People also use the phrase "premature optimization" to make a perfectly accurate observation about eager, costly optimizations which may be obviated by changes

People also use the phrase "premature optimization" to make a perfectly accurate observation about eager, costly optimizations which may be obviated by changes or have no meaningful relationship to any business need. Is there really any good reason not to focus optimization on areas it will make a difference, and guide it with actual profiling results?