Các tiêu chí quyết định sự lựa chọn trang thương mại điện tử sử dụng phương pháp đánh giá Fuzzy-Delphi: Nghiên cứu trường hợp tại Việt Nam

 SỐ 63 (8-2020) 
KHOA HỌC - CÔNG NGHỆ 
95 
TẠP CHÍ ISSN: 1859-316X 
KHOA HỌC CÔNG NGHỆ HÀNG HẢI 
JOURNAL OF MARINE SCIENCE AND TECHNOLOGY 
THE DETERMINANTS OF E-COMMERCE WEBSITE SELECTION USING 
DELPHI - FUZZY EVALUATION METHOD: A CASE STUDY OF VIETNAM 
CÁC TIÊU CHÍ QUYẾT ĐỊNH SỰ LỰA CHỌN TRANG THƯƠNG MẠI ĐIỆN TỬ 
SỬ DỤNG PHƯƠNG PHÁP ĐÁNH GIÁ FUZZY-DELPHI: NGHIÊN CỨU TRƯỜNG 
HỢP TẠI VIỆT NAM 
LÊ THỊ NGỌC NGÀ1, PHẠM THỊ NGA1, NGUYỄN THỊ HƯƠNG TRÀ1, 
ĐÀO THỊ HƯƠNG LINH1, TRỊNH THỊ HỒNG MINH1, PHẠM THỊ YẾN2* 
1International School of Education, Vietnam Maritime University 
2Faculty of Maritime Business, Vietnam Maritime University 
*Email: phamyen@vimaru.edu.vn 
Abstract 
Vietnam is one of the developing countries that is 
approving and bringing e-commerce into an 
important element. It leads to the changing the 
behavior of users in Vietnam in which people can 
use electronic device features to do shopping 
activities. Along with the occurrence and 
development of E-commerce, customers can have 
benefits from diversity of choices but it causes a 
fiercer competition among these e-retailers at the 
same time. Understanding the criteria that 
influence the choice of e-commerce websites is 
vital. Hence, the study aims to examine the criteria 
affecting the selection of websites of customers by 
applying a modified Delphi method and fuzzy 
theory. The results illustrate that among 11 factors, 
product or service quality is prioritized as 
selecting a website, followed by feedback from 
previous customers. 
Keywords: E-commerce, competition, modified 
Delphi method and fuzzy theory. 
Tóm tắt 
Thương mại điện tử tại Việt Nam đang phát triển 
mạnh mẽ. Thương mại điện tử dẫn tới sự thay đổi 
thói quen của người tiêu dùng ở Việt Nam khi mà 
các hoạt động mua sắm thông qua việc sử dụng 
các thiết bị điện tử. Với sự ra đời và phát triển của 
các trang thương mại điện tử, người tiêu dùng 
ngày càng được hưởng nhiều lợi ích hơn từ việc có 
nhiều sự lựa chọn hơn. Đồng thời, chính sự gia 
tăng nhanh chóng này cũng tạo ra sự cạnh tranh 
gay gắt giữa các kênh thương mại điện tử. Nhận 
biết được các tiêu chí quyết định đến sự lựa chọn 
các trang website là rất quan trọng. Chính vì vậy, 
nghiên cứu thực hiện nhằm mục đích đánh giá các 
tiêu chí ảnh hưởng đến việc lựa chọn trang website 
của khách hàng bằng cách áp dụng phương pháp 
Delphi và lý thuyết Fuzzy. Kết quả cho thấy rằng 
chất lượng sản phẩm/ dịch vụ được ưu tiên hơn khi 
chọn một trang website, tiếp theo là phản hồi từ 
các khách hàng trước đó. 
Từ khóa: Thương mại điện tử, cạnh tranh, 
phương pháp Delphi và lý thuyết Fuzzy. 
1. Introduction 
E-commerce has emerged as an indispensable 
part of modern lifestyle in which buying and selling 
of products and services are supported by internet. 
