CONTEXT: Scholars have argued that empirical studies of adoption in agriculture should consider adoption as a dynamic process rather than a binary choice, but many empirical studies continue to be based on cross-sectional surveys in which adoption is treated binarily. In general, surveys put more emphasis on investigating adoption drivers (i.e. independent variables) at the expense of defining complete adoption measures (i.e. dependent variables). OBJECTIVE: In this study, we present, demonstrate and illustrate a method - adoption pathways analysis – as an approach to better represent and analyse the dynamics and diversity of adoption. METHODS: The approach consists of conducting a survey to define individual decisions at different stages of adoption and producing proportional flow diagrams representing the collective results of adopters moving through these various stages. The method is illustrated for four well-established practices in New Zealand pastoral farming using responses from 138 farmer surveys. RESULTS AND CONCLUSIONS: Findings show how the current use status for each practice was the result of individual adoption journeys, converging in distinct pathways. For example, the current population of farmers can be broken down into those who have maintained or increased use of a practice over the medium or long term, those who have decreased their use of the practice since first adopting it, those who are still trialling the practice, those who adopted and then dis-adopted the practice, those who are aware of the practice but have never adopted it, and those who are not aware of it. The pathway to adoption may or may not have included trialling of the practice. Anticipating future pathways, we identified that farmers may intend to increase, maintain or decrease their adoption, and that current non-adopters may or may not be interested in future adoption. For different practices, different proportions of the farm population followed different adoption pathways. Observing these differences provides insights into adoption, and adoption barriers, for each practice. SIGNIFICANCE: Our approach provides a method for adoption research with a highly informative way to unpack the diversity of dynamic adoption pathways for agricultural practices, addressing the current imbalance in survey design that puts more emphasis on potential drivers of adoption at the expense of adoption measures. We discuss the potential uses of adoption pathways analysis to agricultural researchers and extension agents, and its potential to contribute to better explaining past adoption or predicting future adoption.