An analysis of inertia effects in controlled stress rheometry is presented. For Newtonian fluids, this study takes exact account of both the instrumental inertia and the discontinuous stress change in this type of rheometer. In fact, supposedly smooth stress ramps are always approximated by a series of stress steps. The significance of this limitation, which is sometimes critical, is pointed out for the first time. The error induced by assuming that the stress ramps are smooth in Krieger's method for inertia correction [1], is quantitatively evaluated. An improved calculation for non-Newtonian, inelastic fluids is described.