Disasters continue to have a dramatic impact on lives, livelihoods and environments communities depend on. In response to these losses, the global community has developed various theories, assessment methodologies and policies aimed at reducing global losses. A contemporary outcome of these interventions is to build the disaster resilience. However, despite the disaster resilience-building endeavours espoused by policies, theories and methodologies, very little progress is being made in reducing disaster losses. This paper argues that a possible reason behind the limitations of current resilience-building policies and methodologies could be that most of these policies are based a mechanistic scientific paradigm that places an emphasis on system components that are perceived to build resilience and not the function of systems as a whole. This often leads to resilience-building initiatives that are based on a ‘one-size-fits-all’ approach. This paper argues for the use of a complex adaptive systems approach to building resilience. This approach argues that contextual factors within different social systems will have a nonlinear affect on disaster resilience-building efforts. Therefore, it is crucial to move away from ‘one-size-fits-all’ approaches to more flexible approaches to building resilience. These hypotheses are tested by means of a correlation statistical analysis of agricultural communities in Southern Africa. Results of this analysis indicate that unique resilience profiles are evident in almost all of the communities studied. This indicates that resilience is not the same for everybody, and that resilience-building endeavours should be flexible enough to be adapted for different contexts.