Presumably most of the DNA in honey is coming from pollen? If so, is it worth looking at it the other way round and back calculating what coverage of regiosn you need to be able to discriminate based on the likely number of taxa present and frequency of occurrence? There is a lot of data about pollen density and diversity in honey.