The first thing I do when I get a casting log is draw a line from top to bottom, halfway through the area designated for comments.
That's not what makes this trick so powerful (or reliable). The part that makes it powerful (and reliable) is how I label the sections. One side is "Look"; the other side is "Performance". What this does is allow me to evaluate each actor for both how well he or she suits the role superficially and how convincingly he or she plays the part.
Sure, the guy may look like the quintessential matador, but if he's a lousy actor we're better off going with someone who doesn't look quite as convincing, but can actually perform. (I'm amazed that in 16 years of working with casting directors, producers, copywriters, art directors, and creative directors as a director –– plus all those years of working as a copywriter on shoots with other directors –– I never once came across anybody else who broke it down like this.)
I encourage the ad agency people I'm working with to break down their evaluations the same way. This does two things. The first is to keep the conversation focused –– if we don't agree about a particular person, we can more easily figure out why. The second is a bit more subtle, and that's so I can get a clearer picture of how the agency people see their spot. We all have our own pictures of what a shaman looks like, and it's often only when we each respond to the actors who come in that I can see clearly what the art director and copywriter had in mind when they came up with the concept. Which, by the way, are often wildly different from each other.
I'll give you another tip. The evaluations are relative, not absolute. What I mean is that you could go insane trying to find the perfect space hooker when all you need to find is the best space hooker.
There's a neat implication to this that I discovered about ten years ago. I rate each actor on a one to ten basis, with ten being the highest, so the first person –– no matter how good or how bad –– gets a five for both look and performance.
If the next person is better, he or she might get a six. Nobody ever gets a ten because I always want to leave open the possibility that somebody else better might come along. I've passed on people who scored 9.875. And I've cast people who scored a five.
Sometimes, I can't help but give a zero, because I can't imagine how anybody could possibly be worse. Invariably, somebody is, which only goes to show that my system, as good as it sounds in theory, is pretty hard to put into practice.
At the end of the session, I take the top scorers for both looks and performance and figure out which contestant meets both criteria the best. If the ones that look the best aren't the same as the ones that perform the best –– and this is the case a lot –– I make a judgement call.
But then, that's what I get paid for. To make judgement calls. My particular way of evaluating casting is only one way for me to make the best decision.
If it works for you, great. If it doesn't, find another way that does.