A recommender system uses past data to learn patterns and an optimization target to rank content. The target determines which characteristics help content rise in recommendations. section 10.1.2
Choosing engagement as the target is not just a technical decision: it sets incentives for creators and can shape online and public discourse. Competition for attention may encourage increasingly extreme tactics; engagement optimization can also enable misinformation or expose vulnerable users to harmful content. section 10.1.2
When YouTube changed its target
When YouTube shifted from ranking videos by raw clicks to promoting them based on length, creators changed their videos’ structure and length to improve their chances of being recommended. section 10.1.2
