearity, diversification, and openness [1]. As a result, it has garnered significant attention in both academic and practical domains.
In essence, scientific crowdsourcing is an activity where participants engage in interactive collaboration and co-innovative processes to create value [2], with value co-creation being its fundamental objective [3,4]. Consequently, the study of scientific crowdsourcing must ultimately address the critical issue of “value co-creation”. In recent years, the rapid and dynamic development of the internet has blurred organizational boundaries, redefining stakeholder groups through “smart connectivity” and enabling more stakeholders to become active participants in value co-creation. From the perspective of contemporary complex networks, the service ecosystem has become a vital framework for understanding value co-creation [5,6]. However, current scholars predominantly focus on the antecedents and outcomes, emphasizing the driving mechanisms of value co-creation in scientific crowdsourcing [7,8] and the nature and types of output values [9]. There is a lack of consensus on the specific processes of value co-creation in scientific crowdsourcing. Existing studies often outline the interaction and resource integration as a general process model [10,11], lacking a detailed characterization of the value co-creation process. Additionally, some scholars approach this from a more micro perspective, describing the specific implementation procedures of scientific crowdsourcing as a process of value co-creation [12], yet fail to integrate and refine the detailed steps. Currently, there is a notable deficiency in theoretical guidance on the specific procedural pathways and motivational enhancements required to achieve value co-creation in scientific crowdsourcing.
It is noteworthy that the key element of innovation is knowledge; thus, the essence of innovation lies in knowledge creation. Scientific crowdsourcing activities are predicated on the distribution of knowledge among diverse entities [13]. Essentially, the process of scientific crowdsourcing involves the continuous flow of knowledge. Consequently, knowledge flow and value co-creation are core characteristics of research crowdsourcing [14]. Is there an interlinked relationship between these two? Can the reorganization of knowledge resources through knowledge flow lead to the emergence of value? Could knowledge flow be the key to unlocking the “grey box” of the value co-creation process?
In light of this, this paper aims to utilize the Kaggle case to deeply analyze the realization process and intrinsic mechanisms of value co-creation in scientific crowdsourcing from the perspective of knowledge flow, focusing on the following questions:
(1)
What are the stages of the knowledge flow process in scientific crowdsourcing?
(2)
How does the value co-creation process evolve in scientific crowdsourcing?
(3)
What is the intrinsic logical relationship between knowledge flow and value co-creation?
By addressing these questions, this study seeks to complement the process logic of value co-creation in scientific crowdsourcing, and provide theoretical support and practical references to enhance the stability and sustainability of value co-creation in scientific crowdsourcing.
The remainder of this paper is structured as follows: Section 2 provides a comprehensive literature review; Section 3 outlines the research design, including research methods, case selection, data collection, and a brief introduction to the case; Section 4 presents the case analysis and discussion; and Section 5 summarizes the research conclusions and theoretical contributions, offering management insights and future research directions.

2. Literature Review

2.1. Scientific Crowdsourcing

Crowdsourcing has received close attention in both commercial and academic fields since it was first proposed [15,16]. Along with the development of practice, the contents of crowdsourcing have become increasingly diverse, leading to the emergence of various forms of crowdsourcing models. Among them, scientific crowdsourcing, which has evolved from traditional crowdsourcing within the realm of scientific research, represents a novel crowdsourcing innovation model [17]. Scientific crowdsourcing refers to the outsourcing of research tasks to the public by social and economic entities such as individuals, enterprises, and research institutions with the help of online platforms, so that the research and technological innovation problems can be accomplished collaboratively at low cost and high efficiency by harnessing the wisdom of the online public [18].
Compared with traditional crowdsourcing models, scientific crowdsourcing presents new characteristics: (1) Professionalism—scientific crowdsourcing has a high threshold for professional knowledge. Both the initiators and solvers require certain skills or experience in specific technical fields. (2) Durability—scientific crowdsourcing projects involve multiple steps and intensive sub-tasks, so the implementation process is somewhat longer [19]. (3) Proactivity—the role of solvers evolves from an auxiliary role to that of a collaborator in the scientific crowdsourcing model. The solvers’ participation typically spans the entire project, with a high degree of interaction with the initiators [20]. (4) Stickiness—the complexity and specialization of scientific crowdsourcing tasks will, to a certain extent, attract solvers to participate in the challenge. In addition, since more complex tasks correspond to higher rewards, this in turn motivates solvers to participate with enthusiasm and effort, resulting in relatively high stickiness among solvers. Although scientific crowdsourcing follows the open innovation logic of traditional crowdsourcing, its uniqueness necessitates that its organization, design, and implementation cannot simply replicate the processes and methods of traditional crowdsourcing models. Consequently, numerous scholars have focused on the three core elements of scientific crowdsourcing, i.e., initiators, solvers, and scientific crowdsourcing platforms. They have explored different aspects, including the classification of scientific crowdsourcing types based on different problem-solving objectives [21], the classification and selection of scientific crowdsourcing organizational models based on different platforms (self-built, third-party, etc.) [22], and the operational processes and mechanisms of scientific crowdsourcing [23]. In terms of further enhancing the performance of scientific crowdsourcing, some scholars have focused on the behavioral characteristics of the bilateral market [24,25,26].

2.2. Knowledge Flow

Knowledge serves as the catalyst for innovation, playing a pivotal role in establishing organizational performance and core competitive advantage [27]. Consequently, the acquisition and development of knowledge resources are of paramount importance. Open innovation, as a distributed innovation paradigm, fundamentally relies on the unfettered flow of knowledge [28,29].
The flow of knowledge encompasses three key attributes: direction, content, and medium [30]. Traditionally, knowledge flow refers to the linear or chain-like transfer process of knowledge from the source unit to the receiving unit within a specific context [31]. Open innovation, as an organizational model for multi-agent collaborative innovation, employs the interactive relationships among diverse entities as carriers of knowledge flow between nodes, thereby forming a knowledge flow network to facilitate iterative knowledge-driven innovation [32,33]. In fact, knowledge flow is a complex dynamic process, yet scholars have not reached a consensus on its description, distinguishing mainly between narrow and broad perspectives. The former primarily outlines the transmission process from the knowledge source to the receiver but neglects the stages of knowledge generation and creation [34], whereas the latter extends and supplements this by emphasizing the iterative and progressive innovation evolution of knowledge flow [35]. To grasp the fundamental principles of knowledge flow, scholars have further abstracted its complex mechanisms and constructed diverse models of knowledge flow for exploration. Nonaka [36] proposed the famous SECI model based on traditional business organizations, which suggests that explicit and tacit knowledge is transformed through four basic modes: socialization, externalization, combination, and internalization. With the proposal of a knowledge ecosystem, scholars combined with this with the life cycle theory to reveal the evolutionary law of knowledge flow within the system. For instance, Wang and You [37] proposed a knowledge flow lifecycle model containing four different manifestations of knowledge generation, knowledge diffusion, knowledge transfer, and knowledge decay. Chu and Li [38] divided the process of knowledge evolution into the initial state, flow state, interactive state, and fading state, and thus constructed a K-SEIR model to analyze knowledge evolution from a quantitative perspective.

