I really enjoy comparisons between music and food. Making music is a bit like cooking—you have to balance the flavour profile, but that’s not something you can do at the end; you have to taste along the way.
Because I am a word-oriented person, I am finding that starting with words can send things out of whack: the flavours of the words tend to dominate the dish (the song), and like too much salt, they can be tricky to pull back.
It seems useful to think in terms of the ways in which our compositional context (including the materials to hand) exert a kind of contextual gravity. What elements are exerting the strongest gravitational pull in an unfolding creative process? The words? The sounds? Something else?
Once the mise en place has been set up (the context), an implicit set of rules are in play. If you’re cooking a steak dinner and decide halfway through that you’d rather make a dessert, you’re unlikely to beat the context (unless you can find a way to turn the work in progress into a convincing dessert). Better to see the steak through and then loop back to make dessert. You could throw the steak out halfway through, but that would be a waste, of resources and of a practical learning opportunity.