You don't. It is not even agreed upon that evolution optimizes <<something>>.

And even if you agree that evolution is an optimization process, attempts to define its objective function run into problems with tautology and circular definitions once you move beyond simple examples.

For a cross-species comparison, you'd usually measure evolutionary success in terms of biomass. More is better.

It is described by Dawkins et al.

More successful == gets to have more surviving offspring.

That is the TL;DR. He, she or it that has the most babies, wins.