Download files
Abstract
Post-purchase customer-centric indices – customer perceived value (PV), satisfaction (CS), loyalty (CL), and retention (CR) – affect corporate performance. Hence, this research investigated how these indices were formed in the Australian mobile-services industry. Its primary objective is to advance the knowledge about the formation of these indices, with emphasis towards (1) identification of a manageable suite of pivotal attributes affecting the indices, (2) development of a research model to explore the importance of these attributes, and (3) improvement of the model’s predictive power by incorporating the moderating effect of customer usage.
As to the attribute determination, the relevant Literature was reviewed and thirty customers were interviewed. Afterwards, threshold tests were used to finalise the Principal Categories which, together with additional pivotal categories from the customer discussion, resulted in ten attributes to be built into the research model.
Regarding the structural analysis, the model was evaluated with the data surveyed from 440 users using partial least square (PLS). The conceptualisation of perceived service quality (PSQ) as a second-order formative construct was examined, and PSQ was built into the model using the repeated indicator approach with Mode B (RI-B). The comparison of this with the other two models – the model with the first-order quality dimensions, and the model with the first-order reflective PSQ – suggested that all three models provided satisfactory explanatory power and similar path significance. Therefore, the conceptualisation of PSQ should depend on research aims.
With respect to the predictive power, the existence of the moderating effect of usage on the links from performance attributes to the indices was studied using a two-step approach. The outcomes confirmed the positive moderating effect of usage on the PSQ-CS link and suggested that, for heavy users, PSQ contributes more strongly to CS.
Consequently, this research contributed to the Literature in many ways. First, the conceptualisation of PSQ as a formative hierarchical construct is insightful. Second, the identified moderating effect explains how usage contributes to the formation of the indices. Third, the use of RI-B and the two-stage analysis shows how to unify both approaches and provides a basis for further studies on hierarchical constructs and moderating effects.