INfo Sys Infra-Article Summary in own words
ankituhmc
ContentServer2.pdf
RESEARCH ARTICLE
Assessing the application of big data
technology in platform business model: A
hierarchical framework
Xiaomin Du1, Yang Gao2, Linlin Chang2, Xiangxiang Lang2, Xingqun Xue2, Datian BiID 3*
1 Department of Economic Management, Yingkou Institute of Technology, Yingkou, China, 2 School of
Business, Dalian University of Technology, Panjin, China, 3 School of Management, Jilin University,
Changchun, China
Abstract
This research aims to create a hierarchical framework for the development of a platform
business model based on big data. However, this hierarchical framework must consider
unnecessary attributes and the interrelationships between the aspects and the criteria.
Hence, fuzzy set theory is used for screening out the unnecessary attributes, a decision-
making and trial evaluation laboratory (DEMATEL) is proposed to manage the complex
interrelationships among the aspects and attributes, and interpretive structural modeling
(ISM) is used to divide the hierarchy and finally construct a hierarchical framework. The
results reveal that (1) value proposition and community building in value production are fun-
damental links; (2) information technology and information management in value production
are technical supports; (3) customer development in value marketing is the power source;
and (4) value acquisition is the last link, which is established on the basis of and influenced
by value marketing and value network. This hierarchical framework aims to guide the plat-
form toward the application of big data. This study also proposes engagement of stakehold-
ers for promoting value creation and establishing a sound business model from multiple
levels and links.
1. Introduction
With the development, popularization and application of information technology, the use of
big data has penetrated the daily lives of ordinary people and created unprecedented opportu-
nities for enterprises to use their data assets to conduct market activities. Google, Amazon,
Facebook and other companies have undertaken great efforts in industrial operations by col-
lecting and utilizing big data [1]. Determining how to use big data to achieve their own leap-
frog development over competitors has gradually become the main task of platform
enterprises.
Currently, big data is such a popular topic that many scholars have devoted themselves to
studying it. Research on the application of big data in platform enterprises has mostly focused
on the impact on a certain link in the value chain of platform enterprises, such as the
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 1 / 21
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Du X, Gao Y, Chang L, Lang X, Xue X, Bi
D (2020) Assessing the application of big data
technology in platform business model: A
hierarchical framework. PLoS ONE 15(9):
e0238152. https://doi.org/10.1371/journal.
pone.0238152
Editor: Jacopo Soldani, University of Pisa, ITALY
Received: May 12, 2020
Accepted: August 10, 2020
Published: September 24, 2020
Copyright: © 2020 Du et al. This is an open access article distributed under the terms of the Creative
Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in
any medium, provided the original author and
source are credited.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
file.
Funding: This work was supported by China Post-
Doctoral Project (2019M651124). The funder had
no role in study design, data collection and
analysis, decision to publish, or preparation of the
manuscript.
Competing interests: The authors have declared
that no competing interests exist.
application of big data in inventory management [2], transportation logistics management [3]
and supply chain management [4]. Studies on the combination of big data and customer man-
agement are more common. Platform enterprises can use product search information to ana-
lyze consumer preferences to conduct personalized precision marketing [5]; consumers can
visit different online stores to compare prices for the same product and thus influence product
pricing [6]. Enterprises use big data to analyze the impact of customer online reviews on prod-
uct experience to predict trends in product design innovation [7]. Although the above studies
have different emphases, they ignore the mutual coordination and limitations among different
links of the value chain, which is not conducive to the overall grasp of the operation of plat-
form enterprises and is more detrimental to the overall design of the platform business model.
In addition, in terms of the application of big data in platform enterprises, most studies have
adopted the method of case analysis and focused on well-known platform enterprises, such as
Amazon [8], JD [9] and Taobao [10], lacking a universal integration framework. E-commerce
platform is a virtual network used to carrying out transactions over the Internet and a manage-
ment environment for ensuring a smooth business operation. It is an important platform for
the orderly coordination and integration of information flow, goods flow and capital flow.
Enterprises and merchants can make full use of the network infrastructure, payment platform,
security platform, management platform and other shared resources provided by E-commerce
platform to carry out their own business activities more effectively and at a lower cost [11].
Therefore, it is of great significance to explore the influence of big data on the E-commerce
platform business model, explore the mutual cooperation and limitations between different
modules of the platform business model, and finally reveal the hierarchical path of platform
enterprises under big data to compensate for the shortage of such research in the extant
literature.
Based on this, the research mainly solves the following problems. First, it need to clarify the
difference between the platform business model and the traditional business model. By means
of literature collection, this paper follows the description of the traditional business model by
Teece [12] and explores the differences between various value sectors of the platform business
model and traditional enterprises from the value perspective. Then, the existing research on
the business model of platform enterprises is systematically sorted, and the value plates are
sorted into five categories: value proposition, value production, value marketing, value acquisi-
tion and value network. Second, this paper must address the upgrades and changes brought
about by big data to the platform business model, in order to build a new business model sys-
tem of platform based on big data. According to the value chain theory, this paper systemati-
cally sorts the data collected by platform enterprises into different links of the value chain and
analyzes their specific applications to further summarize the impact of big data on the platform
business model. Then, the new criteria system of the platform business model based on big
data have been obtained. Finally, this paper must explore the development path of the platform
business model under big data and build a hierarchical framework. In this process, the depen-
dence among criteria should be identified. Then, this research considers the mutual restric-
tions and influences among criteria and finally makes multi-attribute decisions under multiple
criteria. This paper uses fuzzy set theory and decision-making and trial evaluation laboratory
(DEMATEL) to evaluate the cause-and-effect relationships among various criteria and probes
into the degree of comprehensive influence among various criteria. At last, this study uses an
interpretive structure model (ISM) to divide the hierarchy and finally construct a hierarchical
framework of the platform business model under big data.
The following conclusions are found in this paper. First, the influence of big data on the
platform business model is mainly reflected in five aspects: value proposition, value produc-
tion, value marketing, value acquisition and value network. Second, we find that these five
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 2 / 21
aspects are in different positions and have complex relationships with each other: (1) value
proposition and community building in value production are at the first level, and both have
the most important impact on the later levels. (2) Information technology and information
management in value production are at the second level, which becomes the subsequent point
for platform enterprises to attract customers. (3) The criteria of value marketing and value net-
work are almost at the third and fourth levels; they influence and restrict each other and are
the last links before value acquisition. (4) Value acquisition at the last level requires the joint
promotion of the first four levels. This paper not only makes up for the shortage of research on
the use of big data to improve the competitiveness of platform enterprises but also constructs a
hierarchical framework for the platform business model based on big data for the first time,
which provides guidance for platform enterprises to use big data to establish a sound business
model and obtain competitive advantage.
2. Literature
2.1 Big data
Over the past few decades, technological development has dramatically expanded the amount
of data available to an organization, enhancing the importance of data and information to
enterprises competitiveness [13, 14]. Therefore, scholars have paid more attention on how
enterprises can use big data to create value [15]. Wamba et al. [16] define big data as an integral
approach to manage and analyze five Vs (i.e. volume, variety, velocity, veracity and value) so as
to create feasible insights for sustainable value delivery and establishing competitive advan-
tages. Big data can change competition among enterprises by “transforming processes, altering
corporate ecosystems, and facilitating innovation” [17]. Besides, it can unlock organization
business value by unleashing new organizational capabilities and new value, which is useful to
tackle their key business challenges [18]. At a moment when the survival of E-commerce plat-
form is threatened by the highly volatile economic conditions, Big Data can remodel the busi-
ness model and provide a much-needed competitive edge which can improve profitability and
the chances of survival [19, 20].
2.2 Traditional business model
There are many studies of traditional business models among scholars, and scholars have
given definitions of business models from various perspectives. The business model is under-
stood as an overall description of how an enterprise creates value through interdependent
activities in its business ecosystem [21]. Timmers [22] proposed that the business model
includes products, the architecture of service, information flow, business participants and their
roles, participants’ interests, revenue sources and marketing strategies. Osterwalder [23] ini-
tially identified the nine elements that constitute the business model. Johnson and Christensen
[24] stated that the business model is composed of four interacting factors: customer value
proposition, key resources, key processes and profit model. Teece [12] described the business
model as "the design or architecture of the value creation, delivery, and capture mechanisms".
