Convective self-aggregation, the spontaneous clustering of convection in idealised simulations of radiative-convective equilibrium, is sometimes dismissed as being a model artefact or a result of conditions that are too simplified to be applicable to the real world. Here, I will argue that we have already learned some important things from these simulations about feedback processes (including radiation and circulation feedbacks) that have been shown to play a role in tropical phenomena including the MJO, tropical cyclones, the ITCZ, and both shallow and deep convective clustering. I will also look ahead to some remaining questions and avenues for further research.