Telecommunications Policy 39 (2015) 627641

Contents lists available at ScienceDirect

Telecommunications Policy
URL: www.elsevier.com/locate/telpol

Effect of the customer experience on satisfaction

with smartphones: Assessing smart satisfaction index
with partial least squares
Dong-Hee Shin n
Sungkyunkwan University, B307 International Hall, 53 Myeongnyun-dong 3-ga, Jongno-gu,
Seoul 110-745, South Korea

a r t i c l e i n f o abstract

Available online 13 November 2014 With the rapid diffusion of a wide variety of smartphones, quality issues have become
Keywords: central to consumers. While customer satisfaction of most goods and services has been
Smartphone well researched, little research seems to exist on satisfaction and loyalty with respect to
Customer satisfaction index advanced mobile services, such as smartphones. This study applied a customer satisfac-
Smart service CSI model tion index (CSI) model to the smart mobile sector in order to derive a smart-service CSI
South Korea (SCSI). The SCSI model and its hypotheses were then tested using partial least square
analysis and index calculation. The findings showed that the perceived value and
customer satisfaction are key variables mediating the relationship between quality and
customer loyalty. The proposed model demonstrated strong explanatory power, with
satisfactory reliability and validity. The SCSI model establishes a foundation for future
smart service categories on the basis of providing a powerful tool for quality assessment.
The results of this study provide useful insights for the telecom industry and policy-
makers, for the forging of effective policies and competitive strategies.
& 2014 Elsevier Ltd. All rights reserved.

1. Introduction

As smart technology has become almost a necessity these days, consumer satisfaction when using smart services, in
terms of perception and behaviors, has become a topic of considerable interest.1 This invigorated interest is due in large part
to the rising importance of smart services, and the need to understand how to improve customer satisfaction and loyalty in
the smart era. Smart devices are designed to support a variety of services, and to support a range of properties pertaining to
ubiquitous computing. In recent years, the rapid development of mobile app markets and of mobile commerce has been
drivers of smartphone adoption.
With the drastic increase in the development of smart devices, it is even more important for service providers to
measure the level of satisfaction among customers, as this would help the providers to identify and reform any areas in need

n
Tel.: 82 02 740 1864; fax: 82 02 740 1856.
E-mail address: dshin@skku.edu
1
While there is no single accepted definition, in general, smart technology is defined as technology possessing one or more advanced characteristics
such as Internet access, capturing still or moving images, and communicating via phone, SMS, Bluetooth, or other applications, or via programs that allow a
channel of data share or communication. Examples of smart technology include mobile phones, tablet devices, smart watches, ipods, and ipads. Smart
service refers to the service provided through smart technologies, and is characterized as easy to use, user-centered, and creative services.

http://dx.doi.org/10.1016/j.telpol.2014.10.001
0308-5961/& 2014 Elsevier Ltd. All rights reserved.
628 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

of improvement. Despite the increasing need for an objective measurement of smart service quality, no common agreement
has been made on how smartphone satisfaction should be conceptualized and measured (Chun, Lee, & Kim, 2012). Issues
related to the improvement of customer satisfaction have long been a major area of research, and the smartphone is no
exception. While the American Customer Satisfaction Index (ACSI) has provided a comprehensive framework for numerous
existing index models (Fornell, Johnson, Anderson, Cha, & Everitt 1996), related studies have focused mostly on calculating
customer satisfaction indices for entire industry sectors. The resulting rankings have been used for marketing and
advertising purposes. Only a few previous studies in this field have explored the perceived quality factors that have an actual
impact on customer satisfaction, and none have sought to develop strategies for quality improvement. In light of this gap,
the present study proposes a smart-service customer satisfaction index (SCSI) model, modified from ACSI to better suit the
emerging field of smart technologies. By measuring customer satisfaction index, the findings from this study offer service
providers practical suggestions which would result in the improved quality of service, in advance of full-fledged rollout.
In addition, this study offers policymakers and researchers fundamental guidelines for assessing the performance of mobile
network operators, in terms of user satisfaction, and continuing loyalty.
This study contributes to the existing literature in three ways. First, by providing the first customer satisfaction index
(CSI) for smart services, it provides useful insights, particularly pertaining to the disentangling of possible sources of
variation in consumer satisfaction (or related variables) in smart services. Second, the results should prove valuable to
market research practitioners engaged in smartphone customer satisfaction measurement programs, as these researchers
are increasingly tasked with the development of smartphone-specific factors and satisfaction, making vital decisions based
on this information. Finally, this study provides insights and guidance to the smart industry for their international market
strategies, seeking global growth through improved cross-national customer satisfaction.

2. Theoretical background

2.1. Prior models of the customer satisfaction index

Customer satisfaction models normally include the attributes that describe a product or service, the benefits of
consequences these attributes provide customers, a customer's overall evaluation of their purchase and using experience,
and the intentions. The CSI model is a structural model based on the theory that customer satisfaction is created by factors
such as quality, value, expectations of customers, and the image of a company (Fornell et al., 1996). These factors are the
antecedents of overall customer satisfaction. The model also estimates the results when a customer is satisfied or not. These
results of customer satisfaction are consequence factors, such as complaints or loyalty of the customer (Fornell, 1992). Each
factor in the CSI model is a latent construct, which is operationalized by multiple indicators (Turkyilmaz & Ozkan, 2007).
The Swedish customer satisfaction barometer (SCSB) of 1989 was the first national CSI (Fornell, 1992). It was applied to
130 companies from 32 Swedish industries. However, while Fornell (1992) described the marketing foundations of SCSB in
great detail, he only examined the statistical aspects of the problem. To overcome the limitations of the method, the German
Customer Satisfaction Barometer (German Customer Barometer, 1995) was introduced. It covers 42 different industries, for
which telephone interviews are conducted with the public by industry. Standardized questions about the levels of and
reasons for satisfaction/dissatisfaction, recommendation of products and services, and complaint handling are all included.
The ACSI model, which builds upon the original SCSB model, was launched in 1994 and has served as the basis for other
CSI models developed in many countries around the world (Fornell & Larcker, 1981). The model is composed of six factors:
perceived quality, customer expectations, perceived value, overall customer satisfaction, customer complaints, and customer
loyalty. Each factor is linked to the others through a causal relationship (Fornell et al., 1996). The higher the customer
expectations are, the higher the perceived quality is; in turn, higher customer expectations and higher perceived quality
lead to higher perceived value, which ultimately results in higher customer satisfaction. Likewise, a high level of customer
satisfaction tends to reduce customer complaints, while also increasing customer loyalty. Thus, the causal model explains
the inverse proportional relationship between customer complaints and customer loyalty (ACSI, 2010).
The European customer satisfaction index (ECSI) was introduced in 1999 in 11 European countries. The ECSI model
(1999) employs six constructs including image, customer expectations, perceived qualities of hardware and software,
perceived value, customer satisfaction, and customer loyalty. These six factors are linked through causal relationships;
image has a determining influence on customer expectations, while customer expectations, in turn, affect the perceived
quality of hardware or software (O'Loughlin and Coenders, 2004).
The National Customer Satisfaction Index (NCSI) was developed by the Korea Productivity Center and has been in use since
2009 (NCSI, 2014). The NCSI gauges the satisfaction level of a product, and the results are synthesized into data in the following
categories: company, industry, economic sector and national levels. Specifically, the NCSI model assesses the expectancy level,
quality, recognition of value, total satisfaction, complaint rate, customer loyalty, and customer maintenance rate. The strong point
of the index is that it analyzes the cause and effect of the findings in its reports. Despite the performance of the NCSI, however,
there are some measures which could be improved in the aspect of degree of influence on consumers.
While the validity of previous CSI models has been proven and they have been widely applied in various fields, most of
the frameworks have essential weaknesses in different contexts and cross-sectors. Most of the frameworks are sector-
agnostic, applying multiple sectors indiscriminately along with macro-indicators such as market-based performance
measures, which allow for comparisons between companies, sectors, and/or nations (Park, Heo, & Rim, 2008; Shin, 2014).
 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 629

Table 1
Research on mobile quality.

