What methods should we use to answer ontological questions? In particular, should we favour parsimonious theories? In this paper I will discuss two arguments that can be used to favour parsimonious theories in science and show how they can be applied to ontology—specifically, to the debate about material composition. The primary aim of this paper is to argue that considerations regarding parsimony in philosophy of science can be applied to metaphysics. The secondary aim is to assess the effect of these arguments and make explicit the assumptions needed for the arguments to support parsimonious theories about material composition. The overall verdict will be that a preference for parsimonious theories can be justified, but that this justification is far from straightforward, and is sensitive to our answers to a number of controversial questions e.g. the size of simples and the necessity of scrutability.