This is a technique used by Amazon to suggest items of interest, whether CDs or books, to their customers. The software keeps track of all the kinds of material everyone looks at and purchases. It is then possible, based on interest and buying patterns, to suggest new material based on what appear to be common and collective interests. This is described well in Andy Clark's Natural Born Cyborgs (pp. 243-248) in a chapter titled Global Swarming. He likens it to slug trails and ant paths that aggregate colony interests in food as pherimone trails to and from food sources. The idea is that instead of asking people to explicitly categorize their interests, the software keeps track of their interests, no matter how diverse and tangential, just to make certain that commonalities are always noted. What results is a massive database of relationships among subject matter and the people who have expressed interest in it.
That it works is indisputable. I've found many new works by simply taking the trouble to browse through Amazon's suggestions for me. If you've not availed yourself of Amazon's site (choosing, perhaps, to stick to Chapters or some other seller), I recommend it highly. Besides that, their prices, even at Amazon.ca, are much better on average than anywhere else. Plus, they've created a network of used book sellers that can really save you a lot of time (and money) searching for out-of-print volumes.