Can Customer Intimacy Strategy Generate Intention to Buy?

Creating customer intention to buy is obviously a major task of every marketer and/or firm. Many tactics are exercised to generate the intention, in which a buying behavior is hopefully occured. A customer intimacy strategy supposedly be a particular way to do it. However, its power to generate the intention hypothetically is not straight forward, but through other variables. It is commonly known, in accordance with the Theory of Planned Behavior (TPB), the intention could be predicted by consumer attitude and subjective norm. Meanwhile, the attitude and subjective norm theirself are frequently in-line with the product’s performance. Therefore.the purpose of the study is to investigate the power of customer intimacy strategy in creating the customer intention to buy through the product’s brand equity and both the consumer attitude and subjective norm. A 108 sample is withdrawn from those who recognize, are interested of and want to buy Dagadu products. Amos 16.0 and SPSS 16.0 are employed in analyzing data. The result shows that the customer intimacy strategy has significant effects to the brand equity, attitude and subjective norm. In addition, the brand equity also has a significant influence to the intention.


H8:
The more favorable the Subjective Norm is, the greater the Behavioral Intention will be.
f. Effect of the Hypotheses already Formulated: an intervene position of the Attitude and Subjective Norm It is hypothesized that brand equity affects the behavioral intention. Further, it is hypothesized that brand equity affects both attitude and subjective norm. While it is hypothesized as well that whether attitude or subjective norm affects behavioral intention, consequently, both attitude and subjective norm likely post as mediator. Therefore, next hypotheses can be drawn as follows: H9: Attitude mediates the relationship between brand equity and behavioral intention H10: Subjective norm mediates the relationship between brand equity and behavioral intention

Research Model
Based on the hypotheses a research model can be developed as follows in Figure 1:

Methods
The population of the study is consumers who know Dagadu, are interested of and want to buy the products, and live at Central Java, Indonesia. A sample is drawn using the convenience and judgment technique (Cooper & Schindler, 2008). Data are collected by questionnaires, which consist of five items for the customer intimacy variable, four items for the brand equity variables, six items for the attitude variables, six items for the subjective norm variables, and four items for the behavioral intention. They are distributed to respondents who live at Semarang, Yogyakarta, and other cities at Central Java. After examining the forms for the data's completion, 108 out of the 110 questionnaire forms are accepted which supposed meet the sample adequacy (Ghozali, 2004(Ghozali, , 2007Hair et al., 1995).
A Likert scale is operated corresponding to a five-point scale ranging from 1 (=completely disagree) to 5 (=completely agree). The instrument, which denotes to indicators, will firstly be justified through confirmatory factor analysis, Construct Reliability and Variance Extracted. Further, data are analyzed by employing Amos 16.0.

a. Confirmatory Factor Analysis
The confirmatory factor analysis is not simultaneously carried out, but done in phases. The first phase contains two variables, i.e., Customer Intimacy (CI) and Subjective Norm (SN). The second phase examines two variables, attitude (Ab) and Behavioral Intention (BI). The third phase considers one variable, i.e., Brand Equity (BE). The process illustrated at Appendix A, while its result exemplified at

b. The Structural Equation Model
The model has one initial independents variable (CI) and four dependent variables (BE, Ab, SN, BI) in which the three dependent variables (BE, Ab, SN) at some extent are treated as independent variables as well. Since the purpose of the study is eagerly to know the relationship between the one initial independents variable (CI) and the primary dependent variables (BE, Ab, SN, BI), likewise among the four dependent variables separately and simultaneously, a Structural Equation Modelling (SEM) is employed (Hair et al., 1995). In addition, the use of SEM will give advantages such as fast, accurate and more detail. It is possible since the method performs a unification of factor analysis and path analysis (Ghozali, 2004(Ghozali, , 2007).
An initial structural equation model is drawn by connecting all variables as hypothesized. This model is likely not thoroughly appropriate to expectancy, since all indicators, i.e., Chi-Square/Prob, Cmin/df, GFI, AGFI, TLI, RMSEA, do not meet the criteria (Appendix B). Consequently, a modification model is generated by connecting e1↔e2 and e3↔e4, This modification model seemingly produces better scores than before (Table 2, Figure 2).  Evaluation of Normality. Evaluation of normality is carried out by univariate test (Ferdinand, 2002;Ghozali, 2004). It is exercised by scrutinizing the skewness value whether its critical ratio values are less or equal to ±2.58. As a matter of fact, there is one variable, i.e., SN, whose c.r of the skewness value are more than ±2.58. As a consequent, it indicates that univariately the data distribution is not normal. To check further, a multivariate test is executed. The result of the data analysis shows up that the multivariate critical value is 18,937. It is more than 2.58 as required (Appendix C) . As a result, the normality test needs a bootstrap analysis.