2.3. Value Co-Creation

The idea of value co-creation originated from the co-production theory [39] and has since evolved into two important schools of thought: one is the value co-creation theory based on customer experience [40]; the other is the value co-creation theory based on service-dominant logic [41]. The former believes that the customer experience is built through continuous interaction and dialogue between the customer and the company, and the value co-creation is in the process of forming the customer experience. Based on this, scholars further proposed the classical “DART” model, pointing out that dialog, access, risk assessment, and transparency are important practices to promote the creation of a value co-creation environment [42]. The latter group of scholars believes that service is the root of all economic transactions, and operative resources are the fundamental source of service, emphasizing that customers realize value co-creation in the process of integrating and exchanging knowledge, skills, and other resources by participating in the design, production, delivery, and consumption of enterprises.
In recent years, the rapid and dynamic development of the internet has blurred organizational boundaries, redefining stakeholder groups through “smart connectivity” and enabling more stakeholders to become active participants in value co-creation. From the perspective of contemporary complex networks, the service ecosystem has become a vital framework for understanding value co-creation [5,6]. This perspective extends the early focus of service-dominant logic on the binary interactions between firms and customers to a broader, more intricate, and loosely coupled dynamic network system. It underscores that all socio-economic actors are essential components of value co-creation, collaboratively creating value within this dynamic system through service exchange and resource integration.
However, whether focusing on binary or pluralistic relationships, interaction is always a decisive factor in the realization of value co-creation [43]. Gummesson and Mele [44] divided the whole process of value co-creation into two major stages: interaction and resource integration. Ramaswamy and Gouillart [45] further subdivided it into four steps: identifying stakeholders, understanding the interaction, sharing experiences, and proposing solutions. Aarikka and Jaakkola [46] proposed a theoretical framework of value co-creation that includes needs diagnosis, designing and proposing solutions, organizing processes and resources, managing value conflicts, and implementing solutions, while Zhou et al. [47] interpreted the value co-creation process in three dimensions: conceptual consensus, value symbiosis, and co-win with value. To sum up, the development of practice and changes in environmental conditions constantly bring a new connotation to value co-creation, and the participants in value co-creation become more diversified and dynamic. The question of how to clarify subject interactions and promote sustainable value co-creation in complex contexts has become a critical issue.

2.4. Research Gaps and Contributions

Overall, the aforementioned studies exhibit the following limitations:
(1)
The flow of knowledge is pivotal in facilitating the emergence and dissemination of innovation. While scientific crowdsourcing represents a novel and important pathway for collaborative scientific innovation, existing literature has yet to deeply analyze the evolutionary dynamics of knowledge flow in this context.
(2)
Value co-creation highlights the importance of participant interactions in generating value. However, existing studies on process models of value co-creation through resource integration and service exchange, along with their extended theoretical models, have not systematically elucidated the specific processes of interaction and value realization among participants. Particularly in the context of scientific crowdsourcing, the mechanisms by which users, platforms, and other entities engage in value creation remain theoretically ambiguous.
To bridge these research gaps, this study constructs a service ecosystem for scientific crowdsourcing, delineating the processes of knowledge flow within it. It investigates the interplay between knowledge flow and value co-creation, thereby revealing the underlying mechanisms of value co-creation in scientific crowdsourcing. This study aims to provide definitive theoretical guidance for advancing value co-creation in scientific crowdsourcing.

3. Research Design

3.1. Research Method

Unlike empirical research, the primary purpose of case studies is not to test theories but to build them [48]. The purpose of this paper is to investigate how value co-creation can be realized through scientific crowdsourcing. Due to the lack of relevant theoretical studies on value co-creation in scientific crowdsourcing, and the fact that little is known about its process and mechanism, a single-case study can provide insight into a complex vertical evolution process of value co-creation from a dynamic perspective, and has the advantage of revealing the internal logic and laws of value co-creation in scientific crowdsourcing comprehensively and systematically. Hence, a single-case study analysis approach is adopted in this study to decipher the research issues and enrich the research theory.

3.2. Case Selection and Data Collection

This study adheres to several principles in case selection:
Firstly, the typicality and representativeness of the case. Kaggle, founded in 2010, is an online community comprising data scientists and machine learning practitioners, now attracting nearly 800,000 data scientists. Recognized for its substantial commercial potential, it was acquired by Google in 2017. Compared to other scientific crowdsourcing platforms such as Zooniverse or Hypothes.is, Kaggle is dedicated to research in the fields of machine learning (ML) and artificial intelligence (AI). It holds a significant level of recognition and influence within the industry [49].
Secondly, the alignment of the case with the theoretical objectives. This study investigates the process and intrinsic mechanisms of value co-creation in scientific crowdsourcing. Kaggle enables users to find and publish datasets, explore and construct models in a web-based data science environment, and collaborate with other data scientists and machine learning engineers to tackle data science challenges, thereby co-creating value. Consequently, Kaggle effectively exemplifies the comprehensive process of value co-creation in scientific crowdsourcing, aligning well with the theoretical aims of this study.
Lastly, the availability of data from the case. The research team can access relevant information from Kaggle’s official website, social media reports, related books, and academic literature. Furthermore, team members have actively participated in various projects on the Kaggle platform, tracking project processes and progress, thus ensuring the reliability of the research data.
The data for this study are primarily sourced from the following two channels:
(1) Participatory Observation. Given the challenges of conducting field research and interviews with Kaggle, an overseas internet company, this study relies on first-hand data through personal experience. To gain a comprehensive understanding of the operational processes and mechanisms of Kaggle’s scientific crowdsourcing projects, we conducted two phases of personal experience in 2022 and 2024. In 2022, seven members of our research team participated in Kaggle projects, three of them were registered as initiators, while the other three were registered as solvers, and the remaining one as a recorder. In 2024, five members participated as solvers in the platform’s data science competitions. (2) Second-hand data. Second-hand data were obtained using Python to scrape information from Kaggle’s official website, social media news reports, relevant books, and academic literature.

3.3. Case Introduction

Kaggle, as an online community comprising data scientists and machine learning practitioners, enables users to discover and publish datasets, explore and develop models within a web-based data science environment, collaborate with other data scientists and machine learning engineers, and participate in competitions to address data science challenges. The subsequent sections will provide an overview of the key participants, including initiators, solvers, and the Kaggle crowdsourcing platform.
(1)
Initiators
The initiators of scientific crowdsourcing are diversified and can be enterprises, research institutions, universities, or individuals, who are mainly responsible for the initiation of projects and proposing tasks. There are generally external and internal attributes associated with the initiator. The former refers to the credibility and influence of the initiator; the latter refers to the experience and knowledge of the initiator in its research field [50]. Kaggle (https://www.kaggle.com, accessed on 1 July 2024) has attracted many well-known companies and organizations to settle in as initiators. Table 1 depicts the top 20 initiators that carry out scientific crowdsourcing on Kaggle, including well-known large companies such as Google, Microsoft, Walmart, and Facebook, as well as universities, sports clubs, government organizations, and other organizations, such as the National Football League, University of Nicosia, Radiological Society of North America, etc. In fact, Kaggle itself is the initiator that has carried out the most scientific crowdsourcing competitions. As for Kaggle, a total of 63 crowdsourcing competitions have been held, attracting more than 50,000 solvers to participate.
(2)
Solvers
The solvers, as innovation subjects, and the actual providers of research services, are characterized by diversity, independence, and decentralization. Motivation is a prerequisite for the solver to actively engage in scientific crowdsourcing, which is formed by a combination of internal and external motivation [51]. The solvers on the Kaggle platform consist of a large number of data scientists, who master proficient programming skills such as Python, Perl, C++, Java, SQL, etc., and build data models to resolve practical issues. Senior data scientists also have experience and skills in machine learning (ML) and artificial intelligence (AI) [52].
(3)
Crowdsourcing Platform
As well as being the interface between the initiators and solvers, the platform also functions as the architecture’s nerve center. Assisting in the acquisition of research and innovation results and facilitating their utilization comprise the responsibilities of the platform. The Kaggle platform consists of five core modules: Competitions, Datasets, Code, Discussions, and Courses, which are referred to as the “C-D-C-D-C” (Figure 1). “Competition” module fosters the formation of bilateral markets. “Datasets”, “Code”, “Discussions”, and “Courses” are auxiliary modules that provide a friendly “innovation tools box” for the “Competitions”, which greatly enhances the stickiness of Kagglers. These are the unique advantages that distinguish Kaggle from its competitors.

