On the macro I mean, how much of the reward pool is being allocated to it currently in total, and how much will it scale up in the future?
Less than 1/2 of one percent. And we have been shifting from reliance on delegation to self-power over the last few months. Now with communities the VP is drawn down some, so we will start leasing delegation again. Still, I don't expect it to ever grow large enough to be a significant portion of rewards pool. Any single program supported by misterdelegation has much bigger impact... like drugwars self-enriching itself through fundition votes for many months... that dwarfs the entire history of SBI voting.
This includes both the voting itself, and policing of the votes for abuses.
SBI has followed @steemcleaners and @buildawhale blacklists for more than a year. There are a few exceptions (members can appeal their blacklist status, but we encourage them to address directly first) and sometimes members build up a sizeable vote and then becoming abusive, but most aggressive abusers of Steem aren't comfortable with locking up STEEM in a long-term subscription.
What percentage of people that meet both of our standards (or a reasonable standard of quality/effort/PoB) actually getting these votes?
Quality is incredibly subjective. We try to police for plagiarism, spam, and fraud, but aside from that we think that crowdsourcing our vote list works really well!