Development of a Decision Support Model to Inform an Organization's Marketing and 'Decision to Bid' Strategies

Important items for the decision to bid process are identified. A questionnaire to elicit numerical assessment of the items is developed and validated with a representative sample of the decision makers in a UK construction company. Data are collected from the historical projects of the company, based on the questionnaire and the decision to bid option that had been adopted for the projects. Statistical tools are used to gain a better understanding of the data characteristics and modelling of the process. A regression model, which is inherently a linear function-mapping tool, is used to model the decision to bid process. This model fails to achieve the desired prediction accuracy. This is possibly an indication that the relationships between items and the decision to bid options are non-linear. In the second stages of analyses, neural networks which are non-linear function mapping tools, were employed. The focus of this paper, however, is the identification of the important items and the statistical modelling process. Detailed information related to the neural networks approach is provided in a subsequent paper.