The relationship between website functionality/usability and online customer satisfaction and the moderating role of product involvement in e-retailing

The Relationship between Website Functionality/Usability and

Online Customer Satisfaction and the Moderating Role of

Product Involvement in E-retailing

Bingcong Zhang 10979972

Date of Submission: 24th _{March, 2017}

Supervisor: Drs. ing. A.C.J. Meulemans

MSc. in Business Administration – Marketing Track

1 Statement of Originality

This document is written by Bingcong Zhang who declares to take full responsibility

for the contents of this document.

I declare that the text and the work presented in this document is original and that no

sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of

2

Acknowledgment

With this thesis my academic study in the Netherlands comes to an end, and I would

like to show my gratitude to the people who helped me in this process.

Firstly I would like to thank my supervisor Drs. Meulemans for his support, guidance

and patience throughout the whole process. He helps me with all the questions and doubts,

and support the thesis with great patience.

Secondly I would like to thank my parents and family for their endless love and support

throughout all my study years. Knowing that they are always by my side gives me great

courage and confidence to finish my study.

Lastly I am grateful to all the participants of my survey, they gave me good insights and

3

Table of Contents

2.1.1 The Expectation (dis)Confirmation Theory ... 12

2.1.2 Online Customer Satisfaction ... 15

2.2 Website Functionality and Website Usability ... 19

2.3 Product Involvement ... 26

2.4 Conceptual Framework and Hypotheses ... 33

2.4.1 Conceptual Framework ... 33 2.4.2 Hypothesis Development ... 34 3. Methodology ... 36 3.1 Method ... 36 3.2 Pretest ... 37 3.3 Main Survey ... 39 3.4 Measurement of Variables ... 39 4. Results ... 40 4.1 Descriptive Statistics... 41 4.2 Data Preparation ... 42 4.2.1 Recoding Items ... 42 4.2.2 Reliability Check ... 43 4.2.3 Correlation Analysis ... 44 4.2.4 Multicollinearity Test ... 45 4.3 Hypotheses Testing ... 46 4.3.1 Regression Analysis ... 46 4.3.2 Moderation Analysis ... 47 4.4 Results Summary ... 52 5. Discussion ... 52 6. Conclusion ... 56 7. References ... 57 8. Appendices ... 62

4

Abstract

Purpose – The purpose of study is to investigate the relationship between e-retailing website

functionality and website usability and customer satisfaction in the online context. In addition,

high and low involvement product will be used as the moderators of this relationship.

Design/methodology/approach – Firstly a pre-test is conducted in order to identify low/high

involvement products for respondents. Then the main questionnaire is carried out. In total there

are 160 samples used using correlation analysis.

Findings – The testing results indicate that both website functionality and usability are

positively related with online customer satisfaction. The moderating effect of product

involvement is proven as well that the relationship is stronger under low involvement products.

Value – The managerial contribution of this study is to help managers better allocate their

resources in order to boost online customer satisfaction. The academic contribution is that it

takes a segmented view on product categories instead on an aggregate industrial level, and thus

the findings are more applicable to further product-level studies.

Keywords – Online Customer Satisfaction, Website Functionality, Website Usability, Product

5

1. Introduction

With the proliferation of information technology (IT) and the internet, online shopping

is becoming more and more popular among customers. Only a decade ago online shopping was

only adopted by a small group of “venture seekers” because it carried high risks and uncertainties given the technology condition. However, during the last few years online

shopping has been gradually overtaking the dominant role in the merchandise market where

(potential) customers range from teenagers who buy make-ups online to adults purchasing

home supplies from e-retailers. Examples can be seen in developing and developed countries

all over the world. The largest e-retailing merchandiser Ablibaba in China reached an

astonishing sales of $17.8 billion within one day on the 11th November 2016, where the Black Friday and Cyber Monday in the U.S. also received $2.74 billion and $3.07 billion sales

respectively in the same year (Cheang Ming, 2016). For the Netherlands, it is expected that the

share of online sales would triple from 9% to 27% in four categories by 2020 (Health and

Beauty, Consumer Electronics and Appliances, Toys and Games, Apparel). Meanwhile, offline

sales will decline as more customers are shifting towards the online shopping environment

(Marco & Marc, 2013). All these figures prove the point that online shopping is becoming

increasingly important and it is pivotal that retailers keep up their pace in e-commerce in order

to attract and maintain customers.

With the unstoppable development of online shopping and e-commerce, retailers are

either forced to or activity choosing to be part of it and striving to master in e-commerce. This

trend can be easily observed from the growing number of e-retailers and also the unprecedented

6 platform in various ways, aiming to enhance customer experience and hence generate more

sales. Meanwhile, traditionally offline food retailers such as Albert Heijn is also setting up

digital channels where they offer website ordering and home delivery. Therefore, it is crucial

that businesses, no matter born-digital ones or traditional ones to find innovative ways to boost

performance and better meeting customer needs by building new communication, distribution

and promotion channels (Silva and Goncalves, 2016). In this digital era, this new channel will

be internet-based and customer-oriented under the development of social media and internet

technology.

As the most important e-channels to enhance performance, online communication tools

and social media platforms are facilitating companies to learn more about their customers. For

instance, personal information, habits, purchase behavior and preferences are all available

online if companies know how to acquire them. Meanwhile, the very same things are

empowering customers to a large extent as well: a Facebook fan is said to worth 174 USD for

business on average (Wasserman, 2013), the power of eWOM, a comment on the purchasing

website or a simple Tweet with less than 140 words would greatly impact business reputation

and more importantly, influence other customers. In addition, online customers can easily

compare products and acquire information across different merchandisers, thus making them

even more powerful in the online environment than offline. Therefore, in the e-business world

the game changer seems not to be merely “being the best”, but more of “being perceived as the

best by customers”.

