AbstractMany development initiatives fail because of the way the basic concepts of development relating to the stimulation of agricultural production are generated. The problem with the conventional approach to project identification is that it fails to take into account key decision‐makers (local farmers and other interested groups) and to address the problem of rationality. The accurate identification of farmers' constraints is crucial to the design of projects if these are to have the desired results. Problems resulting from the current approach to project identification are illustrated by the group farming projects in Uganda, and tubewell projects in Bangladesh. The activities of the typical project cycle are unfavourably compared with a decision paradigm, and are related to professional bias, semantic confusion and ‘groupthink’. Nine workable prescriptions are of