It's not only about big thingys. It's any small difference in behavior where some developer tests only on the browser with 90% usage* and ignores the small niche of Firefox users, leaving them accidentally with a broken site. Some CSS property, some JS API whatever.

*) The stats are somewhat wrong, I guess there are more Firefox users with different privacy blockers than Chrome, thus hiding from Google analytics and similar, which people use for stats in higher percentage