Case study

A US electricity utility

 A regression analysis and cost modeling project

PA's client, a US electric utility, wanted to develop a market-defined basis for segmenting its population of agricultural customers. PA designed a market research study to collect the data and develop a segmentation strategy. PA's first step was to design and implement a mail survey, which resulted in 600 completed surveys representing the population of agricultural customers.

PA then developed a segmentation model, using discriminant analysis, factor analysis and cluster analysis techniques to assign survey respondents to segments according to needs, values and other characteristics. Based on the final needs-based segmentation model, a second 'classification' model was developed. This classification model is used to assign the remaining members of the agricultural population (those who were not surveyed) to the needs-based segments. The assignment is done on a probabilistic basis, based on similarities of each customer's characteristics (using information available from customer records) to the needs-based segment profiles.