At the broadest scales, species
durations seem just as good a measure of "success" as clade size.
I disagree. Species duration depends entirely on the species
concept... you have at least 25 to choose from.
Your point is well taken. However, even though using duration is
difficult, it is still a good measure in theory. In practice, it will
only be viable for some groups and in certain situations. Comparing
durations can still be informative (within certain bounds). Comparing
clade size is quite informative, as well, but it probably says as much
about isolation tendency as it does about mortality reduction or other
aspects we might associate with 'success'. I generally do not consider
a clade successful simply because that clade is large. Other factors
must also be considered.