Given the competitive environment of e-commerce and empowered customers, customer

7 on retaining existing customers, attracting new customers, customer loyalty and eventually

purchase intention (Tandon, Kiran and Sah, 2015). Conversely, unsatisfied customers may not

only behave negatively but also express their disappointment through eWOM, which would

eventually harm the profitability and reputation of the business (Hussain. R, Al Nasser and

Hussain,Y. K., 2015). Evidence shows that unsatisfied customers would share with 9 people

about their disappointed purchasing experience and on the other hand, once the problem is

resolved they would communicate to 5 other people about the treatment (Hussain et al. 2015).

Therefore, online customer satisfaction and how it is constructed/perceived is a pivotal topic

in online marketing and also one of the research objectives of this study.

Online customer satisfaction needs to be investigated separately from offline satisfaction.

Due to the distinct online and offline channel attributes, customer power and satisfaction

formation are different in these two scenarios. Customers under offline context are unable to

acquire information or only limited information about the product pre-purchase. During the

purchasing process, there is usually a face-to-face interaction between customer and the

business, and product information will be mostly delivered from the sales force. Whereas after

purchase, customers can only express their feelings and satisfaction to a limited number of

others, either positively or negatively. In this case, customers usually have less power than

businesses where information is selectively given and WOM only plays a minor role. However,

this relationship between customers and retailers is fundamentally different in the online

environment where customers usually have more power. It is widely accepted that online

shopping behavior is a more complex matter and customer expectations are also changing

8 customers because of the virtual environment, but at the same time they are able to acquire

more information pre-purchase and compare them cross businesses. During purchasing process,

customers have access to more alternatives and product varieties with no or very little

additional costs because of the online channel. Their e-WOM is also more powerful in this case

where customers can leave comments on business websites or any social media, and these

user-generated information is highly valued by other potential customers. All these attributes of

online channels are empowering customers to a great extent, which means that customer

satisfaction is even more important in this case. Hence, before taking a deeper look at online

customer satisfaction it is necessary to draw the line between online and offline shopping and

understand their distinctiveness in order to better understand satisfaction formation.

One of the most distinct features of online customers is that they are “both purchasers of

products/services, and users of web-based technologies” (Shankar, Smith, & Rangaswamy,

2003; Teo, 2006). As the most important web-based technology and the most frequently chosen

online shopping channel, merchandise website is the touch point where a potential customer

first encounters with. It is also the platform where customer searches for information, seeks for

advice, makes the purchasing decision and asks for after-sale service if necessary. Thus the

website of the business is of vital importance for customer purchasing experience and customer

satisfaction in virtual environment, and e-retailers should attach great attention to their websites.

In an e-commerce report about the Netherlands, 95% of Dutch citizens are internet users and

93% of them have made an online order in 2015, with an average of 13 purchases per shopper

(Laura, 2016). During the past decade many studies have looked into the influencers of

9 as website design, ease of use, convenience of ordering, security and privacy and payment

options, etc. are the most frequently mentioned ones. However, these features are scattered in

various studies which makes it difficult to formulate any systematic conclusions about website.

It was until 2015 when Tandon, et al. concluded these website features into two explicit

categories: Website Functionality and Website Usability. Thus all the attributes are classified

and clearly presented to readers. Therefore, this study will focus on the relationship between

website features and online customer satisfaction because they play the prominent role in

purchasing process. In addition, the integrated view of Tandon (2015) will be adopted to gain

a systematic view where website features are categorized into two: website functionality and

website usability.

Lastly, moderators in the relationship between website functionality and website usability

and online customer satisfaction should also be considered because they tend to influence the

explanatory power of independent variables. Product involvement is commonly regarded as

one of the most significant moderators that influence purchasing behavior (Celsi & Olson,

1988), and is generally defined in terms of perceived personal relevance as either short-lived

or permanent (Celsi & Olson, 1988; Richins & Bloch, 1986). During purchasing, consumers in

high product involvement category tend to search for more product information and make more

comparisons to ensure quality and value (Nijssen, Bucklin and Uiji, 1995) and thus reduce the

perceived risk. On the other hand consumers in low product involvement situation may rely

more on salient cues to make product selection (Nkwocha, Bao, Johnson and Brotspies, 2005).

Therefore, we have reasons to believe that under different levels of product involvement,

10 abundant studies look into the moderating role of product involvement, some of them focusing

only on highly involved products as they tend to trigger stronger satisfaction fluctuation. There

are less studies comparing e-satisfaction for high and low involvement products at the same

time, and as a result the behavior of influence factors of e-satisfaction is unclear. Thus it is

another research objective of this study to test for the moderating role of product involvement,

in both high and low involvement levels.

In summary of the above-mentioned interests of online customer satisfaction and gaps

between previous studies on moderators, the research question of this paper is:

What is the relationship between website functionality/usability and online

customer satisfaction in the e-retailing context, and how the relationship is moderated by

product involvement level?

The practical contribution of this study is that, as all e-retailers are all striving to enhance

online customer satisfaction and since website is the most important channel to achieve this

goal, businesses need to know how website features such as functionality and usability are

influencing e-satisfaction and therefore how to make use of them. In addition, it is also worthy

investigating how this relationship is moderated by product involvement levels because then

the retailers can better improve websites according to the type of products they sell. It is

possible that food shoppers and shoes shoppers have diverse involvement levels with the

products they purchase, and hence they attach different degree of importance to website

functionality and website usability and eventually their satisfaction level is formulated

accordingly. Businesses need to be able to empathize with their customers in order to enhance

11 issue and thus help managers to further improve their websites.

The theoretical contribution is twofold. First, this study aims to enrich the market

research on online customer satisfaction formation and the impacts of website attributes. This

could help further academics to better understand how the e-channels are influencing

satisfaction, and the conclusions could be tested for other online shopping channels, such as

mobile devices or tablets. In addition, many studies focus on solely high product involvement

or on service satisfaction (Richins and Bloch, 1991; Prenshaw, Kovar and Gladden, 2006), little

attention is paid to both high and low involvement on website features for e-retailers. This

paper thus fills the gap of website attributes as one of the e-channels and the moderating role

of high and low product involvement on online customer satisfaction.