System quality
Definition The users' evaluation of system performance in delivering information and meeting users' needs.
Related studies Cheong and Park (2005), DeLone and McLean (2003), Hau et al. (2012), Kuo et al. (2009), Lee, Shin,
and Lee (2008), Lin and Lu (2000), Shin (2009), Suki (2012)
Index development Islam (2012), Shin (2011), Shin (2014)

Access quality
Definition The users' ability to use service and ability to obtain information/content
Related studies Park et al. (2008), Shin (2011)
Index development Park et al. (2008), Shin (2011)

Interaction quality
Definition The ease of interacting with the other party in a telephone call or communications
Related studies Choi and Kim (2013), Negash, Ryan, and Igbaria (2003)
Index development Choi and Kim (2013), Negash et al. (2003)

Content quality (Information quality)

Definition The relevance, reliability, and timeliness of knowledge provided by the smartphone services.
Information quality is interchangeably used with content quality describing the quality
of the content
Related studies Cheong and Park (2005), DeLone and McLean (2003), Landor (2003), Lin and Lu (2000),
Mazzoni et al. (2007)
Index development DeLone and McLean (2003), Negash et al. (2003), Seddon (1997), Wang and Liao (2007),
Lai (2004), Park et al. (2008)

Service quality
Definition Assessment of how well a delivered service conforms to the customer's expectations
Related studies Aghdale and Faghani (2012), DeLone and McLean (2003), Kuo et al. (2009), Samen et al. (2013)
Index development Negash et al. (2003), DeLone and McLean (2003), Shin (2011), Shin (2014), Wang and Liao (2007),
Lai (2004), Park et al. (2008)

Experience quality (Quality of experience)

Definition Overall acceptability of an application or service, as perceived subjectively by the end user
Related studies Lauhari and Connelly (2012), Deng et al. (2010)
Index development Alben (1996)

This method bears inherent limitations to understanding customer satisfaction, especially in new technologies where
satisfaction is shaped every time customers encounter new brands, their products or their services. In this light, Morgeson,
Mithas, Keiningham, and Aksoy (2011) argued that satisfaction, as perceived by the customer, is formed through the whole
experience process from recognition and the decision to purchase to continuing use. Thus, the measurement of customer
satisfaction should not be based on a one-time impression, or a single service or product. Too often, industry tends to make
the over-generalization that customer satisfaction index can be applied in any sector. It is too easy for academic scholars to
assert that certain customer satisfaction models can be applied to wide contexts. As Turkyilmaz and Ozkan (2007) argued,
satisfaction should be flexibly and contextually viewed in response to a changing environment. It is necessary to develop a
new heuristic index in order to effectively reflect the fast-moving technological environment, particularly when it comes to
ever-changing smart services. The question of how to define quality in the smart era is the most open aspect of the new
landscape, as smartphones are comprised of different, complicated layers of technology. They are highly service-based, and
it is therefore ineffective to apply the traditional CSI models which focus on firms and their customers. Instead, the
smartphone CSI should be constructed from a consumer point of view, concentrating on services and customers.

2.2. Smartphones and quality of services

More and more smartphone applications evolve as a service, and the issue of quality has become the most important
factors in such services. Decisions regarding the adoption of mobile services revolve around consumer perceptions of their
potential use, as well as the expected benefits and costs (Bouwman, Carlsson, Walden, & Molina-Castillo, 2008). Such
decisions are dependent on the characteristics of the technology and service. In this regard, it is necessary to include the
service characteristics of mobile phones, which can affect various aspects of the quality factors. Researchers have added
quality factors into acceptance models, and various aspects of quality, such as information quality, system quality, service
quality, and content quality have been examined (Table 1).
Despite the growing market for smartphones, research on customer satisfaction and quality is still in the initial stages.2
Access to mobile broadband has changed people's priorities, along with the expectations they have of their service

2
Although there have been a few commercial studies examining purchasing behaviors for smartphone devices, studies focusing on smart services and
satisfaction have been scarce. CSI models have examined the satisfaction with certain services. Likewise, SCSI mainly addresses the satisfaction of smart
services, which are provided through smartphones. This study focuses on the services and satisfaction, while the issue of acquiring smartphones is not the
main point.
630 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

providers. Many operators around the world are taking action to improve the quality of experience for their subscribers.
Quality factors of smartphones can be broadly categorized into the areas of content quality, service quality and system
quality (Shin, 2012). One of the most important perceived quality factors for customers of smart service is the content
quality of a device, also called the applications. Smartphones offer various contents and personalized applications.
The market for smartphone applications has been increasing at an exponential rate. More smartphones can access the
Web at higher speeds, thanks to the growth of 3G and 4G data networks. In order to provide various services and
applications, smartphones are required to have a stable system. Smartphones provide users with various contents via
applications, thus content quality is a unique feature of smartphones.

3. Research model and hypotheses

To increase the level of satisfaction among customers of smart services, to reduce complaints, and thereby to enhance
customer loyalty, we identified a series of perceived quality factors that influence customer satisfaction and proposed a CSI
model, as depicted in Fig. 1. The perceived quality included system quality, service quality and content quality, as derived
from validated research from IS and IT, as well as consumer research. Customer satisfaction, customer complaints, and
customer loyalty form a causal chain of relationships. In other words, a high level of customer satisfaction decreases the
amount of customer complaints, which in turn results in enhanced customer loyalty.

3.1. Content quality

In this study, content quality is defined as the relevance, reliability, and timeliness of knowledge provided by the
smartphone services. Content quality has been interchangeably used with information quality, as numerous studies have
shown information quality to be a determinant of utility and ease of use (DeLone & McLean, 2003). Numerous research
efforts have been focused on developing content quality as a discrete determinant of system quality, and it was argued that
information quality variables were useful predictors of perceived ease of use and perceived usefulness (Lin & Lu, 2000;
Mazzoni, Castaldi, and Addeo, 2007). DeLone and McLean (2003) showed that content quality impacts both a user's
satisfaction with the system and the user's intentions to use the system. These, in turn, impact the extent to which the
system is able to yield benefits to the user.
As smart systems have become sufficiently complex to include various contents, numerous studies have used perceived
content quality in lieu of perceived information quality (Agboma & Liotta, 2010). For example, Landor (2003) showed the
mobile content quality to play a decisive role in mobile commerce by affecting customers' perceived utility. It can be said
that subscribers' perception of smartphone utility is determined by the quality of mobile content, as subscribers heavily
consume important content via their smartphones. In another study, Beyah et al. (2003) measured perceived content quality
along with other constructs for assessing web-based reference systems. They found perceived content quality to be one of
the significant determinants in the usage of the reference systems. Most recently, Cheong and Park (2005) applied perceived
content quality to the acceptance model of mobile Internet. Their factor analyses demonstrated content quality to be a valid
predictor, and it was concluded that it was a significant factor in the adoption of that technology.

Content Service System

quality quality quality
H1 H5
H3 H4

H2 H6

Perceived Perceived
utility hedonicity

H7 H8
Customer
satisfaction
H10 H9

H11
Customer Customer
Compliant loyalty

Fig. 1. SCSI casual model.

