Each major conference (for instance in Theoretical Computer Science) requires 3 referees per paper, and each major journal at least 2, so count the number of papers you submit per year, normalized by number of authors and try to make your ratio refereed /submitted above 3:1 (or 3:1 for conference submissions and 2:1 for journal submissions).

If you don’t do your share of work, it will be done by somebody else, of course, meaning that you cannot complain if your conference submissions are briefly reviewed by a PC member or if you receive an uninformative journal review clearly written by an unexperienced referee.