The next section is literature review which addresses the most relevant and influential

works in the field, explaining the key terms of this study such as online customer satisfaction,

product involvement and its moderating role. It ends with the conceptual framework and

hypotheses of this study. Then section 3 describes the research design and methodology,

including description of samples, pre-test and measurements of variables. Section 4 presents

the analytical process and the testing result, followed by section 5 which concludes the study

and answers the research question. Lastly section 6 provides a discussion of strength and

12

2. Literature Review

2.1 Customer Satisfaction

2.1.1 The Expectation (dis)Confirmation Theory

Many academic studies as well as practical reports on online shopping have confirmed

that customer satisfaction is a crucial player in business operation, positively influencing

purchase intention, customer retention, customer loyalty and eventually leading to positive

word-of-mouth, improved brand image and better financial results, etc. (Menorca, Ortiz,

Lombardo and Emeterio 2016). It is thus essential for researchers to fully understand how

customer satisfaction is formulated and influenced in order to draw conclusions and

suggestions. Also for online businesses customer satisfaction is crucial for a positive business

reputation and also for boosting sales.

The Expectation (dis)Confirmation Theory (ECT) is the first cognitive theory to describe

customer satisfaction formation, proposed by a series of two studies written by Oliver in 1977

and 1980. He describes satisfaction as a function of pre-purchase expectation, post-purchase

performance evaluation and disconfirmation between the two (Oliver, 1980). It is stated that

“satisfaction is an emotional state arising from the non-confirmation of positive or negative initial expectations for the experience of possession or consumption” (Oliver, 1977&1980). In addition, Oliver suggests that expectation effects and disconfirmation perceptions are additive.

Expectation creates a reference point about which customers make a judgment on the

product/service, and performance evaluation is based on such expectation. Performance being

evaluated as poorer than expectation is rated below the reference point and thus creating

13 and leading to positive disconfirmation. During satisfaction formation process, expectation

about product performance is influenced mainly by three phenomena: (1) the product itself

including previous experience, brand image or symbolic elements; (2) shopping context

including communication with salesperson and social references; and (3) customer

characteristics such as persuasibility and perceptual distortion. Post purchase disconfirmation

is then caused by the degree to which product/service meets, exceeds or falls short of such

expectation. While these studies are the first to develop a consistent model to describe the

cognitive theory of satisfaction formation, they are restricted to the time of written when

shopping are only conducted offline. Shopping context is fundamentally different in the online

environment, e.g. the communication with salesperson is replaced by the encounter with

website, and product features are also communicated online. In addition, the ECT model also

receives some critiques and discussions about its applicability when measuring or predicting

customer satisfaction in other studies.

While performance is universally agreed upon by most researchers as the primary

determinant of customer satisfaction, there are discussions about the operationalization of

disconfirmation and expectation. In a later study by Wu, DeSarbo, Chen and Fu (2006) the

authors discuss two beliefs about disconfirmation. One belief is that disconfirmation is merely

a comparison between post-purchase performance evaluation and pre-purchase expectation,

thus operationalize the concept as the algebraic differences (Wu et al., 2006). The other belief

states that disconfirmation is different from the algebraic gap, because expectation shifts when

disconfirmation is measured and therefore should be employed independently in the model.

14 should not be treated as an independent determinant: (1) disconfirmation already implicitly

incorporate expectation; (2) pre-purchase expectation is difficult to measure in commercial

studies; and (3) expectation has the weakest importance on overall satisfaction. Yet there are

many studies suggests that expectation should be adopted as an independent determinant.

Despite of these controversial views, expectation, disconfirmation and performance are all used

frequently in studies concerning customer satisfaction as either explanatory factors or

antecedents. Understanding the ECT could help readers to further understand how satisfaction

is formulated.

In summary, the ECT model has four crucial components: expectation, disconfirmation,

performance and satisfaction. Expectation is the beliefs of customers before purchasing the

product or service, and it would affect purchasing behavior to some extent. It usually depends

on previous experience of usage, word of mouth or advertisements, etc. (Song Zhu, Jon Kuo

and Munkhbold, 2016). Perceived performance is the actual quality and usefulness of the

product or service received, regardless of expectation. Thus WOM and advertisements can

hardly influence performance and it is evaluated based on facts and actual attributes of products.

Disconfirmation, is evaluated as the difference between performance and expectation, whereas

performance better than expectation leads to positive disconfirmation and higher satisfaction,

otherwise negative disconfirmation will be formulated and thus lower satisfaction level.

ECT is the foundational cognitive theory to explain and predict customer satisfaction

formation. Despite of the discussions concerning the concept, the application of ECT model

can be seen in numerous studies, from traditional satisfaction researches in an offline context

15 in explaining and predicting customer satisfaction is firmly proved, and therefore this study

will also use the ECT model to explain customer satisfaction variation under the influence of

website functionality and website usability.

2.1.2 Online Customer Satisfaction

With new technology develops and e-commerce accelerates, online customer satisfaction

in particular is receiving more and more attention compared to offline satisfaction, and

therefore needs to be considered separately. There is evidence showing that customer attitudes

and behaviors are systematically different and changing in the online environment, given that

competition is fiercer online when alternatives are just a few clicks away (Shankar et al. 2003).

It is thus worthy to investigate the diverse attributes of online and offline contexts and how

customer satisfaction is formulated accordingly. In this section the differences and comparisons

between them will be made, including shopping environment and customer attributes.

Shankar et al. (2003) provide one of the first influential works that compare customer

satisfaction in online and offline environments and is then adopted by many further studies.

According to the authors, there are three possible reasons why online context affects customer

satisfaction differently. (1) Online environment allows customers to better sorting, grouping,

comparing information and accessing more alternatives. With more pre-purchase information

acquired, customers are more likely to make high quality choices which will deliver higher

satisfaction when purchase is made online. (2) From the perspective of ECT model, online

context could also influence customer expectation in the way that as more information is

acquired customers are less likely to be surprised as the product delivered. This implies that

16 therefore avoid false expectation. As a result, negative disconfirmation is lower because

performance tends to align with expectation and hence satisfaction would be higher. (3)

Negative impacts on satisfaction could also be generated in online environment, if customers

perceive higher financial risk, lack of privacy and interaction, poor digital channel design etc.

