I have a campaign that primarily focuses on bottom 50%. A coworker uses a low income zip code exclusion into paid social campaigns. However I have been leery about adding something like this in Search because there will be targets in low income zips. I have been relying on Google's ML to make these decisions on who to show ads to based on maximize conversions bid strategy. However yesterday I attended a presentation by another ppc expert who said that by adding the exclusions you teach the machine learning faster and then can exclude the list. That makes sense. Would love to get other experts' opinion before I mess with my campaigns in this way.