4.4.1 Analyzing relationship between demographics factors with service providers.
To conduct the analysis of demographic factors that affect customer loyalty, the author used Chi-square test to assess the relationship combined with Cross Tabulation procedures. The relationship between demographic
factors with customer loyalty makes sense if the alpha level of significance of Chi-Square test is less than 0.05. In SPPS, Chi-Square will have meaningful conclusions when Sig value <0.05.
Gender
Results of statistical analysis showed that the percentage of male subscribers are higher than female subscribers. Specifically, Viettel subscribers accounted 57.6% for male and 42.4% for female. Vinaphone subscribers are 57.6% for male and 42.4% for female. Mobifone subscribers
Table 4.10
Relationship Between Gender with Service Providers
Gender
Service Providers
Total Viettel Vinafone Mobifone Others
Male Count 99 72 43 17 231
% 57.6% 57.6% 58.1% 58.6% 57.8%
Female Count 73 53 31 12 169
% 42.4% 42.4% 41.9% 41.4% 42.2%
Total
Count 172 125 74 29 400
% 100.0% 100.0% 100.0% 100.0% 100.0%
Chi-Square Tests
Value df Asymp. Sig.
(2-sided)
Pearson Chi-Square .017(a) 3 .999
Likelihood Ratio .017 3 .999
Linear-by-Linear Association .014 1 .906
N of Valid Cases 400
are 58.1% for male and 41.9% for female. Othe providers (Gmobile, Sfone and Vietnammobile) are 58.6% for male and 41.4% for female. Thus, there is not difference between the gender of subscribers in choosing service providers. The result of Chi-square test indicated that, there is not relationship between the gender of subscribers in choosing service providers due to the significance of the test is greater than 0.05 (Sig = 0.999 ).
Education level
The analytical results showed that subscribers who are educated from college or higher are using the big service providers (Viettel, Vinaphone, Mobifone). Result of Chi-square test also showed that there was relationship between the education of subscriber in choosing service providers due to the value of the Chi-square test was less than 0.05 (Sig = 0.000). In the process of gathering information, the author found that subscribers with high education who have a stable job need regular contact. Therefore, they usually choose big service providers who have the best quality service such as large coverage, high call quality, etc. Other service providers (Sfone, Gmobile and Vietnammobie) are small providers and new entering mobile communication market. These service providers have advantage in providing value – added servicers, cheap rates and many promotions but they have limitations such as small coverage and call quality is not stable, so the subscribers who have high education are less selective of these service providers.
Table 4.11
Relationship Between Education of Subscribers with Selection Service Providers
Education level Service Providers
Total Viettel Vinafone Mobifone Others
Below than high school diploma
Count 3 4 0 4 11
% 1.7% 3.2% .0% 13.8% 2.8%
High school diploma
Count 9 11 12 8 40
% 5.2% 8.8% 16.2% 27.6% 10.0%
Associate degree Count 78 59 29 2 168
% 45.3% 47.2% 39.2% 6.9% 42.0%
Bachelor degree Count 72 43 30 11 156
% 41.9% 34.4% 40.5% 37.9% 39.0%
Master degree an upper
Count 10 8 3 4 25
% 5.8% 6.4% 4.1% 13.8% 6.2%
Total Count 172 125 74 29 400
% 100.0% 100.0% 100.0% 100.0% 100.0%
Chi-Square Tests
Value df Asymp. Sig. (2-sided)
Pearson Chi-Square 45.810(a) 12 .000
Likelihood Ratio 43.026 12 .000
Linear-by-Linear
Association 4.110 1 .043
N of Valid Cases 400
Occupation
During the survey, author found that many customers have been owning two or more subscriptions. But, only one of the subscriptions was actually active and the remaining are subscribing to promotion (SIM “junk”).
Subscribers often operated to concentrate in big service providers (Viettel, Vinaphone and Mobifone). While subscribers used one time only concentrate in the small service providers (Vietnammobile, Gmobile and Sfone). The reason was that small suppliers have promotions for new subscribers than big service providers. However, the infrastructure of small service providers were limited so the subscribers did not continue to use their services after the
promotion ended. Therefore, there were many customers who become new subscribers of small service providers. These subscribers did not have a different occupation. The statistical result showed that there is not relationship between occupation of subscribers with selection of service providers due to value of Chi-square is 0.546.
