Here's an approach editors typically take in high-affiliate-spam areas: go in and look at the site, to see if it shows signs of (1) being unique, or (2) being non-unique. (After reviewing a few hundred sites in an area, an editor will have "a good feel" for those signs, even if they can't describe that verbally.)
If the site looks "non-unique", the editor will probably then pick a product "more or less at random" and try to see if it is already represented. If it is, that's two strikes (one called, and one swing). In this game, that's all you get. It's OUT! -- Umpire's call, and very few instant replays.
If you'll think about it a bit, you'll realize that no other approach is practical. Do you really think an editor is going to look at every single product on a site that bears all the earmarks of being an affiliate site -- knowing full well that many affiliate programs offer thousands or tens of thousands of products? It will never happen.
In order to allow editors to distinguish between "unique" content and "affiliated/linked" content, it will be necessary to make that distinction _very_ clearly on the website (otherwise they won't know), and make the "unique" content _very_ prominent in the site design (otherwise they won't care).
This isn't dictating how you design your website; It's describing how we build ours (a subject, BTW, on which we do not accept dictation.) It may give a little bit more insight into what we consider a site "useful" to surfers. (Note that we have no interest in whether sites "attract" surfers --after all, roach hotels attract roaches -- but that is not exactly the same thing as being useful to them.)