It varies a lot. For many conditions, like cancers, the placebo effect is basically random noise (cancers sometimes just reduce or go away on their own, regardless of treatment or even of the appearance of treatment ; we don't know the natural rate of this, because it's unethical to not treat people who you know have cancer). In addition, in non-blind trials, there is a "placebo effect" that amounts to mistakes or lies by those involved in the research in favor of a positive outcome. This is not a real effect in patients at all, just an artifact in the reported data.

Then, for conditions linked to our psyche, including pure psychological conditions but also things like pain, blood pressure, heart rate, nausea, and some others - the placebo effect is more real, but usually temporary. Some people who have been living in some amount of despair at their condition experience a positive surge of hope once treatment starts, and they can ignore the pain, or feel some push to get out of their depression, or calm their anxiety which was exacerbating, say, the high blood pressure etc. This effect almost always tapers off if the treatment is not doing anything more fundamental.

Coupled with the fact that we don't understand how psychological disorders work at the chemical level at all, especially in relation to the conscious mind and interventions on that (e.g. therapy, but also various religious practices), this means it's very hard to account for this without a double-blind RCT.