Table 4.12
Relationship Between Selections of Service Provider with Occupation of the Subscribers
Occupation Service Providers
Total Viettel Vinafone Mobifone Others
Manage Count 18 9 6 6 39
% 10.5% 7.2% 8.1% 20.7% 9.8%
Worker Count 35 18 11 8 72
% 20.3% 14.4% 14.9% 27.6% 18.0%
Office worker Count 59 43 27 9 138
% 34.3% 34.4% 36.5% 31.0% 34.5%
Pupil, student Count 3 5 2 1 11
% 1.7% 4.0% 2.7% 3.4% 2.8%
Self – employed Count 38 35 19 4 96
% 22.1% 28.0% 25.7% 13.8% 24.0%
Others Count 19 15 9 1 44
% 11.0% 12.0% 12.2% 3.4% 11.0%
Total Count 172 125 74 29 400
% 100.0% 100.0% 100.0% 100.0% 100.0%
Chi-Square Tests
Value df Asymp. Sig.
(2-sided)
Pearson Chi-Square 13.730(a) 15 .546
Likelihood Ratio 13.558 15 .559
Linear-by-Linear Association .409 1 .523
N of Valid Cases 400
Per average income
The analytical results show that there is relationship between income of subscribers with service providers due to the significance of the Chi-square
test is less than 0.05 (Sig = 0.000). Specifically, big service providers’
subscribers having an income from 5 million or more have higher rates than small service providers. This reflects the real situation. Because subscribers with high incomes (manager, office worker, self employee, etc) who often have to communicate much due to the requirements of their jobs. Pupils, students and worker are low – income subscribers or no income so they often choose small services providers which have many promotions and low cost services.
Table 4.13
Relationship Between Selection of Service Provider With Per Average Income of Subscribers
Per Average Income Service Providers
Total Viettel Vinafone Mobifone Others
Under 3 Count 8 15 5 6 34
% 4.8% 12.5% 6.9% 21.4% 8.8%
3-5 Count 55 47 24 9 135
% 32.7% 39.2% 33.3% 32.1% 34.8%
5-7 Count 79 45 19 8 151
% 47.0% 37.5% 26.4% 28.6% 38.9%
Over 7 Count 26 13 24 5 68
% 15.5% 10.8% 33.3% 17.9% 17.5%
Total Count 168 120 72 28 388
% 100.0% 100.0% 100.0% 100.0% 100.0%
Chi-Square Tests
Value df Asymp. Sig.
(2-sided)
Pearson Chi-Square 31.653(a) 9 .000
Likelihood Ratio 29.167 9 .001
Linear-by-Linear
Association .454 1 .500
N of Valid Cases 388
4.4.2 Relationship between factors affecting loyalty versus demographic factors
Gender
Analytical result showed that gender of subscribers have relationship with “value – added service” and “adapting costs” due to values of Chi-square test are less than 0.05. The remaining factors affecting loyalty have no relationship with the subscriber's gender due to values of Chi-square test are higher than 0.05. (Appendix 3.1)
Specifically, the value of Chi-square on the relationship between gender of subscribers with value added services of the provider is 0.017. This suggests that the perceived quality of the value-added services of providers with subscriber's gender has a relation. Assessable quality value – added services of male subscribers is higher than female subscribers. Rate disagreement about the quality value-added service of providers is 7.8% for male and 14.2% for female (in which; strongly disagree is 2.4% and disagree is 11.8%).
Table 4.14
Relationship between Value - Add Service with Gender of Subscribers
Gender
Value - added service
Total Strongly
disagree Disagree Normal Agree Strongly agree
Male Count 0 18 121 88 4 231
% .0% 7.8% 52.4% 38.1% 1.7% 100.0%
Female Count 4 20 72 65 8 169
% 2.4% 11.8% 42.6% 38.5% 4.7% 100.0%
Total Count 4 38 193 153 12 400
% 1.0% 9.5% 48.2% 38.2% 3.0% 100.0%
Chi-Square Tests
Value df Asymp. Sig.
(2-sided)
Pearson Chi-Square 12.015a 4 .017
Likelihood Ratio 13.409 4 .009
Linear-by-Linear Association .106 1 .744
N of Valid Cases 400
The value of Chi-square on the relationship between gender of subscribers with adapting cost of the provider is 0.012. This suggests that the perceived adapting cost of subscribers with subscriber's gender is related.
