Improved prediction of the recovery of oil-in-place in basin-floor fan reservoirs requires accurate characterization and modelling of multiscale heterogeneities. The use of outcrop analogues is a key tool to augment this process by documenting and quantifying sedimentary architecture, hierarchy and sedimentary facies relationships. A 3D geological modelling workflow is presented that tests the impact of fine-scale heterogeneities within basin-floor lobe complexes on reservoir connectivity. Construction of geological models of a basin-floor lobe complex allows realistic depositional architecture and facies distributions to be captured. In addition, detailed models are constructed from channelized areas within a basin-floor lobe complex. Petrophysical modelling and streamline analysis are employed to test the impact on reservoir connectivity between lobe models with: (i) vertically stacked facies with coarsening- and thickening-upwards trends in all locations; and (ii) lateral facies changes with dimensions and distributions constrained from outcrop data. The findings show that differences in facies architecture and, in particular, lobe-on-lobe amalgamation have a significant impact on connectivity and macroscopic sweep efficiency, which influence the production results. Channelized lobe areas are less predictable reservoir targets owing to uncertainties associated with channel-fill heterogeneities. The use of deterministic sedimentary architecture concepts and facies relationships have proven vital in the accurate modelling of reservoir heterogeneities.