Figure 2. The Modification Model
Outliers. Evaluation of the outliers can be carried out by either a univariate test or a multivariate test (Ferdinand, 2002). The univariate test is successfully employed by firstly converting the data to Z-scores, which should be less than ±3.0 (Hair et al., 1995). The result indicates that most of the variables' Z-scores are less than ±3.0, except BE1, ev3, NB2, and MC3,, which their scores are more than ±3.0 (Appendix E). Therefore, the existence of outliers is indicated.
To check further, a multivariate outliers test is needed. It determines the chi-square value which subsequently is used as the upper limit, which could be calculated by searching on a chi-square     Table 3 shows that the influence of CI to BE is denoted by p=0.000. It means that the influence of CI to BE is significant. Likewise the influences of CI to AB and CI to SN belong to be significant as well, since their probabilities are less than 0.05 (p=0.001 and p=0.004). The probabilities of BE to Ab, BE to SN, and BE to BI are also less than 0.05, indicating that the influence of the variables are significant (p=0.000, p=0.002, and p=0.006). While the influence of SN to BI is positively less than 0.05 (p=0.042), the influence of Ab to BI has probability more than 0.05 (p=0.088). However,it can be categorized to be significant when the threshold is altered from 0.05 to 0.10.

Discussion
Testing of intervene position indicates that whether the indirect effect of BE to BI through Ab, or the indirect effect of BE to BI through SN, is bigger than the direct effect. Consequently, both Ab and SN post as mediators.

Conclusion
The hypotheses of, i.e., "Customer intimacy influences brand equity (H1)", "Customer intimacy affects one's attitude (H2)" and "Customer intimacy affects one's subjective norm ( (2015), that support the theory of planned behavior, in which attitude and subjective norm are predictors of behavioral intention. This can be explained by the intention to buy, while being determined by attitude (Fishbein & Ajzen, 1975), and likewise shaped by the subjective norm, obviously suggests that whatever happens to the attitude or the subjective norm, the intention to buy apparently also follows, and the alteration of intention to buy is in accordance with the change of them.
The hypotheses of Ab and SN as mediators (H9, H10) are also supported. As a matter of fact, all hypotheses are successfully proven. The consequences of the study carries out two things, firstly that the findings contribute as a bridge of other previous studies. Secondly the study justifies the theory of Treacy and Wiersma (1997).
Back to the title of the manuscript, i.e., "Can Customer Intimacy Strategy Generate Customer Intention to Buy?" The answer is, yes and not. The meaning of yes is, that the effect of the customer intimacy strategy later on generates the behavioral intention, particularly intention to buy. Whereas the meaning of not is, the stategy could not directly generate the intention. However, it is empirically supported that customer strategy leads to the creation of behavioral intention, particularly intention to buy, whether through brand equity, through both brand equity-attitude and brand equity-subjective norm, or through both attitude and subjective norm.

Limitations and Future Directions
There are some limitations of the study, firstly, the customer intimacy is supposed operated by indicators as follows, the diversification of the product is in line with consumers' taste; the product's message is personal; customer oriented; managers, staffs and employees are responsive; and personalized program. They are fully genuine which are not employed in such topic beforehand.
Though based on CFA test they belong to valid indicators (Table 1), it is not impossible that other indicators might be employed which might contribute better results.
Secondly, it likely the model is not in accordance with the title. The customer intimacy variable is not directly regressed to the behavioral intention variable. It is based on such point of view as follows.
Since an intention to buy does not likely arise spontaneously but through something impressively, the brand equity variable is supposed worthy to trigger the intention. However, it might be possible, under such assumption, that the customer intimacy is regressed directly to the intention.
Thereby, it is recoomended to carry out such study which firstly, exploring other indicators of the customer intimacy variable. Secondly, developing other model that leads to regress directy the customer intimacy to the intention.

Appendixes
Appendix A