We got a lot of great medicine whose biomechanics we don’t really understand and whose trials don’t prove but definitely safely and significantly help people. A study just seeking to prove that your drug does the biomechanics it claims measures a placebo effect of zero if it is done correctly, because if you observe a placebo effect you’re not dealing with a biomechanically constrained outcome anymore.
There isn’t really a viable alternative to evidence based medicine, it’s the only game in town. We just have to carefully decide what outcomes we care about and whether or not those outcomes (1) discourage scientific understanding of the underlying biomechanics of our drugs, by virtue of making black-boxedness acceptable and (2) are co-opted, via enemy action, by either conspiracy theorists sowing doubt in things like vaccines or moneyed interests who make mediocre drugs look good.
However that is not to say that the placebo effect never helps you. As http://protomag.com/articles/the-placebo-problem shows, people who are receiving a placebo will experience medically significant effects, and which effects they experience depend on which placebo they got. Furthermore the reason why double blind studies became standard is that it turns out that the doctor's knowledge has an impact even when the doctor is trying to not communicate it. That is, if doctors are told that one group is receiving a placebo and another is not, the group that is not will do better EVEN THOUGH both groups are in fact receiving the same placebo! (I heard about this some time ago as a classic study, but I can't seem to track it down easily.)