[ericb]

Median funding might be interesting (and less skewed by larger follow-ons).

[smt88]

True, but still wouldn't solve the problem: the best outcome and worst outcome look the same in that data set ($0 raised).

[solve]

That's what quantiles are for.