Documente Academic
Documente Profesional
Documente Cultură
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
ABSTRACT
Managing customer satisfaction has become a crucial issue in fast-food industry. This study aims at
identifying determinant factor related to customer satisfaction in fast-food restaurant. Customer data are
analyzed by using data mining method with two classification techniques such as decision tree and neural
network. Classification models are developed using decision tree and neural network to determine
underlying attributes of customer satisfaction. Generated rules are beneficial for managerial and practical
implementation in fast-food industry. Decision tree and neural network yield more than 80% of predictive
accuracy.
Keywords: customer satisfaction, classification model, data mining, fast-food restaurant, rules extraction
18
Journal of Theoretical and Applied Information Technology
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
managerial implication of the findings for fast-food order to provide responsive decision making
restaurant operators. continuously.
2. LITERATURE REVIEW Table 1: Attributes Related to Customer Satisfaction
2.1 Customer Satisfaction in Fast-food Authors Methodology Result
Restaurant Qu [13] Regression Food and environment
Nowadays, marketing strategy has shifted from analysis have the largest effect
strategy that focuses on the brand (brand-centered) on intention of customer
to customer-focused strategy (customer-centered). to come back.
Many enterprises attempt to identify the factors that Arora & Regression Food quality and
support customer satisfaction since it is a one of Singer [14] analysis service have significant
crucial business decision. As the heart of marketing, influence to customer
the ability to satisfy customers is great importance satisfaction
Baek, et al Conjoint Menu price are the key
for many reasons [10]. Customer satisfaction is the
[15] analysis attributes of choosing
proportion between what is customers expectation fast-food restaurant.
and perceptions [11]. In business perspective, Gupta, et Descriptive Food performance is
customer satisfaction is a measure of how products al [16] statistic considered as crucial
and services provided by an enterprise meet attribute attributes of
customer expectations [12]. improvement for
guests come back
In the context of fast-food industry, it is Park [17] Correlation Food quality, staff
necessary to understand determinant factors driving analysis kindness, and quick
customer satisfaction. For the restaurant managers, service are the most
an understanding of underlying factors that important attributes in
influence customer satisfaction is a guidance to fast-food restaurant
design services and offering [10]. Prior studies Jang, et al Factor Service and food quality
revealed the important attributes of customer [18] analysis and are key attributes
satisfaction in restaurant. Since only few studies Structural influencing customer
equation intention.
discussed customer satisfaction attributes in fast-
modeling
food restaurant, thus we present and summarize
attributes affect customer satisfaction in several
types of restaurant as shown in Table 1. State of the art technological developments
related to information management such as data
Tangible (e.g. food quality and food presentation)
warehouse and data mining enable enterprises to
and intangible (e.g. staff behavior and restaurant
capture and analyze customer data up to the level of
atmosphere) attributes were considered. For
individual customers [20]. Confluence of data
instance, Qu [13], Arora & Singer [14], Baek et al
mining and machine learning in customer
[15], and Gupta et al [16] considered attributes such
relationship management (CRM) domain have been
as food quality (taste), food range, and food price
widely employed as tools to discover the
are the most important attributes in restaurant.
relationship among attributes in the data. In
Several studies also indicate that nonfood-related
customer-related research, DT and NN have been
attributes such as restaurant environment and staff
considerably used in closed cycle CRM dimensions
friendliness were the determinant factors
such as customer identification, customer attraction,
contributing to customer experience in restaurant
customer retention, and customer development [12].
[17], [18].
For instance, DT and rule extraction algorithm,
2.2 Data Mining and Rule Extraction called NeuroRule [21], had been applied to predict
Techniques consumer preference of fast-food franchise.
As stated by Hogan et al [19], the ability of an Classification rules that were generated from these
enterprise in acquiring and managing customer techniques could achieve predictive accuracy of
information is the key to maintain competitive more than 80% [22]. An empirical experiment that
advantage; but, managing customer relationships is was conducted by Hung, et al [23] shows these two
hard to perform since customer, either as individual techniques effectively enabled telecom service
or in group, has different preferences and providers to make more accurate and effective
expectations. Hence, to overcome such problems, churner prediction. Neural network also
information-based marketing strategy is needed in outperformed the Logit Regression in identifying
potential churners [24].
19
Journal of Theoretical and Applied Information Technology
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
3. RESEARCH PROCESS
Model Usage
In this section, we depict the research process of
this study as in Figure 1. Details of each steps is
described below.
Figure 1: Diagram of Research Process
3.1 Data Collection and Pre-Processing
According to Kotler and Armstrong [28], four
tools are employed to measure customer satisfaction
3.2 Data Mining and Model Evaluation
as well as providing complaint systems, satisfaction
To extract the rules, we adopted well-known
survey, ghost shopping, and losing customer
C4.5 program [9] was implemented as Rattle in R
analysis. Hence, we collected data samples through
packages [29]. The interface (Figure 2) was
a survey relating to customer satisfaction.
developed based on REANN algorithm to provide
Empirical data were collected through a self- researchers, restaurant manager, or decision makers
administered, survey of customers in big retail a simple and easy tool in performing marketing
chain of fast-food restaurant in Palembang, research.