4. Analysis and Discussion

4.1. Analysis of the Operation Process of Scientific Crowdsourcing Based on Knowledge Flow

Scientific crowdsourcing based on the Kaggle platform is an activity in which all participants jointly invest in superior resources and capabilities, and then achieve collaborative innovation through division of labor and cooperation. Differences in knowledge attributes and knowledge reserves among participants create a potential gradient, leading to multidirectional knowledge flow. This dynamic stimulates iterative cycles of knowledge-driven innovation [53]. By its very nature, knowledge flow accompanies the whole process of scientific crowdsourcing. Therefore, this section comprehensively deconstructs the operational processes of scientific crowdsourcing on the Kaggle platform by analyzing sources from Kaggle’s official website, social media news reports, relevant books, and literature, alongside the research team’s extensive discussions based on their participation experiences. From the perspective of knowledge flow, we categorize Kaggle’s scientific crowdsourcing into five progressive evolutionary stages: knowledge sharing, knowledge innovation, knowledge dissemination, knowledge application, and the formation of knowledge advantage (Figure 2).
(1)
Knowledge sharing
Before the implementation of the scientific crowdsourcing competition on the Kaggle platform, the initiators need to carry out the preparation work of the competition project (Figure 3). Initiators are first required to specify the problem they want to solve, i.e., to determine the competition type and theme, and to develop relevant implementation details using Kaggle’s competition project creation toolbar, including uploading competition data, participation methods, solution submission format, deadline, evaluation criteria, rewards, etc., to form a document detailing the requirements of the competition project. Subsequently, the initiators will submit the document of requirements to Kaggle for review, and if the review passes, the project can be officially launched and promoted. In this stage, the initiators refine and encode knowledge based on their own knowledge storage to create the document of competition requirements, which contains the professional background knowledge needed to complete the competition. After the release of competition items, knowledge from the initiators is successfully transferred to the Kaggle platform, where more heterogeneous public knowledge has been compiled and improved. The solvers, it should be noted, are not yet formed and are hidden among the many Kagglers, but they can still keep an eye on the project through accessing the information about competition projects on the platform. The knowledge is therefore shared between the participants at this point.
(2)
Knowledge innovation
With the official release of competitions on the Kaggle platform, many Kagglers have open access to competition details, thus matching their capabilities with task requirements and transforming into solvers driven by different motivations for participation. As the innovative subjects of scientific crowdsourcing, the solvers devote their knowledge, skills, and experience, and integrate other necessary resources (“Datasets”, “Code”, “Discussions”, “Courses” and other auxiliary modules) in the Kaggle platform to develop novel solutions. At this stage, solvers, comprising multiple individual actors, initially perceive and understand the explicit knowledge contained in the competition projects. Through internal learning, discussion, and interaction facilitated by the crowdsourcing platform, they internalize this explicit knowledge into new tacit knowledge. Building on this foundation, solvers can make their tacit knowledge, which encompasses critical technical know-how, explicit through specialized tools or methods (such as models, algorithms, etc.). They then integrate other explicit knowledge scattered across the platform or external environments, thereby processing and forming new, systematic, and organized knowledge.
(3)
Knowledge dissemination
Upon the completion of task execution, solvers submit their project solutions to the Kaggle platform. Kaggle swiftly evaluates the efficiency and accuracy of the submitted solutions based on pre-established criteria set by the initiators, providing corresponding scores and rankings (Figure 4). Under this incentivization mechanism, solvers can further refine and improve their solutions before ultimately submitting a satisfactory final version. At this stage, the numerous heterogeneous solvers exhibit a state of both competition and cooperation. Initially, solvers who complete preliminary solutions can disseminate new knowledge through the platform’s “Discussion” functional modules. This operation allows other solvers to absorb and assimilate the newly available explicit knowledge within the platform, transforming it into new tacit knowledge. This process facilitates the iteration and optimization of solutions. Through multiple dynamic cycles, the solutions are continuously refined until they are finally submitted.
(4)
Knowledge application
In the later stages of Kaggle’s scientific crowdsourcing competition, the scores and rankings of the submitted solutions are determined. The initiators can identify and target final research results based on the ranking, which can be incubated for application and thus gain social utility. Also, an agreed incentive will be offered to the solver who wins the bid. In this stage, initiators are required to digest and assimilate solutions embedded with new explicit knowledge in conjunction with innovation requirements and previous cognition. Thus, it can be internalized into their own tacit knowledge and eventually transformed into concrete practices and actions.
(5)
Knowledge advantage formation
Knowledge advantage formation is the stage in which scientific crowdsourcing competition projects on the Kaggle platform go through the progressive links of knowledge sharing–knowledge innovation–knowledge dissemination–knowledge application, and finally obtain a competitive advantage. In essence, it is a dynamic integration process of continuous accumulation and optimization of knowledge flow advantage and knowledge stock advantage [53]. In particular, in the knowledge-sharing stage, the Kaggle collaborative innovation field and the toolbar for creating projects stimulate the willingness of the solvers to share knowledge, increasing knowledge flow and an expansion of each subject’s knowledge stock. In this sense, knowledge sharing is the basis for forming knowledge advantages. In the knowledge innovation stage, the “Innovation toolbox” on the Kaggle platform greatly enhances the aggregation and reorganization of knowledge and accelerates the emergence of new knowledge, which is vital for forming knowledge advantage. In the stage of knowledge dissemination and application, the increase in interaction frequency among subjects drives the overflow and diffusion of new knowledge and then realizes the utility and value of knowledge, which is the key link in the formation of knowledge advantage.
It can be observed that in general, each stage of knowledge flow in Kaggle’s scientific crowdsourcing competition can be optimized and upgraded quantitatively and qualitatively. Through the close connection, interaction, and interpenetration of each stage, the overall knowledge innovation ability is ultimately realized, facilitating a leap up.

