the data says nobody buys it when it's available
so the percentage of people it saves more people money is vastly larger than the percentage it solves problems for
basically just learn to use ANSI you giant nerds
I've seen this point before and I do get it. In a world with so much competition and so little margin, optimising sets really does matter in order to gain enough popularity to meet MOQ. You say
vastly, but the split we tend to see in high end kits is about 80% ANSI users and 20% ISO users, so there's no doubt the custom end of the hobby skews American but there is a significant community outside of the US+China.
What we're talking about here is a 3-4% increase in costs (assuming $5 for adding physical ISO support in a base kit costing $130) in order to enable 20% of the community. In addition to that, about 8-12% of users typically buy NorDeUk kit, clearly showing there's a minority of users who want the option to have legends that are regionally correct.
For many outside the US/China ISO is "the standard" and the custom side of the hobby includes a lot more ISO boards than someone from an ANSI-using country might expect. Not only that, but there's a significant population of people on non-custom mechanical keyboards looking to get into the hobby.
As this hobby grows, if the overall "pie" is getting bigger we should support ISO as the 20% of people using ISO will want those keys in order to get physical compatibility. ANSI isn't the global standard for keyboard layouts, we have two. The one from the American National Standards Institute and the one from the International Standards Organisation and both get about 50/50 usage in the "real world".
At the end of the day, they might just seem like different shapes for keys, but for a lot of people these shapes are fundamental to what a keyboard is.