 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 631

In addition, a number of research efforts have found a significant relation of quality and hedonic-related factors, such as
enjoyment and fun. Emerging smart technologies add a hedonic aspect to the technology use experience, as they are
increasingly used not only for utilitarian, but also for hedonic purposes, in all aspects of users' personal lives (Chun et al.,
2012). As numerous studies have shown perceived hedonicity to be influenced by quality factors (such as Deng, Turner,
Gehling, and Prince, 2010), the present study hypothesizes the evaluation of hedonic attributes of the smartphone to be
direct antecedents of content quality. Content quality is related to hedonic performance (Van der Heijden, 2004), whereas
system quality can be said to be associated with utilitarian performance. Therefore, it was hypothesized that content quality
has a positive impact on perceived value, because better content results in a more enjoyable experience.

H1. Content quality has a positive effect on the utility of smartphones.

H2. Content quality has a positive effect on the hedonicity of smartphones.

3.2. System quality

System quality in this study was defined as the user evaluation of system performance when delivering information and
meeting the needs of the users (Shin, 2009). Smartphones are operated by operating systems which manage both their
hardware and software resources. Such operating systems determine the system quality, which is important in the context
of smartphones. Consumers are reluctant to use smartphones when they experience frequent delays in response, frequent
disconnection, lack of access, or poor security (Suki, 2012).
DeLone and McLean (2003) examined the relationship between system quality and user satisfaction, incorporating
perceived usefulness as a measure of user satisfaction, and perceived ease of use, perceived usefulness and information
quality as determinants of user satisfaction. That is, system quality indirectly impacts the extent to which the system is able
to deliver benefits by means of mediational relationships through the usage intentions and user satisfaction constructs.
Since the milestone work by DeLone and McLean (2003), numerous studies have followed up with system quality
research. Hau, Kim, and Kim (2012) investigated the significant effects of system quality of perceived usefulness and
perceived enjoyment in mobile services. Similarly, Lin and Lu (2000) examined information quality, response time, and
system accessibility as IS qualities. They argued that these three variables are useful predictors of perceived ease of use
and perceived usefulness. Since response time and system accessibility, along with other factors such as system reliability
and security can be understood as attributes explaining system quality, IS quality can be comprehensively identified by
system and information quality. Cheong and Park (2005) found positive causal relationships between the perceived system
quality and perceived usefulness.

H3. System quality has a positive effect on the utility of smartphones.

H4. System quality has a positive effect on the hedonicity of smartphones.

3.3. Service quality

Service quality is defined as an assessment of how well a delivered service conforms to the customer's expectations.
DeLone and McLean (2003) argued that service quality directly impacts usage intentions and user satisfaction with the
system, which, in turn, impact the benefits produced by the system. Similarly, Kuo, Wu, and Deng (2009) discussed the
relationships among service quality, perceived value, and customer satisfaction in mobile services. The service quality of
smartphones is particularly important, as most applications are provided through a form of service. Responsiveness,
reliability and assurance have been considered as critical in smartphone services. Recently, numerous studies applied these
elements in the mobile sector, as service quality has become increasingly important in mobile service (Aghdale & Faghani,
2012; Samen, Akroush, & AbuLail, 2013). For example, Negi (2009) showed a causal relationship between service quality
dimensions and overall service quality, and for the identification of service quality gaps as experienced by subscribers of the
mobile services.

H5. Service quality has a positive effect on the utility of smartphones.

H6. Service quality has a positive effect on the hedonicity of smartphones.

3.4. Customer value: utility and hedonicity

It has been widely accepted that consumers evaluate mobile services along both utilitarian and hedonic dimensions
(Hoffman, Novak, and Duhachek, 2003; Van der Heijden, 2004). While traditional IT systems are mostly work-related and
thus utilitarian in nature, hedonic dimensions have become increasingly important in the design and adoption of smart
technologies (Chun et al., 2012). The present study proposes perceived utilitarian performance and hedonic performance as
two primary evaluative dimensions of smart services. Perceived utilitarian and hedonic performances are derived from the
traditional Technology Acceptance Model of Davis (1989), and from subsequent works (Van der Heijden, 2004). From the
632 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

viewpoint of satisfaction as an evaluative outcome (Deng et al., 2010), this study suggests that evaluation of the utilitarian
and hedonic performances of smart services are direct antecedents of satisfaction.

H7. Utility (the utilitarian performance) has a positive effect on satisfaction with smartphones.

H8. Hedonicity (the hedonic performance) has a positive effect on satisfaction with smartphones.

3.5. Satisfaction, customer complaints, and loyalty

As reported by Fornell et al. (1996), customer satisfaction improves a company's performance. While customer
satisfaction has become a topic of great interest to organizations and researchers alike, its definition has been widely
debated as organizations have increasingly attempted to measure it. Customer loyalty is defined as a long-term commitment
to repurchasing, which involves both repeated patronage and a favorable attitude towards a product or service (Kim et al.,
2004). Gerpott, Rams, and Schindler (2001) argue that service loyalty results from consumer commitment to a service due to
an explicit and extensive decision-making process. Customer loyalty frequently operates as a conscious evaluation of the
price/quality ratio, as willingness to pay a premium price, or alternatively, as price indifference (Eshghi, Haughton, and Topi,
2007). Cner and Gngr (2002) argued that customer loyalty is affected by product quality, service quality, and retailer
image. In addition, quality is directly related to customer satisfaction, which leads to customer loyalty. This marketing
research is in line with the ACSI model. Customer complaints result from customer dissatisfaction. According to Zeithaml,
Parasuraman, and Malhotra (2002), dissatisfaction leads to complaint behavior, which can be expressed in voice responses,
private responses, or third-party responses. These causal relationships have been empirically confirmed by Turel and
Serenko (2006). Therefore, the following hypotheses were proposed:

H9. Customer satisfaction has a positive effect on customer loyalty to smartphones.

H10. Customer satisfaction has a negative effect on customer complaints about smartphones.

H11. Customer complaints have a negative effect on customer loyalty to smartphones.

4. Study design

4.1. Measurement development

The goal of development of the scale was to reduce measurement error. The variables included in the present study have
been frequently used in previous research. In this study, the measures of perceived usefulness and perceived enjoyment
were adapted from previous studies relating to the TAM model of Van der Heijden (2004). Satisfaction and usage measures
were derived from Roca, Chiu, and Martinez (2006). Measures of customer loyalty, complaints, and satisfaction were
adapted from Fornell et al. (1996), and measures of service quality and content quality were adapted from Fornell et al.
(1996) and Shin (2009), respectively. To address the elements of service quality, the measures of Parasuraman, Zeithaml, and
Berry's (1988) SERVQUAL were employed. The final scales used in this study consisted of 24 items, with three items per
factor. A pilot test was conducted prior to further analysis. Participants indicated their agreement with a set of statements
using a ten-point scale, as suggested by Fornell et al. (1996). As the items in the survey were adapted from previously
validated work, the content validity for these two constructs was established through literature review.
A survey questionnaire was developed based on several rounds of comments from an expert panel consisting of
professors, researchers, and industry experts. The expert interviews were conducted according to the Delphi method (Green,
Armstrong, and Graefe, 2007), a structured communication technique based on a systematic, interactive forecasting method,
involving a panel of experts. The experts answered questionnaires in two or three rounds. A pretest was carried out to
examine the test-retest reliability, and to construct reliability indices before conducting the fieldwork. Forty current users
with interest in smartphones and advanced smart services participated in two pretests at two-week intervals.

