I would love to see a set of (enforced) policies that are blind to the user in question, if such a thing could be. For example:

1. If a post is flagged for review, the account name/owner, social score, social graph, ad revenue, all of it, is hidden from the review person, team, or algorithm. The only history they would have access to is violation history, influencing some kind of recidivism penalty.

2. The duration and severity of the consequences of the violation is applied to the account automatically, and the "judge" is unaware of any backlash from the user / community.

3. Any protest or appeal is sent through the process, blind again.

I realize this is isn't nuanced or realistic in the slightest. What about mistakes? What about gamification of moderation processes? What about revenue impact? How could #1 be cleaned enough to fool anyone...

Still.