Just half?
I'd argue MOST of the Job changes (and game changes in general) are due to players asking for things, that later players then complained about. For years I've seen people complain WAR's AOE was cone shaped. As soon as it was finally changed to a circle, the complaints have been endless that it's not unique anymore. While Cleric Stance existed, people complained endlessly about it, how people wouldn't use it, how people used it and got the party killed, etc etc. So it was changed into a mere damage boosting CD. Then everyone complained (people that loved and hated Cleric Stance before that) that it was boring and pointless and they might as well just bake the boost into the base healer potencies. So then they did that, and people have complained ever sense that Cleric Stance was removed.
People complained Relics were too grindy, then people complained that ShB's Relic starting step wasn't too grindy enough, THEN turned around and complained that it was too grindy in the mid-to-later steps that required a lot of grinding, resulting in it being nerfed, at which point people promptly complained about the nerf and how it was too braindead easy.
Now, we'd be fools to think these are the same people (someone wishing WAR had two circles instead of a cone and then a circle are PROBABLY not the same people complaining about the cone removal), but the fact remains that changes are made due to the community saying things, that then result in a different part of the community being upset. Or, in some cases, EVERYONE being upset. There's literally no winning with some people, so the goal is just how do they piss off the fewest people and keep even the upset people GENERALLY interested/somewhat happy.
The problem is the Steam figures were pointed out as potentially not representing the larger game trend - something no one contested (instead just insisting they were valid and that no one should question them) - and that the overall Lucky Bancho numbers indicate EW has had the greatest growth of any expansion in FFXIV's history and that the post-expansion bump drop is on par with past expansions on a percent basis.