Identifying homogeneous target groups of dairy producers: An application of combined principal component and cluster analysis

Research aimed at developing appropriate interventions to assist smallholder dairy producers requires a clear understanding of the overall systems environment. Smallholder dairy systems are shaped by production characteristics and technologies, but also by farm/household resource endowments and the market environment. For example, farms in agro-ecological zones of low productive potential, yet close to urban centres, may compete effectively for markets against distant farms located on better land, and their production strategies will reflect that market orientation. The appropriate characterisation of dairy producer farm/households should thus consider this multiplicity of factors, so that resulting research priorities will be more likely to address the factors to which producers themselves are responding, and ultimately to yield the desired impact. This paper describes the characterisation method used to identify patterns among smallholder dairy farm/households in Kiambu District in the highlands of central Kenya, and then to classify them into relatively homogenous target groups. The approach differs from previous methods by combining principal component and cluster analysis to allow incorporation of many variables