Materials research involves many areas for which a proper understanding of multi-component mass transport is essential. Examples include sintering and transport-limited reaction in syntheses. In addition, materials may be principally designed for direct manipulation of mass transport, as in membrane materials. Macroscopic descriptions for mass transport are available, but physical interpretation of related transport parameters is generally not straightforward and often relies on microscopic considerations. We will show that, even for diffusion in a simple ideal Langmuir type lattice, macroscopic theories should be used with caution. Differences in mobilities of dissimilar species can set off percolation behavior, causing the flux of the more mobile species to vanish. Such behavior is, for instance, observed for zeolite membranes and cannot be predicted by commonly accepted macroscopic transport theories. Correlations between successive movements of molecules cause a decrease in the self-diffusion coefficient, DS. For non-equilibrium transport it can be shown that correlation effects in most cases disappear in which case non-equilibrium transport becomes related to the component diffusion coefficient D, instead of the smaller DS.