4.2. Value Co-Creation Process of Scientific Crowdsourcing Based on Knowledge Flow

As scientific crowdsourcing activities on the Kaggle platform deepen, the dynamic linkages among participants through resource integration and service exchange have formed an effective networked collaborative organizational structure—a scientific crowdsourcing service ecosystem. Value co-creation not only serves as the primary driving force behind this ecosystem, but also as its ultimate goal. Considering the specific characteristics of scientific crowdsourcing activities on the Kaggle platform and building on existing research [3,54,55], a multi-level interactive structure encompassing “macro-meso-micro” levels has been constructed for the Kaggle scientific crowdsourcing service ecosystem (Figure 5). Specifically, the macro layer includes wider social participants such as the government, intermediary agencies, enterprises, scientific research institutions, and the public. They are potential participants and beneficiaries of value creation, and they can penetrate into the meso-level driven by different motives, so as to realize role change. The meso layer contains the Kaggle platform and the numerous professional users (i.e., potential initiators) and general users (i.e., potential solvers) on the platform. Under the “connection and aggregation” logic of the Kaggle platform, many users gather around data science challenges or goals. The microlayer is a binary interaction between the specific initiators and the heterogeneous solvers, where a competitive and cooperative relationship exists among the heterogeneous solvers. The above three layers are nested and related, and evolve dynamically over time. It is worth noting that in the service ecosystem of Kaggle-based scientific crowdsourcing, the initiators, the solvers, and the Kaggle platform are the core constituents and key actors. With the Kaggle platform as the intermediary, the initiators and the solvers carry out resource integration and service exchange around the data crowdsourcing competition, and the value emerges and unfolds in the connection and aggregation of resources, which constitutes the core value co-creation model of the service ecosystem of Kaggle-based scientific crowdsourcing. Participants in the meso-layer and macro-layer, on the other hand, act only as facilitators and receivers of value creation, providing the necessary resources to assist in core value creation activities while enjoying value spillovers. The outermost layer encompasses a collection of various environmental factors essential for the value creation activities in scientific crowdsourcing. The results of value co-creation in scientific crowdsourcing are continuously fed back into the broader political, economic, social, and cultural external environments, thereby inducing changes in these external conditions. This feedback loop supplies additional crucial resources for scientific crowdsourcing activities, thus accelerating the interaction and collaboration among participants and enhancing value co-creation. This cyclical process ultimately facilitates the dynamic iterative evolution of the scientific crowdsourcing service ecosystem.
As previously discussed, the operation of scientific crowdsourcing competitions on the Kaggle platform essentially represents a cyclical and progressive process of knowledge flow. During this process, knowledge flow not only achieves the reorganization and innovation of knowledge resources but also transforms knowledge-embedded solutions into instruments of service exchange. The primary activities of value co-creation are resource integration and service exchange [44]. Therefore, there is a high degree of alignment between knowledge flow and value co-creation. This section provides a detailed analysis of the value co-creation process based on the stages of knowledge flow in Kaggle’s scientific crowdsourcing competitions.
(1)
Value proposition
The starting point of the value co-creation of scientific crowdsourcing is the existence of scientific research-based collaborative innovation problems. Driven by the demand for innovation, the initiators clearly express their value appeals by submitting the document of the competition project requirements to the Kaggle platform. Solvers who are driven by different motivations to join the competition must tailor their problem solutions to fit the individual requirements of the initiators. The solution has only potential use value until it is accepted and implemented, and is therefore essentially a value proposition put forward by the solvers in response to the value appeals of the initiators. The value proposition is the necessary prerequisite for value co-creation. In this stage, the solver acts as a resource integrator, absorbing the knowledge transferred from the initiator on the one hand and integrating the existing public resources available on the Kaggle platform on the other, to improve their own knowledge structure and propose a more competitive solution. Hence, knowledge sharing and knowledge innovation are essential elements of developing solutions embedded with new knowledge. In other words, knowledge sharing and knowledge innovation support value proposition activities.
(2)
Value communication
Kaggle, based on the predefined evaluation metrics of the competition, can initially evaluate the potential value of the solvers’ solutions and publish the ranking correspondingly on the Public Leaderboard. The solvers benefit from this service, since it allows them to see the gap between individual expectations and actual results in the solutions. The solvers must optimize the initial value proposition to achieve the engagement objective. First, the solvers need to develop a value statement for the initial solution. The Kaggle platform provides a series of “Innovation Toolbox” services, such as “Datasets”, “Code”, “Discussions”, “Courses”, etc., which support many solvers in communicating actively and inclusively about the shortcomings of their solutions based on specific competition situations. Through reflection and learning, the solvers can reshape and upgrade their respective value propositions, thereby continuously enhancing their potential for value in use. Value communication is an important basis for value co-creation. It is only through the free flow of knowledge within the system that ideas can collide and integrate, generating new sparks of insight. The dissemination of knowledge thus creates the conditions and possibilities for value communication.
(3)
Value consensus
The solvers, who are limited in rationality, ultimately determine a satisfactory value proposition through iterative value communication. Meanwhile, the initiators understand and identify value propositions that meet their value expectations with the help of the support tool (Private Leaderboard) on the Kaggle platform. Once the initiators admit the solution and apply it to a specific practice, it reflects the initiators’ recognition of the potential value of the solution. A consensus on the value of the solution is reached at this stage, and the accepted solution is then used as an input to the resource integration process of the initiators, which ultimately yields practical utility through its continuous digestion and absorption. Considering that the value of a solution is closely tied to its specific context, the initiator is the ultimate determiner of the value, while the use of new knowledge can be considered feedback to the consensus on value. Consequently, the knowledge application stage is an inherent representation of value consensus formation among multiple subjects.
(4)
All-win value
Based on the value consensus, a service relationship has been formally established between the initiators, the solvers, and the Kaggle platform. The three parties can achieve mutual benefits and win-win results through in-depth interaction and resource integration. Specifically, the initiator, with the assistance of the Kaggle platform and the solvers, can achieve a powerful solution to data science challenges at a lower cost in terms of time, manpower, and money, while simultaneously gaining social utility by incubating research results. The solver winning the bid can not only obtain the corresponding reward but also, more importantly, expand the knowledge system and enhance personal competitiveness. The Kaggle platform, as the interactive interface between the initiators and solvers, is essential to facilitate the formation of loosely coupled associations among participants. Heterogeneous supplementary resources and services, which are provided by the Kaggle platform, are integrated into the crowdsourcing process. After matching supply and demand among users, the Kaggle platform can, on the one hand, exchange service revenue for investment recovery, which is vital for the platform; on the other hand, innovative ideas and clever methods generated by user interaction help expand the platform database. In addition, the feedback arising from the interaction process can also incrementally promote the optimization and innovation of Kaggle platform services. As the purpose of obtaining competitive advantage is to create value, the initiators, the solvers, and the Kaggle platform accumulate knowledge advantages through knowledge sharing, innovation, dissemination, and application evolution in the scientific crowdsourcing competition process, and then enhance their knowledge ability, which is regarded as the source of potential competitive advantage. Accordingly, knowledge advantage formation is an important driving force for the tripartite participating parties to achieve all-win value.
In summary, the value co-creation of Kaggle-based scientific crowdsourcing is a complex dynamic evolution and continuous optimization process of multi-subject interaction. The four links of value proposition, value communication, value consensus, and all-win value are interconnected to form a complete value co-creation chain. The flow of knowledge is the “undercurrent” under the value co-creation activities, which activates a series of Kaggle-based scientific crowdsourcing value co-creation activities through “surging”. In addition, the successful implementation of scientific crowdsourcing competition on the Kaggle platform will serve as an eminent sample of value co-creation, bringing positive social impact output, attracting more users to register and participate, and paving the way for the next round of value co-creation. Such a benign circle continuously promotes the sustainable development and operation of scientific crowdsourcing on the Kaggle platform.

4.3. Value Co-Creation Realization Mechanism of Scientific Crowdsourcing Based on Knowledge Flow

Essentially, a complete value co-creation process is a knowledge flow loop, and the efficiency and effectiveness of the knowledge flow directly affect the success or failure of value co-creation. The knowledge flow in the scientific crowdsourcing competition project based on the Kaggle platform involves the interaction between multiple knowledge subjects and knowledge objects. To reduce obstacles and risks in the process of knowledge flow, it is necessary to standardize various knowledge behaviors through a certain system. Kaggle has designed and provided a relatively mature resource aggregation mechanism, user evaluation mechanism, and transparent incentive mechanism to control, guide, and regulate the running activities of various stages of knowledge flow, so that the initiators, the solvers, and the Kaggle platform can cooperate in an orderly and standardized manner in the service ecosystem of Kaggle-based scientific crowdsourcing, and finally realize value co-creation.

