To me this is similar to asking, “why should we identify to the species level instead of genus?”. Certainly some subspecies have little to no visible differences (and this can even be true at the species level, e.g., Eastern/Gray Ratsnake Complex in northern parts of their range), and we shouldn’t add lower taxon IDs beyond what we can actually ID. But for the ones that are discernible, of course it is useful to identify them.
Some subspecies are already of conservation concern (e.g., Ammodramus maritimus mirabilis is listed as endangered under the ESA while other subspecies are not). Differentiating between subspecies of Palm Warblers could provide critical data if one subspecies started to decline suddenly.
It is important to remember that taxonomy is an organizational tool and does not represent true hard lines between taxonomic groups. Preserving diversity, whether it’s between species, subspecies, or even populations, is important.