Mediating Effect of Brand Trust in the Relationship Between Marketing Activities and Purchase Intention Among the Malaysian Made Car Owners: A Pls-Sem Approach

As a citizen of a country, there is a need to support any local industry that will enhance the country’s economic situation. The execution of the country's first national project by Perusahaan AutomobilNasionalBerhad (PROTON) in 1985 and Perusahaan AutomobilKedua (Perodua) was an important step towards the improvement of the motor vehicle industry in Malaysia. However, local car manufacturers are struggling to sustain its sales and profit and the need to compete with the foreign made cars. Thus, this study aims to find out how Marketing Activities (i.e. Trendiness, Interaction, and Entertainment) could affect customers’ Purchase Intention and also Brand Trust. A survey was employed to collect data from 253 respondents who owned local made cars or have intention to purchase local made cars around Klang Valley. Majority of the respondents were from Petaling Jaya and Shah Alam area where judgmental sampling was used. A questionnaire which was developed and adapted from previous studies was used to collect data from the respondents using self-administered and online survey method. Data was developed and analysed using structural equation model where results of this study showed that Marketing Activities and Brand Trust does affect Purchase Intention. The analysis also confirms that Brand Trust partially mediates the relationship between Entertainment and Interaction toward Purchase Intention, whereas Brand Trust fully mediates the relationship between Trendiness and Purchase Intention.


Introduction
The execution of the country's first national project by Perusahaan AutomobilNasionalBerhad (Proton) in 1985 and Perusahaan AutomobilKedua (Perodua) was an important step towards the improvement of motor vehicle industry in Malaysia. However, the local car industry is struggling to sustain and need to compete with the foreign made cars. On the one hand, local automotive companies face a very challenging competition in automotive sector since there are many competitors who had launched their amazing products and attracted many consumers in Malaysia. On the other hand, local automotive industries should make enhancements to recapture the market in Malaysia and one of the approaches is by improving marketing strategies. In the rapid development of new technologies, Malaysian made car companies should never be left behind to experience new marketing strategies that can enhance their product or services. In the domestic market, Proton recorded sales of 100,010 units in financial year 2016 and DRB-HICOM's automotive sector reported vehicle sales growth of 227,223 units in financial year 2015 from 225,162 units in financial year 2014 (DRB-HICOM Berhad, 2015). While Perodua sold around 213,000 vehicles in 2015. With the weaker market conditions in 2016, the volume of Perodua dropped by 3% to around 207,000 units (UMW. Holdings Berhad, 2015).
Consumers will make a purchase when preferred brands can fulfil their needs to accomplish a particular lifestyle. There are certain findings from previous studies which demonstrated that customers are loyal to the brand when the brand experience fits well with their ways of life and social identity (Janghyeon et al., 2011). Those findings show that trendiness play an important role in developing consumer loyalty and it will make the purchase happen. In another study, by building up a brand that matches with the way of the consumer's life, it will make the consumer satisfied with that brand (Solomon, 2013). This study defines the trendiness characteristics such as different, new, innovative, trendy, up-to-date, imaginative, and contemporary.
In order to create the brand that may lead to higher levels of brand relationship quality for target customers, a research should be conducted to find the best social interactions to approach the customer because study by (Lovelock and Wirtz, 2010) stated that interaction as a key in conventional services marketing has been most obviously defined in the concept of the service encounter. This shows that interaction is one of the contributing factors of purchase behaviour. Interaction can be defined by two-way communication amongst company and clients and also among clients themselves (Goldfarb and Tucker, 2011).All the medium to approach customer were established by both managements of Proton and Perodua and it shows that the interaction between them and the customers was a very important phase in catching their target customers. This study defines interaction as information sharing and opinion exchange with others.
Entertaining advertisement that are perceived to be fun have a positive effect on attitude toward the advertisement (Taylor et al., 2011) and it proved that entertainment can give an attraction to customer when they are using the website of local made car. In addition, entertainment element is believed to be able to lead people towards consuming brand-related content online (Muntinga et al., 2011).According to UMW website, the UFirstPerodua Passport Mobile App is an application which offers entertainment and other general benefits, and it is the first automotive mobile app in Malaysia. It shows entertainment was one of the factors that can affect purchase intention since after PeroduaAxia was launched on 15 th September 2014, it was Perodua's top-selling model in 2015 (UMW. Holdings Berhad, 2015). Given the above discussion, this study defines entertainment as an event, performance, or activity designed to entertain others.
Brand trust is a facet of brand value that bonds consumers with the brand (Shin et al., 2016) and can be defined as the willingness of consumers as trustors to depend on the trustee, to cooperate with the trustee, and, more generally, the adoption of trusting behaviour (Pauline et al., 2017). A study by Yap et al. (2012), Kim A. J. and Ko (2010) stated that trust need to be obtained first to ensure customers have lasting relationships with a brand and when one party trusts another, it can exist the purchase intention and their study found that trust was influenced by entertainment and trend. The discussion in this paper therefore refers to brand trust as the willingness of consumer to rely on the stated function of the brand.
Buying behaviour of consumer can give some insight related to three things which are what product will survive, which industry to invest and which company to succeed (Wang et al., 2014). Study by Lin and Lekhawipat (2014); (Sanchez-Garcia et al., 2012) have proved that satisfied customer affected on repurchase behaviour. Referring to those statement, repurchase intention may be influenced by the brand trust and not by first time purchase. The companies need to guarantee their customers' trust to them with great marketing approach since it can affect sales to be higher. Besides that, study by Kim A. J. and Ko (2010) found that purchase intention was influenced by entertainment and interaction. In addition, companies should make their client satisfied and attracted to their product since customer trust plays a significant role in the decision making for purchasing the product (Gupta and Kim, 2010). In the light of this discussion, this study expects the marketing activities to have an impact on purchase intention.
Thus, this study wishes to investigate whether marketing strategies such as trendiness, interaction, and entertainment have an effect on the purchase intention of the buyers. This study wishes to focus on the trust of buyers towards the local made car industry due to its lack of competitive edge compared to the foreign made cars. The aim of this study is to investigate the effect between marketing activities, brand trust, and purchase intention, and to identify factors that could affect the purchase intention and the mediating effect of brand trust towards local made car.

