This paper follows the research thread of “motivation → mechanism → value”, and adopts a dual case study of Douyin and Netflix. It induces the value drivers of recommendation algorithms driving content platforms, and analyzes the mechanisms of value creation driven by different value drivers. Then, this paper proposes the theoretical framework for value creation driven by recommendation algorithms on content platforms. The study finds that the value drivers for content platforms using correlation-based recommendation algorithms are “people divided by groups” and “items clustered by categories”, which promote traffic value creation by generating learning effects and economies of scope; and the value drivers for content platforms using causal recommendation algorithms are “making the best use of talents” and “making the best use of things”, which enhance retention value creation through compounding effects and economies of speed. This paper enriches the research on value creation of algorithm-driven business model, and is instructive for further exploration of intelligent business model.