Mackenzie, John (1998) A distribution model for the UK timber industry: wood-based panels. (PhD thesis), Kingston University, .
Abstract
The research presented here is firmly embedded within the broad area of design of channels of distribution, and in particular, the functional perspective of the subject matter. The conceptual foundations of the research can be found in the model developed by Rangan et al. (1992) (referred to in the thesis as the RMM model) and the prioritisation methodology proposed by Barich and Srinivasan (1993). Rangan and his co-authors posit that the design of channels of distribution for new industrial products can be based on the following eight determinant functions: product information, product customisation, product quality assurance, lot size, assortment, availability, after-sales service and logistics (these functions formed the dependent variables of this study). They go on to suggest that an appropriate channel structure can be predicted by examination of the importance of these functions as specified by industry experts, and weighted by their level of confidence in their estimates. Barich and Srinivasan (ibid.) used conjoint measurements in developing a methodology designed to offer guidelines as to the prioritisation of managerial actions. The main aim of this research has been the development of a normative model for the design of distribution channels in the UK trade of WBPs. Because of the wide range of wood-based panels (WBPs) traded in the UK market, a calibration method has been developed that, based on two boundary defining panels (i.e. standard-grade OSB and exterior-grade MDF), allows the interpolation of the functional profile of all other panels. A third WBP, namely particleboard, has been used to test the predictive powers of the model. Examination of actual flows of distribution lead to the specification of the following two independent variables: type of WBP (studied at three levels - OSB, MDF and particleboard) and type of channel intermediary (also studied at three levels - overseas manufacturers, domestic manufacturers and endusers). In order to fulfil the research aim eight hypotheses have been formulated and presented under the following conceptually related headings: • Justification for Weighting the Data H [sub]A: There is no association between expert respondents' 'level of confidence in response' and 'type ofWBP'. H[sub]B: There is no correlation between the pattern of solutions obtained from weighted and unweighted data . • Predictive Powers of the RMM Model H[sub]e: There is no significant difference in predictive accuracy between MDF and OSB as the base of the calibration method. H[sub]D: There is no significant difference in predictive accuracy between unweighted and weighted data. • Model Stability H[sub]E: There is no correlation between the rank order of channel function importance ratings obtained from the RMM model and the conjoint measurement approach. • Explanatory Powers of the RMM Model H[sub]F: The main effects of the 'type of expert respondent' variable are not significant in explaining variation in the importance placed on the eight channel functions (either collectively or individually). H[sub]G: The main effects of the 'type of WBP' variable are not significant in explaining variation in the importance placed on the eight channel functions (either collectively or individually). H[sub]H: The interactions between 'type of expert respondent' and 'type of WBP' are not significant in explaining variation in the importance placed on the eight channel functions (either collectively or individually). Data were collected through a series of four cross-sectional surveys of industry experts, and analysed using a variety of classical inferential statistics and conjoint analysis. On the strength of the results, all ofthe above hypotheses were rejected. The research findings cast doubt on the use of weighted data in the design of channels of distribution for established products. In addition, it is indicated that when predicting channel structures, the level/degree of similarity between the base and predicted products is an important consideration, thus justifying the use of boundary defining products in the development of the proposed model. Furthermore, during the development of the final model, the analysis has provided evidence to suggest that some of the function-related heuristics specified in the literature are context-specific, thus questioning their generalisability. Finally, the findings are synthesised into a normative model and a method designed to enable providers to prioritise adjustments to channel structures if proposed. Finally, it is believed that the research presented in this thesis contributes to the general understanding of issues related to the design of channel structures by: (a) offering empirical evidence that questions some of currently accepted principles, such as functional heuristics, (b) modifying and, at least partially, validating the RMM model, (c) making suggestions as to calibration methods that may be used to interpolate channel structures, (d) proposing a new method of predicting channel structures based of the functional requirements of endusers, and (e) offers guidelines as to the prioritisation of adjustments that need to be made in existing channel structures in order to bring them in line with endusers requirements.
Actions (Repository Editors)
 |
Item Control Page |