In transitioning to fivetran, an ongoing frustration I have is that the column name translations are not preserved.  I know that you essentially word-split the names with underscores, and also condense double underscores, but that creates issues in matching/translating to other true api-name centric applications and communicating across teams as well as translating queries.

it also seems that you use the true API names in your capture of salesforce formula fields vs the names you store as a result of the syncs.  Without this translation, we would have to manually parse and translate formulas.