When I'm asked to read a document as an editor, the first step is to review and comment. I read through the document, mark all the passages that strike me funny, and comment on each of them. After that, I go through each of my notes with the author (or whoever is in charge).
If it is a rough draft for a document, there may be a good long discussion to go over. If it is a polished document, it may only have a couple of minor comments per page. In the latter case, your job as an editor is different than what people expect.
My job when working with an author is to get them to write something that they are happy with (this is true no matter what). It is not my job to take control of the document or change things for the sake of changing them.
So when I am working on an already polished document, the job is to test the strength of it. Is there ambiguous wording? Is a particular synonym saying what you want it to? Does your document cover everything that it needs to cover? Should sections be added? Can any sections be removed?
If you think of your documents as buildings, you need to test the structure of it to see what it can stand up to. This is what I am doing.
Sometimes, the author agrees that some changes could be made. Other times, they disagree with everything I say. Every time they disagree, I ask them about it. Are they sure? What if they looked at it a different way? What if somebody misunderstands a passage?
If the author completely disagrees with me and after my review, leaves the document unchanged, was there any point in having me look it over?
Hell yes. I am a trained, sensitive reader. I have seen a lo of crap and a fair amount of gold. I am a natural arguer and will find every possible way to tear apart an argument. In a sense, my goal is to destroy this document, and if I fail, then the document succeeds.
At the end of the day, my review becomes a stamp of approval. The document has passed all of the rigours it can. On top of that, the author has had it confirmed that the document is solid. To be a good editor, you need to be a firm confirmer. Make sure that the author believes that they have the best version of their document that can exist.