4.3.1. Resource Aggregation Mechanism

Kaggle positions itself as an open data competition platform based on the crowdsourcing model, leading the analysis and mining of big data. It is convenient for organizations or individuals with innovative needs and data scientists to quickly select and join, thus forming a two-sided market based on the Kaggle platform, which is the starting point for resource aggregation. Kaggle connects heterogeneous users and builds bridges for cooperation and communication between them. Users with complementary knowledge resources are gradually transformed into interactive relationships under the connection of the Kaggle platform, which leads to knowledge flow. Kaggle skillfully screens and classifies the huge amount of knowledge resources on the platform, and finally forms a platform service framework containing five core modules of “C-D-C-D-C” and multiple subdivisions. This service framework is a concentrated expression of the resource aggregation mechanism. First of all, the project-assisted creation service under “Competitions” can quickly issue the competition project requirements document, which greatly shortens the running-in period between the initiators and the platform, effectively improves the experience of the initiators, and enhances their willingness to share knowledge afterward, thereby consolidating knowledge flow at the initial stage. Subsequently, Kaggle utilizes technical support to process and store the transferred knowledge in different types of platform-shared knowledge bases, such as “Datasets”, “Code” and “Courses”, effectively shortening the distance of knowledge and reducing the cost of knowledge search for the solvers. In addition, the “Discussions” module is designed to create a barrier-free and friendly learning and communication space for many solvers. By conducting a correlation analysis of the points information in the Kaggle ranking section, as shown in Figure 6, we discovered a positive correlation between the points solvers earn from participating in discussions and those from participating in competitions. This indicates that the more solvers engage in discussions, the higher their competition capability becomes. Thus, interaction and collaboration among solvers not only facilitate the transfer of explicit and tacit knowledge but also accelerate the intertwining and non-linear recombination of knowledge through real-time feedback [56,57,58]. This results in the emergence of more new knowledge on the platform, leading to higher-order knowledge transitions within the platform due to scale and integration effects. Consequently, the close coupling and coordination among participants gradually stimulate collaborative innovation enthusiasm, thereby accelerating the realization of value co-creation.

4.3.2. User Dynamic Evaluation Mechanism

Kaggle has innovatively designed a user evaluation mechanism to standardize the evaluation of data scientists’ knowledge levels. This evaluation mechanism is based on four categories, including “Competitions”, “Datasets”, “Notebook” (a sub-section under the Code module), and “Discussions”. There are corresponding evaluation criteria in each category, which give data scientists medals and points according to the quantity and quality of their scientific achievements in the corresponding category. It is worth noting that medals can only increase but not decrease, while points will gradually decrease over time. On the one hand, this reflects Kaggle’s full recognition of the efforts of data scientists on the platform. On the other hand, the change in points leads to the real-time and dynamic change in ranking, which maintains the intense competitive environment on the Kaggle platform.
By crawling data from Kaggle website with four key words, which are “Competition”, “Dataset”, “Notebook” and “Discuss”, the solvers’ point distributions in these four areas are ascertained, and are shown in Figure 7. The X axis is the solver’s ranking order, and the y axis is the solver’s points for Competition, Dataset, Notebook and Discussion, accordingly. As shown in Figure 7, in the rankings and point distributions across different categories of solvers, a few solvers obtain significantly higher points than others, indicating that top-performing solvers dominate.
In addition, the dynamic scoring system can stimulate the willingness of data scientists to participate in knowledge innovation, so as to maintain the continuity and stability of knowledge flow on the platform [52]. For the Kaggle platform, the platform can build a user ability information database to improve the accuracy of project recommendation and matching on the platform, so as to start the knowledge innovation stage, advancing and accelerating the stage evolution of knowledge flow. To some extent, it can also reduce the knowledge gap between participants and prevent the interruption of knowledge flow in the competition process.

4.3.3. Reasonable and Transparent Reward Mechanism

Kaggle has elaborately designed the reward categories, reward range, and standards to give full play to the role of the reward mechanism. Taking into account the diverse motivations of different data scientists to participate in scientific crowdsourcing competitions, Kaggle has carefully guided the initiators to set up matching rewards for the competitions. Using “prize” as a keyword, this study utilized Python scripts to web-scrape incentive information from the competition section of Kaggle. The data reveals that the rewards encompass various categories, including “Monetary”, “Knowledge”, “Jobs”, “Swag” and “Kudos”, with “Monetary” prizes accounting for 81% of the total (Figure 8). These types of rewards greatly satisfy the psychology of the solvers and stimulate their enthusiasm for new knowledge creation [59]. In terms of the scope and criteria of rewards, the Kaggle platform guides the initiators to decompose the task packages and provide reasonable reward levels, to avoid the “reward paradox” as much as possible. In addition, Kaggle respects and recognizes the hard work and effort contained in each submitted solution, so the reward is set to be multi-gradient, not just simply selecting the optimal solution as the award target. Notably, Kaggle has innovatively set up efficiency awards based on accuracy, which gives the solvers a greater chance of winning and enhances the platform’s ability to compensate for innovative knowledge. Kaggle realizes the complete transparency and openness of the whole award selection process from the announcement of award standards and the evaluation of competition results to the determination of winners, minimizing opportunistic behaviors. It not only shows the platform’s emphasis on knowledge innovation but also reinforces the user’s sense of trust in the platform, and enhances the enthusiasm and effort of the solvers to participate in the long run. Ultimately, more solvers voluntarily seek self-improvement, spontaneously form coupling and interconnection, and independently break through barriers, thus opening up knowledge flow channels, realizing knowledge integration and innovation, and giving rise to value co-creation.

5. Research Summary

5.1. Research Conclusions

At present, scientific crowdsourcing has become a frontier model for scientific research cooperation and innovation, and its fundamental goal is to realize value co-creation. Focusing on the core issue of the “value co-creation mechanism of scientific crowdsourcing”, this study takes knowledge flow as the entry point, selects Kaggle as a typical case, and deeply analyzes the realization process and internal mechanism of value co-creation of scientific crowdsourcing, thus extracting the theoretical model, as shown in Figure 9.
Based on the model, the following research conclusions can be drawn:
(1)
Due to the knowledge differentiation and complementarity between the participants in scientific crowdsourcing, knowledge flow permeates the whole process of scientific crowdsourcing. The knowledge flow process of scientific crowdsourcing is divided into five evolutionary and progressive stages: “knowledge sharing–knowledge innovation–knowledge dissemination–knowledge application–knowledge advantage formation”. Knowledge realizes level upgrades through effective flow and accumulation, and finally, promotes scientific research collaboration and innovation.
(2)
In the context of scientific crowdsourcing, the participants are coupled, transformed, and complementary to each other, thus forming a service ecology system of scientific crowdsourcing. Among them, the initiators, the solvers, and the crowdsourcing platform are the core elements and key actors of the system. They carry out resource integration and service exchange around the scientific crowdsourcing project and eventually create common value. Value co-creation in scientific crowdsourcing is a dynamic evolution and continuous optimization process, including value proposition, value communication, value consensus, and all-win value, within which value proposition is the premise, value communication is the foundation, value consensus is the core, and value win-win is the goal. The four links complement each other and jointly promote the realization of value co-creation.
(3)
Knowledge flow is the deep-level logic of value co-creation in scientific crowdsourcing, and the two have a high degree of fit. Specifically, knowledge sharing and knowledge innovation support value proposition, knowledge dissemination creates conditions for value communication, knowledge application is the internal expression of value consensus, and knowledge advantage is an important driving force for value win-win. Thus, the value co-creation activities of scientific crowdsourcing are gradually activated under the flow of knowledge.
(4)
The resource aggregation mechanism, user evaluation mechanism, and transparent incentive mechanism are indispensable in the implementation of scientific crowdsourcing, which can control and guide the running activities of each stage of knowledge flow, to open up the channels of knowledge flow and accelerate the realization of value co-creation.