Through the summaries of the business model by existing scholars, it can be found that the
key elements of the business model mainly focus on "value proposition", "value creation",
"value delivery" and "value acquisition".
2.3 Platform business model
Platform business model can be thought of as an open business model, with the openness of
the multilateral networks including platform users, platform infrastructure, platform providers
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 3 / 21
and so on [25, 26]. Multilateral networks allow platform enterprises to link various groups of
participants through a highly adaptable and permeable infrastructure, which makes it possible
to enable information and knowledge flow throughout the network of multilateral participants
[27]. In the network era, benefits are increasingly generated through platforms, which allow
various participants to engage with one another [28]. This novel form of participant-to- partic-
ipant service exchange challenges the idea of one firm managing an entire activity system—an
idea nested in traditional business model [29].
There are some differences between the platform business model and the traditional busi-
ness model. The model expression of traditional enterprises attaches importance to the induc-
tion and refinement of the components of the business model but ignores the analysis of
internal relations among the components of the business model as a whole [30]. Weiblen [31]
claims that the business model of platform enterprises will be more open, collaborative, com-
petitive and networked in the future, thus forming a more sustainable and stable platform net-
work ecosystem. Evans [32] state that the scope and depth of the bilateral network effect and
the differentiated distribution of bilateral consumers are the key factors that determine the suc-
cess of platform enterprises’ operations. The description of the constituent elements and the
relationship between these elements are two basic levels for understanding the platform busi-
ness model [8]. This paper finally concludes the five major parts of the platform business
model, namely value proposition, value production, value marketing, value acquisition and
value network.
2.4 The influence of big data on the platform business model
Recently, big data has attracted extensive attention from academic circles [33, 34]. The emer-
gence of big data has made a significant impact on the development of high-tech platform
enterprises [35] and has promoted the transformation of enterprises from being traditional-
factor driven to being innovation driven [36]. Platform enterprises can use the large amount of
content created by the platform to excavate hidden opportunities, continuously develop new
products, new technology and new services, and thus enhance their competitive advantage
[37].
2.4.1 Value proposition. McKinsey initially defined value proposition as "the benefits
provided to the customer community and the price that the customer will pay" [38]; that is,
value proposition is the description of the content of the value provided by the target customer
[39]. Dibb [40] introduced the concept of "market segmentation." The essence of market seg-
mentation is to aggregate convergent consumers in the market environment and define the
target consumer group [41]. Against the background of big data, platform enterprises can
understand customer needs according to their browsing and purchasing conditions and deter-
mine target customers based on the enterprise’s business strategy [42]. In addition, as technol-
ogy advances, the digital transformation of the business model is reshaping consumer
preferences [43], so enterprises must adjust their value content according to the dynamic
changes in the industry to enhance their competitiveness. The platform can use the customer
evaluation content to determine customer preferences and decide whether to develop its own
products or its own logistics system to make adjustments to its own value content [44]. This
paper divides the value proposition of platform enterprises based on big data into two second-
ary criteria: market segmentation and value content.
2.4.2 Value production. Value production involves a company’s value structure and
mechanism, which are reflected in the arrangement of enterprise resources and processes [45].
Platform enterprises use electronic processing and information technology as the bases of
their information transmission tools and use electronic means to engage in business
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 4 / 21
operations and sales activities [46]. In the context of big data, technical support becomes more
important. Platform enterprises use web or mobile application technology to collect accurate
and effective data to better analyze customer preferences [47]. Currently, the key to attracting
a large number of customers involves the different experiences brought about by different
interface structure designs and the functions of different platforms. Based on big data, plat-
form enterprises use web mining technology to expand the keywords input by users effectively
and quickly to improve the accuracy of commodity information retrieval. In addition, enter-
prises dynamically adjust the layout and classification of the entire platform interfaces accord-
ing to customers’ consumption habits so that customers can find the consumer goods they
want at a glance to achieve the goal of meeting customers’ personalized needs efficiently [48].
Moreover, it is more convenient to use big data to manage the information content resources
of suppliers, consumers and other stakeholders [49]. Enterprises can use big data to precisely
classify user data, thus facilitating precision marketing [50]. This paper divides the value pro-
duction of platform enterprises based on big data into three secondary criteria: technical sup-
port, community building and information management.
2.4.3 Value marketing. The emergence of big data requires platform enterprises to trans-
form their extensive marketing models into effective precision marketing models [50]. Relying
on modern information means, precision marketing comprehensively and systematically ana-
lyzes the specific needs of users through the application of massive data on the platform to
accurately locate customer groups and build user labels that predict customer needs [51], ulti-
mately forming a personalized service system to attract customers. Vendrell-Herrero argues
that the digital transformation of business models is reshaping consumer preferences and
behavior [43], thus also changing customer relationships [52]. Then, the information of con-
sumer groups is managed, and based on this fact, the platform has an in-depth understanding
of the consumer psychology of users, which can help the platform develop personalized mar-
keting plans to attract customers and drive existing customers to recommend the platform to
potential customers [49]. This paper divides the value marketing of platform enterprises based
on big data into three secondary criteria: customer attraction, relationship management and
customer development.
2.4.4 Value acquisition. Value acquisition describes how an enterprise converts the value
it delivers to its customers into revenue and profit [12, 53, 54]. Against the background of big
data, one important income stream comes from platform enterprises using their own plat-
forms to attract advertising investors [55]. However, platform enterprises should not ignore
the consumer experience purely for the sake of advertising fees. Too much advertising will
make consumers feel bored and will be counterproductive [56]. Pricing strategy will also affect
the revenue source of the platform, and the platform can develop personalized pricing meth-
ods based on consumer preference prices [6]. With the widespread application of big data,
mobile payment services have brought about great development opportunities to platform
enterprises for value acquisition. Platform enterprises can take advantage of consumer pay-
ment habits to develop financial products and strive to provide consumers and retailers with
greater value than that provided by traditional payment providers (such as banks) [57].
Through the use of big data, on the one hand, enterprises can perform qualitative and quanti-
tative assessments of buyers’ credit according to their consumption behaviors; on the other
hand, enterprises can make the same assessments of seller credit by means of tracking the sales
volume and service quality of merchants. With the help of this two-way credit rating system,
we can better maintain and optimize the rules and regulations of buyers and sellers in the
transaction process [58]. This paper divides the value acquisition of platform enterprises based
on big data into three secondary criteria: media advertising, pricing and financial products.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 5 / 21
2.4.5 Value networks. Value networks constitute a unique attribute of platform enter-
prises. The ultimate goal of platform enterprises is to build a value network ecosystem and
gather all participants in the platform to form an economic community [59]. Currently, many
platforms present the phenomenon that new services or new businesses are increasing, and the
scale of platforms is expanding continuously. Platform enterprises adopt the services provided
by new partners with an open attitude and create new value [60]. However, data resources are
now in "rampant expansion", so when platform enterprises are building the value network,
legitimacy construction is necessary for almost every important decision [11]. Enterprises
should strengthen their own data security construction and establish multidimensional protec-
tion measures and trustworthy protection mechanisms to ensure the confidentiality of user
resources in the process of big data platform cooperation and to prevent data leakage [61]. In
addition, since platform enterprises have a full and accurate grasp of data and information,
they should take the initiative to assume social responsibility and combine big data with the
needs of the public to better serve society as a whole. This paper divides the value network of
platform enterprises based on big data into three secondary criteria: symbiosis, legality and
privacy security. A detailed explanation of each criterion is shown in Table 1.
3. Method
The objective of this paper is to provide a hierarchical framework for the application of big
data in the field of E-commerce platform. DEMATEL determines the importance and cause-
and-effect relationships between criteria from a micro point of view. However, it cannot dis-
play the intrinsic relationship and the division of the hierarchical structure, making it is diffi-
cult to effectively manage and control these criteria [62]. ISM is macroscopically oriented and
Table 1. Proposed attributes.