4.2. Instrument validity and reliability

Partial least squares (PLS) modeling does not provide specific goodness-of-fit criterion. Instead, there are three general
criteria for evaluating the reliability and validity of the measurement models: construct validity, convergent validity, and
discriminant validity.
This study used the two-step approach suggested by Anderson and Gerbing (1998). It first assessed the convergent
validity and reliability. Construct validity confirms the extent to which the results obtained are compatible and parallel to
theoretical or conceptual values. The question here is does the instrument tap the concept as theorized? This can be
evaluated through convergent and discriminant validity. Convergent validity was assessed through composite reliability
(CR), factor loadings and average variance extracted (AVE). As shown in Table 2, all constructs had CR values of more than
0.88, which is beyond the recommended value of 0.7. The average variance extracted (AVE) measures the variance captured
by the indicators relative to measurement error, and should be greater than 0.5 to justify using a construct. The average
 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 633

Table 2
Convergent validity and internal consistency reliability.

Variable Items Factor loadings CR Cronbach's alpha

CQ CQ1 0.873 0.93 0.867

CQ2 0.832
CQ3 0.832

SEQ SEQ1 0.713 0.92 0.899

SEQ2 0.945
SEQ3 0.916

SYQ SYQ1 0.855 0.90 0.902

SYQ2 0.824
SYQ3 0.832

PU PU1 0.695 0.88 0.887

PU2 0.764
PU3 0.808

PH PH1 0.697 0.89 0.923

PH2 0.695
PH3 0.892

CS CS1 0.877 0.94 0.910

CS2 0.835
CS3 0.832

CC CC1 0.724 0.91 0.922

CC2 0.945
CC3 0.913

CL CL1 0.852 0.90 0.890

CL2 0.822
CL3 0.829

CR (square of the summation of the factor loadings)/{(square of the summation of the factor loadings) (square of
the summation of the error variances)}.
CQ: content quality, SYQ: system quality, SEQ: service quality, PH: perceived hedonicity, PU: perceived utility,
CS: customer satisfaction, CC: customer complaint, CL: customer loyalty.

Table 3
Correlation and Q2 value.

Variable CQ SEQ SYQ PU PH CS CC CL Q2 value

CQ 0.84 0.013
SEQ 0.310n 0.86 0.056
SYQ 0.410nn 0.39 0.88 0.136
PU 0.339n 0.310n 0.426nn 0.92 0.346
PH 0.459nn 0.399n 0.415nn 0.305n 0.89 0.432
CS 0.381n 0.471nn 0.422nn 0.321n 0.422n 0.85 0.005
CC 0.416nn 0.423nn 0.301n 0.392nn 0.332n 0.58n 0.90 0.091
CL 0.311n 0.341n 0.390n 0.311n 0.48nn 0.401nn 0.310n 0.89 0.184

AVE {sum of (standardized loadings squared)}/{[sum of (standardized loadings squared)] (sum of indicator measurement errors)}.
Diagonal elements are the square roots of the AVE between constructs and their measures. Off-diagonal elements are correlations between constructs.
Discriminant validity was confirmed, as the diagonal values were all larger than the off-diagonal elements in the same row and column.
n
p o 0.05.
nn
p o 0.01.

variances extracted were in the range of 0.840.92. With the calculated values, it was shown that the use of convergent
validity has allowed the use of the literature criteria and the constructs can be used for data collection.
The discriminant validity of the measures was assessed by examining the correlations between the measures of
potentially overlapping constructs. Items should load more strongly on their own constructs in the model, and the average
variance shared between each construct and its measures should be greater than the variance shared between the construct
and other constructs. As shown in Table 3, the squared correlations for each construct were less than the average variance
extracted from the indicators measuring that construct, indicating adequate discriminant validity. Overall, the measurement
model demonstrated adequate convergent validity and discriminant validity.
Finally even one's model consists of latent variables with high levels of internal consistency, in order to be consistent
with the causal-predictive goal of PLS, greater focus should be paid on the predictive relevance of a model. The extent to
which this prediction exercise is successful can be measured by the Q2 statistic. This study picked an omission distance (D) 8.
Q2 40 indicates predictive relevance, and Table 3 shows the Q2 values.
634 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

Table 4
Characteristics of respondents (N 485).

Age N User experience N Education N Gender N

Under 20 85 Less than 6 months 16 High school or less 92 Female 246

2130 194 16 months 98 College 311 Male 239
3140 180 712 months 120 Graduate school or higher 72 Chi-square 9.83
Over 41 26 Over 1 years 251 Others 10 p-Value 1.4

The reliability and validity of PLS was reconfirmed in the present study with Chins method (1998), in which a two-step
process was proposed: (1) assessment of the outer model, and (2) assessment of the inner model.

4.3. Survey administration

The final survey was administered online and via smartphones between January and May 2014.
To acquire good quality data, the survey was administered by a professional marketing firm specializing in survey development,
data collection, analysis, and reporting. The company possesses a robust panel of data related to various customers. The specific
topic of smartphones made it necessary to use a professional marketing firm to ensure the quality of the data. Over a five-month
period, a total of 822 users viewed the survey and 506 questionnaires were submitted. Of the submitted questionnaires, 21 were
excluded because of incomplete answers, leaving 485 usable responses (Table 4). In addition to the online survey, mobile calls were
placed to smartphone users. Researchers placed calls to 300 smartphone users, and 89 usable responses were collected. Although
this might present an issue of data aggregation due to the different methods used, it was necessary to perform the telephone
interviews given that the nature of the study was to investigate service experiences by smartphone users. In order to control for
country-specific effects, only residents of Korea were surveyed. Surveying this sample population yielded statistical results that are
generalizable to the entire Korean user population since all respondents used smartphones for at least three months, which is
sufficient to establish reliable perceptions and opinions regarding the service. In addition, this sample is an actual representation of
the entire Korean user population, based on comparison of the demographic data. A Chi-square test for goodness of fit revealed that
market shares in our sample and in the Korean market did not differ significantly at the 1% level (Chi-square test statistic of 9.83
with three degrees of freedom, for which the p-value was 1.4%).

5. Results

As Fornell (1992) stated that PLS is a powerful estimation method for CSI, PLS regression was used to estimate the model.
PLS is a method of causal modeling, principal component analysis and multiple regressions. In the ASCI model, it extracts
principal components from different measurement variables of latent variables by using the data in the regression model to
find relationships between independent (exogenous) and dependent (endogenous) variables.

5.1. Mean and standard deviation

Through data collection and data compilation from the survey, the mean value and standard deviation were summarized
(Table 5). The mean value shows the average level of customer evaluations. The results indicated customer expectations of
the quality to be high: they liked good quality service, content, and systems. At the same time, the mean value of complaints
shows high value that customers do not like some aspects of smartphones. The standard deviation (SD) measures the
dispersion of a data set. Low standard deviation indicates all of the data points are very close to mean value, while high
standard deviation indicates that the data are spread out over a large range of values. All values of SD obtained in this survey
were almost equal to 2, which indicates that there were only slight differences in the valuations.

5.2. Coefficient calculation

Using Simca-p, the coefficients for each latent variable and R2 of the data from the survey are shown in Fig. 2 (ranged
from 0.293 to 0.419). With PLS, the structural model (inner model) is used to examine the relations among variables. The
equation is as follows:

j oj ji 1;i a j ji i j 8 j 1; ; J
j constant term; _ji regression coefficient; _j residual term
Relations of manifest variables and their respective latent variable: Each latent variable is indirectly describable by a set of
manifest variables X and each manifest variable relates to its respective latent variable through a simple regression.
Xjh jh0 jh jh jh
jh has a standard deviation of one.
 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 635

Table 5
Mean and standard deviation.