5.2. Theoretical Contributions and Management Insights

5.2.1. Theoretical Contributions

First, this study depicts the interactive relationship between the participants of scientific crowdsourcing, deeply analyzes the knowledge activities in different links of scientific crowdsourcing, and puts forward the operation process of scientific crowdsourcing based on knowledge flow. Second, it expands the application of value co-creation theory in the field of scientific crowdsourcing and reveals the process logic of value co-creation of scientific crowdsourcing. Building on the value co-creation process model of resource integration and service exchange proposed by Gummesson and Mele [44], this study further refines and identifies four specific value activities: value proposition, value communication, value consensus, and value win-win. These refinements provide a clear understanding of the dynamic evolution and continuous optimization of the value co-creation process in scientific crowdsourcing. Third, it clarifies the logical relationship between knowledge flow and value co-creation, extracts the specific path of knowledge flow to promote value co-creation, and reveals the key role played by institutions in the process, to provide a new perspective for further mining and enhancing the value co-creation capabilities of scientific crowdsourcing and a more theoretical basis for solving the deep-level difficulties and pains in the process of scientific crowdsourcing value co-creation.

5.2.2. Management Insights

As a leader in crowdsourcing platforms for data science research, Kaggle has relied on the successful implementation of one scientific crowdsourcing competition after another since its establishment in 2010 to realize the layer-by-layer superposition of value co-creation and impact output, thereby consolidating its popularity and recognition in the industry. Therefore, it is necessary to analyze the value co-creation mechanism of Kaggle, summarize the replicable behavior logic, and provide more clear action guidance for the management practice of scientific crowdsourcing value co-creation.
Given that knowledge flow is the deep-level logic of scientific crowdsourcing value co-creation, promoting knowledge flow is the key to achieving value co-creation. Specifically:
(1)
It is essential to clearly define the nature of the platform and the crowdsourcing projects within it, thereby identifying the necessary steps for project implementation and eliminating redundant processes. Ensuring that each step is seamlessly integrated will minimize the waste of human, material, and financial resources, fostering tighter coupling among participants. This approach reduces the temporal and spatial distance of knowledge flow, thereby facilitating more efficient knowledge exchange and integration.
(2)
It is imperative to accelerate the development of a competitive integrated platform service architecture, providing user-friendly and convenient project support tools. These tools should cater to all stages of scientific crowdsourcing projects, from creation, publication, matching, management, and maintenance to completion. This approach will stimulate the proactivity of participants, enhancing their willingness to share knowledge at each stage, especially the transfer of tacit knowledge.
(3)
It is necessary to establish sound matching mechanisms, incentive mechanisms, and regulatory mechanisms to foster a fair and trustworthy crowdsourcing competition environment. These mechanisms effectively control and guide knowledge activities, better eliminating barriers to knowledge flow. Promoting cooperative, shared, and mutually beneficial relationships among participants reduces free-riding behavior and prevents the destruction of value.
(4)
As the value co-creation in scientific crowdsourcing is a dynamic, evolving process that encompasses numerous value activities and continuous optimization, it is crucial to focus on the status of key nodes in each value activity. Effectively leveraging the impact output of value co-creation, enhancing publicity, and attracting user participation are essential to increase the user base. This approach fosters continuous value co-creation, thereby maintaining the platform’s vitality and sustainability.

5.3. Future Prospects and Limitations

This study has certain limitations. Firstly, as Kaggle is an international internet company, the lack of field research and interviews means that the quantity and quality of first-hand data are insufficient. Additionally, the study’s conclusions lack external validity, necessitating an increase in the number of case studies in the future to enhance the reliability of the conclusions through replication and comparative analysis. Furthermore, since Kaggle focuses solely on specialized fields, such as artificial intelligence and machine learning, it would be beneficial to further verify and elucidate the applicability of these conclusions to other scientific crowdsourcing platforms. Finally, empirical and quantitative research is needed to explore and enhance the value co-creation capabilities in scientific crowdsourcing, such as developing efficiency measurement models for knowledge flow in scientific crowdsourcing and identifying influencing factors.

Author Contributions

Conceptualization, R.Q. and G.W.; methodology, R.Q.; validation, R.Q., G.W., L.Y. and H.Y.; investigation, Y.X.; resources, H.Y.; data curation, G.W.; writing—original draft preparation, R.Q.; writing—review and editing, L.Y.; supervision, L.Y.; project administration, L.Y. and H.Y.; funding acquisition, H.Y. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Shanghai Municipal Education Science Research Project for the Year 2021 (Grant No. C2021152).

Data Availability Statement

Data for this study were obtained from the Kaggle official website “https://www.kaggle.com (accessed on 1 July 2024)”.

