The problem is that conventional medicine includes loads of perfectly sensible things for which evidence would be sparse, especially double-blind evidence. Physiotherapy. Counselling. Nutritionists.
The problem is that 'perfectly sensible' does not equate to 'efficacious', but is it cheaper to run a physio service based on plausible ideas or to run double blind trials on a phsyio service.
I'd say that SCAM is defined by its inherent implausibility rather than its quality of evidence. Very few implausible things would ever prove useful. So, if they were rejected on the basis of common sense we would not be missing out on much.
What is the source of plausibility? The rest of science, I suppose. Homeopathy and chiropractic fail this test. Reflexology and iridology do. Acupuncture is a mixed example. It's these things on the boundaries of plausibility where research money should be invested to resolve the issue- maybe acupuncture works for acute pain but not to stop smoking. Money spent offering homeopathy or researching homeopathy is just being p8ss8d away.