The question rarely asked is, "Can your team go from zero to sixty on this sort of thing when this happens?"

Technically optimizing as you go might be a 'waste' of time, until it isn't, because your team can handle a serious performance problem in days instead of months (or IME, never).