Acknowledgments

The authors would like to thank the anonymous referees for their valuable comments and suggestions.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Lenart-gansiniec, R.; Czakon, W.; Sułkowski, Ł. Understanding crowdsourcing in science. Rev. Manag. Sci. 2023, 17, 2797–2830. [Google Scholar] [CrossRef]
  2. Kit, C.W.; Zolkepli, I.A. Crowdsourcing and service-dominant logic: The interaction effect of brand familiarity, customer experience and loyalty on value co-creation. Asia Pac. J. Adv. Bus. Soc. Stud. 2018, 4, 154–162. [Google Scholar]
  3. Xue, X.; Zhao, Y.X.; Zhu, Q.H.; Ying, J.; Hao, S.B. Construction of open data service ecosystem for major public health events based on public science model. Libr. Inf. Work. 2022, 66, 33–44. [Google Scholar] [CrossRef]
  4. Mishra, D.; Maheshwari, N. Crowdsourcing a wellspring of value co-creation: An integration of social capital and organizational learning mechanisms. Kybernetes 2024, 53, 424–450. [Google Scholar] [CrossRef]
  5. Vargo, S.L.; Lusch, R.F. From Repeat Patronage to Value Co-creation in Service Ecosystems: A Transcending Conceptualization of Relationship. J. Bus. Mark. Manag. 2010, 4, 169–179. [Google Scholar] [CrossRef]
  6. Vargo, S.L.; Lusch, R.F. Institutions and axioms: An extension and update of service-dominant logic. J. Acad. Mark. Sci. 2016, 44, 5–23. [Google Scholar] [CrossRef]
  7. Zhao, D.H.; Zhang, C.N.; Sun, X.B. A case study of collaborative incentive mechanism among creative class crowdsourcing participants. Technol. Econ. 2021, 40, 133–145. [Google Scholar]
  8. Al-kumaim, N.H.; Alhazmi, A.K.; Ramayah, T.; Shabbir, M.S.; Gazem, N.A. Sustaining continuous engagement in value co-creation among individuals in universities using online platforms: Role of knowledge self-efficacy, commitment and perceived benefits. Front. Psychol. 2021, 12, 637808. [Google Scholar] [CrossRef]
  9. Reypens, C.; Lievens, A.; Blazevic, V. Leveraging value in multi-stakeholder innovation networks: A process framework for value co-creation and capture. Ind. Mark. Manag. 2016, 56, 40–50. [Google Scholar] [CrossRef]
  10. Liu, Q.; Zhao, X.; Sun, B. Value co-creation mechanisms of enterprises and users under crowdsource-based open innovation. Int. J. Crowd Sci. 2018, 2, 2–17. [Google Scholar] [CrossRef]
  11. Meng, Q.L.; Hang, Y.; Chen, X.J. Value co-creation mechanisms of multi-agent participation in crowdsourcing innovation: A grounded theory study. J. Intell. Fuzzy Syst. 2021, 41, 4995–5006. [Google Scholar]
  12. Galateanu, E.; Avasilcai, S. Value co–creation through crowdsourcing: The case of Squadhelp competition platform. In MATEC Web of Conferences. EDP Sci. 2018, 184, 04011. [Google Scholar]
  13. Beck, S.; Brasseur, T.M.; Poetz, M.; Sauermann, H. Crowdsourcing research questions in science. Res. Policy 2022, 51, 104491. [Google Scholar] [CrossRef]
  14. Wang, G.H.; Yu, L.Y. Differential game analysis of scientific crowdsourcing on knowledge transfer. Sustainability 2019, 11, 2735. [Google Scholar] [CrossRef]
  15. Xu, W.; Duan, H.; Chen, X. Blockchain-based multi-skill mobile crowdsourcing services. EURASIP J. Wirel. Commun. Netw. 2022, 2022, 55. [Google Scholar] [CrossRef]
  16. Wang, H.; Liu, W.; Liu, A.; Wang, T.; Song, H.; Zhang, S. SQCS: A sustainable quality control system for spatial crowdsourcing via three-party evolutionary game: Theory and practice. Expert Syst. Appl. 2024, 238, 122132. [Google Scholar] [CrossRef]
  17. Uhlmann, E.L.; Ebersole, C.R.; Chartier, C.R. Scientific utopia III: Crowdsourcing science. Perspect. Psychol. Sci. 2019, 14, 711–733. [Google Scholar] [CrossRef] [PubMed]
  18. Saez-Rodriguez, J.; Costello, J.C.; Friend, S.H. Crowdsourcing biomedical research: Leveraging communities as innovation engines. Nat. Rev. Genet. 2016, 17, 470–486. [Google Scholar] [CrossRef]
  19. Zahay, D.; Hajli, N.; Sihi, D. Managerial perspectives on crowdsourcing in the new product development process. Ind. Mark. Manag. 2018, 17, 41–53. [Google Scholar] [CrossRef]
  20. Kou, L.L.; Zhu, Z.M.; Bai, L.L. Research on the collaboration model of scientific crowdsourcing projects—Crowd Research as an example. Libr. Intell. Work 2018, 62, 146–152. [Google Scholar] [CrossRef]
  21. Lenart-Gansiniec, R. Towards a typology development of crowdsourcing in science. J. Inf. Sci. 2022, 0, 01655515221118045. [Google Scholar] [CrossRef]
  22. Qin, S.F.; Van der Velde, D.; Chatzakis, E.; McStea, T.; Smith, N. Exploring barriers and opportunities in adopting crowdsourcing based new product development in manufacturing SMEs. Chin. J. Mech. Eng. 2016, 29, 1052–1066. [Google Scholar] [CrossRef]
  23. Bassi, H.; Lee, C.J.; Misener, L. Exploring the characteristics of crowdsourcing: An online observational study. J. Inf. Sci. 2020, 46, 291–312. [Google Scholar] [CrossRef]
  24. Gao, L.; Gan, Y.; Zhou, B. A user-knowledge crowdsourcing task assignment model and heuristic algorithm for expert knowledge recommendation systems. Eng. Appl. Artif. Intell. 2020, 96, 103959. [Google Scholar] [CrossRef]
  25. Tang, J.; Zhou, X.; Zhao, Y. How the type and valence of feedback information influence volunteers’ knowledge contribution in citizen science projects. Inf. Process. Manag. 2021, 58, 102633. [Google Scholar] [CrossRef]
  26. Djelassi, S.; Cambier, F. Creative crowdsourcing: Understanding participation barriers and levers from a heterogeneous crowd perspective. J. Mark. Manag. 2023, 39, 585–614. [Google Scholar] [CrossRef]
  27. Che, Z.; Wu, C.; Qu, W. How does multidimensional interaction of knowledge transfer affect digital innovation capability? Knowl. Manag. Res. Pract. 2023, 21, 1–13. [Google Scholar] [CrossRef]
  28. Fu, L.; Liu, Z.; Zhou, Z. Can open innovation improve firm performance? An investigation of financial information in the biopharmaceutical industry. Technol. Anal. Strateg. Manag. 2019, 31, 776–790. [Google Scholar] [CrossRef]
  29. Chiu, M.L.; Cheng, T.S.; Lin, C.N. Driving open innovation capability through new knowledge diffusion of integrating intrinsic and extrinsic motivations in organizations: Moderator of individual absorptive capacity. J. Knowl. Econ. 2023, 15, 3685–3717. [Google Scholar] [CrossRef]
  30. Hai, Z. Knowledge flow network planning and simulation. Decis. Support Syst. 2006, 42, 571–592. [Google Scholar] [CrossRef]
  31. Szulanski, G. Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strateg. Manag. J. 1996, 17, 27–43. [Google Scholar] [CrossRef]
  32. Kong, J.; Zhang, J.; Deng, S. Knowledge convergence of science and technology in patent inventions. J. Informetr. 2023, 17, 101435. [Google Scholar] [CrossRef]
  33. Aliasghar, O.; Haar, J. Open innovation: Are absorptive and desorptive capabilities complementary? Int. Bus. Rev. 2021, 32, 101865. [Google Scholar] [CrossRef]
  34. West, J.; Bogers, M. Leveraging external sources of innovation: A review of research on open innovation. J. Prod. Innov. Manag. 2014, 31, 814–831. [Google Scholar] [CrossRef]
  35. Zhang, S.; Wang, X.; Zhang, B. Innovation ability of universities and the efficiency of university–industry knowledge flow: The moderating effect of provincial innovative agglomeration. Chin. Manag. Stud. 2022, 16, 446–465. [Google Scholar] [CrossRef]
  36. Nonaka, I. The knowledge-creating company. In The economic Impact of Knowledge; Routledge: London, UK, 2009; pp. 175–187. [Google Scholar]
  37. Wang, L.; You, Z.J. Simulation study on the value evolution of innovation ecosystem based on knowledge flow. Chin. Sci. Technol. Forum 2019, 6, 48–55. [Google Scholar] [CrossRef]
  38. Chu, J.W.; Li, J.X. Research on knowledge evolution and wisdom creation in knowledge ecosystem--a case study of Zhihu community. Intell. Theory Pract. 2022, 45, 51–58. [Google Scholar] [CrossRef]
  39. Ranjan, K.R.; Read, S. Value co-creation: Concept and measurement. J. Acad. Mark. Sci. 2016, 44, 290–315. [Google Scholar] [CrossRef]
  40. Prahalad, C.K.; Ramaswamy, V. Co-opting customer competence. Harv. Bus. Rev. 2000, 78, 79–90. [Google Scholar]
