You should only be looking at 99th percentile metrics for comparison because that shows you how each job performs at their intended ceiling. 75th is not 'top end', and if you are falling there you are making frequent errors. The people performing even lower than that are barely playing the same game at all.
I don't really like the idea of balancing jobs based on how they perform at the lower end too. You're basically asking SMN/SAM/NIN players to pay an 'idiot tax' because of low performing players benefiting from their high performance floor. Conversely, jobs like BLM get a difficulty rebate - their job has to be overtuned because the people bad at black mage are really bad at black mage. That's not a very satisfying way to balance the jobs to me, it's basically a caste system with jobs split into 'expert' and 'beginner' jobs. I'd know which one I'd want to play in that case.
It's also important to note that how jobs perform when piloted poorly is not the only factor behind inter-job metrics at low percentiles. The worse the clears you look at, the more you find things like red mages getting embolden-shafted by having extra casters in their party, poor usage of uptime strats that hurts melees and black mages, poor coordination of burst damage that punishes bards and dancers. I'd be careful about drawing conclusions from that data in a vacuum.
Besides, if you rag on summoner you can't really justify any numerical buffs to the physical ranged at all. Unfun gameplay is one thing, but raw numbers are another, and between their mobility and general 'ease of use', they are everything people love to hate about summoner turned up to 11.