How Collaborative Filtering Powers Recommendations on Netflix, Spotify and Amazon

Collaborative filtering is a recommendation technique dating back to the early 1990s, originating from a research project called GroupLens, which found that people who agreed on preferences in the past tend to agree again in the future. The method works by identifying users with similar behavioral patterns — such as ratings or viewing history — rather than analyzing the content of movies, music, or products themselves. Platforms like Netflix, Amazon, and Spotify adopted and scaled this approach, making it a core component of modern recommendation engines. There are two main variants: user-based filtering, which finds people with similar tastes, and item-based filtering, which groups products frequently enjoyed by the same audiences. Similarity between users is calculated mathematically, often using measures like cosine similarity or Pearson correlation, to generate ranked recommendations.