We could imagine cases where people underwent homeopathic treatments and saw improvements in their symptoms for other reasons. For example, colds usually stick around for 3-4 days and dissipate without treatment, so you take a homeopathic medicine and two days your cold vanishes and you think "It worked." The correlation-causation error that might seem obvious to skeptics, but it isn't to the homeopath believers.

As I interpret the Franklin quote, you provisionally accept (don't immediately and explicitly challenge) the claim that the homeopathic medicine made the cold go away, so you can establish a further dialogue with some chance (let's just say 10%) of causing doubt in the other person. If you immediately say "There is no way that the homeopathic medicine had any effect," the person will get angry at you. You'll probably have a smaller chance of changing their mind, and they won't like you, which generally doesn't help you accomplish goals.

With Franklin's approach, I think it doesn't even matter that there are no merits to a homeopaths treatments (or insert whichever group); you need to cede some ground to keep negotiations open and to get people to like you because it's helpful later.

I'm not sure I know what point you meant to make by this.

I read Franklin's advice as applying, and intending to be applied, quite readily in those cases where one's interlocutor is totally and clearly wrong. The idea is that you take a certain roundabout approach to telling them that they're wrong, without quite coming out and saying it straight out. The fact that they are wrong need not be in question; it's merely a matter of which tactics are effective in convincing them. (The assumption, of course, is that you're interested in convincing them.)

In any case, I am unsure in what sense your comment is a response to what I said... could you clarify?