In general, it's bogus. There can be NO mechanism of action. There are a ton of studies claiming many of these alternative medications work. Fact is, these are either not peer-reviewed, have terrible design or lack of statistical control, and otherwise consist of studies run by people who have a very poor grasp of basic science.
You have anecdotes at best that are not, in any statistical, significant sense, causal, nor reliable or valid indices of anything. Many people mistake correlation and coincidence for causation (the "regression fallacy"). Causation is kinda important in terms of efficacy, but also for safety reasons. The "appeal to nature" fallacy does not always hold true, particularly, regarding homeopathy.
We also exist in a quick fix society, so anecdotes are latched onto as a function of simple "availability heuristics". Just too bad these are elevated to the status of real evidence, knowledge, and truth.
Anecdotes are not entirely different, though, from what is termed "clincical evidence", seen in conventional medicine. Many MD's rely on what has worked for his/her patients, often times this is favoured over empirical evidence. Too often. Anecdotes to do with alternative medicine as well as clinical anectodes used regarding conventional medicine are both exemplars of heresay; each nothing even close to hard, unequivocal, supportive evidence.
Based on the law of thermodynamics to describe chemical action, energy must be created to provoke a cellular response. This energy comes from the binding of a ligand molecule to a receptor or other enzyme. Not possible with alternative medications and demonstrative of the fact that there CAN be NO mechanism of action with alternative medicines.
Placebo responses are typically 30% (give or take) in even well-designed, controlled, empirical RCT's. With poor design, "efficacy" whether conventional or alternative can be chaulked up to, you guessed it, placebo-- aka, wishful thinking, the power of the mind.
Pharmaceutical companies do sponsor many studies of many conventional meds, so the validity and reliability of results may lend themselves to the same problems seen in "testing" alternative "medications". But, with alternative medicine, real science is consistently ignored, all together--essentially, pseudoscience.