Aspects Criteria Explanation
Value
proposition
Market Segmentation
(C1)
Determine target customers according to customers’ browsing and purchasing conditions
Value Content(C2) Decide whether to develop their own products or own logistics systems according to the evaluation of customers on the
purchasing of goods
Value
production
Information
Management(C3)
Management and systematic analysis of consumer information and store information
Community Building
(C4)
Design pages according to customers’ browsing habits and select suppliers according to customers’ complaint rates
Technical Support(C5) Improve the platform technology according to customers’ platform experiences
Value
marketing
Customer Attraction(C6) Select the optimal advertising strategy according to the promotion data of the cooperation platform and make
recommendations based on the products that customers browse, click, add to the shopping cart and purchase on the
platform
Relationship
Management(C7)
Use big data to establish a membership system for customer relationship management and improve corresponding
services to maintain customer relationships according to customer feedback information
Customer Development
(C8)
Existing customers recommend the business to potential customers
Value
acquisition
Pricing(C9) According to the feedback of customers on product preference and price information, product pricing can be conveyed
to merchants or self-operated products
Media Advertising(C10) Arrange advertising space according to product popularity and charge a commission in proportion to sales volume
Financial Products(C11) Develop financial products according to payment methods and establish credit compensation systems according to the
default rates and repayment delay times of financial products
Value network Privacy Security(C12) Use authentication techniques to ensure the authenticity and confidentiality of partners
Symbiosis(C13) Select platform cooperation products and promotion modes according to platform market segmentation data
Legality(C14) Establish effective information security for a large amount of user information held by the platform to prevent
information leakage
https://doi.org/10.1371/journal.pone.0238152.t001
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 6 / 21
used to decompose complex systems into subsystems. ISM can transform complex thoughts
and ideas into an intuitive model of structural relationships to understand the relationship
between the variables. The combination of the two methods can make the results more accu-
rate and intuitive [63]. In addition, by using fuzzy set theory, triangular fuzzy number is used
to replace the original accurate value of expert evaluation, which can improve the credibility of
the analysis results and provide a more valuable reference for managers to make decisions
[62]. The following sections of this paper provides corresponding formulas to help to under-
stand the integration method.
3.1 Fuzzy DEMATEL
DEMATEL was first proposed by Gabus &Fontela, scholars in Battelle Laboratory. DEMA-
TEL is a systematic analysis method by using graph theory and matrix tools to handle com-
plex and difficult problems. Through the logical relationship and direct influence matrix in
the system, we can calculate the causality and centrality of criteria as the basis for construct-
ing the model and determine the causal relationship between criteria and the position of each
criterion in the system. Triangular fuzzy number (TFN) provides an effective means of quan-
tifying human linguistic preferences into computable form. Fuzzy-DEMATEL method
retains the practical and effective advantages of traditional DEMATEL method in factor rec-
ognition, while fuzzy concepts allow the capture of artificial deviations and uncertainties that
DEMATEL cannot handle in the data. Therefore, Fuzzy-DEMATEL was used in this study to
explore the causal relationship between the criteria and the degree of influence. The steps are
as follows.
Step 1: For the problem under study, build a system of influencing factors set to F1, F2, . . ., Fn.
Step 2: Determine the influence relationship between two factors by an expert scoring method
and express the relationship in matrix form. Invite experts to use the language operators
"no impact (N)", "very weak influence (VL)", "weak influence (L)", "strong influence (H)",
and "very strong influence (VH)". The relationship between the two factors is assessed.
Convert the original expert evaluations into triangular fuzzy numbers via a semantic table
wkij ¼ ða k 1ij; a
k 2ij; a
k 3ijÞ to represent the extent to which k experts consider the influence of the
i-th factor on the j-th factor, as shown in Table 2.
Step 3: Using the Converting the Fuzzy data into Crips Scores (CFCS) method to defuzzify the
initial values of the expert scores, the nth order directly affects the matrix Z, and the direct
influence matrix reflects the direct effect between the factors, including the following four
steps:
Table 2. Semantic transformation table.
Linguistic variables TFN
N (No influence) (0,0,0. 2)
VL (Very low influence) (0,0. 2,0. 4)
L (Low influence) (0. 2,0. 4,0. 6)
H (High influence) (0. 4,0. 6,0. 8)
VH (Very high influence) (0. 8,1,1)
https://doi.org/10.1371/journal.pone.0238152.t002
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 7 / 21
(1) Normalize triangular fuzzy numbers:
xak 1ij ¼ ða
k 1ij � min a
k 1ijÞ=D
max min ð1Þ
xak 2ij ¼ ða
k 2ij � min a
k 1ijÞ=D
max min ð2Þ
xak 3ij ¼ ða
k 3ij � min a
k 1ijÞ=D
max min ð3Þ
(2) Normalize the left value (ls) and right value (rs):
x lskij ¼ xa k 2ij=ð1þ xa
k 2ij � xa
k 1ijÞ ð4Þ
x rskij ¼ xa k 3ij=ð1þ xa
k 3ij � xa
k 2ijÞ ð5Þ
(3) Calculate the clear value after defuzzification:
xkij ¼ ½x ls k ijð1 � x ls
k ijÞ þ x rs
k ijx rs
k ij�=½1 � x ls
k ij þ x rs
k ij� ð6Þ
zkij ¼ min a k 1ij þ x
k ij � D
max min ð7Þ
(4) Calculate the average clear value:
zkij ¼ ðz 1
ij þ z 2
ij þ � � � þz k ijÞ=n ð8Þ
Step 4: Normalize the direct influence matrix Z to obtain the standardized direct influence
matrix G:
l ¼ 1=max 1�i�n
Xn
j¼1
zij;G ¼ lZ ð9Þ
Step 5: According to T = G + G2 +� � �Gn or T = G(E − G)−1, E is the identity matrix, and the comprehensive influence matrix T is obtained.
Step 6: Analyze the comprehensive matrix to reveal the internal structure of the system. The
elements in matrix T are added by row as the influence degree Di, which represents the
comprehensive influence value of the row factor on all other factors. The elements in matrix
T are added as the affected degree Ri by column, indicating the comprehensive influence
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 8 / 21
value of all other factors in that column. The formulas are as follows:
Di ¼ Xn
j¼1
tijði ¼ 1; 2; � � � ; nÞ ð10Þ
Ri ¼ Xn
i¼1
tijði ¼ 1; 2; � � � ; nÞ ð11Þ
The sum of the influence degree and affected degree is called centrality, which indicates the
position of the factor in the system and the size of its role. The difference between the influence
degree and the affected degree is called causality, which reflects the causal relationship between
the influencing factors. If the causality is greater than 0, the factor has a great effect on other
factors and is called the factor of cause. If the causality is less than 0, the factor is greatly
affected by other factors and is called the factor of result. The formulas are as follows:
mi ¼ Di þ Riði ¼ 1; 2; � � � ; nÞ ð12Þ
ni ¼ Di � Riði ¼ 1; 2; � � � ; nÞ
H ¼ Ti � Riði ¼ 1; 2; � � � ; nÞ ð13Þ
3.2 ISM
ISM was developed in 1973 by Professor Walter Felter as a method of analyzing problems
related to complex socio-economic systems, which could make full use of people’s practical
experience and knowledge to decompose the complex system into several subsystems, and
finally construct the system into a multi-level hierarchical structure model. As a conceptual
model, ISM can transform ambiguous ideas and views into intuitionistic structural model. It is
suitable for systematic analysis with many variables, complex relationships and unclear struc-
tures. The formula used to divide the hierarchy is shown below.
The comprehensive influence matrix T above reflects only the mutual influence relationship
and degree between different factors and does not consider the influence of factors on itself.
Therefore, it is necessary to calculate the overall influence relationship reflecting system fac-
tors, i.e., the overall influence matrix. The calculation formula is:
H ¼ T þ E ¼ hij ð14Þ
Next, a threshold λ is introduced to eliminate redundant information and obtain the most simplified matrix. According to the trial calculation, the most suitable threshold calculation
model is obtained.
l ¼ aþ b ð15Þ
where α and β are the mean and standard deviation of all elements in the comprehensive influ- ence matrix T, respectively.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 9 / 21
The threshold λ is used to remove the redundant factors, and the reachable matrix is obtained.