The first and third arguments about online and offline satisfaction are all caused by the

attributes of website and purchasing channel, and thus also in line with the research question

of my study that looks into the role of website in customer satisfaction. This work by Shanker

sheds light in understanding and explaining the major differences between online and offline

context, which could help further researches to better interpret customer satisfaction formation

under each scenario. In particular, the arguments point out that website as the online purchasing

channel would be the decisive factor that differentiate online and offline shopping, and thus

greatly impact customer satisfaction in the online context. Website attributes become an

essential topic in further studies on e-satisfaction. However, one limitation is that it does not

look into the attributes of online and offline customers but only the shopping environment. In

addition, no multiple-item measures were taken due to data limitation, and this could affect the

testing results. Therefore, in my study multi-item measures will be used for all the variables to

avoid such problem.

Later in the study by Teo (2006), it fills the gap of online shopper attributes in the virtual

environment. The perceptions of online shopping adopters and non-adopters in terms of their

demographic profile, consumer expectation, advantages and disadvantages of online shopping

are investigated. They propose a valuable insight that in addition to the differences between

17 shopping behavior as well (Teo, 2006). Positive attitudes towards e-commerce and the

perceived usefulness impacts the adoption of e-retailing, and the perceived risk and ease of use

influence attitudes towards online shopping. The findings are that Internet users are mostly

male, young and educated, and adopters of online shopping tend to be older and have higher

income than non-adopters. It is also shown in the study that online customers have different

expectations than offline counterparts, meaning that they have higher expectations on ease of

use of the online channel, ensured security and privacy, etc. The advantages of online shopping

include convenience, time and cost saving, ease of comparison, etc.” (Teo, 2006), while disadvantages are security problems and the risk of leaking personal information. The

important implication of this study is that it facilitates us to understand the different

characteristics between adopters and non-adopters. In addition, it shows that not only

e-satisfaction is influenced by website attributes but also the decision of whether to purchase

online is greatly influenced as well. If the perceived risks are too high of using a website, it

may prevent potential customers to order online at all and shift to offline purchasing. On the

other hand, if the website is regarded to be easy to use and highly useful (e.g. provide abundant

information and convenient delivery) the customers are more likely to give it a try. While this

research offers good understanding of online customers, it is limited in terms of research

method that it does not propose any model or hypotheses testing. The finding that Internet

adopters are mostly male is also worthy of reconsideration. Given the time of this study (2006)

online shopping was not fully popularized and could be considered as a risk, therefore more

male is detected as adopters. Also the target country Singapore was at an initial stage of online

18 why online adopters are perceived as bold and risk-seekers. Nowadays online shopping is a

common phenomenon and not a risk any more as technology is mature, it is expected that there

will be at least as much of female online shoppers as male.

In the more recent study by Pereira, Salgueiro and Rita (2016) on e-satisfaction in the

tourism industry, the authors also explain some distinctive features of online customer

satisfaction. First customer expectations change rapidly due to the technology development

and availability of cutting-edge features and services (Pereira, et al., 2016). This means that

expectations are growing higher and higher over time, because with the emergence of new

technology customers would only demand more from online shopping. This indicates that

businesses should constantly improve their performance just to remain the same level of

attraction to their customers, otherwise they will be quickly left out by the fierce competition

in the online retailing market. Also, if expectations are always growing higher it is thus more

difficult for e-retailers to main customer satisfaction not to mention improve it. This would

mean that performance needs to be enhanced even further than the growing expectation in order

to boost customer satisfaction. The advantages of online shopping (such as new technology

and more options) are also risks for businesses to loss customer satisfaction. In addition,

customer motivations for shop online are different during purchasing process. There is study

showing that what induces customers to shop online for the first time is different from the

reason for repurchasing (Pereira, et al., 2016). The implication is that businesses need to fully

understand their customers’ satisfaction formation during different stages of purchasing, and thus better target and fulfill their needs. One example is that a first-time shopper would be more

19 the ease of using the website is more important because he/she wants to be able to place the

order quickly and conveniently. Another key concept proposed by the authors is that online

shopping is not a replacement for offline, but more of a complement. This indicates that

customers may use different channels for different purposes during their purchase. The most

common use of online channel is to search for production information and users’ reviews, and whether to shop online or visit a physical store is to be determined by further elements. The

contribution is that, after discussing all the differences between online and offline shopping it

is pivotal for researchers and also business owners to realize that online and offline channels

are not necessarily competitors but could be supplements and thus creating synergy effects.

However, given the frame of my study this aspect will not be further investigated.

2.2 Website Functionality and Website Usability

After a deep understanding of customer satisfaction and in particular e-satisfaction, now

the focus will move to the explanatory and influential elements of online customer satisfaction.

As mentioned in the previous sections, website is the most essential e-channel for customers to

either search for product information or place the order online. Therefore, website attributes

are expected to place great influence on customer satisfaction formation in the online shopping

environment, and they are also the research subject in my study.