Methodology
According to literature review, it's clear there are several types of marketing activities which affecting on purchase intention. Due to that, the following hypotheses have been developed: Basically, the target population for this study is a respondent that used a Malaysian local car either PROTON or Perodua car brands, since both car brands were classified as the Malaysian local cars (DRB-HICOM Berhad, 2017). This study applied self-administered questionnaires which were distributed to respondents at available showrooms of Perodua and Proton in Shah Alam, since this centre having service rooms for their car models. Online survey was also used in order to cover the respondents in Petaling Jaya area which this study intends to cover. The duration of data collection was up to almost 6 weeks. Judgmental Sampling was used and deemed suitable since there is no sampling frame criteria for choosing the respondent. This study developed survey items to measure the existing latent variables based on a 7-point Likert scale and the measurement items are adapted from previously validated constructs in journal about marketing research and this study re-validates these measures (Straub, 1989). The questionnaires consist of five sections. For Section A, the questionnaire asks about the demographic profile of respondent such as gender, age, monthly income, and current car used. For Section B, Section C, and Section D, the questionnaire asks questions related to marketing activities and adapted from (Kim A. J. and Ko, 2012). While, for Section E, the questionnaire asks questions related to mediating variable adapted from Kim J. H. and Hyun (2011) and Lau and Lee (1999). Lastly, for Section F, the questionnaire asks about purchase intention adapted from (Park et al., 2007). Figure 1 shows the conceptual framework analysed using SmartPLS 2.0 software.

