In this short review, we provide an account of a number of computational studies of catalytic reaction mechanisms carried out in our groups. We focus in particular on studies in which we came to realize during the course of the investigation that the active catalytic species was a bimetallic complex, rather a monometallic one as previously assumed. In some cases, this realization was in part prompted by experimental observations, but careful exploration based on computation of the speciation of the metal precursor also provided a powerful guide: it is often possible to predict that bimetallic species (intermediates or transition states) lie lower in free energy than a priori competitive monometallic species. In this sense, we argue that in organometallic catalysis, the rule whereby two is better than one turns out to be relevant much more often than one might expect.