One of the earliest works that have investigated e- satisfaction separately from offline

satisfaction is from Szymanski and Hise (2000). During that time (around 2000) the

antecedents of customer satisfaction have been extensively studied in the classical contexts, as

also mentioned before the study by Oliver (1977, 1980). However, there is no systematic

20 gap by initiating a qualitative research to identify possible influence factors and determinants

of e-satisfaction for the first time (Szymanski & Hise, 2000). The research design of this study

is twofold: first a qualitative study followed by a quantitative one. The reason for the qualitative

research is because this type of method is frequently used in formulating a new model for

subjects that are ill-defined, under-researched or relatively new (Szymanski & Hise, 2000). It

is conducted with focus-group interviews on topics regarding online purchase behavior,

formation of satisfaction and elements that make e-retailing more/less satisfying. The result of

this qualitative study shows that there are four influence factors of e-satisfaction being

documented from the interviews: convenience (time and effort saving resulting from browsing

online and no need to leave home), merchandising (product offerings and information available

online), site design (fast, structured and easy to use of the website) and financial security

(security of online transaction). In the second step the authors then conduct a quantitative

research to testify the validity of the four factors in influencing e-satisfaction. The testing

results an online survey reveal that convenience, site design and financial security are all

positively related with e-satisfaction while merchandising is being left out. Among the three

factors, convenience is perceived to have the greatest positive impact on e-satisfaction followed

by site design and financial security. One implication from these resting results is that, all the

testified influential factors of e-satisfaction are website-related: convenience, side design and

financial security. While merchandising is more about online exclusive product offerings, it is

not closely related with channel attributes (in this case the website) and its impact is on

e-satisfaction is also the weakest. This would once again prove the point that online customer

21 Szymanski & Hise is the first one to research e-satisfaction determinants explicitly, and it

serves as the foundation for studies in this field. These positive influencers are being adopted

repeatedly in further literature as well, however the authors do not make a classification on

these website attributes and thus the conclusions are not systematic and easy to generalize.

Another influential work that looks into e-satisfaction influencers is from Devaraj, Fan

and Kohli (2002), in which another research method of studying e-satisfaction is taken. The

theoretical foundation of this study is an inclusive and integrated adaptation of three existing

models: Technology Acceptance Model (TAM), Transaction Cost Analysis (TCA) and Service

Quality (SERVQUAL) model. TAM is generally used to predict intention and attitude towards

an information system with a focus on technology applications. Specifically, perceived ease of

use and perceived usefulness of the e-commerce channel determine the attitude towards the

information technology, and eventually lead to satisfaction and channel preference. TCA is an

extension of TAM in the way that it incorporates the transaction cost aspect of online

purchasing. Specifically, three aspects of transaction cost: perceived ease of use, time

efficiency and price saving are included in this model. Perceived ease of use is also an element

of TAM, measuring the efforts imposed on online shopping including information searching

and monitoring, etc. Time efficiency is another concern of customers where they face with not

only economic constrains but also time constrains. If shopping online saves time, for example

the reduced travel time to a physical store or the delivery period is short, then customers will

be more likely to shop online and therefore fell gratifying about it. Price saving is generally

caused by the reduced operation cost of e-retailers such as managerial cost and store rent. In

22 is originally proposed byParasuraman, Zeithaml and Berry (1988) to measure service quality,

and is then adopted by many studies in predicting customer satisfaction. Four dimensions of

SERVQUAL are: reliability, responsiveness, empathy and assurance. They are all expected to

be positively related with e-satisfaction and channel preference. The results show that, while

all the factors except reliability and responsiveness from the SERVQUAL model are

explanatory for e-satisfaction, ease of use tested in both TAM and TCA receive the highest

explanatory power (0.51 and 0.74 respectively). Therefore, using the existed classical theories

also points to the same direction that website attributes is the most influential factor to predict

and explain online customer satisfaction. In addition, by looking into the concepts beyond these

antecedents some common points with the study by Szymanski & Hise (2000) can be spotted

here: convenience generally measures the same thing as ease of use, and security is similar to

the concept of assurance. Although these studies depart from different theory basis, their results

converge at the points regarding website features.

Later on in a study by Liu, He, Gao and Xie (2008), the authors incorporate most of the

website attributes that have been identified by previous studies that might impact customer

satisfaction, including the ignored ones. The aim is to develop the satisfaction model in

e-commerce and identify key factors that influence e-satisfaction. The authors take a perspective

of the entire purchasing experience and measure customers’ satisfaction during the whole

process instead of a segmented view. The purchasing process is divided into 3 phases:

pre-purchase stage of information search and alternative evaluation, pre-purchase stage and

post-purchase stage, each containing the website factors that customers encounter with during that

23 design and merchandise attributes; purchase stage involves four elements: transaction capacity,

response time, security/privacy and payment; post-purchase stage has two factors: delivery and

customer service. These 9 factors in total are tested for their relationship with online customer

satisfaction. Using a multiple regression analysis with 1001 respondents in total, authors find

that among all these website features 8 of them significantly influences online customer

satisfaction: information quality, website design, merchandise attributes, transaction capability,

security/privacy, payment, delivery and customer service. Delivery has the greatest impact on

e-satisfaction while website design has the weakest impact. The only factor being left out is

response time. The findings contradict with previous studies in the way that website design is

said to have the least impact on e-satisfaction, however it is stated as the strongest influencer

in previous studies. This could be the result of the authors’ failing to control product category

in their study, while customers are tested for general opinion on online shopping but not limited

to product category (Liu, et al., 2008). This limitation suggests that further studies should take

product category into account when measuring online customer satisfaction, and hence my

study will regard it as a moderating in order to draw more consistent conclusions. The

contribution of this study is also prominent: it is a comprehensive study that incorporates a

wide range of website attributes concluded from previous studies, and they are further

distinguished based on different purchasing stages.

Next the study by Tandon et al. (2015) is discussed here to provide a way of

systematically classify website attributes. Since online shopping is mostly conducted via

website channel, the focus of this study is how website attributes influence online customer

24 usability. “Website functionality is the extent to which the website operates in the way that it is structured and is expected to perform as users’ desire” (Tandon et al., 2015). In this sense the

authors summarize from previous studies 5 dimensions of website functionality: security and

privacy, website design, experiential features, navigational characteristics and consistency

features. In addition, website usability is also considered as a crucial factor in online shopping

because it is important in gaining trust and satisfaction. The authors conclude 4 dimensions

under usability: ease of use, ease of understanding, ease of ordering and ease of purchase. From

the dimensions and definitions we can conclude that website functionality is about the

functioning and operating of the website, while usability is mostly about the ease of using the

website as a shopping channel. The analysis result indicates that website navigation, followed

by website design and security and payment are the most significant components of website

functionality, and they in turn have positive relationship with customer satisfaction (Tandon et

al., 2015). Meanwhile, ease of use and ease of understanding of website usability are tested to

have positive impact on customer satisfaction in synchronization with similar studies. In

addition, ease of ordering is validated for the first time to also have an influence on satisfaction.