M ¼ ½mij�n�n ; ði ¼ 1; 2 . . . :n; j ¼ 1; 2 . . . :nÞ ð16Þ
mij ¼ 1; h � l
0; h � l ði ¼ 1; 2 . . . :n; j ¼ 1; 2 . . . :nÞ
(
ð17Þ
1 means there is a direct effect between the two factors, and 0 means there is no direct effect
between the two factors.
The reachable set L(fi), antecedent set P(fi) and common set
CðfiÞ ¼ LðfiÞ \ PðfiÞ ð18Þ
are obtained by hierarchical processing.
Finally, the ISM is determined by the reachable set and common set.
4. Results
Based on the literature reviews and analysis, this paper summarized 14 criteria. In order to
standardize the application of platform business model and ensure the embedding of big data,
it is necessary to evaluate the rationality and standardization of these criteria through the
expert committee which is composed of 7 experts. We have clear requirements for experts. We
define the scope of experts in university scholars and e-commerce platform staffs. For univer-
sity scholars, professors who have studied the same area for at least eight years were selected.
For e-commerce platform staffs, we mainly select middle and senior leaders who have good
knowledge of business model and rich experience in practice.
Before evaluating the attributes proposed in this study (including aspects and criteria), the
expert committee should prove these attributes can reflect the real situation of the platform
enterprises. Once one expert disagrees with the proposed attributes, the committee needs to
discuss this problem until all experts reaching an agreement. Therefore, several rounds of dis-
cussion would be needed to ensure the reliability of this research. Once the criteria are con-
firmed, the questionnaire used to evaluate the importance of the criteria was conducted. Then,
we sent out questionnaires to each expert individually to prevent his/her judgment effected by
other experts and the research purpose which does not have any conflict of interest will be
informed. Then we explained the connotation of the 14 criteria (see Table 1 for details) and
the significance of each blank in the questionnaire. After that, the experts will fill out the ques-
tionnaires and translated their subjective assessment about the importance of the criteria into
numbers 1–5. We can have some auxiliary questions and answers during the process. Finally,
the original results of seven experts were obtained, one of which were presented in Table 3.
Then, the original data are processed by utilizing formula (1)–(8) according to the CFCS
method, and finally, the direct influence matrix for the influencing factors of big data on plat-
form business model is determined, as shown in Table 4. The number in the direct influence
matrix shows the degree of direct influence between corresponding elements.
Normalization is the normal operation of standardizing things. The direct influence matrix
is standardized to obtain the standardized direct influence matrix by using formula (9). Then,
according to the formula T = G(E-G)-1, MATLAB software is used for matrix calculation, and the comprehensive influence matrix was obtained, as shown in Table 5. The comprehensive
influence matrix increases the indirect relation between criteria and can accurately reflect the
comprehensive relation between criteria.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 10 / 21
According to formulas (10)–(13), influence degree, affected degree, centrality and causality
are calculated as shown in Table 6. Influence degree is the sum of the rows in the matrix T that
represents the comprehensive influence value of the corresponding row factor on all other fac-
tors. Affected degree is the sum of the columns in the matrix T that indicates the comprehen-
sive affected value of the corresponding column factor on all other factors.
Causality represents the influence degree of the criterion on the other criteria. Depending
on whether causality is greater than or less than 0, 14 criteria are divided into a cause set and a
result set. As seen from Table 5, there are 7 criteria in the cause set, including market segmen-
tation(C1), value content(C2), information management(C3), community building(C4), tech-
nical support(C5), customer attraction(C6) and customer development(C8). Among them,
market segmentation(C1), value content(C2) and community building(C4) are the main driv-
ers. C1, C2 and C4 have corresponding influence degrees of 1.2174, 1.3537 and 1.5601, respec-
tively, which are the three factors with the largest influence degrees, indicating that these three
Table 3. Sample assessment for criteria of one expert.
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
C1 0 2 4 3 4 4 1 4 3 3 2 0 3 2
C2 2 0 4 3 2 4 3 4 2 3 2 3 3 3
C3 2 2 0 3 3 4 4 4 2 2 2 4 1 2
C4 3 4 4 0 4 2 4 3 2 3 0 1 4 3
C5 1 2 2 3 0 4 4 4 0 1 2 3 3 0
C6 1 2 3 3 3 0 3 2 4 1 2 4 1 2
C7 2 2 3 3 0 1 0 3 1 1 2 4 4 4
C8 2 2 1 2 0 1 4 0 4 2 1 4 2 4
C9 1 1 0 0 2 1 1 1 0 2 2 1 0 0
C10 3 2 2 1 0 2 2 1 1 0 0 1 0 1
C11 2 0 0 1 0 2 1 1 2 1 0 1 3 1
C12 3 3 2 1 2 0 1 1 0 4 4 0 1 4
C13 2 2 2 3 0 0 2 1 2 4 4 0 0 3
C14 2 3 0 0 2 0 0 0 4 4 4 4 3 0
https://doi.org/10.1371/journal.pone.0238152.t003
Table 4. The direct influence matrix of big data on platform business model.
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
C1 0.0000 0.2296 0.4609 0.2704 0.4609 0.4609 0.1071 0.4881 0.4065 0.2840 0.2704 0.0119 0.2160 0.2160
C2 0.2160 0.0000 0.4609 0.2704 0.2568 0.4609 0.2160 0.4609 0.2704 0.2840 0.2704 0.2296 0.2296 0.2296
C3 0.1752 0.2432 0.0000 0.2704 0.2704 0.4881 0.4337 0.4609 0.1480 0.1480 0.0391 0.4881 0.1888 0.2160
C4 0.2568 0.2840 0.4609 0.0000 0.4337 0.2704 0.4881 0.2840 0.2024 0.2840 0.0119 0.1888 0.4337 0.2840
C5 0.1752 0.2024 0.2568 0.2840 0.0000 0.4609 0.4609 0.4337 0.0119 0.0391 0.2704 0.2432 0.2296 0.0119
C6 0.2024 0.2024 0.2432 0.1888 0.1752 0.0000 0.2704 0.0527 0.4065 0.1071 0.1888 0.4609 0.2296 0.2024
C7 0.1616 0.1888 0.2024 0.2024 0.0119 0.1752 0.0000 0.2160 0.1616 0.1071 0.2024 0.4881 0.4337 0.4609
C8 0.1752 0.1480 0.1071 0.1480 0.0119 0.1207 0.4337 0.0000 0.4609 0.1888 0.1616 0.4337 0.2432 0.4881
C9 0.1207 0.0527 0.0119 0.0119 0.1888 0.1071 0.1071 0.0391 0.0000 0.1888 0.2704 0.1071 0.0119 0.0255
C10 0.2432 0.1616 0.1616 0.1071 0.0119 0.1616 0.1616 0.1071 0.1071 0.0000 0.0119 0.1207 0.0119 0.1071
C11 0.1616 0.0119 0.0119 0.1071 0.0119 0.1616 0.1071 0.1071 0.1888 0.1071 0.0000 0.1071 0.2160 0.1071
C12 0.2296 0.1888 0.2024 0.0935 0.1616 0.0119 0.1344 0.1344 0.0119 0.4881 0.4609 0.0000 0.1207 0.1752
C13 0.1480 0.1480 0.1480 0.2568 0.0119 0.0119 0.1752 0.1616 0.1752 0.4609 0.4881 0.0799 0.0000 0.2432
C14 0.1752 0.2296 0.0255 0.0119 0.2160 0.0119 0.0119 0.0119 0.4609 0.4881 0.4609 0.2568 0.2296 0.0000
https://doi.org/10.1371/journal.pone.0238152.t004
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 11 / 21
factors have the greatest influence on other factors. This is because the construction of a sus-
tainable development system of platform enterprises under big data is inseparable from value
proposition and community building, and value proposition will greatly affect its implementa-
tion and development. Community building is conducive to the establishment and improve-
ment of a new model of information resource sharing mechanisms to facilitate the
management of the platform. Therefore, the relevant measures should be taken into consider-
ation when exploring the applications of big data on the platform business model. To make the
causal relationship between the criteria clearer, a causal relationship diagram is made, as
shown in Fig 1.
