Hybrid web recommender systems combine multiple recommendation techniques to improve performance. This chapter surveys two-part hybrid systems, comparing four recommendation techniques and seven hybridization strategies. It examines 41 hybrid implementations, including novel combinations, and finds that cascade and augmented hybrids perform well, especially when combining components with differing strengths. Recommender systems are personalized information agents that provide suggestions for items likely to be useful to a user. They differ from information retrieval systems in user interaction semantics and personalization. Techniques include collaborative, content-based, knowledge-based, and demographic methods, each with known limitations, such as the cold-start problem for collaborative and content-based systems. Hybrid systems combine techniques to achieve synergy. For example, a collaborative system and a knowledge-based system can complement each other: the knowledge-based component can address the cold-start problem for new users, while the collaborative component can find similar users with unexpected preferences. Recommendation techniques are distinguished by their knowledge sources: user preferences or domain knowledge. Previous work identified four classes of recommendation techniques based on knowledge sources: collaborative, content-based, demographic, and knowledge-based. Each has been extensively studied since the mid-1990s, with well-known capabilities and limitations. This chapter explores the landscape of possible hybrid systems, investigating various hybridization methods and demonstrating quantitative results for comparison.Hybrid web recommender systems combine multiple recommendation techniques to improve performance. This chapter surveys two-part hybrid systems, comparing four recommendation techniques and seven hybridization strategies. It examines 41 hybrid implementations, including novel combinations, and finds that cascade and augmented hybrids perform well, especially when combining components with differing strengths. Recommender systems are personalized information agents that provide suggestions for items likely to be useful to a user. They differ from information retrieval systems in user interaction semantics and personalization. Techniques include collaborative, content-based, knowledge-based, and demographic methods, each with known limitations, such as the cold-start problem for collaborative and content-based systems. Hybrid systems combine techniques to achieve synergy. For example, a collaborative system and a knowledge-based system can complement each other: the knowledge-based component can address the cold-start problem for new users, while the collaborative component can find similar users with unexpected preferences. Recommendation techniques are distinguished by their knowledge sources: user preferences or domain knowledge. Previous work identified four classes of recommendation techniques based on knowledge sources: collaborative, content-based, demographic, and knowledge-based. Each has been extensively studied since the mid-1990s, with well-known capabilities and limitations. This chapter explores the landscape of possible hybrid systems, investigating various hybridization methods and demonstrating quantitative results for comparison.