By now, online shoppers are accustomed to getting these personalized suggestions. Netflix suggests videos to watch. TiVo records programs on its own, just in case we’re interested. And Pandora builds personalized music streams by predicting what we’ll want to listen to.

All of these suggestions come from recommender systems. Driven by computer algorithms, recommenders help consumers by selecting products they will probably like and might buy based on their browsing, searches, purchases, and preferences. Designed to help retailers boost sales, recommenders are a huge and growing business. Meanwhile, the field of recommender system development has grown from a couple of dozen researchers in the mid-1990s to hundreds of researchers today–working for universities, the large online retailers, and dozens of other companies whose sole focus is on these types of systems.

…Our experience, therefore, gives us a lot of insight into what’s going on behind the scenes at Amazon and other online retailers, even though those companies seldom speak publicly about exactly how their recommendations work. (In this article, our analysis is based on educated observation and deduction, not on any inside information.) Here’s what we know.