The others are 7 result elements, including relationship management(C7), pricing(C9),
media advertising(C10), financial products(C11), privacy security(C12), symbiosis(C13) and
legality(C14). These result elements have a weaker impact on the platform business model
based on big data but are more likely to be affected by other factors to make changes.
Table 5. The comprehensive influence matrix of big data on platform business model.
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
C1 0.1552 0.1976 0.2721 0.2034 0.2440 0.2884 0.2244 0.2938 0.2876 0.2574 0.2532 0.2123 0.2225 0.2248
C2 0.2051 0.1445 0.2714 0.2013 0.1941 0.2819 0.2442 0.2837 0.2539 0.2622 0.2545 0.2623 0.2255 0.2307
C3 0.1895 0.1966 0.1611 0.1952 0.1904 0.2753 0.2849 0.2745 0.2150 0.2267 0.1988 0.3165 0.2125 0.2250
C4 0.2201 0.2189 0.2821 0.1488 0.2404 0.2473 0.3148 0.2562 0.2402 0.2729 0.2069 0.2609 0.2834 0.2525
C5 0.1721 0.1701 0.2035 0.1872 0.1117 0.2545 0.2746 0.2527 0.1640 0.1739 0.2268 0.2390 0.2084 0.1617
C6 0.1623 0.1534 0.1791 0.1451 0.1432 0.1273 0.1992 0.1423 0.2296 0.1773 0.1954 0.2553 0.1797 0.1740
C7 0.1561 0.1536 0.1688 0.1487 0.1036 0.1613 0.1359 0.1781 0.1823 0.1912 0.2079 0.2639 0.2315 0.2426
C8 0.1567 0.1406 0.1410 0.1304 0.1010 0.1450 0.2302 0.1223 0.2500 0.2042 0.1952 0.2494 0.1836 0.2461
C9 0.0766 0.0545 0.0522 0.0447 0.0843 0.0800 0.0816 0.0628 0.0569 0.1022 0.1239 0.0848 0.0542 0.0561
C10 0.1240 0.1010 0.1126 0.0846 0.0641 0.1155 0.1177 0.1033 0.1080 0.0860 0.0838 0.1160 0.0760 0.1015
C11 0.0933 0.0526 0.0597 0.0730 0.0499 0.0953 0.0887 0.0839 0.1139 0.0991 0.0740 0.0934 0.1116 0.0875
C12 0.1487 0.1293 0.1461 0.1049 0.1157 0.1095 0.1417 0.1395 0.1135 0.2340 0.2223 0.1170 0.1293 0.1440
C13 0.1293 0.1188 0.1304 0.1403 0.0804 0.1039 0.1498 0.1401 0.1549 0.2299 0.2286 0.1336 0.1033 0.1616
C14 0.1304 0.1304 0.0944 0.0775 0.1238 0.0995 0.1007 0.0990 0.2105 0.2304 0.2238 0.1611 0.1421 0.0890
https://doi.org/10.1371/journal.pone.0238152.t005
Table 6. Comprehensive influence matrix analysis.
Factor Influence degree Affected degree Centrality Causality
C1 3.3368 2.1194 5.4561 1.2174
C2 3.3155 1.9618 5.2773 1.3537
C3 3.1620 2.2745 5.4365 0.8875
C4 3.4453 1.8852 5.3306 1.5601
C5 2.8002 1.8467 4.6469 0.9536
C6 2.4632 2.3848 4.8480 0.0784
C7 2.5256 2.5885 5.1141 -0.0629
C8 2.4956 2.4322 4.9278 0.0634
C9 1.0147 2.5805 3.5952 -1.5658
C10 1.3941 2.7474 4.1415 -1.3533
C11 1.1759 2.6950 3.8709 -1.5191
C12 1.9954 2.7655 4.7609 -0.7701
C13 2.0049 2.3635 4.3684 -0.3586
C14 1.9127 2.3970 4.3097 -0.4843
https://doi.org/10.1371/journal.pone.0238152.t006
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 12 / 21
Therefore, proper attention and control should be paid to actual management to help improve
the management effect.
Centrality indicates the position of the criteria in the system and the role they play in the s
system of e-commerce platform business model based on big data. According to the degree of
centrality, the criteria are C1, C3, C4, C2, C7, C8, C6, C12, C5, C13, C14, C10, C11 and C9 in
descending order. More attention should be paid to criteria with higher centrality, such as
market segmentation(C1), information management(C3) and community building (C4).
ISM is used to construct the system into a multi-level hierarchical structure model. The
overall influence matrix H obtained by formula (14) is shown in Table 7, already considering
the influence of criteria on itself.
Fig 1. DEMATEL causal diagram.
https://doi.org/10.1371/journal.pone.0238152.g001
Table 7. Overall influence matrix.
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
C1 1.1552 0.1976 0.2721 0.2034 0.2440 0.2884 0.2244 0.2938 0.2876 0.2574 0.2532 0.2123 0.2225 0.2248
C2 0.2051 1.1445 0.2714 0.2013 0.1941 0.2819 0.2442 0.2837 0.2539 0.2622 0.2545 0.2623 0.2255 0.2307
C3 0.1895 0.1966 1.1611 0.1952 0.1904 0.2753 0.2849 0.2745 0.2150 0.2267 0.1988 0.3165 0.2125 0.2250
C4 0.2201 0.2189 0.2821 1.1488 0.2404 0.2473 0.3148 0.2562 0.2402 0.2729 0.2069 0.2609 0.2834 0.2525
C5 0.1721 0.1701 0.2035 0.1872 1.1117 0.2545 0.2746 0.2527 0.1640 0.1739 0.2268 0.2390 0.2084 0.1617
C6 0.1623 0.1534 0.1791 0.1451 0.1432 1.1273 0.1992 0.1423 0.2296 0.1773 0.1954 0.2553 0.1797 0.1740
C7 0.1561 0.1536 0.1688 0.1487 0.1036 0.1613 1.1359 0.1781 0.1823 0.1912 0.2079 0.2639 0.2315 0.2426
C8 0.1567 0.1406 0.1410 0.1304 0.1010 0.1450 0.2302 1.1223 0.2500 0.2042 0.1952 0.2494 0.1836 0.2461
C9 0.0766 0.0545 0.0522 0.0447 0.0843 0.0800 0.0816 0.0628 1.0569 0.1022 0.1239 0.0848 0.0542 0.0561
C10 0.1240 0.1010 0.1126 0.0846 0.0641 0.1155 0.1177 0.1033 0.1080 1.0860 0.0838 0.1160 0.0760 0.1015
C11 0.0933 0.0526 0.0597 0.0730 0.0499 0.0953 0.0887 0.0839 0.1139 0.0991 1.0740 0.0934 0.1116 0.0875
C12 0.1487 0.1293 0.1461 0.1049 0.1157 0.1095 0.1417 0.1395 0.1135 0.2340 0.2223 1.1170 0.1293 0.1440
C13 0.1293 0.1188 0.1304 0.1403 0.0804 0.1039 0.1498 0.1401 0.1549 0.2299 0.2286 0.1336 1.1033 0.1616
C14 0.1304 0.1304 0.0944 0.0775 0.1238 0.0995 0.1007 0.0990 0.2105 0.2304 0.2238 0.1611 0.1421 1.0890
https://doi.org/10.1371/journal.pone.0238152.t007
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 13 / 21
Then, the threshold λ by formula (15) can be calculated and the research will get λ = α+β = 0.1678+0.0677 = 0.2355. The value higher than λ in the overall influence matrix means row factors affect column factors and the corresponding position is marked as 1, while the value
lower than λ means row factors do not affect column factors and the corresponding position is marked as 0. Finally, the reachability matrix can be obtained by formula (16) and (17). The
reachability matrix in the Table 8 determines whether the two factors impact each other.
