Jane Holland’s comments on how she goes about fixing ‘failed’ poems are very instructive.
The time element is important for me. Leaving poems and not looking at them for months often helps me to identify the bits that don’t work.
I also identify with the section on conclusions. Often poems don’t work well because the ending isn’t good enough. I very rarely get my conclusion right on the first draft or two and I’m amazed at poets who start writing their poems already knowing how they’re going to finish off! In one way, I feel envious. In another way, I wonder if that prior knowledge is restrictive and liable to tie down poems too much. I guess people’s brains work differently when it comes to writing a poem.
For me, the most interesting part of Jane’s commentary was this:
‘So what did I change, in these 'laid-aside' poems, to make them good enough to enter in a poetry competition? Well, first of all, I read each one through several times so that I could hear the rhythm of the poem before I started carelessly hacking at it. Once I felt comfortable with the rhythm, I began making notes on the poem itself with a pencil…’
and later on:
‘…the most useful thing seems to be arriving early at a sense of the poem's original rhythm and purpose. That's why I advocate reading it through several times and letting the poem sink into your psyche, bad lines and all, before beginning any salvage work. For its rhythm is the poem, and without grasping that fully within yourself, you will only destroy the 'good' parts of the poem by making cuts and revisions which don't take rhythm into account.’
It’s well worth checking out the full essay.