Establishing competitive differentiation to attract and 
satisfy customers is paramount importance to 
e-commerce websites. Understanding customer 
desires can be considered as a key component of 
effectiveness and success. Many empirical studies 
have conducted to evaluate factors affecting the 
decision of selecting e-commerce websites. Samira, 
et al, [1] used Analytic Hierarchy Process method to 
identify 6 factors effecting the decision of selecting 
e-commerce websites in Bangladesh including ease 
of usage, past experience, privacy and security, 
customer service, product variety and brand image. 
Meanwhile, the success factors for e-commerce in 
Thailand were contributed by industry KSFs, 
well-designed websites, Internet connection, IT 
capability, large product selection, online security, 
brand name recognition, competitive prices and 
promotion, customer support and relationship, and 
order fulfillment under matrix method [2]. In Nigeria, 
Folorunso, et al, [3] used a correlation matrix 
approach to prove that the cost of implementation, 
accessibility, data security and citizen’s income are 
the most important factors. Factors affecting the 
selecting e-commerce websites vary from country to 
country, from region to region as a result of 
differences in cultures, needs, and customers 
KHOA HỌC - CÔNG NGHỆ 
96 SỐ 63 (8-2020) 
TẠP CHÍ ISSN: 1859-316X 
KHOA HỌC CÔNG NGHỆ HÀNG HẢI 
JOURNAL OF MARINE SCIENCE AND TECHNOLOGY 
behaviour. The determinants of e-commerce website 
selection in Vietnam are definitely different from 
those in a certain country. 
Vietnam’s e-commerce sector is witnessed strong 
growth in the region with revenue of approximately 
3 billion US$ and more than 50 million users in 2019 
by Statista, though it is just as an emerging market 
behind Singapore, Indonesia, and Thailand. Owning 
a range of comparative advantages such as young 
population, increasing middle class, high internet and 
smart phone users is considered the key factor 
driving Vietnam becoming one of the most 
promising e-commerce market attracting both 
domestic and foreign investors. The top five most 
successful e-commerce platforms in Vietnam 
including Vietnamese e-commerce platforms namely 
The gioi Di Dong, Sendo, Tiki and the two 
international co-operations Lazada Viet Nam and 
Shopee Viet Nam by iPrice Group have made 
improvements to attract more e-commerce customers. 
However, the number of researches about 
determinants of e-commerce website selection in 
Vietnam is exceedingly limited. Therefore, to fill the 
gap, the study aims to examine the criteria affecting 
the selection of websites of customers by applying 
modified Delphi method and fuzzy theory. The study 
here aims to identify factors affecting e-commerce 
website selection with the scope of Vietnam, thereby 
provides both academic and practice implications to 
improve the services of the E-commerce websites. 
2. Methodology 
E-commerce website selection is a multiple 
criteria decision making problem. In many cases, the 
preferential model of decision making is uncertain, 
and it is difficult for decision makers to provide 
exact numerical values for comparative ratios [4]. 
This study hence proposes using fuzzy theory to 
resolve the uncertainly and imprecision of 
performance evaluations, in which the comparison 
judgments of a decision maker are presented as fuzzy 
triangular numbers. To more accurately reflect the 
original opinions of decision makers, a Fuzzy - 
Delphi methodology, which is able to handle both 
the quantitative and qualitative elements of a 
problem, is used. 
Fuzzy - Delphi is a methodology combining the 
Fuzzy method and Delphi method for optimal 
decision making strategies. The Fuzzy Delphi 
method can resolve uncertainly regarding decision 
space and combine the advantages of statistical 
methods [5]. It has four advantages: to decrease the 
times of questionnaire surveys, to avoid distorting 
individual expert opinions, to clearly express the 
sematic structure of predicted items, and to consider 
the fuzzy nature during the interview process [6]. 