  41. Vargo, S.L.; Lusch, R.F. Evolving to a new dominant logic for marketing. J. Mark. 2004, 68, 1–17. [Google Scholar] [CrossRef]
  42. Prahalad, C.K.; Ramaswamy, V. Co-creation experiences: The next practice in value creation. J. Interact. Mark. 2004, 18, 5–14. [Google Scholar] [CrossRef]
  43. Pera, R.; Occhiocupo, N.; Clarke, J. Motives and resources for value co-creation in a multi-stakeholder ecosystem: A managerial perspective. J. Bus. Res. 2016, 69, 4033–4041. [Google Scholar] [CrossRef]
  44. Gummesson, E.; Mele, C. Marketing as value co-creation through network interaction and resource integration. J. Bus. Mark. Manag. 2010, 4, 181–198. [Google Scholar] [CrossRef]
  45. Ramaswamy, V.; Gouillart, F. Building the co-creative enterprise. Harv. Bus. Rev. 2010, 88, 100–109. [Google Scholar] [PubMed]
  46. Aarikka, S.L.; Jaakkola, E. Value co-creation in knowledge intensive business services: A dyadic perspective on the joint problem-solving process. Ind. Mark. Manag. 2012, 41, 15–26. [Google Scholar] [CrossRef]
  47. Zhou, W.H.; Cao, Y.; Zhou, Y.F. Consensus, co-production and win-win: A process model of value co-creation. Sci. Res. Manag. 2015, 36, 129–135. [Google Scholar] [CrossRef]
  48. Eisenhardt, K.M.; Graebner, M.E. Theory building from cases: Opportunities and challenges. Acad. Manag. J. 2007, 50, 25–32. [Google Scholar] [CrossRef]
  49. Bojer, C.S.; Meldgaard, J.P. Kaggle forecasting competitions: An overlooked learning opportunity. Int. J. Forecast. 2021, 37, 587–603. [Google Scholar] [CrossRef]
  50. Chen, Y.Q.; Zhao, Y.X.; Zhu, Q.H. Research on task matching model for public science projects from the perspective of scientific crowdsourcing. Libr. Intell. Knowl. 2018, 3, 4–15. [Google Scholar] [CrossRef]
  51. Acar, O.A. Motivations and solution appropriateness in crowdsourcing challenges for Innovation. Res. Policy 2019, 48, 103716. [Google Scholar] [CrossRef]
  52. Dissanayake, I.; Zhang, J.; Yasar, M.; Nerur, S.P. Strategic effort allocation in online innovation tournaments. Inf. Manag. 2018, 55, 396–406. [Google Scholar] [CrossRef]
  53. Tu, Z.Z.; Gu, X. Research on the collaborative innovation process of industry-university-research based on knowledge flow. Sci. Technol. Res. 2013, 31, 1381–1390. [Google Scholar] [CrossRef]
  54. Chandler, J.D.; Vargo, S.L. Contextualization and value-in-context: How context frames exchange. Mark. Theory 2011, 11, 35–49. [Google Scholar] [CrossRef]
  55. Beirão, G.; Patrício, L.; Fisk, R.P. Value cocreation in service ecosystems: Investigating health care at the micro, meso, and macro levels. J. Serv. Manag. 2017, 28, 227–249. [Google Scholar] [CrossRef]
  56. Ren, R.; Yan, B.; Jian, L. Show me your expertise before teaming up: Sharing online profiles predicts success in open innovation. Internet Res. 2020, 30, 845–868. [Google Scholar] [CrossRef]
  57. Wang, X.; Khasraghi, H.J.; Schneider, H. Towards an understanding of participants’ sustained participation in crowdsourcing contests. Inf. Syst. Manag. 2020, 37, 213–226. [Google Scholar] [CrossRef]
  58. Javadi, K.H.; Hirschheim, R. Collaboration in crowdsourcing contests: How different levels of collaboration affect team performance. Behav. Inf. Technol. 2022, 41, 1566–1582. [Google Scholar] [CrossRef]
  59. Javadi, K.H.; Vaghefi, I.; Hirschheim, R. Contribution to team and community in crowdsourcing contests: A qualitative investigation. Inf. Technol. People 2024, 37, 223–250. [Google Scholar] [CrossRef]
Systems 12 00295 g001
Figure 1. The core modules and components of the Kaggle platform. Source: Kaggle official website (https://www.kaggle.com, accessed on 1 July 2024); compiled by authors.
Figure 1. The core modules and components of the Kaggle platform. Source: Kaggle official website (https://www.kaggle.com, accessed on 1 July 2024); compiled by authors.
Systems 12 00295 g001
Systems 12 00295 g002
Figure 2. The operation process of Kaggle’s scientific crowdsourcing competition based on knowledge flow.
Figure 2. The operation process of Kaggle’s scientific crowdsourcing competition based on knowledge flow.
Systems 12 00295 g002
Systems 12 00295 g003
Figure 3. Competition creation. Source: Kaggle official website (https://www.kaggle.com, accessed on 1 July 2024); compiled by authors.
Figure 3. Competition creation. Source: Kaggle official website (https://www.kaggle.com, accessed on 1 July 2024); compiled by authors.
Systems 12 00295 g003
Systems 12 00295 g004
Figure 4. Solution submission and evaluation. Source: https://www.kaggle.com/competitions/leaderboard (accessed on 1 July 2024).
Figure 4. Solution submission and evaluation. Source: https://www.kaggle.com/competitions/leaderboard (accessed on 1 July 2024).
Systems 12 00295 g004
Systems 12 00295 g005
Figure 5. Composition of Kaggle-based scientific crowdsourcing service ecosystem.
Figure 5. Composition of Kaggle-based scientific crowdsourcing service ecosystem.
Systems 12 00295 g005
Systems 12 00295 g006
Figure 6. Correlation between solvers discussion points and competition points. Source: https://www.kaggle.com/rankings (accessed on 1 July 2024).
Figure 6. Correlation between solvers discussion points and competition points. Source: https://www.kaggle.com/rankings (accessed on 1 July 2024).
Systems 12 00295 g006
Systems 12 00295 g007
Figure 7. Solver’s points distribution by (a) Competition; (b) Dataset; (c) Notebook; (d) Discussion. Source: https://www.kaggle.com/rankings (accessed on 1 July 2024).
Figure 7. Solver’s points distribution by (a) Competition; (b) Dataset; (c) Notebook; (d) Discussion. Source: https://www.kaggle.com/rankings (accessed on 1 July 2024).
Systems 12 00295 g007
Systems 12 00295 g008
Figure 8. Solver’s prizes on Kaggle’s scientific crowdsourcing platform. Source: https://www.kaggle.com/competitions (accessed on 1 July 2024).
Figure 8. Solver’s prizes on Kaggle’s scientific crowdsourcing platform. Source: https://www.kaggle.com/competitions (accessed on 1 July 2024).
Systems 12 00295 g008
Systems 12 00295 g009
Figure 9. A theoretical model of the evolutionary process and intrinsic mechanisms of value co-creation in scientific crowdsourcing under knowledge flow.
Figure 9. A theoretical model of the evolutionary process and intrinsic mechanisms of value co-creation in scientific crowdsourcing under knowledge flow.
Systems 12 00295 g009
Table 1. Major initiators on Kaggle.
Table 1. Major initiators on Kaggle.
InitiatorsNumber of CompetitionsTotal Number of Solvers
Kaggle6353,480
Google Research104855
Google Cloud94943
Google915,663
Fine-Grained Visual Categorization93060
Booz Allen Hamilton610,522
Facebook53087
Avito43114
Banco Santander420,173
Google Brain41577
Jigsaw/Conversation AI39336
The National Football League32038
Two Sigma37473
Allstate Insurance34712
Walmart32215
TalkingData25626
Radiological Society of North America22844
Quora27341
Microsoft22803
University of Nicosia26467
Source: Python 3.9 is used to collect 404 contests held on the Kaggle platform in the past 10 years, eliminate the contests with missing relevant information, and keep 382 valid contest datasets.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Qiu, R.; Wang, G.; Yu, L.; Xing, Y.; Yang, H. How Can Scientific Crowdsourcing Realize Value Co-Creation? A Knowledge Flow-Based Perspective. Systems 2024, 12, 295. https://doi.org/10.3390/systems12080295

AMA Style

Qiu R, Wang G, Yu L, Xing Y, Yang H. How Can Scientific Crowdsourcing Realize Value Co-Creation? A Knowledge Flow-Based Perspective. Systems. 2024; 12(8):295. https://doi.org/10.3390/systems12080295

Chicago/Turabian Style

Qiu, Ran, Guohao Wang, Liying Yu, Yuanzhi Xing, and Hui Yang. 2024. "How Can Scientific Crowdsourcing Realize Value Co-Creation? A Knowledge Flow-Based Perspective" Systems 12, no. 8: 295. https://doi.org/10.3390/systems12080295

APA Style

Qiu, R., Wang, G., Yu, L., Xing, Y., & Yang, H. (2024). How Can Scientific Crowdsourcing Realize Value Co-Creation? A Knowledge Flow-Based Perspective. Systems, 12(8), 295. https://doi.org/10.3390/systems12080295

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop