FACTORS AFFECTING CUSTOMER SATISFACTION IN GROUP BUYING IN VIETNAM

As the review above, there is no study focus on service quality of group buying even it is the critical factor affecting customer satisfaction.. Therefore, this research will try to figu

UNIVERSITY OF ECONOMICS HO CHI MINH CITY

International School of Business

-

Nguyen Thi My Chi

FACTORS AFFECTING CUSTOMER SATISFACTION IN GROUP BUYING IN

VIETNAM

MASTER OF BUSINESS (Honours)

Ho Chi Minh City – 2013

UNIVERSITY OF ECONOMICS HO CHI MINH CITY

International School of Business

-

Nguyen Thi My Chi

FACTORS AFFECTING CUSTOMER SATISFACTION IN GROUP BUYING IN

VIETNAM

ID: 60340102

SUPERVISOR: Dr Tran Ha Minh Quan

Ho Chi Minh City – 2013

ABSTRACT

Recently, group buying model is developing rapidly and impressively in Vietnam as a type of online shopping This study aims at investigating the motivations that influence customer satisfaction as well as service quality of group buying This research will find out that service quality, brand, trust and price are critical factors influenced customer satisfaction with service quality is the most important factor

Keywords: group buying, group buying site, customer satisfaction

TABLE OF CONTENTS

ABSTRACT i

TABLE OF CONTENTS ii

LIST OF FIGURES iv

LIST OF TABLES v

I INTRODUCTION 1

1 Research background 1

2 Research motivation 3

3 Research questions 4

4 Research objectives 4

II LITERATURE REVIEW 4

1 Group buying 4

2 Service quality 6

3 Online service quality 8

4 Service quality and customer satisfaction 11

5 Price discount and customer satisfaction 12

6 Trust and customer satisfaction 13

7 Brand and customer satisfaction 13

8 Research framework 14

III RESEARCH METHODOLOGY 16

1 Research process 16

2 Research sampling method 18

3 Measurement scales 19

4 Data collection procedure 21

5 Data analysis method 21

5.1 Descriptive analysis 21

5.2 Reliability analysis 22

5.3 Exploratory Factor Analysis (EFA) 22

5.4 Multiple regression analysis 23

IV FINDING AND ANALYSIS 24

1 Sample description 24

2 Evaluation and refinement of measurement scales 27

2.1 Cronbach Alpha Reliability Analysis 28

2.2 Exploratory Factor Analysis (EFA) 30

2.2.1 EFA for independent variables 30

2.2.2 EFA for dependent variable 37

2.2.3 Pearson Correlation Coefficient 38

3 Hypotheses testing 39

3.1 Testing assumptions of Multiple Regressions 39

3.2 Regression analysis 41

3.3 Summary of hypotheses testing results 42

4 Discussions of the findings 43

V CONCLUSIONS AND IMPLICATIONS 45

1 Conclusion 45

2 Implications of this research 45

3 Research limitations and implications for future research 47

REFERENCES 49

APPENDIX A 54

APPENDIX B 60

APPENDIX C 66

APPENDIX D 68

APPENDIX E 70

APPENDIX F 71

LIST OF FIGURES

Figure 2.1 The research model and hypotheses 14

Figure 3.1 Research process 17

Figure D.1 Histogram of SATIS dependent variable 68

Figure D.2 Normal P-P Plot of regression standardized residual 68

Figure D.3 Scatterplot 69

LIST OF TABLES

Table 3.1 Survey item summary 20

Table 4.1 Sample characteristics 24

Table 4.2 Cronbach’s alpha reliability test result 28

Table 4.3 KMO and Bartlett’s Test for independent variables 30

Table 4.4 Rotated Component Matrix 31

Table 4.5 Removal process during EFA 32

Table 4.6 Total Variance Explained 33

Table 4.7 Rotated Component Matrix 34

Table 4.8 Cronbach’s alpha summary after EFA 36

Table 4.9 KMO and Bartlett's Test for dependent variable 37

Table 4.10 Factor loading of satisfaction variable 37

Table 4.11 Pearson correlation coefficients between SATIS and other independent variables 38

Table 4.12 Coefficientsa 40

Table 4.13 The summary of overall value 41

Table 4.14 ANOVAa result 41

Table 4.15 The result of testing hypotheses 42

Table C.1 Pearson correlations 66

Table E.1 Descriptive Statistics 70

Table F.1 Measurement scales 71

I INTRODUCTION

In the first chapter, the background of the chosen research subject will be presented, along with problem background and the purpose of this thesis Finally, the research question will be addressed with the research objectives which will be clarified clearly in later parts

1 Research background

In the past, consumers must go to the store and buy commodities directly It takes much times and also brings the inconvenience to consumers, especially who are far away from the store Nowadays, as the Internet has been developing, more buyers are enjoying online shopping because the Internet services are as a convenient way to shop Consumers can buy goods, even services, online, where makes more options for them Consumers who have a want for some favor services

or goods could be gathered when making an online shopping (J Chen, Chen, & Song, 2002; Pi, Liao, Liu & Lee, 2011) Additionally, it offers many advantages, such as facility, rapidity, a 24-hour opening and unrestricted geography

Due to the advantages of Internet, in the context of electronic commerce, online group buying becomes more and more popular According to Rezabakhsh et

al (2006), “central idea of online group-buying is the aggregation of geographically dispersed consumer purchasing power from all over the countries that have a common interest in a certain product and come together on the group-buying sites to get price discounts” In addition, joining online group buying, consumers not only get competitive advantages of quantity system price and delivery service, but also have chances to share shopping experiences with other buyers

In Vietnam, according to electronic information website Maskonline, website phagia.com appeared officially on June 2010, making an underground river to the field of electronic commerce in the country Some months later, many other group buying sties were born to compete each other, as cungmua.com, muachung.vn, nhommua.com, etc According to Vu Hong Quang, a representative of the Southern

VC Corp which owns muachung.com, after just 7 months of operation, his site already has 100.000 members Muachung.com’s main competitor, nhommua.com has the same number of registered users but even some of the newer sites have 60.000 members According to uncompleted statistics, up to now, Vietnamese market has about 97 sites that offer group buying services (E-Money, 2011)

Deals sites can create a winning situation all around that give benefit to three sites: business, coupon site and consumer Businesses get some much needed traffic, the coupon sites get a commission for pulling in that traffic and consumers get a good deal And Vietnamese are loving this model

Group buying model adopts “tipping point” mechanism This mechanism requires a minimum number of participants for each deal There are two conditions make a transaction happens: the number of actual participants exceeds the minimum number and the potential buyers have prepaid for the deal (Cohen, 2009) A deal usually includes price discount of the product or service When the number of people reaches a certain limitation, customers will get direct discounts on goods or services or they can buy vouchers with attractive prices (E-Money, 2011)

According to action.vn, group buying model is as a balloon inflated round two years ago, but it is now flat and gradually has to transform to survive (May 2012) Even Groupon in the U.S., the first group buying model in the world, also had bitterly announced dismal sales declines in the last 6 months In Vietnam, the

group buying model is facing to changes to overcome this difficult period In the past, the majority of the initiators were main buyers in online group buying However, because of the impact of global economic recession, many initiators are aggressive being the roles of vendors holding the products or services to supply to consumers This research doesn’t focus on analyzing the reason why initiator acts as vendor, however, vendor role of initiator is considered as a part of group buying mechanism to evaluate service quality later

in recent difficult environment

In previous studies, some of them discussed online group buying in operator point of view: Wei et al (2011) proposed a management model in online group buying’s process, Matsuo (2009) proposed an effective pricing mechanism for traders, Anand and Aron (2003) using analytical economic models to study the phenomenon of online group buying The other discussed from consumer point of view: Liao et al (2012) focused on consumers’ demands and preference of stuff in online group buying, Tsai et al (2011) found out that trust in virtual community and website quality are the determinants of online group buying’s intention, etc

As the review above, there is no study focus on service quality of group buying even it is the critical factor affecting customer satisfaction Therefore, this research will try to figure out how to measure service quality of a group buying service toward customer satisfaction It also provides the main factors along with service quality to enhance customer satisfaction of group buying service

3 Research questions

According to the research background and motivations outlined before, this research aims to investigate the following questions from the experiences of buyers

in online group buying:

(1) What are online service quality dimension in group buying service and how do they affect customer satisfaction?

(2) What are factors that affect customer satisfaction in Vietnamese group buying service and how do they affect to customer satisfaction?

4 Research objectives

The main objective of this research is to investigate the usefulness and applicability of theoretical model in measuring online group buying quality and customer satisfaction:

- Analyze the method used to measure service quality of group buying and the influence of the measure to customer satisfaction

- Analyze factors influencing customer satisfaction in group buying in Vietnam

II LITERATURE REVIEW

In this chapter, some literatures are reviewed concerning the online service quality This research will continue by presenting some popular scales such as SERVQUAL and E-S-QUAL Finally, the purposed research model for the adoption

of the group buying will be presented

1 Group buying

Nowadays, multiple types of electronic commerce have been developing strongly Group buying is one of those diversified types, and has made an impressive speed of progression in recent years (Matsuo, 2009) Group buying (i.e collective buying, group purchase) has existed in marketplaces for several decades

It is defined as the purchase from a group of consumers who seeks to get products

or services at the cheapest price but in the best quality Group buying model can help many buyers gather into collective buying groups, even though they are not near or they do not know each other (Kittipong et al., 2011)

In the past, members of group buying were restricted They could only ask for the person who they are familiar with to join group buying (Liao et al., 2012) Yuan (2004) also pointed out that the main reason of this issue is the difficulty in finding a group of people with the same demand all the time in traditional markets Therefore, one of advantages of online group buying is that many buyers can form groups on the Internet in a short period of time and buy goods or services at a discounted price even if they don’t know each other (Matsuo, 2009) Besides, geographical limitations are also a problem in traditional market (Liang & Lim, 2011) However, with the growth of Internet, consumers with the same interests can converge together easily through Internet without any difficulty as they meet

directly (Yuan, 2004) Thus, the other advantage of online group buying is that online group buying can overcome the geographical limitations that the traditional group buying could not do like that

Erdogmus and Cicek (2011) mentioned that there are two different types of online group buying systems: dynamic pricing and stable pricing In the dynamic pricing type, the discount prices can be changed basing on the price-quantity tables defined by the sellers If in a predetermined time, a group of buyers could be gathered, then consumers could get the product or service at the same discounted price

In the second type of the online group buying, stable pricing, “the seller offers a certain product or service at a large discount price, but the price is static, does not get lower as the number of the buyers increases (Erdogmus & Cicek, 2011, p.309)” However, the discount price is only applied when the total number of the buyers who buy the deal is greater than the predetermined limit of the minimum required number of buyers Due to the second type of online group buying is a new form of marketing at the connection of promotion and pricing and grows rapidly (Erdogmus & Cicek, 2011), as well as it is more popular in Vietnam than the first type, this research will focus on the context of the second type of online group buying

According to Zhou (2005) and Chen (2007), there are two roles in the online group buying chain One is main buyer or it is also called initiator The responsibilities of main buyer are that he/she posts the issue about the requirement

of group buying, and then calculates the qualities of demand as well as price to order the products which are buyers expected The other role is follower who is

gathered in a group by main buyer Follower is mostly consumer in online group buying chain

Lai (2002) pointed out one new facet of group buying model Recently, due

to the economic situation on the world, the role of initiator in group buying is able

to be a main buyer or a seller Initiator can buy good or service from supplier or vendor, then sells vouchers to consumers, or initiator can sell good or service directly to consumers As a result, correspond to the second type of group buying, initiators here are the voucher suppliers, sometimes are the product vendors also

Consequently, in this research, service quality of group buying will be examined throughout the process from buyers (consumers) start to buy vouchers/services from initiators till consumers finish using services or get the products

2 Service quality

In this part, some definitions and dimensions of service quality are reviewed for the purpose of choosing the suitable measurement for the research Firstly, Parasuraman et al (1988) defined service quality as the ability of the organization

to meet or exceed customer expectations However, there is the difference between customer expectations of service and perceived service (Zeithaml et al., 1990) Zeithaml et al (1990) explained that perceived service quality derives from comparisons of customer expectations with their perceptions of service delivered by the suppliers If expectations are greater, then perceived quality is less than satisfactory and hence customer dissatisfaction occurs (Parasuraman et al., 1985; Lewis and Mitchell, 1990)

In the past, many studies have attempted to discover the global or standard dimensions of service quality such as Parasuraman et al., 1985, 1988; Paulin and Perrien, 1996; Pitt et al., 1999; Sasser et al., 1978 Among them, Parasuraman et al.’s (1985) research has been the most prominent because of its affect on other researches later on Through a series of focus group sessions, that study uncovered ten dimensions of service quality that are generic and relevant to services in general:

tangibles, reliability, responsiveness, communication, credibility, security, competence, courtesy, understanding the customer, and access

Later, these ten dimensions were improved and reduced to five dimensions:

tangibles, reliability, responsibility, assurance, and empathy (Parasuraman et al.,

1988) Then, a new global measurement for service quality, SERVQUAL was

developed by Parasuraman et al (1988) basing on those five dimensions After that, many researchers use the SERVQUAL scale to measure the service quality not only

in a specific industry but also in various service industries such as hospital, food, train ticket, etc

Anyway, SERVQUAL measurement is not a perfect scale in all cases Because the SERVQUAL measurement does not include the unique facets of online service quality as one prior study pointed out, such as customer-to-Web-site interactions, since this instrument was constructed based mainly on customer-to-employee interactions At the beginning, the five dimensions of SERVQUAL were found to be valid in the traditional service market However, when some academic researchers applied the SERVQUAL scale to measure e-service quality, it has been considered problematic and may not be quite appropriate The reason is that there are some differences between e-service and traditional service in some aspects, such

as the absence of sales staff, the absence of traditional tangible element, and service of customers As a result, the SERVQUAL is not suitable for measuring e-

self-service quality, and it is meaningful to renovate SERVQUAL scale into a new valuable instrument for measuring e-service quality Below part will review some online service quality dimensions which are developed from this scale

3 Online service quality

Parasuraman et al (2002) defined online service quality as the extent to which a website facilitates efficient and effective shopping, purchasing, and delivery We can recognize that it includes broadly all phases of a customer’s interactions Santos (2003) argues that online service quality is an overall customer assessment and judgment of online service delivery in the virtual marketplace It is much easier for customers to compare different service offerings than through traditional channels (Santos, 2003) In this study, service quality provided by group buying sites – a special type of retail web sites – is focused on Because the definition of e-service quality by Parasuraman et al (2002) cares almost enough phases of online service, so it has been used by many researchers, and it is also applied for the context of this study in building a conceptualization of online service quality

Yang (2001) considered service quality as an important factor to decide the success of online retail businesses because it influences to consumers’ online shopping experiences The result of one empirical evidence of Kim et al (2006) proved that poor service quality would affect online retailers negatively When online shoppers don’t believe in shopping and handling charges, over 60 percent of them will exit before completing the buying transaction For example, incomplete product information (e.g missing links and non-working buttons) leads to customer frustration and in turn to existing Nonetheless, some online retailers lose a valuable

opportunity to build customer loyalty because of poor service quality (Wachter, 2002)

Besides online service quality becomes a hot topic in research field, measuring and monitoring online service quality in the virtual world are also considered as an importance Many researchers have already built online service quality measurement basing on analyzing its dimensions For example, Dabholkar (1996) conducts a research work on the dimensions of e-service quality focusing on website design, reliability, delivery, ease of use, enjoyment and control Next, Yoo and Donthu (2001) develop SITEQUAL scale to measure online service quality of website including the four dimensions: ease of use, aesthetic design, processing speed, and interactive responsiveness Compare with the traditional dimensions of service quality, Cox and Dale (2001) proposed six new dimensions of online retailing service quality, including website appearance, communication, accessibility, credibility, understanding and availability

As we can recognize, these scales just focus on evaluating web site quality such as design, usage and performance It lacks of information about service attributes completely In addition, web site quality is just a small part in the service quality chain of group buying model Group buying must pay attention to other attributes to satisfy customers besides web site quality For example, the four dimensions of SITEQUAL by Yoo and Donthu (2001) do not capture all aspects of the buying process, and thus do not provide a comprehensive evaluation of a site (Parasuraman et al., 2005)

From limitations of measuring online service quality in above analyses, in

2005, it is the first time the meaning of service is comprehensive in a research of Parasuraman et al because it contains all aspects of online service, in which pre-

and post- web site service are described About the measure of online service quality, the one used in that research of Parasuraman (2005) named E-S-QUAL includes all phases of a customer’s interactions with a website There are four dimensions of E-S-QUAL: efficiency, fulfillment, system availability and privacy

In addition, Parasuraman et al (2005) also developed an e-recovery service quality scale as long as E-S-QUAL named E-RecS-QUAL as a subscale for problem resolution with three dimensions: responsiveness, compensation and contact

From that time, E-S-QUAL and E-RecS-QUAL become an instrument to use popularly because of their strong effects on customer satisfaction However, according to Kim et al (2006), the compensation dimension in this instrument cannot be measured without actual problems to resolve, and then this dimension will be eliminated from the model used in this research Summarily, this research will measure service quality of group buying by totally six dimensions:

- Efficiency: the ease extent that customers can access and use the website in purchasing transactions

- Fulfillment: the extent that the site can keep its promises about order delivery and item availability

- System availability: the availability of the website whenever customers would like to access without functioning problems such as missing links, non-working buttons

- Privacy: the safety extent that the site can protect customer information

- Responsiveness: the effective control ability of the site when facing up to problems and returns from customers

- Contact: the support availability of the site whenever customers need to contact

4 Service quality and customer satisfaction

According to Spreng and Mackory (1996), service quality and customer satisfaction are inarguably the two concepts that are so difficult to explain in marketing theory and practice In 2001, Yang argued that when comparing the performance of a product or service with their expectations before purchasing, if it exceeds the expectation, the satisfaction of customers would be increased However, four years later, Saha and Zhoa (2005) gave out a new definition of satisfaction like

“a collective outcome of perception, evaluation and psychological reactions to the consumption experience with a product or service”

The appraisal-emotional response-coping attitudinal framework of Bagozzi (1988) thought that service quality is a construct oriented to awareness, while satisfaction is an emotional construct It is the reason that many researchers suggested service quality is an antecedent of satisfaction After that, in 2003, Woldfinbarger and Gilly discovered a positive relationship between overall service quality and satisfaction in case of fixed web retailing; also, Montoya-Weiss et al found out an evidence to prove the relationship between service quality and satisfaction by analyzing the determinants of online channel preference for a multi-channel service provider

One of the targets of research question in this study is that to define how service quality dimensions affect to customer satisfaction Yang et al (2004) has a study about online service quality dimensions and their relationships with satisfaction At the result, it confirms that the most often-cited service quality dimensions leading to customer satisfaction are: responsiveness, ease of use (efficiency) and service reliability (fulfillment) Yang and Tsai (2007) also have a research that measure the affect level of E-S-QUAL scale on customer satisfaction

and confirmed that the factors representing efficiency, fulfillment and responsiveness have the strongest effects, followed by contact and then privacy Moreover, according to Kim et al (2006), a website needs to contain many different browser functionalities although system availability dimension is found to make a strong impact on customer evaluation, otherwise it may lead to customer frustration and dissatisfaction From these points of view, we can extract the summary that there is relationship between E-S-QUAL scale of online service quality and customer satisfaction and can build following conceptual hypotheses:

H1a: Efficiency has a positive impact on online customer satisfaction

H1b: Fulfillment has a positive impact on online customer satisfaction H1c: System availability has a positive impact on online customer satisfaction

H1d: Privacy has a positive impact on online customer satisfaction

H1e: Responsiveness has a positive impact on online customer satisfaction H1f: Contact has a positive impact on online customer satisfaction

Especially, in this research, customer satisfaction was defined as the overall evaluation in the consumption experience of product or service after the customer used the group buying website

5 Price discount and customer satisfaction

The most benefit that group buying service can give customers is discounted price Group buying sites must provide buyers with rare resources which the buyers could not get from any other online purchasing sites (Kittipong et al., 2011) People come to group buying sites and buy product or service if they could find a cheaper and more reasonable price compared to other sites When the buyer perceives that

with the cheaper price, they could still get the valuable product or service, the price satisfaction of the buyer is increased, and he or she is more likely to join and purchase on group buying sites Thus, price is considered as a hypothesis to analyze:

H2: Price discount has a positive impact on online customer satisfaction

6 Trust and customer satisfaction

Many prior studies considered that trust was a critical factor for customers to create and keep relationships with vendors (Deng, Lu, Wei & Zhang – 2010) Kauffman, Lai & Lin (2010) has a new concept to define trust There are two people that related to trust definition: trustor and trustee When trustor believe that trustee

is sincere and honest, and will not try to harm or trick trustor, leading to tangible or intangible damage, it means that trustor gives trust to trustee For example, when a customer trusts a service or good provider, customer will expect to raise satisfaction and loyalty towards the vendor (Kassim & Abdullah, 2008) As a whole, if customer does not trust the vendor, customer will possibly be dissatisfied with that vendor Deng et al (2010) & Kim et al (2009) think that trust will affect to satisfaction in long term Hence, we have this hypothesis:

H3: Trust will positively affect customer satisfaction toward group buying service

7 Brand and customer satisfaction

Today, brand is the official introduction of the company and the product that company provides customers According to Mowen (2001), brand is a promise or a set of guidelines that reflects the image of a company, it makes customer think of

what the company does Brand gives the specific value to bring the company to the market, therefore it distinguishes the company from others

