Most of life is subjective. Objectivity is very difficult. You are dealing with people who are mostly wacky. I suppose a weighted decision calculation could be used. How do you measure grumpy gus, chatty cathy, or mean marvin? How do handle the unsober?

You can set robust PIP structure:

* clear deliveries which are evaluated by not manager

* collect peers feedback/ratings not visible to manager

This structure can give much stronger metric compared to personal manager feedback.