The first-level decomposition structure is obtained from the reachability matrix and for-
mula (18), as shown in Table 9.
As seen from Table 9, the reachable set and the common set intersect in C9, C10 and C11,
so C9, C10 and C11 constitute the first-level influencing factors. The rows and columns
mapped by influence factors C9, C10, and C11 in matrix M are deleted to obtain a higher-level
decomposition matrix, and the above process is repeatedly performed. After multiple hierar-
chical divisions, the factor set Nq (q = 1, 2, . . ., 9) of each layer is finally obtained: first-level
Table 8. Reachability matrix.
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
C1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C2 1 1 1 0 0 1 1 1 1 1 1 1 1 1
C3 0 0 1 0 0 1 1 1 1 1 0 1 1 1
C4 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C5 0 0 1 0 1 1 1 1 0 0 1 1 1 0
C6 0 0 0 0 0 1 0 0 1 0 0 1 0 0
C7 0 0 0 0 0 0 1 0 0 0 1 1 1 1
C8 0 0 0 0 0 0 1 1 1 1 0 1 0 1
C9 0 0 0 0 0 0 0 0 1 0 0 0 0 0
C10 0 0 0 0 0 0 0 0 0 1 0 0 0 0
C11 0 0 0 0 0 0 0 0 0 0 1 0 0 0
C12 0 0 0 0 0 0 0 0 0 1 1 1 0 0
C13 0 0 0 0 0 0 0 0 0 1 1 0 1 0
C14 0 0 0 0 0 0 0 0 1 1 1 0 0 1
https://doi.org/10.1371/journal.pone.0238152.t008
Table 9. First-level decomposition structure.
i L(fi) P(fi) C(fi) = L(fi)\P(fi) C1 market segmentation 1,3,4,5,6,7,8,9,10,11,12,13,14 1,2,4 1,4
C2 value content 1,2,3,6,7,8,9,10,11,12,13,14 2,4 2
C3 information management 3,6,7,8,9,10,12,13,14 1,2,3,4,5 3
C4 community building 1,2,3,4,5,6,7,8,9,10,11,12,13,14 1,4 1,4
C5 technical support 3,5,6,7,8,11,12,13 1,4,5 5
C6 customer attraction 6,9,12 1,2,3,4,5,6 6
C7 relationship management 7,11,12,13,14 1,2,3,4,5,7,8 7
C8 customer development 7,8,9,10,12,14 1,2,3,4,5,8 8
C9 pricing 9 1,2,3,4,6,8,9,14 9
C10 media advertising 10 1,2,3,4,8,10,12,13,14 10
C11 financial products 11 1,2,4,5,7,11,12,13,14 11
C12 Privacy security 10,11,12 1,2,3,4,5,6,7,8,12 12
C13 symbiosis 10,11,13 1,2,3,4,5,7,13 13
C14 legality 9,10,11,14 1,2,3,4,7,8,14 14
https://doi.org/10.1371/journal.pone.0238152.t009
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 14 / 21
node N1 = {C9, C10, C11}; second-level node N2 = {C12, C13, C14}; third-level node N3 =
{C6, C7}; fourth-level node N4 = {C8}; fifth-level node N5 = {C3}; sixth-level node N6 = {C5};
seventh-level node N7 = {C1}; eighth-level node N8 = {C2}; and ninth-level node N9 = {C4}.
Based on the above analysis, an ISM model is presented in Fig 2.
From the ISM model structure of the influence factor, we can see that market segmentation
(C1), value content(C2) and community building(C4) are the root causes of the impact of big
data on the platform business model. Thus, the way in which to effectively control the tracking
and control of big data becomes a key focus. Platform enterprises should identify target cus-
tomers according to customers’ browsing and purchasing history and decide whether to
develop their own products or own logistics systems according to the evaluation of goods pur-
chased by customers. This also helpful for enterprises to provide more targeted services. While
value acquisition (including pricing(C9), media advertising(C10) and financial products
(C11)) lays at the last level, it reflects the ultimate value flow of platform business model in big
data era.
In conclusion, the factors that affect the application of big data on the platform business
model are very complex, and five aspects interact with each other. However, different factors
have different influence modes, mechanisms and degrees of action, thus forming a systematic
integration framework of platform business model under big data.
5. Discussion
Because the existing framework of platform business model under big data is indistinct, this
paper attempts to explore the path of platform business model based on big data. This study
systematically proposed a set of criteria about the development of platform business model
and constructed a hierarchical model to compensate for the shortage of such research in the
extant literature.
Market segmentation and value content are at the first level in the model framework; there-
fore, value proposition is still the core of the platform business model under big data. The clear
value proposition, an effective method for enterprises to transfer their core identity and values
Fig 2. ISM model structure.
https://doi.org/10.1371/journal.pone.0238152.g002
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 15 / 21
[64], is not only a cornerstone of strategy [65] but also a useful tool for product marketing
[66]. The value proposition of platform enterprises under big data is not different from that of
traditional enterprises, which are not affected by big data and are still the core links of enter-
prises. The difference is that, after big data is embedded, the value proposition of platform
enterprises is richer. Platform enterprises under big data can constantly adjust their value
propositions based on the dynamic needs of customers and seek the blue ocean market to
attract more customer groups, thus improving their comprehensive competitiveness. At the
first level, this conclusion not only unifies the platform business model and traditional busi-
ness model but also verifies that value propositions under big data remain the core of the busi-
ness model, which promotes the consistency of the overall theory of the business model.
At the same time, community building is also at the first level, which has become an impor-
tant new aspect of the development of the business models of platform enterprises under big
data. In traditional enterprises, the importance of the ecosystem is emphasized [67], but insuf-
ficient attention has been paid to it, which has something to do with the preference for product
production. However, in the context of big data, community building has become a key activ-
ity in the value production of platform enterprises [30]. Community building is conducive to
the reasonable integration of various resources [48] and the establishment and improvement
of a new model of information resource sharing mechanisms [68] to facilitate the management
of the platform by enterprises. This conclusion reflects the characteristics of platform enter-
prises under big data and renders the advantages brought about by big data more obvious.
In the entire platform business model system, the first level plays a fundamental role, while
the information technology and information management of the second level are the continu-
ity points for platform enterprises to attract customers, which becomes more important than
in traditional enterprises. In traditional enterprises, information management and informa-
tion technology, as support functions, serve the main functions and become help improve the
efficiency of the main functions. However, for platform enterprises, especially after big data
technology is embedded, the value production process of the entire business model is more
dependent on the improvement of the information system and the application of information
technology. The improvement of the information technology system provides a more ideal
platform browsing experience for customers [48], which is conducive to better maintaining
customer relations [52], thus attracting more customers to the platform [49], consistent with
the results in the existing literature. The technical advantages brought about by big data are
reflected in the front end of value production in platform enterprises, while in traditional
enterprises, there is a time lag [69], so it can see that the two have certain differences.
Value marketing and value network are closely linked. A prerequisite of value network con-
struction is value marketing. Only by gaining the trust of customers and conducting effective
customer relationship management can we build a better and more stable value network [52].
Marketing based on big data is more targeted and can better grasp user behaviors and psycho-
logical preferences to adopt a more effective relationship management approach [52, 70]. In
addition, against the background of big data, customer development in value marketing
becomes more important, which is conducive to the sustainable development and long life of
enterprises. At the same time, new criteria of the value network, including the three features of
symbiosis, legality and privacy security, are developed. The development of criteria is guided
by the construction of ecosystems, emphasizing that the optimization of ecological networks
and mutual recognition among stakeholders are the keys to network optimization. Value net-
work construction is also the last link before value acquisition in the platform business model
under big data. Value marketing and value networks become the only ways to obtain value.
Value acquisition is the last link of the platform business model and is also at the last level.
Compared with traditional enterprises that set prices after developing products and then use
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 16 / 21
marketing methods to promote them [71, 72], the business model of platform enterprises
under big data is different. Reasonable pricing and traffic attraction become the final links,
based on the perfect value network and customer flow. Whether to develop self-supporting
products and how to effectively undertake advertiser injection for platform enterprises also
depend on whether the value network is mature [73]. At the same time, value acquisition
under precision marketing also enables platform enterprises to select different pricing schemes
according to customer characteristics. All of these aspects are very different from those of tra-
ditional business models.
