How can classificatory changes be made in an objective way based on a morphological cladistic/phylogenetic analysis?

Classifications often are based on rather (highly) subjective decisions. Once you have obtained a phylogenetic tree from a parsimony analysis, which ways allow objective decisions regarding classificatory changes suggested by the tree? Rather than arbitrarily labeling clades and branches with family or genus tags, I would prefer an objective way to base my decisions upon.
I add this for clarification: I do not intend to make a mess out of a well-established taxonomy. I am working on a phylogeny of a group that comprises many well-established (and supported) families but also several monotypic families or genera insertae sedis. Their inclusion to the analysis will hopefully improve our understanding of the between-family relationships. It will also lead to new ideas about relationships between well-established families and the small and odd taxa. Furthermore, some genera do not group with the families they had been assigned to originally but form sister-group relationships with other closely-related families. My question is about the decision making: when should taxa be joined and when better not? Is there any published study that used a sensible (guide) line of reasoning?