Grades probably should be less subjective than they are. Rock grades in any country are an attempt to list the routes in an order dependent on the number of climbers able to climb them. So if only 125 out of 1,000 climbers can climb route A and only 30 out of 1,000 can climb route B. B is harder than A. Full stop.
Where we then put the boundaries, 5.12a, 5.12b etc is less clear, but not random, as usually benchmark climbs are used, again with the idea of the fraction of climbers who can climb them. This works well in areas where the routes have been climbed by very large numbers.