A farm typology development cycle: From empirical development through validation, to large-scale organisational deployment

The publication of farm (or farmer) typologies has increased over recent years. The purpose of these studies is usually to differentiate groups of farmers so that they are “targeted” with specific agricultural innovations, or best-bet interventions can be “prioritised”. The degree to which such typologies actually influence development practice is however unclear, and little has been published on that topic. The paper aims to move narratives and practices around farm typologies from theoretical to applied and present a novel methodology for typology validation. We worked with a large-scale development organisation to develop a typology for their use, and report here on the process of enabling the organisation to make use of the typology. The lessons from this process are intended to inform the use of farm typologies in agricultural development work. A typology of farming households was derived from a household survey in Rwanda (previously published), in partnership with a large-scale agricultural development organisation. Responding to the organisation's requests, the researchers created a decision tree tool to rapidly assign households to types; conducted validation exercises to establish confidence in the typology and the decision tree (making adaptations as needed). Validation was with farmers and extensionists and included developing key word and pictorial representations of farm types which were compared against the empirical typology. The decision tree was tested and questions adapted to maximise accuracy. The organisation then used the tools for a period of two years. Finally, the researchers interviewed representatives of the organisation to find out how the tools had been used. The typology validation exercises showed a high level of agreement between farmers and extensionists, and a moderate level of agreement between the empirical typology and the types defined by farmers and extensionists. There was a low level of agreement in the Western province of Rwanda, where the socio-economic situation was radically different to other areas, which had not been accounted for in the empirical typology definition. Establishing the correct questions in the decision tree tool proved important. The organisation reduced the number of farm types, and categorised over 350,000 households, with four use cases developed for the farm typologies: planning for the recruitment of clients (farming households are referred to as clients by the organisation), client needs assessment, intervention adoption rate assessment, and monitoring of farmers along the organisation's conception of their (farmers') journey to prosperity. This study provides lessons on what is required for the application of farm typologies by development organisations. • Farm typology utility for, and deployment by, development actors is under-studied. • Iterative validation and adaptation needed for real-world typology application. • Moderate-high agreement between farmers', extensionists', and empirical typologies. • Development actor categorised 350,000 farmers, supporting tailoring of interventions. • Quicker methods for typology validation are needed for dynamic farmer circumstances