For lack of ideal heuristics, I recently built myself a tool to find books that "people like me" read most. This is similar to the rating heuristic, except it relies on a collaborative filtering & clustering of sorts. The internet is full of people who have reviewed and curated lists of books, books that can thus be associated with each-other. And so now I can ask my tool the simple question: For people who enjoyed these N books, what other books did they most enjoy? This is something I've found incredibly lacking on the likes of Amazon/GR/etc. recommendation mechanisms, hence why I built the tool. I won't plug it here since I don't want to be spammy but I can share directly with people if they'd like.