The algorithm Netflix originally used for recommendations was called Cinematch.
Netflix launched Cinematch in 2000 to recommend movies based on each customer's viewing and rating history. The system helped personalize the DVD-by-mail service by comparing preferences and identifying titles a subscriber might enjoy, including films they had not previously considered. It was an early example of the recommendation technology that later became central to Netflix's streaming business.
In 2006, Netflix created the Netflix Prize, offering $1 million to anyone who could improve Cinematch's ability to predict user ratings by at least 10 percent. The competition attracted teams from around the world and ended in 2009, when BellKor's Pragmatic Chaos won the grand prize. BellKor was therefore a contest-winning team, not Netflix's original algorithm.
Netflix ultimately developed far more complex recommendation systems using viewing behavior, similar users, title information, and other signals. MovieLens is a separate research dataset and recommendation system, while Netflix Prize was the competition itself—not the name of the original algorithm.