This study hence proposes using fuzzy theory to 
measure experts’ perceptions utilizing linguistic 
expressions as ‘strongly unimportant’, ‘unimportant’, 
‘neutral’, ‘important’ and ‘strongly important’ to 
achieve the judgment of decision makers... This is 
shown with the support of membership ability, which 
is evaluated in the unit interval of real [0, 10]. Fuzzy 
sets extend classical sets as the index functions of 
classical sets are special cases of the membership 
ability of fuzzy sets if the latter only have the values 
0 or 10. A classic value set is typically called crisp 
sets in the fuzzy theory [7]. A triangular fuzzy 
number is composed of three parameters, i.e., a1, a2, 
and a3, and the membership function can be 
indicated as shown in equation (1): 
𝜇𝐴 =
{
0, 𝑥 ≤ 𝛼1
(𝑥 − 𝛼1)/((𝛼2 − 𝛼3). 𝛼1 ≤ 𝑥 ≤ 𝛼2
(𝛼3 − 𝑥)/(𝛼3 − 𝛼2), 𝑎2 ≤ 𝑥 ≤ 𝛼3
0, 𝑥 ≥ 𝛼3
 (1) 
Triangular fuzzy numbers between the membership 
function “n” are defined as shown here: 
[�̅� = 𝛼1
(𝑖)
, 𝛼2
(𝑖)
, 𝛼3
(𝑖), 𝑖 = 1,2,3+ 𝑛] (2) 
Fuzzy number is defined as follows: 
[�̅� = 𝐴𝑎𝑣𝑒 =
�̅�+�̅�+⋯+�̅�
𝑛
=
(∑ 𝛼1
(𝑖)
∑ 𝛼2
(𝑖)
∑ 𝛼3
(𝑖)𝑛
1
𝑛
1
𝑛
1 )
𝑛 =(𝛼1,𝛼2,𝛼3)] (3) 
The last step in the fuzzy method is 
de-fuzzification. The aiming of de-fuzzification is to 
convert the results of the whole fuzzy set obtained in 
the previous step into the crisp numbers. The most 
common method of de-fuzzification is the centre of 
gravity. This method solves the centre of the area of 
the binding membership function. 
𝑌∗ =
(𝐴3−𝐴1)+(𝐴2−𝐴1)
3
+ 𝐴1 (4) 
KHOA HỌC - CÔNG NGHỆ 
97 SỐ 63 (8-2020) 
TẠP CHÍ ISSN: 1859-316X 
KHOA HỌC CÔNG NGHỆ HÀNG HẢI 
JOURNAL OF MARINE SCIENCE AND TECHNOLOGY 
Source: Zadeh (1965) [8] 
Figure 1. The triangular fuzzy number 
Table 1. Linguistic variables for the evaluation 
of each factor 
3. Empirical study 
The research process can be found in Figure 2. 
A hybrid Fuzzy - Delphi based methodology divides 
the whole benchmarking process into two stages. 
The first stage includes identification, synthesis of 
the key factors that may affect the e-commerce 
website selection by customers via modified Delphi 
method. The second stage is to set up the fuzzy 
matrix and compute the weights of each KPF using 
the Fuzzy Delphi method to prioritize the key 
factors. 
3.1. Identifying factors affecting the selection of 
e-commerce websites 
In the first stage, determinants deriving from 
previous research related to selecting e-commerce 
websites was circulated among fifteen respondents 
who usually shop online were interviewed during a 
brainstorming session to identify the KPFs. In all, 
within a period of 28 days (from 19 December 2019 
to 15 January 2020), 11 factors were identified in this 
session as shown in Table 2. To determine the crucial 
factors among all of the factors obtained from the 
participants’ opinion more objectively, a 5-point 
scale questionnaire survey was simultaneously 
administered. Cronbach's Alpha was applied to test 
the reliability of the questionnaire before the 
selection of appraisal KPFs. The value of 0.753 that 
was obtained is greater than 0.35 and is therefore 
viewed as reliable. If any Cronbach's Alpha is less 
than 0.35, the corresponding datum is not reliable 
and will be deleted. Those more than 0.35 are viewed 
as reliable [9]. 
Linguistic scale Fuzzy score 
Strongly unimportant ( 1, 1, 2 ) 
Unimportant ( 2, 3, 4 ) 
Neutral ( 4, 5, 6 ) 
Important ( 6, 7, 8 ) 
Strongly important ( 8, 9,10 ) 
Figure 2. Generalized framework through Fuzzy-Delphi based approach 
Level 1: Literature review Open-ended questions, 
Brainstorming 
The first questionnaire 
Level 2: Perform the questionnaire reliability test. 
