Because replica failover is rarely seamless (and often doesn't actually work at all, IME).

Instinctively that's surprising... replica failover should be far simpler technically, shouldn't it?

No? Replication tends to be a bodged-on mess throughout, full of undertested edge cases, of which failover is definitely one. If you build the system so that nodes joining and leaving is a natural and normal part of operation, well, it naturally works a lot better.