That sounds like a true scottsman argument. Naturally any therapy is hard to perform perfectly to get the full placebo effect and a single blind experiment is almost as hopeless as double blind.

https://en.wikipedia.org/wiki/Neuro-linguistic_programming