Level 3: Select of appraisal key performance factor (KPF)s 
Delphi 
Method 
Level 4: The second questionnaire 
Level 5: Establish the triangular fuzzy numbers with each KPF 
Level 6: Rank the preference factors for website selection. 
Fuzzy 
Delphi 
KHOA HỌC - CÔNG NGHỆ 
98 SỐ 63 (8-2020) 
TẠP CHÍ ISSN: 1859-316X 
KHOA HỌC CÔNG NGHỆ HÀNG HẢI 
JOURNAL OF MARINE SCIENCE AND TECHNOLOGY 
3.2. Weighting factors affecting the selection of 
e-commerce websites 
The second survey was conducted with a larger 
number of participants. There were 102 participants 
having experience in buying products in e-commerce 
websites, in which 2 questionnaires missing answer 
were eliminated, so 100 questionnaires were 
successfully returned and validated as shown in table 
2. The evaluation was checked the reliability by 
Cronbach's Alpha which was 0.913. 
To clarify the priority of the criteria, the fuzzy 
method was applied using linguistic variables as 
shown in Table 1. The final fuzzy scores were based 
on equation (3) and the defuzzification was based on 
equation (4). The results are illustrated in Table 3. 
Table 2. General information about responses 
Gender 
Female 72 % 
Male 28 % 
Age 
Less than 18 5 % 
18-25 91 % 
More than 25 4 % 
Occupation 
Student 79 % 
Employee 16 % 
Others 5 % 
Frequency of 
buying 
online 
Always 6 % 
Often 23 % 
Sometimes 58 % 
Seldom 13 % 
Main kind of 
product 
purchased 
Clothes, shoes 53.5 % 
Accessories 10.3 % 
Cosmetic 10.3 % 
Book 8.4 % 
Others 17.5 % 
Table 3. Importance weights of criteria 
Key 
performance 
factors 
Fuzzy score 
Defuzzi
fication 
Rank 
Competitiv
e price 
5.78 6.72 7.72 6.74 3 
The quality 
of product 
and service 
6.17 7.10 8.10 7.12 1 
Shipping 
cost 
5.17 6.12 7.12 6.14 10 
Shipping 
time 
5.50 6.46 7.46 6.47 4 
Feedback 
from 
previous 
customers 
6.09 7.06 8.06 7.07 2 
Warranty 4.97 5.90 6.90 5.92 11 
Customer 
responsive
ness 
5.43 6.42 7.43 6.43 5 
Value-add
ed services 
(discount, 
package,..) 
5.18 6.14 7.14 6.15 8 
Description 
of 
product/ser
vice 
5.43 6.40 7.40 6.41 6 
Refund 
policy 
5.36 6.32 7.32 6.33 7 
Ease of 
checkout 
(payment 
method) 
5.16 6.13 7.13 6.14 9 
The ranking of the determinants of the selection 
of e-commerce websites shows that the quality of 
product and service is ranked as the most important 
factor when choosing an e-commerce website. 
Customers are much interested in websites offering 
clear original, brand name, verified e-retailers and 
high quality products and services. Next, the 
feedback from previous customer factor is ranked 
second. Customers have no have the option of testing 
or checking the product before its delivery, so 
feedbacks from previous customers play an 
important role in the decision to buy products of 
users. They will be more reassuring when purchasing 
products on an e-commerce website with positive 
comments and high rating. “Competitive price” and 
“shipping time” are also among the deciding factors 
when choosing a website, ranked 3 and 4 
respectively. Noticing the high requirement among 
e-commerce for speedy and timely delivery some 
Vietnamese e-commerce platforms such as Tiki 
offered Tikinow or Shopee introduced delivery 
policy in four hours. The lowest priority is given to 
ease of checkout (payment method), shipping cost 
and warranty. Vietnam has low banking penetration, 
so customers could choose the simplest payment 
which is cash on delivery payment method besides 
credit card payment or mobile method. The 
KHOA HỌC - CÔNG NGHỆ 
99 SỐ 63 (8-2020) 
TẠP CHÍ ISSN: 1859-316X 
KHOA HỌC CÔNG NGHỆ HÀNG HẢI 
JOURNAL OF MARINE SCIENCE AND TECHNOLOGY 
“warranty” factor is at the bottom of the rankings. 