Assessable adapting cost of male subscribers is higher than female subscriber’s. Rate of agreement is 44.2% for male (in which, agree is 33.8%
and strongly agree is 10.4%). This rate of female is 38.4% (in which, agree is 19.5% and strongly agree is 18.9%).
Table 4.15
Relationship between “Adapting Cost” with Gender of Subscribers
Gender
Adapting cost
Total Strongly
disagree Disagree Normal Agree Strongly agree Male
Count 1 3 125 78 24 231
% .4% 1.3% 54.1% 33.8% 10.4% 100.0%
Female Count 1 3 100 33 32 169
% .6% 1.8% 59.2% 19.5% 18.9% 100.0%
Total Count 2 6 225 111 56 400
% .5% 1.5% 56.2% 27.8% 14.0% 100.0%
Chi-Square Tests
Value df Asymp. Sig. (2-sided)
Pearson Chi-Square 12.863a 4 .012
Likelihood Ratio 13.060 4 .011
Linear-by-Linear Association .070 1 .791
N of Valid Cases 400
Occupation
The results of analysis showed that there are no relationship between subscriber’s occupation with factors affecting loyalty as follows: “not losing when your messages send and receive”, value – add service and move – in cost due to significance level of Chi-square tests are higher than 0.05. The remaining factors are related with occupational subscribers due to significance level of Chi-square tests are less than 0.05 (appendix 3.2).
For example, significance level of Chi – square test relationship between subscriber’s occupations with “calling quality is clear” of subscriber and “no call dropping when the call is done” are 0.000. This suggests that between subscriber’s occupation with “calling quality is clear” and “no call dropping when the call is done” of subscriber are related. Where by, subscribers who are managers, office workers and self employed assess “call quality clearly” higher than subscribers who pupil and student. Managers, office workers and self employed assess “no call dropping when the call is done” higher than pupils, students and workers.
Per average income
The results of analysis showed that there are no relationship between subscriber’s income with factors affecting loyalty as follows: “not losing when your messages send and receive”; convenience in procedure; value –added service, customer support service and move – in cost due to significance level of these test are higher than 0.05. The remaining factors affecting loyalty are
related with subscriber’s income due to significance level of Chi-square tests are less than 0.05. (appendix 3.3).
For example, there are relationships between subscriber’s income with perception in call quality of service providers of subscribers due to the significance level of Chi-square test is 0.000. Subscribers who have income over 5 million VND assess call quality higher than other subscribers.
Specifically, rate agreement about the call quality of the subscribers who have income under 3 million VND is 14.7% (in which,, agree is 14.7% and strongly agree is 0%). With subscribers who have income from 5 to 7 million VND, the rate is 40.8% (in which, agree is 38.4% and strongly agree is 2.6%).
Subscribers having income over 7 million VND, the rate is 48.5% (in which, agree is 35.3% and strongly agree is 13.2%).The survey showed that subscriber groups having income under 3 million VND so their perception in call quality of the services providers were limited. On the other hand, the provider services in Vietnam attract participation of clients by the promotion:
free roaming charges, plus cash in account for new subscribers, free on-net calls, etc. Hence, the low-income customers will often intend to change providers in order to enjoy more preferential policies from providers for new subscribers.
Meanwhile, the group of high-income customers are often customers who do office work, the manager, etc. These are customers with high levels of education and are more aware of the quality of service mobile communication service providers.
Table 4.16
Relationship Between Calling Quality of Service Providers with Per Average Income of Subscribers
Per average income
Calling Quality
Total Disagree Normal Agree Strongly
agree
Under 3 Count 8 21 5 0 34
% 23.5% 61.8% 14.7% .0% 100.0%
3-5 Count 8 82 42 3 135
% 5.9% 60.7% 31.1% 2.2% 100.0%
5-7 Count 5 84 58 4 151
% 3.3% 55.6% 38.4% 2.6% 100.0%
Over 7 Count 5 30 24 9 68
% 7.4% 44.1% 35.3% 13.2% 100.0%
Total Count 26 217 129 16 388
% 6.7% 55.9% 33.2% 4.1% 100.0%
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Pearson Chi-Square 41.656a 9 .000
Likelihood Ratio 34.201 9 .000
Linear-by-Linear Association 19.237 1 .000
N of Valid Cases 388
Ages
The results of analysis showed that there are relationship between subscribers’ ages with the adapting cost of service providers due to the significance level of Chi-square is 0.003. Remaining factors affecting loyalty are not related with subscriber’s ages due to significance level of Chi-square tests are higher than 0.05 (appendix 3.4).