Figure-1. Conceptual Framework
By using PLS-SEM theory, the measurement model need to be evaluated first before evaluating the structural model (Hair et al., 2017); (Chin, 1998). Assessing the convergent and discriminant validity are two key aspects in determining the quality of the measurement model (Hair et al., 2014); (Hair et al., 2011); (Fornell and Larcker, 1981). Criterion such as indicator loadings, Cronbach's Alpha reliability (α), Composite reliability (ρ), and Average Variance Extracted (AVE) are used in the assessment of convergent validity while Fornell-Larcker discriminant validity is the type of assessment used to evaluate the discriminant validity of the measurement model (Fornell and Larcker, 1981). As for accessing the significant influence of the variables, 5000 replications of samples (i.e. bootstrapping) was used. It is because PLS-SEM analysis basically is a non-parametric analysis and the estimation of probability values can't directly compute via standard procedure like Maximum Likelihood Estimation (Hair et al., 2017). Hence, based on the bootstrapping procedure, Bootstrap-t as well as t-statistics was computed. In addition, for measuring the effect of mediating, the following procedure for deciding the mediating effect was used (Zhao et al., 2010); (Iacobucci et al., 2007). The procedures are: 1) If the path of independent variable to dependent variable was not significant and the indirect effect is significant, hence the mediating effect was a full mediation effect. 2) If the path of independent variable to dependent variable was significant and the indirect effect is significant, hence the mediating effect was a partial mediation effect. Table 1 shows the result of respondent's profile. In implementing the process of collecting data, the researcher has collected around 253 respondents. This study reveals that 44% of the respondents were male and 56% were female. Besides that, the majority of them were between 18 to 27 years old (78%), 14% for 28 to 40 years old, and 8% for age above 40 years old. In income perspective, most of the respondents have an income of RM2001 to RM4000 (36%) followed by 30% from income below RM2000, 16% from income RM4001 to RM6000, 13% from income RM6001 to RM10000, and 5% from income of more than RM10000. Furthermore, all respondents indicated that the distribution was almost equal where 51% respondents own Perodua car and 49% own Proton car.  Table 2 shows the summary results of the convergent validity assessment for the measurement model based on theoretical framework. All indicators that were used to measure targeted constructs meet the minimum requirement of the loading value above .70 (Iacobucci et al., 2007) but one indicator was maintained in the analysis since the loading values were above .60 (Hair et al., 2017). Besides that, the assessment of Average Variance Explain (i.e. AVE) for each construct was above .50 (Hair et al., 2017), as well as both reliability tests (i.e. Composite Reliability and Cronbach's Alpha) for each targeted construct were also above .70 (Hair et al., 2017). Hence, it meets the validity from the aspects of unidimensionality concept. As for discriminant validity, Table 3 shows the result of Fornell-Larcker analysis for accessing the discriminant validity of the model. The analysis confirms that, each latent variable was totally discriminate to each other's since the diagonal values were greater than the off-diagonal values (Hair et al., 2011); (Hair et al., 2017). Therefore, the indicators that were used to measure targeted construct were totally used for the respective construct. Note: The value in the diagonal (bold) is a square root of the AVE of each latent variable and the element off-diagonal value is the inter-correlation value between latent variable.

