I want to say that librarians really do use this document as a guide for the de-acquisition process (though not only this guide). It's a good way to create reports (i.e. show me everything beyond a certain copyright date or that hasn't checked out in more than "x" years) so it saves having to judge thousands of items individually, and it provides excellent guidelines for what to look for when you undertake a weeding project (side note: ideally, weeding would be an ongoing project, but...).
But each responsible librarian (generally of a subject matter) really does (or should) look at the titles they are getting ready to weed and decide if it indeed needs to go, needs to get updated, or still has value and needs to be promoted in some way. And they do (or should) regularly examine their area for items that are in poor condition or "MUSTIE" (one thing I always say to selectors is "If in doubt, toss it and replace it. If it's worth keeping, it's worth buying new.").
CREW also provides an excellent explanation to lay people like community members and library boards for why the process is necessary. People have a strange hagiography of the book, but this report makes it very clear why libraries get rid of items and that they endeavor to replace them if necessary. It's very helpful to have a few copies around to hand out.
EDIT: edited for clarity. Just trying to offer some real world experience on how we use CREW.
> weed books by celebrity chefs and television
cooks once their popularity has waned
The simple fact is that space is limited. If it ain't gonna check out and it ain't of huge culture importance, it can make way for something that's one of those two things.
In addition to physical space, the literature is pretty clear that weeding has a long term positive effect on circulation. A physical collection that looks good and seems current is more likely to get used than one that looks beat up and who's best material is buried among rows of books about the OJ Simpson trial. We don't really have data or experience yet on electronic collections, but I'm genuinely curious if the same would be true.
The OJ Simpson story is from actual experience. At a conference I went to a local library that had no fewer than 5 books about OJ Simpson with copyright dates in the mid 1990s. They looked like they hadn't been checked out for some time...