With adapting cost, the higher age of the subscribers, the higher the assessment on adapting cost are because they have the ability to update the information of service providers. So, they were having difficulty in selecting a
service provider when they want to use mobile communication services.
During the survey process, the author found that subscribers who were over 55 of age received guidance on choosing service providers from young people. Analytical result showed that agree and strongly agree rates in assessment about adapting cost of subscribers are over 55 of age to be 60%
(in which agree rate is 20% and strongly rate is 40%). These rates of subscribers from ages 18 to 24 are 37% (in which agree rate is 33.3% and strongly rate is 3.7%).
Table 4.17
Relationship Between the Adapting Cost with Subscriber’s Ages
Ages
Adapting cost
Total Strongly
disagree Disagree Normal Agree Strongly agree 18-24
Count 1 2 14 9 1 27
% 3.7% 7.4% 51.9% 33.3% 3.7% 100.0%
25-34 Count 0 1 96 41 16 154
% .0% .6% 62.3% 26.6% 10.4% 100.0%
35-44 Count 0 0 70 35 24 129
% .0% .0% 54.3% 27.1% 18.6% 100.0%
45-54 Count 1 3 39 23 9 75
% 1.3% 4.0% 52.0% 30.7% 12.0% 100.0%
Over 55 Count 0 0 6 3 6 15
% .0% .0% 40.0% 20.0% 40.0% 100.0%
Total Count 2 6 225 111 56 400
% .5% 1.5% 56.2% 27.8% 14.0% 100.0%
Chi-Square Tests
Value Df Asymp. Sig.
(2-sided)
Pearson Chi-Square 36.362a 16 .003
Likelihood Ratio 30.988 16 .014
Linear-by-Linear Association 5.552 1 .018
N of Valid Cases 400
Education level
The results of analysis showed that there are relationship between subscriber’s education with factors affecting loyalty (calling quality is clear, no call dropping when the call is done, the scope of coverage and the supplier has service packages with different charges to suitable customer demands) due to the significance level of Chi-square tests are less 0.05. Remaining factors affecting loyalty are not related with subscriber’s ages due to significance level of Chi-square tests are higher than 0.05 (appendix 3.5).
For example, when evaluating about “the supplier has service packages with different charges to suitable customer demands”, subscribers having education level as high school diploma, Associate degree, Bachelor degree and Master degree and higher highly agreed.
Specifically, agree and strongly agree rates of subscriber with high school diploma are 82.5%; subscribers with associate degree are 80.4%;
subscribers with bachelor degree are 86.5% and subscribers with master degree and higher are 80%. This rates of subscriber with below than high school diploma are 54.6%. This shows that subscribers with high education subscribers with higher education have better perception about mobile communication service providers of subscribers compared with subscribers with low education level. The result of Chi – Square test showed that, there are relationship between education levels of subscribers with perceptive subscribers about “the supplier has service packages with different charges to
suitable customer demands”. The significance level of test is 0.017 to reflect this relationship have significance.
Table 4.18
Relationship Between Assessment About “The Supplier Has Service Packages With Different Charge to Suitable Customer Demands” With Education Level of Subscribers
Education
The supplier has service packages with different charges to suitable customer
demands Total
Disagree Normal Agree Strongly agree Below than high
school diploma
Count 2 3 3 3 11
% 18.2% 27.3% 27.3% 27.3% 100.0%
High school diploma
Count 2 5 25 8 40
% 5.0% 12.5% 62.5% 20.0% 100.0%
Associate degree Count 2 31 104 31 168
% 1.2% 18.5% 61.9% 18.5% 100.0%
Bachelor degree Count 10 11 95 40 156
% 6.4% 7.1% 60.9% 25.6% 100.0%
Master degree an upper
Count 2 3 15 5 25
% 8.0% 12.0% 60.0% 20.0% 100.0%
Total Count 18 53 242 87 400
% 4.5% 13.2% 60.5% 21.8% 100.0%
Chi-Square Tests
Value df Asymp. Sig.
