Understanding how disturbance shapes the dynamics of ecological systems is of fundamental importance in ecology. One emerging approach to revealing and appreciating disturbance effects involves examining disturbance-driven changes in the variability of ecological responses. Variability is rarely employed as a response variable to assess the influence of disturbance, but recent studies indicate that it can be an extremely sensitive metric, capturing differences obscured by averaging and conveying important ecological information about underlying causal processes. In this paper, we present a conceptual model to understand and predict the effects of disturbance on variability. The model estimates qualitative changes in variability by considering disturbance extent, frequency and intensity, as well as ecosystem recovery, and thereby captures not only the immediate effects of disturbance but also those that arise over time due to the biotic response to an event. We evaluate how well the model performs by comparing predictions with empirical results from studies examining a wide variety of disturbances and ecosystems, and discuss factors that may modify or even confound predictions. We include a concise guide to characterizing and detecting changes in variability, highlighting the most common and easily applied methods and conclude by describing several future directions for research. By considering variability as a response to disturbance, we gain another metric of fundamental system behaviour, an improved ability to identify organizing features of ecosystems and a better understanding of the predictability of disturbance-driven change - all critical aspects of assessing ecosystem response to disturbance.