Lastly, the effect of ease of purchase is not significant in this study. In conclusion of this study

by Tandon et al., the testing results confirm previous researches on the positive impact of

website attributes on online customer satisfaction. In addition, the two categories of website

functionality and website usability cover most of the website attributes and offer an integrated

view on the topic, thus future studies can follow the same direction and discover new attributes

along the two branches. This integrated view of Website Functionality and Website Usability

25 satisfaction. The limitation of this study is also worth noticing: while taking a comprehensive

view on website features, this study seems to overlook the role of online services on customer

satisfaction. Due to the virtue nature of online shopping and websites, lacking of

communication is one of the reasons impeding more online purchases and therefore good

online services could significantly improve online customer satisfaction. Websites are trying to

offer more online services to facilitate the purchasing process for customers, e.g. interactive

chat, online customer service or after-sales services. Their impacts should also be considered.

In the study by Abbaspour and HazarinaHashim (2015) they take another perspective on

website, which share some similarities with the work of Tandon et al., (2015) but more importantly they take into account the role of website service on customer satisfaction and therefore further complement the research on website attributes. In this study, there are three dimensions to describe website quality: system quality, information quality and service quality. System quality concerns the processing quality of the information system, such as convenience, ease of use, website design and interactivities; information quality describes the importance of information accuracy, relevance and usefulness, it is stated that information quality directly influence consumers’ opinion and assessment of the effectiveness of a website; service quality is the overall support delivered by website to facilitate consumers in making the purchase decision, its assessment is generally based on assurance, empathy and responsiveness of the website. From the description it can be seen that system quality corresponds to website functionality mentioned by Tandon et al., where information quality shares some similarities with website usability to extent. Service quality is the component missing from Tandon’s study and therefore is the dimension of website attributes. The analysis shows that information

26

quality is the most significant antecedents of customer satisfaction with security has the strongest influence, followed by system quality. Although service quality also has positive impact on customer satisfaction, the influence power is significantly lower than the other two. The results indicate that online customers attach great value and importance on the information provided by the website, and they also care deeply about the quality of the website. Service provided, on the other hand, is less influential for their satisfaction formation. This study is derived from a different view on website attributes compared to the previous literature, but they draw similar conclusions where system quality, information quality and service quality of the website all have positive impact on customer satisfaction. The contribution is that it includes the role of service quality of the website, and offers a new dimension for further intentions to improve online customer satisfaction. Its limitation lies in the research design where the respondents are all college students who are quite familiar and acquainted with website and online shopping. This could be the reason why service quality is tested to be less influential, because colleague students are generally less inclined to seek for service facilitation when shopping online, compared with elder counterparts.

2.3 Product Involvement

Product involvement has been defined in various ways from previous studies and adopted

as a moderator in the customer satisfaction relationship. In the field of consumer behavior it is

generally defined as the perceived level of personal relevance with the product. To be specific,

it refers to that when consumers perceive something to be self-related or “instrumental in

achieving their personal goals and values” (Celsi and Olson, 1988). It is proved that product involvement influences both of consumers’ overt behavior (such as information acquiring and

27 purchasing) and also cognitive behavior (such as how they interpret information) (Celsi and

Olson, 1988). Therefore, it can be expected that product involvement will also influence the

relationship between website functionality and website usability and online customer

satisfaction in this study, and it should be properly tested in order to untangle its impact.

One of the earliest and most fundamental studies that looks into the role of product

involvement on customer satisfaction is from Oliver and Bearden (1983). The satisfaction

framework adopted by the authors is based on the expectancy disconfirmation model, which is

initially proposed by another of Oliver’s work in 1977 and 1980, as already discussed in the previous literature. The ECT model states that pre-usage expectations, post-usage experience

and the disconfirmation in between these two together affect satisfaction. The view proposed

by the authors concerning product involvement is that, for low importance products the

evaluation towards product may not even be triggered, meanwhile those highly involved

customers are more likely to be aroused at the early stage of purchase but the effect will be

diminished in the later stage. Based on this point of view, Oliver and Bearden propose that “low

involved customers should be less likely to make strong evaluations, if they make them at all”

(Oliver and Bearden, 1983). The object of this study is to testify the disconfirmation model of

satisfaction and examine if involvement level affects the model in any way. Given the time of

the study, it is conducted in a complete offline context and as a result, it partly confirms that

“pre-exposure affect and intention levels were greater for high involvement users” (Oliver and Bearden, 1983). This effect also lasts to the post-usage stage where satisfaction is also higher.

The result shows that customers have higher satisfaction for high involvement products. The

28 in satisfaction model using disconfirmation theory in the offline context. It also serves as the

academic foundation for further studies on product involvement where the disconfirmation

theory is widely adopted to explain the moderating effect. One of the limitation of this study is

the survey design: it was carried out in three waves with a time span of 5 months. There is

reason to believe that the perceived level of satisfaction can be diminished over time after

purchase as the feeling and emotion attached is fading away.

