On the contrary: if it works it works. Just because it's a placebo doesn't mean it only works for one person. The only reason we know that the placebo effect exists because we can show in multiple studies that certain placebos work on a lot of people. In fact, some placebos work better than other placebos[1].
Of course, there are pitfalls. It's important that doctors not lie to their patients, because patients need to trust their doctors, and patients need to be able to make informed choices. And as you pointed out, it's obviously wrong to sell a placebo as if it were not a placebo. But even this doesn't mean placebos are useless: some placebos work even when you know they are placebos[2].
There are many cases where placebo performs better than nothing, and it would be foolish to not use this effect to improve our lives.
From what I've read, the placebo effect is still mysterious, and the only place where it is consistently found to help is in pain perception. Outside of this, there are some studies which show improvements in placebo when compared to no treatment, but other studies ([1][2][3]) find no improvement or only perceived improvement - e.g. the patient feels that they slept better, but in fact they slept the same amount and it took them just as long to get to sleep.
So, overall, it's not clear whether the placebo effect is a real effect that placebos have on patient health, or whether it is an artifact in the data, e.g. due to reporting errors, especially in studies which are not double blind (where the researcher may be biased themselves in the way they are collecting the data). There are also many studies which compare an active substance to a placebo treatment but have no data about patient response under no treatment, which is itself not necessarily 0 [2].
Edit: added a [3] that is explicitly about insomnia - placebo was found to improve subjective perception of sleep quality, but not objective measures (sleep onset latency).
That's interesting, and it's good to know that my sleep improvement might be perceived rather than actual.
That said, this is just evidence that the placebo effect may not apply in this case. It's not evidence that if interventions only work due to the placebo effect, they don't work. You've convinced me that I'm probably not experiencing benefits from the placebo effect in this case, but that conclusion is unrelated to the conclusion of your previous post.
I'll also add that your assertion that pain perception is the only thing that responds to placebo is incorrect. Placebo is effective on a wide variety of things, including weight loss[1], itching[2], asthma[3] and warts[4] (that last one surprised me).
There's also lots of cases where placebo makes things worse, such as people responding to placebo poison ivy[5].
Finally, there's lots of cases where things which were used in studies for placebos turn out to be active. For example, a study of insulin and chromium in sugary foods for diabetics used apple pie as a control source of sugar, but discovered that the cinnamon in the pie produced a larger effect on insulin resistance than the chromium.
Of course, there are pitfalls. It's important that doctors not lie to their patients, because patients need to trust their doctors, and patients need to be able to make informed choices. And as you pointed out, it's obviously wrong to sell a placebo as if it were not a placebo. But even this doesn't mean placebos are useless: some placebos work even when you know they are placebos[2].
There are many cases where placebo performs better than nothing, and it would be foolish to not use this effect to improve our lives.
[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3729687/
[2] https://www.health.harvard.edu/blog/placebo-can-work-even-kn...