Materia, much like teleportation fees, is there to help permanently remove gil out of the in-game economy. Players don't like it, but it's an essential function to keep gil values from endlessly inflating.

Stats and customization are only meaningless if they're designed that way. You're able to math out a universally optimal choice because the present stat design is incredibly safe. Nearly every stat does the same thing (increases damage), except that some do it better than others. The haste stats (skill speed/spell speed) are really the only ones that break this rule, which is why they're unsurprisingly the most fun to mess around with. I would love to max out haste.

You'll actually find a fair amount of variation in games with talent tree systems, Warcraft included. This isn't purely due to lack of knowledge. Sometimes the best build is fight specific. Sometimes the best build is a function of player skill and how much error you can allow yourself as a player. Black Ox Brew may be the mathematically best pick if you are able to use it on cooldown, but the Light Brewing passive is a safer/more forgiving pick if your cooldown usage isn't spot on. You can have a lot of nuance in these systems, but you'll need a little more than four game designers assigned to 19 jobs.

The problem with Tenacity is that it's really just an inferior damage stat with some other random stuff tacked on (+5 to Lalafell detection rate). If you want it to be a 'tank-exclusive stat', then it needs to be at least equivalently desirable as Direct Hit.