Most people may seem to fall tolerably close to one of two stereotypes, but that doesn't make biological sex binary, not does it make it sensible to ignore the fact that it is not in the case of those people who do not, and, surreptitiously enough, there's a significant overlay between people whose experienced gender doesn't match their socially ascribed gender and people whose biological sex-related traits other than external genitalia don't line up with external genitalia in the most common way.
> Biological sex is the most binary thing since zero and one.
It is really not, though for most of human history a reasonably informed person could justifiably believe that; but even that hasn't been true for decades, at least.
For the vast, vast majority of people, it is. Which is what makes it a useful metric.