Variable Items Mean value Average of mean SD Std. error mean Skewness Kurtosis

nn
CQ CQ1 6.5890 6.5013 2.4432 0.14382  0.965 1.093nn
CQ2 6.4082 1.9825  0.617nn 0.691
CQ3 6.5068 2.2791  0.870nn 1.143nn

SEQ SEQ1 6.4945 6.4945 2.0324 0.24933  1.171nn 1.101nn

SEQ2 6.2329 2.1012  0.504nn 0.096
SEQ3 6.3753 2.1792  1.142nn 0.356

SYQ SYQ1 6.3971 6.4019 2.6805 0.24358  1.043nn 0.305

SYQ2 6.3699 2.0582  0.965nn 1.122nn
SYQ3 6.4388 2.4524  1.123nn 0.169

PU PU1 5.9315 6.4096 2.5447 0.20334  0.631nn 0.314

PU2 6.5028 2.3230  1.001n 2.846nn
PU3 6.7945 2.2874  1.478nn 1.877nn

PH PH1 6.2319 5.6596 2.0581 0.16784  1.345nn 2.260nn

PH2 5.3972 2.1015  1.618nn 0.673
PH3 5.3499 2.2655  0.618nn 7.212nn

CS CS1 6.4384 6.4021 2.2114 0.16882  0.871nn 0.070

CS2 6.2329 2.4922  1.170nn 0.585nn
CS3 6.5351 2.4421  0.508nn 1.472nn

CC CC1 6.2973 6.2297 2.4525 0.24423  1.145nn 1.001

CC2 6.1599 2.4235  1.046nn 0.014
CC3 6.2319 2.0904  0.965nn 1.193nn

CL CL1 6.5753 5.7573 2.9241 0.16352  1.124nn 1.232nn

CL2 5.3374 2.2022  0.632nn 0.029
CL3 5.3594 2.0022  0.739nn 0.029

The parameter of standard deviation is each item of variable.

nn
Skewness or kurtosis significant at the 1% level.
n
Skewness or kurtosis significant at the 5% level.

0.49***

0.48***
Content quality
Customer
0.37**
Complaint
Perceived
R2 =0.314
utility R2=0.391 0.40**
-0.39**
0.26*

0.47*** Customer
-0.44***
Service quality Satisfaction
R2=0.419
0.23* 0.40**
Perceived 0.24*
0.36 hedonicity
** R2 =0.293 Customer
Loyalty
System quality R2=0.332

0.40**
0.42**

* p<0.05; ** p<0.01, *** p<0.001

Fig. 2. Results of the smartphone CSI model.

Relations among latent variables: The model is a set composed by linear equations among latent variables. The overall
form of the equations is as follows:

j jo ji :j j
Considering the above formula, the linear equations among latent variables of the model are in Table 6.
636 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

Table 6
Model variables, parameters and equations.

Endogenous variable Exogenous variable Equations

PU CQ, SEQ, SYQ PU C 1CQ 2SEQ 3SYQ

PH CQ, SEQ, SYQ PH C 4CQ 5 SEQ 6SYQ
CS PU, PH CS C 7PU 8PH
CL CS, CC CL C 9CS 10CC
CC CS CC C 11CS

Table 7
PLS standardized path coefficients.

Hypothesis Path coefficient () p-Values T-statistics Support

nnn
H1: CQ-PU 0.48 0.008 4.221 Yes
H2: CQ-PH 0.37 0.040 4.093nn Yes
H3: SYQ-PU 0.26 0.032 2.798nn Yes
H4: SYQ-PH 0.23 0.092 2.973n Yes
H5: SEQ-PU 0.36 0.045 3.201nn Yes
H6: SEQ-PH 0.40 0.039 3.394nn Yes
H7: PU-CS 0.40 0.041 5.491nn Yes
H8: PH-CS 0.24 0.039 3.847nn Yes
H9: CS-CL 0.40 0.091 3.491n Yes
H10: CS-CC  0.39 0.004  3.317nn Yes
H11: CC-CL  0.44 0.001  4.312nnn Yes

S.E. is an estimate of the standard error of the covariance.

All is circumflex (\widehat) beta.
n
p o 0.05.
nn
po 0.01.
nnn
p o0.001.

Having identified the above equations, all the coefficients and parameters can be estimated using PLS. The hypothesized
causal paths were estimated, and all 11 hypotheses were found to be supported (Table 7 and Fig. 2). The results supported
the proposed model well, confirming the key roles of quality, perceived value and satisfaction. Customer expectations, and
the results obtained for all latent variables are satisfactory with great values of Q2 and R2. With regard to the total effects of
satisfaction on loyalty (the direct effect of satisfaction on loyalty plus the indirect effects of satisfaction on loyalty through
complaints), the results were significantly positive. The results highlighted the significant role of perceived value in
determining the satisfaction of users. With regard to the impact of quality on perceived value, all path coefficients were
significantly positive. The high predictive ability of quality on perceived value was evident. Perceived utility and hedonicity
were shown to have significant direct effects on satisfaction (H7&H8). Compared with perceived value, the impact of quality
on satisfaction was generally greater. This is in line with the notion that even though value may be more important in
consumers' initial purchase decisions, quality still plays an important role throughout the usage process (Fornell et al., 1996).
It also supports the findings in ACSI that consumer satisfaction is more quality-driven than value-driven. Given the high
impact of quality, additional paths from quality to satisfaction were tested. With the three newly added paths (content
quality to satisfaction, service quality to satisfaction, and system quality to satisfaction), the overall fit of the model was
improved, with improvement in most of the goodness of fit indices. All three paths showed significant coefficients, implying
the key role played by quality as well as various dimensions.
Average redundancy of the model was estimated to be 0.19 that is equal to the output of some ECSI models. In the model
R2 value ranged from 0.391 to 0.419 and its mean is 0.3498. Goodness of fit index of the model can be calculated as follows:

GIF communality  R2 0:501  0:3498 0:711  0:591 0:420

Since Q2 of all endogenous latent variables are positive (Table 4), it can be concluded that all the latent variables have
been selected correctly. It can be said the model is valid and the paths are reasonable.
Identifying the given relations is a complex procedure in which all latent variables are involved. The equation used is as
follows:
CS 0:291  CQ 0:1681  SEQ 0:187  SYQ 0:091  PU 0:094  PH Constant
The R2 of customer loyalty explained about 41% of the variance, according to two exogenous constructs in the model.
As both R2 values (utility and hedonicity) were fairly high, the model provides possible advancement in terms of theory and
practice. Thus, further tests are necessary to uncover possible underlying effects. Comparison of the direct, indirect, and total
effects among variables (Table 8) showed content quality to have the strongest effect on satisfaction. Content quality showed
not only the greatest direct effects, but also indirect effects on satisfaction. The effects of quality variables on satisfaction are
mediated through perceived values.
 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 637

Table 8
Standardized direct, indirect, and total effects of quality variables on satisfaction.

Ind. variable Direct effects Indirect effects Gross effects

CQ 0.49 0.791
CQ-PU-CS 0.530n0.371 0.21
CQ-PH-CS 0.381n0.2400.09

SEQ 0.47 0.692

SEQ-PU-CS 0.390n0.371 0.13
SEQ-PH-CS 0.390n0.2400.09

SYQ 0.42 0.56

SYQ-PU-CS 0.210n0.390 0.09
SYQ-PH-CS 0.221n0.220 0.05

n
Gross effects Direct effects Indirect effects.

Table 9
ACSI for wireless services by company.
Source: http://www.theaci.org/index, http://www.engadget.com/2013/07/31/acsi-smartphone-satisfaction-2013.