Hammond (1999) supposed that branding is a relationship that was built on understanding and satisfaction For example, when customers know about a famous brand of product well, they tend to be satisfied with that product when purchasing in that group buying site rather than other sites that have other brands of this product Therefore, this research proposes one more hypothesis:

H4: Brand of product or service will positively affect customer satisfaction in group buying service

H1f H1e H1d H1c H1b

H1a

Price discount Efficiency

Contact Group buying

service quality

The hypotheses of this research include:

H1a: Efficiency positively relates to online customer satisfaction

H1b: Fulfillment positively relates to online customer satisfaction

H1c: System availability positively relates to online customer satisfaction H1d: Privacy positively relates to online customer satisfaction

H1e: Responsiveness positively relates to online customer satisfaction

H1f: Contact positively relates to online customer satisfaction

H2: Price discount positively relates to online customer satisfaction

H3: Trust will positively affect customer satisfaction toward group buying service

H4: Brand of product or service will positively affect customer satisfaction in group buying service

III RESEARCH METHODOLOGY

This chapter presents brief description of the research methodology used for the research This includes the research process, research measurement, sample selection methods, data collection methods and data analysis methods

1 Research process

Due to the underlying study, the researcher has simply chosen quantitative approach for achieving the purpose of the study Quantitative research approach is based on the development of testable hypotheses and theory Base on idea of Kenova and Jonnasion (2006), quantitative investigations tend to measure “how often” or “how much” Through utilizing this method, the researcher would measure service quality and the level of customer satisfaction in group buying in Vietnam

To collect the quantitative data, the survey method has been used and eventually the data has been analyzed by using statistical techniques

The research process of this study is described in the figure 3.1

Figure 3.1 Research process

Firstly, research problem was defined, and then research objective and research questions were identified to be the target of solving defined research

Research problem Research objective Research scope

LITERATURE REVIEW

Research model

& hypotheses

Questionnaire development

RESEARCH DESIGN

Pilot study with face-to-face interview (n = 10)

Sampling design

Refine measures &

measurement scale

Refine questionnaire

Main survey

DATA ANALYSIS

Reliability testing (Cronbach’s alpha)

Validity testing (EFA)

Hypothesis testing (regression)

Conclusion and implication

problem After that, this study conducted a literature review to review some relevant theories about factors influencing customer satisfaction to find out the suitable one

to Vietnamese group buying model and built the hypotheses for this study From this, a preliminary questionnaire was developed basing on questions used in some previous studies Next step is research design with 2 sub-steps:

 Pilot study: a study was conducted by interviewing around 10 customers face to face about the content, the number and the structure of questions

in preliminary survey to test the survey and measure before launching the main survey

 Main survey: was conducted in many methods such as sending online survey through mail, social network, or sending the hard copy of survey directly to customer of group buying service Data collection was done two weeks later

After that, collected data was cleaned and used to test reliability of scale and validity of questionnaire through Cronbach’s alpha coefficient and Exploratory Factor Analysis (EFA) method Multiple regression method was used to evaluate the hypotheses which the implication and finding were stated and reported

2 Research sampling method

The respondents are selected through purposive sampling Purposive sampling used is judgment According to Sekaran (2006), judgment involved the choice of subjects who are in the best position to provide the information using non probability methods Therefore, group of people who have knowledge about particular problem can be selected as the sample element The target respondent is group buying customer because after they try service, they can evaluate it exactly from customer perspective

According to DeCoster, J (2004), minimum sample size used in statistic analysis should be is equal to or greater than five times of the number of independent variables, but not less than 100 to generate reliable results:

n >= 100 and n >= 5k (where k is the number of variables)

This research has 47 variables, as a result, the minimum sample size required

to run EFA in this research is:

n = 5 x 47 = 235

Otherwise, Tabachnick & Fidell (2001) stated that the minimum sample size

in case of multiple regression should be:

n = 50+8m (where m is the number of independent variables)

Apply this formula for 9 independent variables of this research, we have the minimum sample size for multiple regression as follow:

n = 50 + 8 x 9 = 122

Summarily, with 47 variables and 9 independent variables, this research needs 235 samples at least to take EFA and regression Hence, 300 surveys were distributed to target respondents to take the data for analysis

