Recent developments in petroleum industry have been witnessing the surge of the use of nuclear magnetic resonance (NMR) log. Despite some remaining problems the NMR technology appears to gain more acceptance as petrophysical tool for evaluating reservoir quality. Comprehensive formation evaluation requires determination of irreducible fluids, movable fluids, and permeability. However, rock heterogeneity introduces complexity in any formation evaluation activities. This can also cause problem for NMR log interpretation. In the presence of clays the most commonly used T2 cut off values, a constant value throughout a formation, seem to eventually yield inaccurate irreducible water saturation estimates, as well as other output such as permeability. This study focuses at finding a solution for finding the best way of choosing the most representative T2 cut off value to be used in NMR log interpretation. This is indeed a common pressing problem for heterogeneous formation rocks such as in the case of Tirrawarra sandstones used in this study. The main part of the study is devoted to comparison between the use of single averaged T2c value and establishment of empirical correlations enabling the provision of T2c for any level of heterogeneity (i.e. various levels of shaliness). The study however surprisingly shows that, in spite of the theoretical soundness of the empirical correlations established, simple averaging of T2c values yielded by a reliable method proves itself adequate. This conclusion therefore helps considerably in reducing complexity in NMR log interpretation.