Indonesia. The field work was conducted from
Rules are extracted from DT. For this purpose, of
January 2011 to February 2011. Some trained
the 340 respondents who participated in survey, 133
volunteers were involved in the survey and the
observations (39.12%) and 207 (60.88%)
author ensured and maintained the integrity of data
observations were assigned to the Class 0
quality. In total, 400 questionnaires with 10 input
(unsatisfied) and Class 1 (satisfied), respectively.
variables were distributed such as taste, serving,
For NN training, Class 0 (unsatisfied) and Class 1
dessert, restaurant environment, washing basin,
(satisfied) cases were assigned target values of (0, 1)
waiting time, staff behavior, menu pricing, discount,
and (1, 0), respectively.
and store location. Final data contains 340 cases
after we performed data pre-processing by To evaluate the model, datasets were partitioned
eliminating duplicate instances or those with many into 2 parts. The first part was for training and rule
inaccurate and missing values. extraction and the second part was applied for
model validation. The first set was selected
randomly and consists of two-third (66%) of the
20
Journal of Theoretical and Applied Information Technology
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
21
Journal of Theoretical and Applied Information Technology
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
In addition, rule R2 can also answer the question service is needed as well to ensure efficient
"What factors in addition to price, service, and a operation.
serving that lead customers to be unsatisfied?" It
Apart from staff behavior, this study also
can be concluded that inconvenient of location is
uncovered convenient location as determinant
significant attribute; therefore R2 might be useful to
factor. It is common that cleanliness is an important
restaurant operators in improving the convenient
aspect of customer experience in the restaurant.
place. It might be of interest to seek further
There is correlation between customer expenditure,
questions to the customers opinion concerning the
customer experience, and sanitary (cleanliness)
restaurant cleanliness and neatness. Through this,
condition.
restaurant operators can then improve the
environment of its outlets accordingly. Finally, food presentation and quality become
essential factors in understanding customer
Moreover, there were no significant difference
satisfaction in fast-food restaurant. In this study it
between the rules generated by C4.5 and REANN.
was also discovered that functional aspects of
Overall, a closer inspection of the rules generated
shaping customer experience i.e. taste [5] has an
by C4.5 and REANN could enable in identifying
association with customer satisfaction. It is
how the fast-food restaurant could strengthen their
recommended that restaurant operators understand
competitive advantage. Mostly, attributes
local taste of food in different areas. For instance, in
dimension such as service and taste appeared in
the certain culture (area), customers prefer to have
most rules, thus the findings suggested that the
high level of spiciness, so that it is notable to make
restaurant operator can enhance customer
food products properly.
satisfaction by offering best service and tasty food.
From business perspective, this experiment revealed
essential recommendation to prioritize attributes by
paying more attention to the best service attributes 5. CONCLUSION
along with taste and price accordingly. However, Data mining approaches based on decision tree
the finding is contrary to [7] that stated food quality and neural network to identify the most determinant
became the most determinant attribute in customer predictor in influencing customer satisfaction in
satisfaction. fast-food restaurant were proposed. Generated rules
from C4.5 and REANN were quite similar, and
4.3 Managerial and Practical Implications REANN performed best accuracy in validation test
This study provides some managerial compared to C4.5. The basic customer service
implications which can be used by restaurant attribute (i.e. staff behavior), restaurants
operators to better understand the factors related to cleanliness, and food presentation was revealed to
customer satisfaction in fast-food restaurant. Fast- be significant predictor in determining customer
food restaurant have typical characteristic and satisfaction. This study provides some findings that
become a trend in Indonesia. However, it is are useful for the marketers, policy makers who
complicated to maintain fast-food customers have interest in customer satisfaction research. For
become loyal customer. In strong competitive practical implication, an understanding of
environment, creating a unique products or brands underlying determinant factors of customer
by potent positioning and differentiating is satisfaction could strengthen the competitive
necessary to enhance competitive advantage. advantage of the fast-food restaurant.
This study revealed underlying factors of In the future we intend to conduct a cross-
customer satisfaction in fast food restaurant such as sectional research by comparing the characteristics
staff behavior. It is recommended that restaurant of fast-food customers in Indonesia and other
operators understand required attitude when serving countries. In addition, it also would be interesting
customers. Several crucial aspects regarding staff by applying the model in other category of
behavior such as staffs attitude, gesture, and restaurants i.e. specialty and/or continental
emotional intelligence in facing and serving restaurant. Finally, it would be useful if more data
customers should be considered. Likewise, samples could be collected in the future.
adequate staff training such as motivational training,
basic customer service training aimed at enabling
fast-food restaurant to offer excellent service to
their customers should be considered. Lastly,
continuous evaluation of the quality of customer
22
Journal of Theoretical and Applied Information Technology
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
23
Journal of Theoretical and Applied Information Technology
th
10 May 2015. Vol.75. No.1
2005 - 2015 JATIT & LLS. All rights reserved.
24