I've been reminded that among the unrealised ideas for my Foldatry tool is to give it a feature of a "backup" copy process that's smart enough to not bother copying various useless parts of the folder/file structures.
The cache folders of web browsers is an obvious example but there are probably others that are just not needed in an archive. Perhaps also image thumbnails?
Particular targets would be those with large numbers of tiny files as these often bog down copy and proof processes.