7000+ copies = you're a star.
4000 - 7000 copies = a damned good job. A "strong seller."
2000 - 4000 copies = sold pretty strongly. You're still in a good position to have your publisher want to take on your second project, or to comfortably find a home elsewhere.
Know what? She's right. And it's not just literary fiction, which she distinguishes from commercial fiction. In my experience, these numbers hold true for trade sales in all book categories.
As of this posting, there were 34 comments on the Editorial Ass post, most expressing surprise at these low numbers. It is pretty easy to see how an author could become disgruntled with the economics of this business when reality is often so far from the expectations (or at least hopes) of most authors. One commenter makes the point that "good sales" depend on the first print run. If you print 7000 and sell 7000 you are doing great. If you print (or expect to sell) 50,000 and you sell 7000. . . uh oh.
Successful book publishing is a magical blend of great content, great design, great packaging, perfect timing, and a lot of luck. It is also a bit like gambling, for the publishers anyway. We may not admit it, but even the realists, skeptics, and cynics hold a bit of hope that a book may break out, and a belief that it could, or we wouldn't--couldn't--stay in the business.