3 Measurement scales

The questionnaire consists of 56 survey questions which are structured into three sections: section A, B and C Nominal scale is used in section A and C with many mutual exclusive and collectively exhausted categories (Sekaran, 2006) Section C (5 questions) includes respondent demographic details such as gender, age, level education, income level Section A includes 4 questions of buying habit such as frequency to buy deal, product/service bought usually About section B, ordinal scale is used as a measurement with multiple choices for respondents to

choose in 47 questions The purpose of the questionnaires is to obtain the major information which is described in part B such as service quality, price, brand and trust In section B, the respondents are asked to tick one answer through questions at the space provided The respondents were also asked to indicate their degree of agreement or disagreement for each item in this section A five point Likert scale will be used with 1 = “strongly satisfied” to 5 = “strongly dissatisfied” Table 3.1 has shown the items of the questionnaire

Table 3.1 Survey item summary

variable Scale

Section A: General information

1 Frequency of group buying purchase 1 Nominal

2 The most frequent online purchase item(s) 1 Nominal

3 Budget for each purchase time 1 Nominal

4 Group buying sites has been joined 1 Nominal

Section B: Evaluation of customer satisfaction

1b System availability 4 Likert – 5 levels

Section C: Personal information

The questionnaire was adapted and developed in English, then was translated into Vietnamese to distribute to respondents

4 Data collection procedure

The executing method of data collection for this research is quantitative The responses of respondents through questionnaire will be the primary data for this research 300 questionnaires were distributed using e-mail, facebook or directly in hard copies on October 15th, 2012 Two weeks later, 287 responses were received in both online survey and hard copies, and finally 270 samples were useable because

17 responses were rejected due to uncompleted answers

5 Data analysis method

SPSS Version 20 (Statistical package for Social Science) was used to analyze the collected data in which Cronbach’s alpha reliability analysis, EFA and multiple regression analysis function were enabled to support this research

satisfaction Using 5-point Likert scale, when the score was equal to 3 referred to acceptable/neutral, the score was greater than 3 is considered good or agreement, whereas when the score was lower than 3, it is considered not good and need to be improved or as an disagreement

5.2 Reliability analysis

Cronbach’s alpha reliability analysis method was used to test the reliability

of the measurement scales The scales are reliable when Cronbach’s alpha coefficient of each scale is equal to or bigger than 0.7 (Pallant 2005) Besides, those items which have the value of Corrected Item-to-total Correlation less than 0.3 could be discarded in order to improve reliability level of the measurement scale

5.3 Exploratory Factor Analysis (EFA)

The primary objectives of EFA are to determine the number of factors affecting the measures and the strength of the relationship between each factor and each observed measure (DeCoster, J., 2004) In this research, Principal Component Analysis (PCA) was used as a method to extract the factors with Varimax as a rotation technique

The result is considered to be acceptable when following conditions are met (Pallant, 2005):

 The sample size should be equal to or greater than 150, and there should

be at least five cases for each variable In this research, 47 variables are used, hence the required minimum sample size for EFA is 235 The valid responses are 270, therefore data of this research are considered valid for EFA

 Factor analysis is appropriate to data if:

 The Kaiser-Meyer-Olkin value (KMO) is 0.6 or greater

 The Bartlett’s test of sphericity is statistically significant: p < 0.05

 The number of factors is determined when:

 The components have an eigenvalue of 1 or more

 The total variance explained by these components should be above 50%

 Factor loading criteria should be 0.5 or more to ensure a practical significance

5.4 Multiple regression analysis

In this research, Multiple Linear Regression method was used to test the research model and hypotheses Pallant (2005) explains the conditions to accept the result are:

 The sample size is: n > 50 + 8m (where m is the number of independent variables)

 No multicollinearity is found

 Normality and linearity should exist

 Heteroskedasticity does not exist

We also use R-square value to express how much of the variance in the dependent variable was explained by the model

IV FINDING AND ANALYSIS

In this chapter, the evaluation of measures and the result of testing hypotheses are described There are four parts in this chapter: sample description, evaluation and refinement of measurement scales, hypotheses testing and result explanation

1 Sample description

300 surveys were sent to target respondents who used to join group buying websites to purchase product or service And there were 287 responses in which 17 were not valid because respondent did not answer over 30% of mandatory questions

or they filled in the same option for all questions As a result, 270 valid answers could be used to analyze, means that validity rate was 90% The sample characteristics are showed in the table 4.1

Table 4.1 Sample characteristics

Description Number of respondents Percentage (%)

Group buying purchasing frequency

Budget for 1 time purchase

of them were 26-35 years old, under 18 and over 35 years old people are few This result showed group buying service just focused on youth Mostly young people are interested in this kind of service

