One of the ways I find new books to take on is to look for references in other books I enjoy. This gives me good leads, especially in the "related books" vein. However, this does contain some inherant biases.
An example: I recently tackled Neil Gaiman's The Graveyard Book, which is a fun, episodic romp following a young boy's journey through the land of the undead. Most notably, it's modeled somewhat on Rudyard Kiplings The Jungle Book, a debt which the author freely acknowledges in the back.
Anyway, the point is that although Gaiman recommends Kipling, Kipling couldn't very well recommend the not-born-at-the-time Gaiman. In fact, all references (or hyperlinks, in 'net lingo) go backwards in time, for the simple reason that one cannot reference something that hasn't yet been written. (Of course, someone adding a preface to a future publication of Kipling's book could add a Gaiman reference, but that's neither here nor there.)
Following hyperlinks is a great way to find material, but it always points a reader back in time. This does preclude finding the "latest and greatest" books, but it also provides a great way to delve into older books. After all, recent bestsellers are always easy to find—it seems everyone is talking about them. But finding the gems of the past is more difficult, especially if they don't make the latest "100 best books of all time" lists.
References can be a great weapon in your arsenal of book-recommending strategies, especially on the older side.