The ones that don't work may have just been experiments that were set up wrong, such as using the wrong plant, harvesting the wrong part of the plant, or otherwise ignoring parts of the herbalist lore that the experimenter thought weren't important. These are all things that happen.

And some herbalism is just bunk, sure. But a failure to reject the null hypothesis isn't "final". It just means that experiment didn't work.

Not everything is easy to test. Not all herbalism, home remedies, etc. are described well; not all practitioners actually understand how it works even if it does work. (And sometimes experimenters test the purported mechanism, rather than the claimed effect, which is just silly.) All of these things can interfere with the transition from folklore to formal methods.