When is a poem not done?

I usually start with a free write or some sort of first draft. Then I revise and revise and revise. After that, I send the poem to a friend or bring it to my poetry group. I gather the impressions, comments, and suggestions—and then I revise again. Finally, I think I have a finished poem. None of this is unusual.

Then at some point, I send that poem out to a publication. Most of the time, the poem and its companions in the self-addressed stamped envelope come back. I send to a different journal or review. They come back just like boomerangs. The sun comes up again and I send poems out again and—I suspect that none of this is very unusual.

My question: How many times do you send these poems out before you stop? Years ago, someone asked me this and I didn't know. Since that time, I've written more poems, had more come back more times, and developed a sizable, or embarrassing, pile of potential rejects. (And all of these poems were at one time new and dear and held a part of me in them.) When do you decide that a poem is, after all, not done?

I've tried looking through my records and counting ("Gee, I've sent that out X times, maybe I need to give it a rest or give up), but I haven't come up with a magic number. At some point, though, I stop sending the work.

And I've gone back to some of those old poems and taken another look at them—more revisions. I find it's a good way to keep my writing muscles working when I'm having a hard time starting something new. With the time distance, it's easier to let go of images or stanzas that once seemed critical but no longer seem to fit. In some cases, a couple of those new-old poems have been published.

But it's pretty depressing to think that every poem will eventually need to be rewritten. Or, as sage advice goes, every poem needs to stay in a drawer for a few years before its revised.

Do you save your work before you work on it and send it out? Do you rework your old poems or put them out to pasture? Do you have a cut-off number—the point at which you won't send a poem out again until you've given it another look, come to new conclusions? How do you work on old work?