Question: In a treatment trial, if a placebo is said to be say 15% effective, and the real treatment is say 20% effective, of that 20% what amount is attributed

Question: In a treatment trial, if a placebo is said to be say 15% effective, and the real treatment is say 20% effective, of that 20% what amount is attributed to the placebo effect? That is, yes they took a legit med, and yes they got a positive result. However, that doesn't necessarily mean the result was from the meds. At that point, it's still correlation. The meds could have been ineffective, but the placebo effect was the true source of the positive result. That is, given the placebo effect how do they sort out correlation from causation?