6. Conclusions
Currently, most studies pay more attention to the impact of big data on a certain link in the
value chain of platform enterprises and tend to ignore the mutual coordination and limitations
of various links in the value chain, which is not conducive to the overall operation of platform
enterprises and the overall design of the platform business model. In addition, in the applica-
tion of big data to platform enterprises, most studies adopt the method of case analysis, focus-
ing on the role of big data in a single well-known platform enterprise, thus lacking a universal
integration framework. Compared with previous studies, this paper systematically explores the
overall impact of big data on the platform business model and considers the synergy and limi-
tations between modules of the platform business model, revealing the hierarchical path of the
platform business model under big data, which is of great significance in compensating for the
shortage of existing research on this topic. This research comprehensively applies fuzzy,
DEMATEL and ISM as integrated research methods, not only eliminating the influence of
experts’ subjective factors but also considering the interactions between different factors and
clarifying the different levels and paths, thus providing important information for the realiza-
tion of the above research work in this paper. In addition, the outbreak of COVID-19 brings
new challenges to E-commerce platforms, putting E-commerce industry in a crisis. At the
same time, it also provides E-commerce enterprises an opportunity to innovate business mod-
els, making "new online models" possible. This study not only sorts out and constructs the plat-
form business model innovation system in the big data environment, but also has reference
significance for the business model innovation of small and medium-sized e-commerce enter-
prises facing current challenges by constructing the hierarchical theoretical framework.
There are still certain limitations to this study. First, although the proposed criteria have
been selected through the extensive literature review, it is still insufficient to cover all possible
attributes; thus, further exploration and improvement in future research are needed. Second,
expert committee was consisted with the experts worked in the platform enterprises, experts in
other fields related to electronic technology, especially big data, should be included in the com-
mittee for increasing the scope and the applicable boundaries. Third, the relationships and
degrees of influence among criteria are processed and analyzed based on the data information
of questionnaires completed by experts. Although fuzzy set theory is used in this paper to solve
the problem of experts’ subjective bias, there are still errors that cannot be completely elimi-
nated, which could have a certain impact on the research results of this paper. In addition, this
study could also use other statistical tools, such as structural equation models, to explore more
influencing factors and perform the statistical verification of the model.
Supporting information
S1 Data.
(XLSX)
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 17 / 21
Author Contributions
Conceptualization: Xiaomin Du.
Data curation: Datian Bi.
Formal analysis: Xingqun Xue.
Supervision: Yang Gao.
Writing – original draft: Linlin Chang.
Writing – review & editing: Xiangxiang Lang.
References
1. Song J, Ma Z, Thomas R, Yu G. Energy efficiency optimization in big data processing platform by
improving resources utilization. Sustainable Computing: Informatics and Systems. 2019; 21(1): 80–
89.
2. Choi T, Wallace SW, Wang Y. Big data analytics in operations management. Production and Opera-
tions Management. 2018; 27(10): 1868–1883.
3. Kamble SS, Gunasekaran A, Goswami M, Manda J. A systematic perspective on the applications of big
data analytics in healthcare management. International Journal of Healthcare Management. 2019; 12
(3): 226–240.
4. Tiwari S, Wee H, Daryanto Y. Big data analytics in supply chain management between 2010 and 2016:
insights to industries. Computers & Industrial Engineering. 2018; 115: 319–330.
5. Jun S, Park D, Yeom J. The possibility of using search traffic information to explore consumer product
attitudes and forecast consumer preference. Technological Forecasting and Social Change. 2014; 86
(1): 237–253.
6. Lim G G, Kang J M, Lee J K, Lee D C. Rule-based personalized comparison shopping including delivery
cost. Electronic Commerce Research and Applications. 2011; 10(6): 637–649.
7. Hu N, Liu L, Zhang JJ. Do online reviews affect product sales? The role of reviewer characteristics and
temporal effects. Information Technology & Management. 2008; 9(3): 201–214.
8. Ritala P, Golnam A, Wegmann A. Coopetition-based business models: The case of Amazon.com.
Industrial Marketing Management. 2014; 43(2): 236–249.
9. Zheng K, Zhang Z, Song B. E-commerce logistics distribution mode in big-data context: A case analysis
of JD.COM. Industrial Marketing Management. 2020; 86(1): 154–162.
10. Ye Q, Cheng Z, Fang B. Learning from other buyers: the effect of purchase history records in online
marketplaces. Decision Support Systems. 2013; 56: 502–512.
11. Kwak J, Zhang Y, Yu J. Legitimacy building and e-commerce platform development in China: the expe-
rience of Alibaba. Technological Forecasting and Social Change. 2019; 139: 115–124. https://doi.org/
10.1016/j.techfore.2018.06.038 PMID: 32287407
12. Teece DJ. Business models, business strategy and innovation. Long Range Planning. 2010; 43(2–3):
172–194. https://doi.org/10.1007/s00191-018-0561-9 PMID: 30613125PMID: 30613125
13. Giudice MD, Peruta MR. The impact of IT-based knowledge management systems on internal venturing
and innovation: a structural equation modeling approach to corporate performance. Journal of Knowl-
edge Management. 2016; 20(3): 484–498.
14. Sumbal MS, Tsui E, Seeto EW. Interrelationship between big data and knowledge management: an
exploratory study in the oil and gas sector. Journal of Knowledge Management. 2017; 21(1): 180–
196.
15. Breidbach CF, Maglio PP. Technology-enabled value co-creation: an empirical analysis of actors,
resources, and practices. Industrial Marketing Management. 2016; 56: 73–85.
16. Wamba SF, Akter S, Edwards A, Chopin G, Gnanzou D. How ‘big data’ can make big impact: findings
from a systematic review and a longitudinal case study. International Journal of Production Economics.
2015; 165: 234–246.
17. Santoro G, Fiano F, Bertoldi B, Ciampi F. Big data for business management in the retail industry. Man-
agement Decision. 2019; 57(8): 1980–1992.
18. Kiron D, Prentice P K, Ferguson R B. The analytics mandate. MIT Sloan Management Review. 2014;
55(4): 1–25.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 18 / 21
19. Loebbecke C, Picot A. Reflections on societal and business model transformation arising from digitiza-
tion and big data analytics: a research agenda. Journal of Strategic Information Systems. 2015; 24(3):
149–157.
20. Yuan C, Wu Y J, Tsai K M. Supply chain innovation in scientific research collaboration. Sustainability.
2019; 11(3): 753.
21. Zott C, Amit R, Massa L. The business model: recent developments and future research. Journal of
Management. 2011; 37(4): 1019–1042.
22. Timmers P. Business models for electronic markets. Electronic Markets.1998; 8(2): 3–8.
23. Osterwalder A, Pigneur Y, Tucci C L. Clarifying business models: origins, present, and future of the con-
cept. Communications of the Ais. 2005; 16(1): 1–25.
24. Johnson M W, Christensen C M, Kagermann H. Reinventing your business model. Harvard Business
Review. 2008; 86(12), 57–68.
25. Ondrus J, Gannamaneni A, Lyytinen K. The impact of openness on the market potential of multi-sided
platforms: a case study of mobile payment platforms. Journal of Information Technology. 2015; 30(3):
260–275.
26. Saebi T, Foss NJ. Business models for open innovation: matching heterogeneous open innovation
strategies with business model dimensions. European Management Journal. 2015; 33(3): 201–213.
27. Gawer A, Cusumano MA. Industry platforms and ecosystem innovation. Journal of Product Innovation
Management. 2014; 31(3): 417–433.
28. Breidbach C F, Maglio P P. Technology-enabled value co-creation: an empirical analysis of actors,
resources, and practices. Industrial Marketing Management. 2016; 56: 73–85.
29. Wieland H, Hartmann N N, Vargo S L. Business models as service strategy. Journal of the Academy of
Marketing Science. 2017; 45(6): 925–943.