While the previous study looks into product involvement and satisfaction from a general

aspect, Richins and Bloch (1991) conduct a research validating the moderating role of

involvement on post-purchase product satisfaction in the durable goods category (cars) and

more importantly, the effect of time on satisfaction is also studied. This study is based on a

similar method as the previous literature in the way that satisfaction is derived from the

difference between product expectation and product perceptions---disconfirmation. Two types

of involvement: enduring involvement (EI) and situational involvement (SI) are identified here,

where EI “represents the on-going, baseline level of concern with the product independent of situational influences” and SI is a concern on the product which occurs during specific transaction situation, such as purchase (Richins and Bloch, 1991). It is argued that for SI, during

a high risk purchase consumers tend to spend more time and effort pre-purchase to avoid

making mistakes, and thus the need to feel satisfied is strong. However as time goes by and SI

reduces, the arousal and attention attached to the purchase declines and invested efforts become

vague, more opportunities for the downside of the choice emerge and therefore perceived

satisfaction declines after purchase. For highly enduring involved (EI) consumers, they are

29 themselves an expertise in the product, consumers in this category have more knowledge and

thinking/talking about the product even after purchase, and admitting a wrong choice would be

embarrassing for them. Thus the authors argue that high EI consumers are more satisfied than

their low EI counterparts. In addition, because these consumers have more information of the

product they are also less likely to experience negative disconfirmation under high EI. The

testing results support the theory in the way that during the entire ownership of the product,

satisfaction level decreases as time passes and customers with high EI appear to be more

satisfied than low EI ones. It also validates that right after the purchase high EI customers show

higher satisfaction, but this satisfaction drops to a level only slightly higher than in the low EI

situation after two months. Customers with high EI are also less likely to experience negative

disconfirmation, partly because they tend to acquire extensive information and knowledge

about the products. It is also suggested by the results that positive and negative

disconfirmations need to be examined separately, at least for durable products. The separation

between SI and EI incorporating time as a factor is also inspirational, but given the scope of

this study and the fact that I intend to measure product involvement from a general perspective,

the distinction between SI and EI is not adopted here. The authors also offer a detailed

explanation to why highly EI customers are more satisfied than low EI counterparts. The

limitation is that satisfaction is measured on an aggregate level, where the respondents are

asked to rate on an 11 point scale instead of taking a separate view on each influence factors of

satisfaction. In addition, both of the above two literature focus on tangible products but ignore

such relationship for services, hence the next study fills this gap by investigating service

30 Prenshaw et al. (2006) examines customer satisfaction also using the ECT for different

product involvement levels, in synchronization with the previous works. All the three elements

including expectation, disconfirmation and performance from ECT are adopted here to predict

customer satisfaction. The authors specifically state that the relationship between the three

variables and satisfaction is dependent on whether goods or services is the context for judgment,

and they choose a new, non-traditional credence-based service offering as the subject. Since

this type of new service (elderly assurance) is difficult or impossible to evaluate even after

purchase, the expectation formation process is particularly affected. Prenshaw et al. (2006)

propose that since the characteristics of this type of new service will influence satisfaction

formation process, so the three factors of ECT will generate different level of impacts on

satisfaction under varying product involvement. The results confirm the hypothesis that

evaluation of performance generates a greater impact on satisfaction under high involvement

level while disconfirmation better assesses satisfaction when involvement level is low. Lastly,

expectation is not influential in both scenarios. The contribution of this study is that it proves

the validity of product involvement as a moderator in the relationship between ECT elements

and customer satisfaction in the service industry. The reason to discuss this literature here is

that online customers purchase not only products but also various services via websites, for

example hotel booking or travel services. Thus it is important to find abundant literature that

validate the moderating role of product involvement for both product and service industry.

The common point of the above mentioned literature is that they all look into the

moderating role of product involvement on customer satisfaction from the ECT perspective.

31 the work by Martin, Camarero and Jose (2011) fills this gap by using consumer involvement

to explain website attributes effect on online customer satisfaction. The theoretical foundation

of this study is the Elaboration Likelihood Model (ELM), which describes two attitude

formation routes of customers determined by level of involvement. The central route occurs

when involvement level is high and customers are more motivated to carefully evaluate or

process information. In this case every relevant information is being elaborated and together

formulate an integrated argument into customer’s attitudinal schema. On the other hand, the peripheral route is triggered when involvement level is low and customers are less motivated

and likely to have elaborate (Martin et al., 2011). In the case of website attributes, some

characteristics are cognitive attributes that affect the cognitive perception of the website, such

as service quality, warranty and security, while some are experiential attributes that affect the

experiences of using the website, such as website design and interactivity. Based on the ELM,

the authors predict that the effect of cognitive website attributes is greater under high

involvement when customers are taking the central route, and the effect of experiential website

attributes is lower under high involvement. The analysis gives inconsistent result that only

“service quality” of cognitive attributes has greater effect on satisfaction under high involvement while warranty and security have significant effects under low involvement

consumers One possible explanation is that the low involved customers are more likely to be

less experienced and less confident about shopping online, and therefore warranty and security

are more important for them. In terms of experiential attributes, website design has significant

effect in both high and low involvement cases, and interactivity has a negative effect under low

32 that the peripheral cue is valued at both high and low involvement levels. Although the analysis

results are inconsistent with the assumptions, this study offers great value to my study because

its research object is website attributes and its theoretical foundation ELM is inspirational in

explaining the moderating effects of involvement level. However, it should be noticed that the

involvement level in the study by Martin et al. (2011) refers to customer involvement with

shopping online, whereas in my study involvement refers to product involvement level.

In conclusion, as one of the most important concepts in marketing studies customer

satisfaction has received many attentions in both academic and practical fields. The studies

initiated from an offline context given the timeline and then gradually transferred into online

context. The influence factors of customer satisfaction are also shifting from offline to online,

and this trend can be spotted from the previously adopted ECT model to the internet-based

technologies. It is thus more challenging to boost customer satisfaction in the digital world

because customer behaviors are usually not stable and constantly affected by numerous reasons

(Pereira, Salgueiro and Rita, 2016). Among all these variables, a majority of them are centered

on website attributes because website is the primary touch point for customers and also the

most selected online shopping channel. While there are abundant studies focusing on scattered

website factors and draw inconsistent conclusions, this study will take the integrated view of

Website Functionality and Website Usability as the main types of website attributes and to

testify their influence on online customer satisfaction. The contribution of this method is that

it categorizes all the website features into two and hence lead to consistent conclusions. Also

this method allows further studies to keep expanding the components of each category without

