The block-by-country feature was added because of a specific problem: password guessing attacks using SMTP AUTH on port 25 from servers in China. Because my MX is an inbound gateway, no legitimate sender should be trying to log into it, so I was able to use it as a honeypot. I stopped counting at 20,000 unique IP addresses that were involved in the attack, which persisted over an extended period. This campaign was detected and reported by multiple users on this forum. All of our other spam problems pale in comparison.
The block-by-country feature may be useful for other contexts, but you have to consider the risks of cloud-based data centers that are replicated all over the world for redundancy. If I remember correctly, one of my vendors has data centers in the US, Ireland, and Singapore. My traffic usually goes through the U.S. data center, but the others are backup if the U.S. center has problems.
My organization also blocks port 443 from foreign countries, for other reasons. So I don't know if China is cranking up similar attacks on port 443 or other ports, but we should probably expect it.