30. Tauscher K, Laudien S M. Understanding platform business models: A mixed methods study of market-
places. European Management Journal. 2017; 36(3): 319–329.
31. Weiblen T. The open business model: understanding an emerging concept. Journal of Multi Business
Model Innovation & Technology. 2014; 1(1): 35–66.
32. Evans D S. Some empirical aspects of multi-sided platform industries. Review of Network Economics.
2009; 2(3): 1–19.
33. Kshetri N. Big data’s role in expanding access to financial services in China. International Journal of
Information Management. 2016; 36(3): 297–308.
34. Mamonov S, Triantoro T. The strategic value of data resources in emergent industries. International
Journal of Information Management. 2018; 39: 146–155.
35. Vera-Baquero A, Colomo-Palacios R, Molloy O, et al. Business process analytics using a big data
approach. IT Professional. 2013; 15(6): 29–35.
36. Porter M E. Strategy and the internet. Harvard Business Review. 2001; 79(3): 63–78. The remote
server returned an error: (404) Not Found.PMID: 11246925
37. Yaqoob I, Hashem I A T, Gani A. Big data: from beginning to future. International Journal of Information
Management. 2016; 36(6): 1231–1247.
38. Bower M, Garda R A. The role of marketing in management. Mckinsey Quarterly. 1985; 3: 34–46.
39. Ballantyne D, Frow P, Varey RJ, Payne A. Value propositions as communication practice: taking a
wider view. Industrial Marketing Management. 2011; 40(2): 202–210.
40. Dibb S. Market segmentation: conceptual and methodological foundations. Journal of Targeting Mea-
surement & Analysis for Marketing. 2000; 9(1): 92–93.
41. Dickson P R, Ginter J L. Market segmentation, product differentiation, and marketing strategy. Journal
of Marketing. 1987; 51(2): 1–10.
42. Srivastava R. Emerging dynamics of labour market inequality in India: migration, informality, seg-
mentation and social discrimination. The Indian Journal of Labour Economics. 2019; 62(2): 147–
171.
43. Vendrellherrero F, Bustinza O F, Parry G, Georgantzis N. Servitization, digitization and supply chain
interdependency. Industrial Marketing Management. 2017; 60: 69–81.
44. Lusch R F, Vargo S L, Tanniru M. Service, value networks and learning. Journal of the Academy of Mar-
keting Science. 2010; 38(1): 19–31.
45. Sharma A, Krishnan R, Grewal D. Value creation in markets a critical area of focus for business-to-busi-
ness markets. Industrial Marketing Management. 2001; 30(4): 391–402.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 19 / 21
46. Huang B, Li C J, Yin C, Zhao X. Cloud manufacturing service platform for small- and medium-sized
enterprises. The International Journal of Advanced Manufacturing Technology. 2013; 65(9): 1261–
1272.
47. Sotoacosta P, Meronocerdan A L. Analyzing e-business value creation from a resource-based perspec-
tive. International Journal of Information Management. 2008; 28(1): 49–60.
48. Choudary S P, Parker G G, Van Alystne M. Platform scale: how an emerging business model helps
startups build large empires with minimum investment. San Francisco: Platform Thinking Labs. 2015.
49. Yang Y, Gong Y, Land L P W, Chesney T. Understanding the effects of physical experience and infor-
mation integration on consumer use of online to offline commerce. International Journal of Information
Management. 2019; 102046.
50. You Z, Si YW, Zhang D, Zeng X, Leung SC, Li T. A decision-making framework for precision marketing.
Expert Systems With Applications. 2015; 42(7): 3357–3367.
51. Wang G, Gunasekaran A, Ngai EW. Big data analytics in logistics and supply chain management: cer-
tain investigations for research and applications. International Journal of Production Economics. 2016;
176(176): 98–110.
52. Dellarocas C. The digitization of word of mouth: promise and challenges of online feedback mecha-
nisms. Management Science. 2003; 49(10): 1407–1424.
53. Wang P, Guo J, Lan Y, Xu J, Cheng X. Your cart tells you: inferring demographic attributes from pur-
chase data. In Proceedings of the Ninth ACM International Conference on Web Search and Data Min-
ing. 2016; 173–182.
54. Abdelkafi N, Tauscher K. Business models for sustainability from a system dynamics perspective.
Organization & Environment. 2016; 29(1): 74–96.
55. Schlie E, Rheinboldt J, Waesche N. Simply seven: seven ways to create a sustainable internet busi-
ness. Springer. 2011.
56. Du S, Wang L, Hu L, Zhu Y. Platform-led green advertising: promote the best or promote by perfor-
mance. Transportation Research Part E: Logistics and Transportation Review. 2019; 128:115–31.
57. Zhou T. An empirical examination of continuance intention of mobile payment services. Decision Sup-
port Systems. 2013; 54(2): 1085–1091.
58. Baghai R P, Becker B. Reputations and credit ratings: evidence from commercial mortgage-backed
securities. Journal of Financial Economics. 2020; 135(2): 425–444.
59. Bagheri S, Kusters R J, Trienekens J J, van der Zandt HV. Classification framework of knowledge trans-
fer issues across value networks. Procedia CIRP. 2016; 47: 382–387.
60. Böhm M, Koleva G, Leimeister S, Riedl C, Krcmar H. Towards a generic value network for cloud com-
puting. International Workshop on Grid Economics and Business Models. 2010; 129–140.
61. Joerling J. Data breach notification laws: an argument for a comprehensive federal law to protect con-
sumer data. Washington University Journal of Law and Policy. 2010; 32(1): 467–488.
62. Lin R. Using fuzzy DEMATEL to evaluate the green supply chain management practices. Journal of
Cleaner Production. 2013; 40: 32–39.
63. Singh P K, Sarkar P. A framework based on fuzzy Delphi and DEMATEL for sustainable product devel-
opment: a case of Indian automotive industry. Journal of Cleaner Production, 2020; 246(10): 118991.
64. Kristensen H S, Remmen A. A framework for sustainable value propositions in product-service sys-
tems. Journal of Cleaner Production. 2019; 223(1): 25–35.
65. Nenonen S, Storbacka K, Sklyar A, Frow P, Payne A. Value propositions as market-shaping devices: a
qualitative comparative analysis. Industrial Marketing Management. 2020; 87(1): 276–290.
66. Mishra S, Ewing M T, Pitt L F. The effects of an articulated customer value proposition (CVP) on promo-
tional expense, brand investment and firm performance in B2B markets: a text-based analysis. Indus-
trial Marketing Management. 2020; 87(1): 264–275.
67. Blomsma F, Pieroni M, Kravchenko M, Pigosso D C, Hildenbrand J, Kristinsdottir A R, et al. Developing
a circular strategies framework for manufacturing companies to support circular economy-oriented inno-
vation. Journal of Cleaner Production. 2019; 241(1): 118271.
68. Casadesus-Masanell R, Halaburda H. When does a platform create value by limiting choice? Journal
Economics & Management Strategy. 2014; 23(2): 259–293.
69. de Camargo Fiorini P, Jabbour C J C. Information systems and sustainable supply chain management
towards a more sustainable society: where we are and where we are going. International Journal of
Information Management. 2017; 37(4), 241–249.
70. Soltani Z, Navimipour N J. Customer relationship management mechanisms: a systematic review of the
state of the art literature and recommendations for future research. Computers in Human Behavior.
2016; 61: 667–688.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 20 / 21
71. Shafer S M, Smith H J, Linder J C. The power of business models. Business Horizons. 2005; 48(3):
199–207.
72. Rappa M A. The utility business model and the future of computing services. IBM Systems Journal.
2004; 43(1): 32–42.
73. Wang W Y, Wang Y. Analytics in the era of big data: the digital transformations and value creation in
industrial marketing. Industrial Marketing Management. 2020; 86(1): 12–15.
PLOS ONE A hierarchical framework
PLOS ONE | https://doi.org/10.1371/journal.pone.0238152 September 24, 2020 21 / 21
Copyright of PLoS ONE is the property of Public Library of Science and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