The highest levels of income were under 5 millions VND per month (38.1%) and from 5 to 10 millions VND per month (36.7%) It leaded to the result that most

of them spend not much money in a purchase time (about 100.000 VND to 500.000 VND), accounting for over 63% of the sample Very few respondents spent more than 500.000 VND for each purchase (over 12%) and next one was under 100.000 VND (23.3%) It showed a tendency towards buying normal and not much valuable products of online customers

The percentage of the respondents who purchased product or service on group buying sites for more than one time in a month was very high, accounting for

79.1% of the sample The remaining of the respondents (about 20%) were join group buying service one time or less than one time in six months or longer This may expressed a tendency towards building a long-termed purchasing relationship with loyal online customers

As shown in Table E.1 (Appendix E), mean scores of 9 variables were ranged from 2.42 to 3.56 on a five-point scale Efficiency was evaluated by customers at the highest mean score of 3.56, followed by brand (3.40), contact (3.35), system availability (3.25), price (3.22) and fulfillment (3.14) Responsiveness had the lowest mean score at 2.42, followed by trust (2.85) and privacy (2.96) These results indicated that brand, contact, system availability and price were considered as the items that need to further enhance Specially, responsiveness, trust and privacy were the things of group buying sites that really need to be improved because of the mean scores were behind acceptable level (3) Ability to offer a safe service to online customers is one of the most difficulties of online purchasing web sites in Vietnam Due to the limited privacy to reduce risk, they also miss customers’ trust and are facing problem to attract customers to join The target set for online purchasing web sites is improving security program for sites to make customers feel safe in transactions Moreover, customer care service after purchasing is still low, it caused responsiveness had the lowest mean score in customers Usually, web sites would no longer take care customers after they sold out product/service or vouchers So responsiveness also needs to be improved much

2 Evaluation and refinement of measurement scales

This research will use Cronbach Alpha Reliability and Exploratory Factor Analysis (EFA) to assess the reliability of measurement scales and the validity of sample data

2.1 Cronbach Alpha Reliability Analysis

This research will use Cronbach’s alpha to examine the reliability of variables in the questionnaire through following coefficients:

 Cronbach’s alpha coefficient: the scale is reliable when this coefficient is 0.6 or more

 Corrected Item – Total correlation: variables are acceptable when this coefficient is 0.3 or more

Base on reliability test results in table 4.2, all the measurement scales have the Cronbach’s alpha coefficient larger than 0.6 In addition, all the Corrected Item-Total Correlation values were over 0.3 Therefore, the scale designed in this research is meaningful in statistic and has the necessary reliability It also means that those four factors of customer satisfaction have enough conditions to be used in linear regression analysis later

Table 4.2 Cronbach’s alpha reliability test result

Observed

ittems

Scale Mean

if Item Deleted

Scale Variance if Item Deleted

Corrected Total Correlation

Item-Cronbach’s Alpha if Item Deleted

Efficiency: Cronbach’s alpha = 0.824

2.2 Exploratory Factor Analysis (EFA)

According to the research model, there are 9 factors with 43 observed variables that influence to customer satisfaction of group buying service After surveying, EFA method with Varimax rotation was used to analyze 43 observed variables of these 9 factors and 4 variables of customer satisfaction factor KMO (Kaiser-Meyer-Olkin) test and Bartlett test were also used to measure the compatibility of sample

2.2.1 EFA for independent variables

Table 4.3 KMO and Bartlett’s Test for independent variables

Kaiser-Meyer-Olkin Measure of Sampling Adequacy .911

Bartlett's Test of Sphericity

Approx Chi-Square 6857.424

It revealed that the KMO values of all factors were bigger than 0.7 and the Bartlett’s Test significance is smaller than 0.05 (.000) This meant observed variables had the correlation with each other and EFA factor analysis was appropriate because these results met the conditions required by EFA method Below table described the factor distribution after the first EFA analysis:

Table 4.4 Rotated Component Matrix

Component

1 2 3 4 5 6 7 8 9 10 TRUST2 741

Extraction Method: Principal Component Analysis

Rotation Method: Varimax with Kaiser Normalization

a Rotation converged in 25 iterations

Numbers in Rotated Component Matrix described the largest weight of factors

or factor loading of each variable In order to EFA factor analysis was considered importantly and had the practical significance, we only keep observed variables which had the factor loading equal or larger than 0.5 As a result, variables that had the factor loading smaller than 0.5 was removed one by one Below is the result of that removal process:

Table 4.5 Removal process during EFA

Time Total of

analyzed

Removed variables KMO

measure Sig Number of analyzed