04 05 06 07 08 09 10 11 12 13 14 Previous year % change First year % change

Apple 83 81 79  2.5  4.8

Motorola Mobility (Google) 70 70 71 72 74 72 76 77 73 77 77 0.0 10.0
Cellular Telephones 69 69 70 70 71 72 76 75 74 76 78 2.6 13.0
Samsung Electronics 73 71 73 70 71 70 76 74 71 76 81 6.6 11.0
Nokia 72 70 69 72 74 74 76 73 75 76 77 1.3 6.9
All others 64 66 68 70 70 73 77 74 73 76 71  6.6 10.9
HTC 75 72 75 4.2 0.0
LG Electronics 75 71 73 2.8  2.7
BlackBerry 69 69 74 7.2 7.2
Kyocera 70 N/A N/A

Smartphone customer satisfaction 2013

Model Manufacturer ACSI score

Galaxy S III Samsung 84

Galaxy Note II Samsung 84
iPhone 5 Apple 82
iPhone 4S Apple 82
iPhone 4 Apple 81
Droid Razr Maxx Motorola Mobility 80
Galaxy S II Samsung 78

6. Discussion

Although numerous CSI models have been developed and used, they are accompanied with a number of weaknesses and
limitations. For example, most index models have only been applied within the national level, and the models have been
applied across industry boundaries in an overarching manner. Thus, mobile-specific CSI models are limited. Furthermore,
there is not yet a standardized measure for advanced mobile services such as smartphones, because these services are
relatively new. While ACSI has surveyed smartphone satisfaction (Table 9), it has been mainly used for comparison purposes
among different smartphone brands. The most important weakness of the current ACSI method is that the index reflects
providers' overall perspective rather than user experience. That is, the current ACSI with regard to smartphones focuses on
non-user issues, such as the technical features, commercial advantages, and product capabilities, instead of on how
customers really experience and how they feel about the product. Thus, the satisfaction index has been used for marketing
and advertising purposes, and has not been properly researched academically until now. There has been a gap in prior
research on customer satisfaction and the use of the ASCI with respect to smartphones and related services. A review of the
literature confirmed that none of the studies citing Fornell et al. (1996) or ACSI focused on smartphone services or customer
perspectives on smartphones.
In addition, there has been a gap in applying ACSI in different countries. As ACSI has been the only uniform and cross-
industry measure of customer satisfaction, it has been problematic when applied to complicated industries such as
telecommunications and mobile services. The table below compares the ACSI for Korean smartphone services with the ACSIs
of other countries.
To address this gap, the present study tested the proposed SCSI model in order to explain how individuals develop
behavioral intentions toward the use of smart mobile services. The results increased our understanding of user perception
638 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

and intentions toward smart mobile services, and clarified implications for the development of effective smart services and
applications. The results of the measurement and structural model tests also supported our proposed research model.
Overall, it was shown that the models demonstrated good predictive powers and explained the index model of smart mobile
services.
The results of this study showed that the modified CSI is accurately able to describe the service perceptions and customer
satisfactions of smartphone users. Consistent with prior research on mobile service satisfaction, the usage of smart services
was found to be determined by perceptions of value and quality, which lead to high customer satisfaction. High customer
satisfaction, in turn, affects customer loyalty. Highly satisfied customers demonstrate a high likelihood of repurchase, as well
as higher tolerance to price increases by their providers, or price decreases by competitors. Furthermore, the findings
confirmed the importance of perceived quality in service, content and system with regard to smartphone services, and also
demonstrated these qualities to contribute to the perceived values of utility and hedonicity. When compared with the initial
model in Fig. 1, certain slight differences appeared, i.e. quality factors also had direct effects on satisfaction. Thus, three new
paths in the model were established through this study. These findings indicate the perceptions and expectations that
customers have of smart features. The potential success of smartphones may be linked to the provision of diverse, useful
applications, and enjoyable services. As smartphones increasingly provide 4 G, smart services are becoming more and more
ubiquitous, and pervasively accessible. Consumers desire seamless interconnections among all kinds of devices and
networks. As indicated in many studies, including the present one, smartphone devices represent digital connections to
friends, family, and resources. This trend will be accentuated and will become popular worldwide, as 4 G continues to
increase in availability. In light of this, smartphone service providers should increase the perception of availability for users.

7. Implications for theory and practice

7.1. Theoretical implications

The empirical findings validate that our smartphone service model is a worthwhile extension of the CSI in the context of
smart services, as the proposed SCSI model allows powerful measurement of customer satisfaction and behavioral intention
with regard to smartphone services. The conventional CSI approaches often miss the measurement of all relevant factors, or
measure them using older, less precise measurement approaches. These limitations have prevented traditional customer
satisfaction designs from producing more powerful new generation outcomes. In light of this limitation, a primary
contribution of this study is that it theoretically conceptualized the notion of quality in smartphones, and established a
relationship between quality and the perceived values held by customers. Previous studies on mobile satisfaction have often
been criticized for their lack of context-specific understanding (Venkatesh & Agarwal, 2006), which leads to satisfaction
models with low explanatory power. Thus, by incorporating smartphone-specific factors into a satisfaction model, how the
factors influence consumers' satisfaction, and how consumers' satisfaction, in turn, affects loyalty can be better explained.
Although studies in mobile literature are plentiful, thus far only few studies have researched the measurement and scale
of smartphone quality. Extensive research on mobile users has previously examined the impacts of perceived quality on
customer intentions for mobile services. However, these studies have neglected to adequately define quality. The concept of
quality in smartphones may refer to different experiences and how consumers perceive them, and it is necessary to
approach quality from different dimensions. The effect of mobile evolution (e.g. 3G, 4G, and 5G) has greatly improved the
number of services offered, as well as the perceived quality. Reflecting the technology evolution, it is important to treat
perceived quality in a more sophisticated way. The present study categorized quality into three dimensions and confirmed
not only the indirect effect of the three dimensions on satisfaction by directly influencing the values, but also the direct
effect of quality factors on satisfaction. In addition, it was verified that the division of smartphone values into utilitarian and
hedonic dimensions is very useful and valid. Previous ACSI models treated perceived value as a single factor, neglecting to
provide an adequate definition of value. Thus, this study offers a heuristic link to TAM, which argues that the usefulness and
ease of use are two fundamental factors in service adoption. This could be a modest but heuristic step toward addressing the
limitations of TAM, and providing a more robust framework for smart technologies.

7.2. Practical implications

Practical implications for the industry can be drawn from these findings in terms of strategies and new models for
advanced smart services. This study measured the level of customer satisfaction with regards to emerging smart services
and identified priorities in quality improvement strategies. Customer satisfaction has a direct effect on the behavioral
intentions of customers. This is even truer for smartphone services. The industry should develop a training program for
customer service representatives in order to present a concerted effort to increase customer loyalty, and also to reduce
customer complaints. Moreover, the industry should formulate a competitive strategy based on the SCSI model in order to
retain current customers, and to enhance the management of customer relationships. The organizations ability to attract
and retain customers and enhance customer relationships over time has become a key aspect in the smartphone industry.
Every interaction a customer has with the products and services of a company is a reflection on quality.
As the focus of this study was smartphones, it has potential implications for the design and implementation of future
smart services. Consistent with prior research in technology acceptance, the two constructs of perceived usefulness and
 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 639

perceived hedonicity continue to play major and important roles in user perceptions, and also on the follow-up behavior
toward products and services. From a consumer perspective, this represents useful service operation and more enjoyable
service through a ubiquitous seamless network, offering significantly attractive features to consumers of smart services.
The results of this study can be used by the smartphone industry to better understand customers and markets, and to
plan their preferred strategies on what quality factors to emphasize. The findings also provide useful insights for the
development of effective marketing strategies to meet customer demands, and further retain and expand a customer base.
From the finding that quality dimensions impact usage behavior through intention, it is suggested that carriers should
establish consumer perception of quality. This can best be achieved by ensuring that the carrier's services are conducted in
accordance with user expectations, namely, that their contents are of high quality, services are reliable, and that promises
and commitments are kept.
In conclusion, the SCSI model provides important information about the purchasing decisions of customers, and leads to
improvements in the quality of goods and services. The smartphone industry can use the SCSI model to formulate a
competitive strategy to retain current customers and to enhance customer relationship management. As users accept smart
technologies as new means of communication, entertainment and functional tools, smart mobile services will become
important applications for the next generation of mobile platform applications. However, in order to become popular, smart
services must overcome several challenges, of which user satisfaction is probably the most crucial. Service developers need
a better understanding of customer perceptions concerning the level of quality, and the impact of satisfaction on customer
loyalty. Indeed, the smart industry landscape changes faster than more traditional industries, and the rules can be rewritten
by new players or new technologies. Disruption will always be a part of smart services, but innovation will likely keep the
sector near the top in customer satisfaction.

