Information lies at the heart of Janet I Lewis' theory on how rebellions survive their initial stages. In the words of an ex-rebel-informant, the key goal was to avoid being 'nipped in the bud' before they really got going. Her model of rebel group formation consists of three actors: The rebels, the state, and civilians, who take centre stage. They observe the formation of a rebel group and must choose whether or not to rat them out to the state. This choice is made based on their expectations about the capabilities of the group, and the justness of their cause. Controlling the information environment of the civilians is therefore the most pressing issue for any aspiring rebel. The implications are manifold. One of my favourites is the discussion of implications for state formation in fragile states. Lewis does an excellent job of spelling out her theoretical model and its implications and provides both quantitative and qualitative evidence. Notably, Lewis has collected new data on the initial stages of rebellion, many of which are 'nipped in the bud' before they make their mark. For this reason, they fail to make it into common datasets such as the UCDP/PRIO Armed Conflict Database and are therefore excluded from most studies of conflict onset. Of Lewis' 83 non-state armed groups in central and eastern Africa between 1997 and 2015, only 32% appear in the UCDP/PRIO dataset. As Lewis points out, more than three-quarters of all articles on conflict onset published in the most prominent journals in our field use this dataset. Both theoretically and empirically, Lewis' book opens a new research frontier for our understanding of how conflicts begin.