(2-sided)
Pearson Chi-Square 24.532a 12 .017
Likelihood Ratio 24.641 12 .017
Linear-by-Linear Association 1.415 1 .234
N of Valid Cases 400
4.4.3 Analyzing the Factors that Affect Customer Loyalty
Theoretical model of the factors affecting customer loyalty of which the author assumed in this study consists of two main factors: customer
satisfaction (factor of 5 components: calling quality, pricing structure, value – add services, convenience in procedures and customer services); and switching barriers (including 5 component factors: loss cost, adapting cost, move - in cost, other attractive provider, customer relationship.
In this study, multiple linear model was used to discuss factors affecting customer loyalty.
CL = intercept + ò1 * (CQ) + ò2 *(PS) + ò3 *(VAS) +ò4 *(CP)+ ò5 * (CSS) + ò6 *(LC) + ò7 *(AC) +ò8 *(MC)+ ò9 * (AOP) + ò10 *(CR) + ε
In that:
CQ Calling quality PS Pricing structure
VAS Value – add services
CP Convenience in procedures CSS Customer support services LC Loss cost
AC Adapting cost MC Move – in cost
AOP Attractiveness of other providers CR Customer Relationship
CL Customer loyalty
Table 4.19 Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .859a .737 .737 .41500
2 .912b .832 .831 .33216
3 .926c .858 .857 .30632
4 .936d .877 .875 .28564
5 .944e .892 .890 .26787
6 .949f .900 .899 .25712
7 .950g .903 .902 .25375
ANOVA
Model Sum of Squares df Mean Square F Sig.
1
Regression 192.451 1 192.451 1.117E3 .000a
Residual 68.546 398 .172
Total 260.998 399
2
Regression 217.198 2 108.599 984.334 .000b
Residual 43.800 397 .110
Total 260.998 399
3
Regression 223.840 3 74.613 795.185 .000c
Residual 37.157 396 .094
Total 260.998 399
4
Regression 228.769 4 57.192 700.969 .000d
Residual 32.228 395 .082
Total 260.998 399
5
Regression 232.727 5 46.545 648.694 .000e
Residual 28.270 394 .072
Total 260.998 399
6
Regression 235.016 6 39.169 592.475 .000f
Residual 25.982 393 .066
Total 260.998 399
7
Regression 235.757 7 33.680 523.068 .000g
Residual 25.240 392 .064
Total 260.998 399
a. Predictors: (Constant), Customer relationship
b. Predictors: (Constant), Customer relationship, Loss cost
c. Predictors: (Constant), Customer relationship, Loss cost, Calling quality
d. Predictors: (Constant), Customer relationship, Loss cost, Calling quality, Attractiveness of other providers
e. Predictors: (Constant), Customer relationship, Loss cost, Calling quality, Attractiveness of other providers, Convenience in procedure
e. Predictors: (Constant), Customer relationship, Loss cost, Calling quality, Attractiveness of other providers, Convenience in procedure
e. Predictors: (Constant), Customer relationship, Loss cost, Calling quality, Attractiveness of other providers, Convenience in procedure
f. Predictors: (Constant), Customer relationship, Loss cost, Calling quality, Attractiveness of other providers, Convenience in procedure, Pricing structure g. Predictors: (Constant), Customer relationship, Loss cost, Calling quality, Attractiveness of other providers, Convenience in procedure, Pricing structure, Value - add service
h. Dependent Variable: Customer loyalty
According to the analysis result, there are 7 model to reflect factors affecting on customer loyalty:
Model 1: The customer loyalty is impacted by “customer relationship”.
The adjuster R2 is 0.737 so variables in this model are explained by 73.7%.
The significant level of the F-test and t-test are less than 0.05 (Sig <0.05) so this model is significant.
Model 2: The customer loyalty is impacted by “customer relationship”
and “loss cost”. The adjuster R2 is 0.831 so variables in this model are explained by 83.1%. The significant level of the F-test is 0.000 (Sig = 0.000).
However, the value test of the “intercept” is higher than 0.05 (Sig = 0.751) so this model is not significant.