8. Limitations and future studies

While the findings of this study are valid and valuable, the results should be approached with caution for several reasons.
First and most importantly, it is not sure whether the sample represents the whole population of Korean smartphone users.
Although a relatively large sample (485) was collected and the number is considered to be large enough to generalize, the
question still remains whether the sample was perfectly or objectively representative, or allowed the in-depth capture of
the best picture of the whole population. Future studies should continue to use a systematic random sampling method to
reduce bias.
Second, the findings reflect only limited aspects of user experiences with smartphones. Because smart mobile services
are not yet a mainstream phenomenon, even in the Korean mobile market, this research is exploratory in nature and was
limited by the fact that the findings cannot be generalized to the overall population of smartphone users. Furthermore,
because current smartphone providers are continually updating the contents and functions of their value-added services, it
is difficult to follow the trail of consumer use outcomes. Therefore, this study used only a one point-in-time method to
explore the use experiences of consumers, to allow easier data collection. A longitudinal study should have been used to
monitor the evolution of customer behavior over changing services. Instead, the circumstances of this study led to limited
generalizability. Since this study focused on smartphone services, it is difficult to generalize the findings to other smart or
mobile contexts. Given that the population of smartphone users is currently greater than several million, the sample size of
485 may be disproportionate to the whole population of mobile users. In addition, the research model is only valid to the
Korean smartphone market due to restrictions in data collection. Generalization of the application scope of the model would
require a global-area data collection process for more thorough validation.
Smartphone consumers evaluate their own needs, and verify the relative interests, compatibility, availability, and other
features of smartphones, thus making a decision about their adoption and continued usage. The fit between personal needs
and the features of products predicts the adoption behaviors of technological innovations. Thus, as argued by Rice and Katz
(2008), future research might include the needs of consumers and the characteristics of innovative products and the content
of media messages, as well as other, more specific dimensions to be tested in order to conduct a more thorough
investigation of the proposed model. A longitudinal study measuring prior expectations and post-use evaluations of these
constructs would increase the theoretical validity of the model.
With such limitations, the extent to which this study reflects the actual phenomena of ongoing smartphone services
must be considered. The purpose of this study was to develop a smart-service CSI index. Due to the limitations of
budget and time, however, we could not conduct a comprehensive study on global perspectives. Future studies will be
able to sample a larger and more diverse cross-section of the population, using stratified sampling or a quota sampling
method to ensure a certain distribution of demographic variables. Generalized application of the extended model would
require a global data collection process for more thorough validation. The first step of future studies should be to test
the SCSI model in other countries, after which a globally accepted universal model can be developed. The development
of a global system of customer satisfaction measurement, founded on common methodology, will allow comprehensive
cross-national satisfaction benchmarking, something that will grow more significant as smart innovation drastically
advances.
In this study, smart services and smart devices were not clearly distinguished. Future studies may consider measuring
customer satisfaction from services and devices differently, as the satisfaction with smart devices is more related to
functionality, device price, the quality of operating systems, and brand image. Per Zeithaml's (1988) argument, consumer
640 D.-H. Shin / Telecommunications Policy 39 (2015) 627641

perceptions of what is received and what is given determine consumers' overall assessment of the utility of a product. Some
of the variables are associated with mobile applications, and others are related to smartphones. For example, content quality
and system quality are related to devices, whereas service quality is related to service itself. Satisfaction may include both
service satisfaction and device satisfaction.

Acknowledgement

This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-
2014S1A5B1014964).

Appendix

Latent variables Manifest variables Source

Content quality CQ1: I think that smartphones provide various information and services. Cheong and Park (2005), Fornell et al. (1996),
CQ2: I think that the services and information I can get from a Shin (2014)
smartphone are valuable.
CQ3: Smartphones provide the information and services that I need.
CQ4: I enjoy content from smartphones. Dropped item

System quality SYQ1: I think that the smartphone provides very reliable service. DeLone and McLean (2003), Fornell et al. (1996)
SYQ2: I think that the speed of the smartphone is sufficiently fast.
SYQ3: I think that smartphones are secure to use.
SYQ4: Overall, smartphone system is reliable. Dropped item

Service quality SEQ1: The response time to a customer request is reasonably fast. DeLone and McLean (2003), Parasuraman, Zeithaml,
SEQ2: Solutions provided to customer requests are pertinent and Berry (1988)
and helpful.
SEQ3: Satisfactory after-sales services.
SEQ4: Customer service from smartphone is professional. Dropped item

Perceived PH1: I evaluate the smartphone service as exciting. Chun et al. (2012), Van der Heijden (2004)
hedonicity PH2: I evaluate the smartphone service as delightful.
PH3: I evaluate the smartphone service as playful.

Perceived utility PU1: I evaluate the smartphone service as useful. Chun et al. (2012), Van der Heijden (2004)
PU2: I evaluate the smartphone service as practical
PU3: I evaluate the smartphone service as functional.

Customer CS1: Overall satisfaction with the smartphone service. Chun et al. (2012), Turel and Serenko (2006)
satisfaction CS2: Degree of expectancy disconfirmation.
CS3: Performance versus ideal smartphone services.

Customer CC1: I complain about the service to customer services. Fornell et al. (1996), Shin (2009)
complaints CC2: I have many complaints that I would like to report.
CC3: I think degree of complaints about the smartphone services
is high.

Customer CL1: I intend to reuse smartphone services. Fornell et al. (1996), Kim et al. (2004)
loyalty CL2: I intend to recommend others to use smartphones.
CL3: I intend to continue using smartphones in the future.

References

ACSI. (2010). The America customer satisfaction index of wireless telecommunications services. Research Report. Ann Arbor, MI: Chaat Butsunturn.
Agboma, F., & Liotta, A. (2010). Quality of experience management in mobile content delivery systems. Telecommunication Systems, 49(1), 8598.
Aghdale, S., & Faghani, F. (2012). Mobile banking service quality and customer satisfaction. International Journal of Management Business Research, 2(4),
351361.
Alben, L. (1996). Quality of experience: Defining the criteria for effective interaction design. Magazine Interaction, 3(3), 1115.
Anderson, J. C., & Gerbing, D. W. (1998). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin,
103(3), 411423.
Beyah, G., Xu, P., Woo, H., Mohan, K., & Straub, D. (2003). Development of an instrument to study the use of recommendation systems. Proceedings of AMCIS,
2003.
Bouwman, H., Carlsson, C., Walden, P., & Molina-Castillo, F. J. (2008). Trends in mobile services in Finland 20042006. Info, 10(2), 7593.
Cheong, J., & Park, M. (2005). Mobile internet acceptance in Korea. Internet Research, 15(2), 125140.
Chin, W. W. (1998). The partial least squares approach to structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research
(pp. 295358). Mahwah, NJ: Lawrence Erlbaum Associates.
Choi, B., & Kim, H. (2013). The impact of outcome quality, interaction quality, and peertopeer quality on customer satisfaction with a hospital service.
Managing Service Quality, 23(3), 188204.
 D.-H. Shin / Telecommunications Policy 39 (2015) 627641 641