33 role of product involvement level will be incorporated. Product involvement is a widely

adopted moderator in satisfaction relationship, but mostly high involvement is tested in an

offline context. This study will then test both high and low involved products, and their

moderating effect on Website Functionality and Website Usability respectively. The

implication would be to provide a more detailed understanding of how online customer

satisfaction is influenced and moderated, depending on the product categories

2.4 Conceptual Framework and Hypotheses

2.4.1 Conceptual Framework

In the previous section the concepts and definitions of the key terms of this study have been

discussed: the dependent variable is Online Customer Satisfaction (Sat), the independent

variables are Website Functionality (WF) and Website Usability (WU), and the moderator is

Product Involvement (PI). The conceptual framework is formulated as following:

Figure 1 Conceptual Framework

(OC H3a H3b H1 Product Involvement (PI) Online Customer Satisfaction (Sat) Website Functionality (WF) Website Usability (WU) H2

34 2.4.2 Hypothesis Development

Website attributes have received abundant attention from researchers on their influence

and impact on online customer satisfaction. From the earlier study by Szymanski and Hise in

2000, it is proved that convenience, site design and financial security are strong influence

factors for e-satisfaction. Later on, based on existing classical theories the study from Devaraj

et al. (2002) also validates that among all these factors, ease of use and perceived usefulness

of the website show the strongest influence power on e-satisfaction. The comprehensive study

by Liu et al. (2008) on the Chinese market also shows somewhat similar results, where delivery

has greatest impact on online customer satisfaction and website design has the weakest power.

It is until the work of Tandon et al. (2015) that the scattered website attributes are being

categorized into Website Functionality (WF) and Website Usability (WU), which enables us to

take an integrated view on this subject and draw systematic conclusions. Website functionality

is the functioning and operating of the website in which customers use to search for information

and place orders and website usability is the general ease of use of the website as means of

purchasing. In the work of Tandon, et al. (2015) they state that both website functionality and

website usability have a positive impact on customer satisfaction and therefore hypothesis 1

and 2 are formulated as a positive relationship. However, based on previous studies that website

attributes have diverse weight and power on online customer satisfaction, this study will also

compare the two categories on their impact on satisfaction.

H1: Website Functionality (WF) positively influences Online Customer Satisfaction

(Sat) in e-retailing

H2: Website Usability (WU) positively influences Online Customer Satisfaction in

35 In the earlier studies from Oliver and Bearden (1983), Richins and Bloch (1991), and

Prenshaw (2006), the results are in general consistent: customer satisfaction is higher under

high (enduring) involvement using ECT as influence factors of customer satisfaction. The study

by Martin et al., 2011 regards website attributes as the explanatory variable for online customer

satisfaction, and the attributes are categorized into cognitive (central route) and experiential

(peripheral route) based the ELM. In my study, WF describes cognitive signals of the website

and thus contains objective attributes such as search function, payment function and website

stability. Customers are taking a central route in this case where they take a deeper and more

conscious analysis of the information received (Martin et al., 2011), thus according to ELM the

effects of WF would be greater on online customer satisfaction under high involvement

products. On the other hand, WU describes the perception and feelings of customers on the

website, and in this scenario users take a peripheral route where emotional evaluation is

triggered. The effects of WU should be greater on e-satisfaction under low involvement

products, based on ELM. Hypothesis 3a and 3b are formulated accordingly.

H3a: The relationship between Website Functionality (WF) and Online Customer

Satisfaction (Sat) is moderated by Product Involvement (PI), so that they are more

positively related under High Involvement (HI) products.

H3b: The relationship between Website Usability (WU) and Online Customer

Satisfaction (Sat) is moderated by Product Involvement (PI), so that they are more

36

3. Methodology

3.1 Method

This quantitative study is explanatory in nature and will conduct an online survey in order

to collect cross-sectional data. This method has been adopted by many previous studies to

measure customer behavior and attitudes toward a product, and it allows for a larger amount of

respondents. Data will be collected from the questionnaires handed out to the respondents, and

statistical testing will be conducted with the data in order to draw conclusions on the research

topic.

The population is all the customers that have previous online shopping experience or

have the intention to shop online. Since the population is expected to be large and extremely

scattered and it is difficult to gain information about any list of online shoppers (e-retailers

might has this but difficult to acquire this information), therefore the sampling frame is also

difficult to identify. In this case a non-probability sampling technique will be conducted in this

study. To be specific, both volunteer sampling and convenience sampling will be adopted here

to collect as much data as possible. I publish the information and description about the survey

online and viewers self-select themselves into taking the survey. These volunteer respondents

are generally more interested in the topic and motivated to go through the whole survey.

Convenience sampling is conducted as well by distributing the survey to social media platforms

or to my personnel connections. These respondents are invited to take the survey, regardless of

their interest and opinion about the topic. Given the number of variables of this study, I intend

to collect as much data as possible with a minimum amount of 200 complete surveys. Noble

37 on the length of the questions. A 10-question survey has a completion rate of 89%, 20-question

surveys are around 87% and a survey with 30 questions is at 85%. Based on these numbers I

set the goal of collecting at least 250 responses in total.

3.2 Pretest

Before conducting the main survey, a pre-test is carried out in order to identify high and

low involvement products. I list multiple products under each category and ask the respondents

to pick out the ones that best fit into their perception of “high/low involved products”.

Examples for high involvement product questions are “which of the products are of highly

importance to you” and “which of the products you would engage in extended thinking before purchase”. Examples of low involvement product questions are “which of the products are

more or less indifferent to you under the same price range” and “which of the products you will not engage in extended thinking before purchase”. This pretest is conducted due to the

consideration that product involvement is not a concept that every potential participant is

familiar with. Asking respondents to directly rate their level of product involvement could

generate biased results if they do not fully understand the concept. Therefore, I provide a list

of products and ask them to choose the ones that fit the best to their own description of high/low

involvement and then adopt the most selected ones as representative of high/low involvement

products.

The pre-test invites 30 participants in total aiming to gain some deep understanding of

the perception of high and low involvement products. In high involvement product category,

mobile phone, personal laptop, shoes and clothing all receive high selection rate with 83.33%