A more relevant metric than admins/professor would be admin staff/scientific staff. Given that a research group under a professor will probably contain numerous associate professors, assistant professors, postdocs, PhDs, and research assistants who all generate some admin workload, 3 admins per professor does not sound outlandish.
I'm not sure what the person meant in the comment you're replying to, but it sounds like in your comment you're reading "professor" as "full professor", which is not how I'd read it. I'd read it as basically "faculty member".