Chun, H., Lee, H., & Kim, D. (2012). The integrated model of smartphone adoption: Hedonic and utilitarian value perceptions of smartphones among Korean
college students. CyberPsychology, Behavior & Social Networking, 15(9), 473479.
Cner, A., & Gngr, M. (2002). Factors affecting customer loyalty in the competitive Turkish metropolitan retail markets. Journal of American Academy of
Business, 2(1), 189195.
Davis, F. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319340.
DeLone, W., & McLean, E. (2003). The DeLone and McLean model of information systems success. Journal of Management Information Systems, 19(4), 930.
Deng, L., Turner, D., Gehling, R., & Prince, B. (2010). User experience, satisfaction, and continual usage intention of IT. European Journal of Information
Systems, 19, 6075.
ECSI Technical Committee (1999). European Customer Satisfaction Index: Foundation and structure for harmonized national pilot projects. Report Prepared for
the ECSI Steering Committee, October 1999.
Eshghi, A., Haughton, D., & Topi, H. (2007). Determinants of customer loyalty in the wireless telecom industry. Telecommunications Policy, 31, 93106.
Fornell, C. (1992). A national customer satisfaction barometer: The Swedish experience. Journal of Marketing, 56, 621.
Fornell, C., Johnson, M., Anderson, E. W., Cha, J., & Everitt, B. (1996). The American customer satisfaction index: Nature, purpose, and findings. Journal of
Marketing, 60, 718.
Fornell, C., & Larcker, V. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research,
18, 3950.
German Customer Barometer. (1995). In Anton Meyer, & Frank Dornach (Eds.), Quality and satisfaction. Yearbook of customer satisfaction in Germany 1995.
German Marketing Association and German Post, 1995.
Gerpott, T., Rams, W., & Schindler, A. (2001). Customer retention, loyalty, and satisfaction in the German mobile cellular telecom market. Telecommunica-
tions Policy, 25, 249269.
Green, K. C., Armstrong, J. S., & Graefe, A. (2007). Methods to elicit forecasts from groups: Delphi and prediction markets compared. The International Journal
of Applied Forecasting, 8, 1720.
Hau, Y. S., Kim, G., & Kim, B. (2012). Antecedents of user satisfaction in the context of mobile data services. International Journal of Mobile Communications,
10(6), 617636.
Hoffman, L. D., Novak, T. P., & Duhachek, A. (2003). The influence of goal directed and experiential activities on on-line flow experiences. Journal of
Consumer Psychology, 13(2), 316.
Islam, A. K. (2012). The role of perceived system quality as educators' motivation to continue e-learning system use. AIS Transactions on HumanComputer
Interaction, 4(1), 2543.
Kim, M., Park, M., & Jeong, D. (2004). The effects of customer satisfaction and switching barrier on customer loyalty in Korean mobile telecommunication
services. Telecommunications Policy, 28, 145159.
Kuo, Y., Wu, C., & Deng, W. (2009). The relationships among service quality, perceived value, customer satisfaction, and post-purchase intention in mobile
value-added services. Computers in Human Behavior, 25, 887896.
Lai, T. (2004). Service quality and perceived values impact on satisfaction, intention, and usage of short message service. Information Systems Frontiers, 6(4),
353368.
Landor, P. (2003). Understanding the foundation of mobile content quality: A presentation of a new research field. Proceedings of the 36th Hawaii
international conference on system sciences, 2003, Big Island, Hawaii.
Lauhari, K., & Connelly, K. (2012). Toward total quality of experience. IEEE Communications Magazine, 50(4), 5865.
Lee, S., Shin, B., & Lee, H. G. (2008). Investigation of factors that have impacts on usage increase and decrease of mobile data services. Paper presented at the
41st annual Hawaii international conference on system sciences (HICSS08), January 710, 2008.
Lin, J., & Lu, H. (2000). Towards an understanding of the behavioral intention to use a web site. International Journal of Information Management, 20, 197208.
Mazzoni, C., Castaldi, L., & Addeo, F. (2007). Consumer behavior in the Italian mobile telecommunication market. Telecommunications Policy, 31, 632647.
Morgeson, F. V., III, Mithas, S., Keiningham, T., & Aksoy, L. (2011). An investigation of the cross-national determinants of customer satisfaction. Journal of the
Academy of Marketing Science, 39(2), 198215.
NCSI (2014). National Customer Satisfaction Index. A report published by Korea Productivity Center. http://www.kpc.or.kr/eng/customer/customer.asp?
c_menu=4s_menu=4_3.
Negash, S., Ryan, T., & Igbaria, M. (2003). Quality and effectiveness in web-based customer support systems. Information & Management, 40(8), 757768.
Negi, R. (2009). User's perceived service quality of mobile communications. International Journal of Quality & Reliability Management, 26(7), 699711.
O'Loughlin, C., & Coenders, G. (2004). Estimation of the European Customer Satisfaction Index: Maximum likelihood versus partial least squares. Total
Quality Management and Business Excellence, 15(9), 12311255.
Parasuraman, A., Zeithaml, V., & Berry, L. (1988). SERVQUAL. Journal of Retailing, 64(1), 1240.
Park, Y., Heo, P., & Rim, M. (2008). Measurement of a customer satisfaction index for improvement of mobile RFID services in Korea. ETRI Journal, 30(5),
634643.
Rice, R., & Katz, J. E. (2008). Assessing new cell phone text and video services. Telecommunications Policy, 32, 455467.
Roca, J. C., Chiu, C., & Martinez, F. J. (2006). Understanding e-learning continuance intention. International Journal of HumanComputer Studies, 64(8),
683696.
Samen, A., Akroush, M., & AbuLail, B. (2013). Mobile SERVQUAL: A comparative analysis of customers' and managers' perceptions. International Journal of
Quality & Reliability Management, 30(4), 403425.
Seddon, P. B. (1997). A respecification and extension of the Delone and Mclean model of success. Information Systems Research, 8(3), 240253.
Shin, D. (2009). Determinants of customer acceptance of multi-service network. Information and Management, 46(1), 1622.
Shin, D. (2011). The influence of perceived characteristics of innovating on 4G mobile adoption. International Journal of Mobile Communications, 9(3),
261279.
Shin, D. (2012). What makes consumers use VoIP over mobile phones?. Telecommunications Policy, 36(4), 311323.
Shin, D. (2014). Measuring the quality of smartphones: Development of a customer satisfaction index for smart devices. International Journal of Mobile
Communications, 12(4), 311327.
Suki, N. (2012). Correlations of perceived flow, perceived system quality, perceived information quality, and perceived user trust on mobile social
networking service users' loyalty. Journal of Information Technology Research, 5(2), 114.
Turel, O., & Serenko, A. (2006). Satisfaction with mobile services in Canada. Telecommunications Policy, 30(5), 314331.
Turkyilmaz, A., & Ozkan, C. (2007). Development of a customer satisfaction index model. Industrial Management & Data Systems, 107(5), 672687.
Van der Heijden, H. (2004). User acceptance of hedonic information systems. MIS Quarterly, 28(4), 695704.
Venkatesh, V., & Agarwal, R. (2006). Turning visitors into customers. Management Science, 52, 367382.
Wang, Y. S., & Liao, Y. (2007). The conceptualization and measurement of m-commerce user satisfaction. Computers in Human Behavior, 23(1), 381398.
Zeithaml, V. A. (1988). Consumer perceptions of price, quality and value: A means-end model and synthesis of evidence. Journal of Marketing, 52(3), 222.
Zeithaml, V. A., Parasuraman, A., & Malhotra, A. (2002). Service quality delivery through websites. Journal of Academy of Marketing Science, 30(4), 362375.