The way in which systems should be decomposed so they can be better understood and better designed remains a fundamental problem in the information systems discipline. A number of different decomposition methodologies have been proposed. However, no methodology has emerged as dominant, presumably because the relative strengths and limitations of each methodology are still unclear. Case study research that has compared the methodologies, for example, has produced only equivocal results. In the absence of a theory of decomposition, it is difficult to make insightful predictions about the merits and failings of a particular methodology. Consequently, it is difficult to undertake empirical research that produces compelling results. Accordingly, in this paper we develop a rudimentary model of decomposition that we hope might form the basis of a subsequent, more complete theory of decomposition.