Measurement Model Analysis
The analysis indicated that, Entertainment, Interaction, and Trendiness were able to explain about 52.8% of variance explained toward Brand Trust, whereas, in the simultaneous concept, these three variables with Brand Trust variable were also able to explain around 48.4% of variance explained toward Purchase Intention. In addition, the assessment of effect size (f 2 ) and predictive relevance (q 2 ) from each respective path can be considered as weak effect. The analysis indicated that, Entertainment (β = 0.478, t = 7.384, p<.05; 95% Bootstrap-t: (0.351, 0.606)), Interaction (β = 0.161, t = 2.304, p<.05; 95% Bootstrap-t: (0.024, 0.298)), and Trendiness (β = 0.208, t = 3.123, p<.05; 95% Bootstrap-t: (0.078, 0.339)) have a simultaneously positive significant effect toward Brand Trust. Therefore, if the average level of Entertainment, Interaction, and Trendiness were at good levels, then the level of Brand Trust will be high. The analysis also confirms that, Entertainment (β = 0.198,t = 2.443,p<.05;0.356)) and Interaction (β = 0.347,t = 4.317,p<.05;0.504)) have positively significant effect toward Purchase Intention, but not for the Trendiness (β = -0.035, t = 4.317, p =.672; 95% Bootstrap-t: (-0.198, 0.128)). Therefore, if the average level of Entertainment and Interaction were at good levels, then the level of Purchase Intention will high, but not for the Trendiness effect. In addition, the analysis also indicated that, Brand Trust (β = 0.313, t = 3.856, p<.05; 95% Bootstrap-t: (0.154, 0.472)) have a simultaneously positive significant effect toward Purchase Intention. Hence, if the average level of Brand Trust was at good levels, then the level of Purchase Intention will also be high. Figure 2 and Figure 3 shows the analysis of PLS-SEM.

Mediating Analysis
The indirect analysis indicated that, Brand Trust simultaneously mediated the relationship between Entertainment (IEC = 0.150,t = 3.156,p<.05;0.244)), Interaction (IEC = 0.050, t = 2.089, p<.05; 95% Bootstrap-t: (0.003, 0.097)), and Trendiness (IEC = 0.065, t = 2.203, p <.05; 95% Bootstrap-t: (0.006, 0.124)) toward Purchase Intention. Since the direct effect of Entertainment and Interaction toward Purchase Intention were significant, hence Brand Trust partially mediated the relationship between Entertainment and Interaction toward Purchase Intention, whereas Brand Trust fully mediated the relationship between Trendiness and Purchase Intention since the direct effect of Trendiness toward Purchase Intention was not significant.

Conclusions
This study had focus on the purchase intention of buyers towards the local made car industry. The results of this study are believed to give significant contribution for automotive companies that usually have various types of marketing activities for approaching their customers. Based on previous stated hypotheses, all the hypotheses was satisfied and only H 1 was not satisfied where Trendiness found have no significant effect on purchase intention. This study found that Marketing Activities which are Entertainment and Interaction affect Purchase Intention. It can be concluded that, when a company provides better interaction and entertainment to their customer it will encourage the customer to purchase their product (Goldfarb and Tucker, 2011). Besides that, interaction is important to ensure that customers are attracted to the company's product while providing some entertaining element in the company's social media platform. The analysis also confirms that, Brand Trust partially mediated the relationship between Entertainment and Interaction toward Purchase Intention, whereas Brand Trust fully mediated the relationship between Trendiness and Purchase Intention (Kim A. J. and Ko, 2012). It can be concluded that Brand Trust is important as it can help create a cooperative environment within the consumer (Kim A. J. and Ko, 2010). Automotive companies can create brand trust by investing in quality products or components. Another good investment by automotive companies that seems important is by providing warranty for products. Also, automotive companies should advice customers to visit their authorised service outlet to service their vehicles because of the genuine vehicle parts that they provide in their outlets, and additionally, the trained experts that will be servicing their vehicle. In doing so, consumers of their vehicles will be more trusting towards the company.
Since this is an initial work on the effect of marketing activities on purchase intention of local made car, researchers can conduct further research to investigate further into the effects of different types of marketing activities and another suitable mediating factor that will give impact towards purchase intention of local made car. Since this study only focuses on general marketing activities approach, hence, recommendation for future study is to focus on both marketing activities which are traditional and online marketing approach in order to differentiate the results between both approaches. Furthermore, since this study used samples from Malaysian made car owners, replicating this study's framework and test on another foreign automotive company is necessary to distinguish the perception of Malaysians towards local made cars and foreign made cars.