|
Craig Schwabe and Michael O'Donovan 1. INTRODUCTION Demographers are often called on to segment populations according to behavioural characteristics such as consumer, reproductive or other types of behaviour. This inevitably involves balancing representational information that may contain contradictions, superfluous or confusing data. Additional challenges include the increasingly unacceptable practice of using "traditional" indicators of behaviour like gender, race and income, since the present political climate demands that categorizations that are unbiased, "value free" and yet functional be used. The changing way in which information is becoming available highlights the need to present information by spatial zones rather than social categories. These presentations simultaneously need to take cognisance of the cultural and economic diversity within such zones or communities. From a planning, business and survey perspective it is important to understand the homogeneous categorizations of society in South Africa. Identifying communities with common socio-economic profiles and, thus, specific concerns and basic needs, state agencies and other service providers can optimally focus their attention and resources. Knowledge of buying patterns and preferences of consumers within different lifestyle segments allows businesses to effectively target their markets. For social scientists the creation of homogeneous socio-economic categories allows researchers to make better assumptions regarding responses to survey questions and may assist in the drawing of samples that are both smaller and geographically representative. Historically lifestyle segmentation has not been used for planning purposes in South Africa. The primary reason for this rests on the lack of information, a lack of understanding of the application of this type of information in the planning process, the unavailability of suitable lifestyle segmentations, and loaded or biased classifications. Lifestyle segmentations likewise have not been used as decision-support information for the South African business sector nor the market survey environment. The development of an objective approach that allows information at the smallest possible spatial unit to be clustered into recognizable groupings may remove some barriers against the widespread use of lifestyle segmentation models. The subtleties of geo-demographic profiles of the South African population need to be highlighted in these models yet the techniques must insure that the information is objectively summarized. A clear indication of the understanding of the dominant descriptors of the population should accompany the stratification. It is from this base that a lifestyle segmentation system was developed for South Africa using Artificial Neural Networks (A.N.N.).
|
