If you're writing for yourself and you find any change weakens the poem, don't change the poem. You have satisfied your audience. You've done what you wanted.
This is a fine way to write poems, as purely personal documents, and the only revisions needed are revisions that reflect changes in yourself as your perspective or insight may change. Sometimes, not even then, because a poem can be a perfectly useful record of who you were at the time you wrote the poem.
These poems may or may not be meaningful to others. If they are, it's coincidence, because it's an accident that they would see them in the first place, and their enjoyment of them is not a concern to you.
If you want them to move other people and that is as important to you as they uses they have for you internally, then a whole different set of questions becomes important in addition to the ones we've spoken about. How is it for you?
For me, I am very much concerned with what other people think. I am very conscious of wanting an audience, and I resent any concessions I feel I may have to make to them. I've learned to take feedback pretty well, even when it's angry, and not to try to defend my poems most of the time. (At least I try not to.) But I find myself blinded to the faults in my poems by the fact that I've written them, and—as Ron says about his software, they don't have bugs, they have features. That makes editing difficult. Often it takes me years between starting a poem and finishing it, 5 years, 20 years or more.
I was looking through a copy of Donald Justice's Selected Poems in 1978 or '79, before he was about to give a reading in New Hampshire. All of these poems had been published before, some of them in Poetry and The New Yorker and magazines at that level. Every poem was filled with penciled notations, crossed out words, substituted phrases. These were changes he wanted to make in poems, many of which were already famous, before they went into another printing. I don't know whether he did or not, but I did see the pencilled revisions.
Robert Graves reported about Thomas Hardy, the English novelist and poet, that Hardy wrote fairly quickly, frequently with only three or four drafts, sometimes as many as seven or eight. Graves said of himself that he wrote more slowly, taking thirty-five drafts or more for a poem. Marvin Bell would write a book of poems in a two month sitting, revise frantically for a month or two and be finished. When Sylvia Plath was "hot," she was able to write two or three poems a day, with very little revision, though I suspect she was able to talk with Ted Hughs about some of them. I don't think he did any revision for her, though.
Is that the sort of feedback you're looking for?