Brand related concepts in the business-to-business sector have received little attention. Nevertheless recent research has shown that brand equity exists in business-to-business markets. Thus, brand-building investments in this sector may be worthwhile. Moreover, the brand loyalty concept appears to be relevant to the business-to-business sector. But little attention has been given to examining the antecedents of brand loyalty in a business service context, although a number of researchers have isolated one or more variables that impact on brand loyalty. However, there appears to be no systematic effort to incorporate the potential antecedents in a single study.
This thesis attempts to address this gap in the literature. Specifically it presents a comprehensive view of brand loyalty for business services and proposes a model of its potential antecedents. Two separate dimensions of brand loyalty are examined: attitudinal brand loyalty and behavioural brand loyalty. Both dimensions are important to the study of brand loyalty and should be considered jointly, rather than in isolation as has occurred in past studies. The antecedents of brand loyalty are drawn from extant business and consumer marketing literature, and include trust, commitment, satisfaction, perceived risk and involvement. Trust and commitment are viewed as 'relational' factors. These constructs are drawn from the business-to-business marketing literature. Satisfaction is described as an 'emotional' construct with roots in both the consumer and business marketing fields. Finally, perceived risk and involvement can be described as 'cognitive' factors. Their importance to brand loyalty is illustrated in the consumer marketing literature. Thus, the model proposed here is comprehensive in scope.
The research context for this study is the business services sector. The data was collected using both field survey research and archival data. The attitudinal data was collected prior to the behavioural data, thus making this research longitudinal in its design. Moreover, the research was undertaken in three stages. The first stage was a pilot study of the survey instrument to test the reliability of the scales used, as many of the scales were adapted from a consumer context. The second stage was the main survey that involved the collection of the cognitive and affective data (capturing the proposed antecedent constructs). The third stage comprised the collection of the behavioural data from company archives, and tested the proposed model. In all, 267 archival records were matched with key informant reports to create the data set for the hypothesis tests. The hypotheses specified in the proposed model were tested using structural equation modelling techniques. A two-stage procedure was adopted that initially involved the estimation of a congeneric model to check measure reliability and validity. The proposed model was then estimated for a test of the research hypotheses.
A main conclusion of this thesis is that attitudinal brand loyalty has the strongest impact on behavioural brand loyalty. The effects of the relational constructs (commitment and trust), buyer emotions (or satisfaction), and cognitive factors (perceived risk and involvement) on behavioural brand loyalty are indirect, through their effect on attitudinal brand loyalty. Of the antecedents studied, satisfaction had the largest relative impact on attitudinal brand loyalty. These results were largely consistent with the conceptual framework. Arguably this thesis makes four contributions. First, it demonstrates the importance of brand loyalty in a business services context. Second, it suggests that attitudinal brand loyalty is a prerequisite for behavioural brand loyalty. Third, it is the first systematic study of direct and indirect antecedents of brand loyalty in a business services context. Finally, the study captures the impact of relational factors on brand loyalty, important in a business services context. To summarise, this thesis offers a comprehensive view of brand loyalty in a business services setting and an empirical test of its potential antecedents
