Small launch vehicles are governed by the same physics as large launch vehicles of course, but due to their small size, some aspects and sensitivities become more important and others less. This paper shows semi-empirical correlations to quantify dry mass fraction for both stage and whole vehicle optimization: mass fraction due to density, mass fraction due to thrust-to-weight, and mass fraction due to size reduction. For single-stage optimizations, a stage performance requirement can be met by a locus of mass fraction vs. specific impulse. Based on the above correlations, this alone can recommend a solid or liquid rocket for a stage. Rocket designs of similar technology levels are compared, focusing on where stages become less mass-efficient as they get smaller. The Mars Ascent Vehicle is shown to exemplify a trade between a two-stage solids vehicle and a one- or two-stage liquids vehicle.