A logistic regression modelling approach to business opportunity assessment

Significant opportunities for improvement exist in the optimisation of business processes. The use of statistically-based methods to characterise business process performance is one way to realise those improvements. Business processes can be difficult to characterise due to variables that are difficult to define, measure and analyse. Outputs in the business process may be qualitative or binary in nature. The cause and effect relationships among inputs, actions and outputs may be difficult to observe and monitor. Sources of variation exist throughout the process, but may be difficult to identify and control. Consequently, statistically-based characterisation methods that have been proven successful for manufacturing processes may not be directly applicable to business processes. Because many of the variables in these business processes are subjective or qualitative in nature, categorical data analysis techniques may be useful. This paper illustrates the use of logistic regression to establish statistically significant relationships between the input and output variables of one complex business process.