This is well explained by the market’s largest 
segment is Fashion followed by electronic and media, 
toy, hobby, furniture and appliances. 
4. Conclusion 
The development of e-commerce is providing a 
great opportunity for e-retailers. Recognizing 
customer needs is crucial to increase the 
competitiveness and success. The study is to analyze 
the determinants of e-commerce websites selection. 
The results using the hybrid modified Delphi - fuzzy 
method proposed criteria in selecting certain 
e-commerce websites in which the quality of product 
and service, feedback from previous customer, 
competitive price and shipping time are important 
factors. This study, thus, have the potential to enrich 
the understanding on how customers select and 
evaluate e-commerce websites. The identification of 
priority factors will be crucial to e-commerce 
websites satisfy customer’s requirement as limited 
resource availability. This leads to useful 
implications for managers of e-retailers and 
e-commerce websites to identify the elements to 
focus and improve. Moreover, the study would a 
timely contribution to the literature on e-commerce 
field. Despite of academic and practical 
implications, the findings should be specified for 
each group customers having different characteristics 
such as ages, income and type of products. It is 
recommended that future studies should replicate and 
develop to examine potential differences of criteria 
to have better insights of each customer 
segmentation as well of each group of e-commerce 
websites for different products and services such as 
consumer goods, hotels, motels, tourism and air 
tickets. 
REFERENCES 
[1] Samira Sahel, Tasmia Anwar, Barnali Nandi - 
University of Asia Pacific, 
“Factors-Affecting-the-Selection-of-an-E-Comm
erce-Website-in-Bangladesh, 2018. 
[2] Laosethakul, Kittipong; Oswald, Sharon; and 
Boulton, William, "Critical Success Factors for 
E-commerce in Thailand: A Multiple Case Study 
Analysis".AMCIS 2006 Proceedings. 423 
[3] Folorunso, Olusegun, Awe Oludare Gabriel, 
Sharma, Sushil K. and Jeff Zhang , 2006. Factors 
Affecting the Adoption of E-commerce: A Study 
in Nigeria. Journal of Applied Sciences, 6: 
2224-2230, 2006. 
[4] Tsai, H.Y., Chang, C.W., Lin, H.L., Fuzzy 
hierarchy sensitive with Delphi method to 
evaluate hospital organization performance. 
Expert Syst. Appl. 37, pp.5533-5541, 2010. 
[5] Ma, Z., Shao, C., Ma, S., Ye, Z. Constructing 
road safety performance indicators using Fuzzy 
Delphi method and Grey Delphi method. Expert 
Syst. Appl. 38, pp.1509-1514, 2010 
[6] Chang, P.C., Wang, Y.W, Fuzzy Delphi and 
back-propagation model for sales forecasting in 
PCB industry. Expert Syst. Appl. 30, pp.715-726, 
2006. 
[7] Jeon, J-w., Yeo, G-T., Thai, V.V. and Yip, T.L, 
An evaluation of the success factors for ship 
management companies using fuzzy evaluation 
method, Int. J. Shipping and Transport Logistics, 
Vol. 8, No. 4, pp.389-405, 2016. 
[8] Zadeh, L.A. ‘Fuzzy sets’, Information and 
Control, Vol. 8, No. 3, pp.338-353, 1965. 
[9] Miao-Ling Wang, Yu-Hsin Lin, 2008, To 
construct a monitoring mechanism of production 
loss by using Fuzzy Delphi method and fuzzy 
regression technique - A case study of IC package 
testing company. Expert Syst. Appl. 35, 
pp.1156-1165, 2008. 
Received: 27 March 2020 
Revised: 10 May 2020 
Accepted: 19 May 2020