Probabilistic Seismic Hazard Analysis (PSHA) is widely accepted as the most robust approach for evaluating the seismic hazard at a given site and provides the basis for seismic loads in most design codes. To obtain more detailed information on the specific earthquake scenarios contributing to the hazard at a site, it is common to include seismic hazard disaggregation results within a PSHA. This is mostly done in terms of exceedance of the intensity level of interest, but for many applications a disaggregation in terms of occurrence of the intensity level of interest is more appropriate. A number of researchers have examined the theoretical differences between the exceedance and occurrence approaches; however, few have provided extensive application examples. This paper therefore compares the approaches for 24 sites across New Zealand. It is shown that the two different approaches can result in moderate differences in the mean magnitudes and site-to-source distances, as well as differences in the relative contributions from different Tectonic Region Types. Whilst some weak trends are identified, it is concluded that generally it is not possible to know a priori whether the difference between occurrence or exceedance approaches would have a tangible impact on disaggregation results or the results of subsequent applications. It is therefore recommended that developers of seismic hazard analysis software and providers of seismic hazard data products make both approaches readily available.