Extensive degradation of coral reefs makes it imperative to create functional models that demonstrate ecological processes which occur in alternative states that persist over time. These models provide important information that can help in decision making regarding management measures for both the prevention of further degradation and the recovery of these ecosystems. Development of these models requires identifying and testing the ecological processes that will impose the reduction of coral cover and, preferably, identifying the disturbance that triggers this phenomenon. For this reason, research programs are a useful tool which allows a focus on the production of information for modelling. It should start with survey investigations and tests of hypotheses concerning the cause of the reduction of coral cover. Subsequently, projects should be guided by the most probable hypotheses, focusing on one guild or functional group at a time until the "trigger" process which unleashed the disturbance is identified. Even if incomplete, these models already provide information for focussing management steps.