Model 3: The customer loyalty is impacted by “customer relationship”,
“loss cost” and “calling quality”. The adjuster R2 is 0.857 so variables in this model are explained by 85.7%. The significant level of the F-test and t-test are less than 0.05 (Sig <0.05). VIF value is less than 10 that it means the model has no multi-collinearity phenomenon. So, this model is significant.
Model 4: The customer loyalty is impacted by “customer relationship”,
“loss cost”, “calling quality” and “attractive other providers”. The adjuster R2 is 0.875 so variables in this model are explained by 87.5%. The significant level of the F-test is 0.000 (Sig = 0.000). However, the value test of the “intercept”
is higher than 0.05 (Sig = 0.055) so this model is not significant.
Model 5: The customer loyalty is impacted by “customer relationship”,
“loss cost”, “calling quality”, “attractive other providers” and “convenience in procedure”. The adjuster R2 is 0.890 so variables in this model are explained by 89.0%. The significant level of the F-test is 0.000 (Sig = 0.000). However, the value test of the “intercept” is higher than 0.05 (Sig = 0.487) so this model is not significant.
Model 6: The customer loyalty is impacted by “customer relationship”,
“loss cost”, “calling quality”, “attractiveness of other providers”, “convenience in procedure” and “pricing structure”. The adjuster R2 is 0.899 so variables in this model are explained by 89.9%. The significant level of the F-test and t- test are less than 0.05 (Sig <0.05). VIF values are less than 10 that is mean the model is not multi-collinearity phenomenon. So, this model is significant.
Model 7: The customer loyalty is impacted by “customer relationship”,
“loss cost”, “calling quality”, “attractive other providers”, “convenience in procedure”, “pricing structure” and “value - added service”. The adjuster R2 is 0.902 so variables in this model are explained by 90.2%. The significant level of the F-test and t-test are less than 0.05 (Sig <0.05). VIF value is less than 10 that is mean the model has no multi-collinearity phenomenon. So, this model is significant.
Thus, Models 1, 4, 6, 7 are high significant. However, the adjuster R2 of the model 7 has the highest value.
The model of analysis affecting customer loyalty in mobile communication services in Hanoi is reflected as follows:
CL = -0.404 + 0.173 * (CQ) + 0.163 *(PS) + 0.097 *(VAS) +0.162*CP+ 0.178
*(LC) - 0.140 * (AOP) + 0.260 *(CR)
According to this model, there are 7 factors affecting customer loyalty.
The factors effecting level are explained as follows:
The regression coefficient of “customer relationship” is 0.260 (ò10
=0.260) so this factor is the strongest influence to customer loyalty. The “loss cost” is the second factor affecting customer loyalty (ò6=0.178). The “calling quality” is third factor affecting customer loyalty. The regression coefficient of
“pricing structure” is 0.163 so this is the fourth factor affecting customer loyalty. The fifth factor is “Convenience in procedure” (ò4= 0.162). The
“attractiveness of other providers” is the sixth factor affecting customer loyalty (ò9 = -0.140). However, the negative regression coefficient reflects inversely proportional relationship between “attractiveness of other providers” with
customer loyalty. The regression coefficient of “value – add services” is 0.097 so this factor is the weakest influence to customer loyalty.
According to Pham Duc Ky and Bui Nguyen Hung (2007), customer loyalty was affected by six factors (calling quality, pricing structure, convenience in procedures, service customer support services, adapting cost and attractiveness of other suppliers). In which, the “calling quality” was the strongest influence customer loyalty and “customer support services” was the lowest factor affecting. However, the “adapting cost” and the “customer support service” are two factors that do not affect customer loyalty in this research. The other factors are calling quality, pricing structure, convenience in procedures and attractiveness of other suppliers will continue to be the factors that affect customer loyalty. But, factors effecting level have changes.
Three factors affecting customer loyalty are added in this research as customer relationship, loss cost and value – added services. In particular, the
“customer relationship” is strongest influence to customer loyalty. This reflects actual situation on mobile communication service in Hanoi where service providers have used new technology to develop value-added services to meet the growing needs of customers. But the value – added services of suppliers are not varying. Customers usually confused in identifying value – added services of suppliers. So, the “customer relationship” will be the factor to help customers who can identify value – added services.