Actually, your approach is not so unusual, it is what GigaPan photographers all the time, at least in terms of combinig multiple images. Though the reason they do it usually has very little to do with composition, but mostly to satisfy the quest for extreme detail and/or large printouts.
In your case, the question is how you know how to crop in post processing stage, but not in viewfinder? I would say some sort of compositional sense in necessary in both cases.