In the digital realm, a key role of a platform is to intermediate between its complementors and users, primarily through algorithmic intermediation. Yet, algorithm is often not the only intermediary on the platform, and third-party human intermediaries often co-exist on the platform. The coexistence of third-party human intermediaries on these platforms prompts a critical question of their purpose and strategic contributions to the platform. Departing from traditional indirect network effects logic, this study contends that human intermediaries, even in their act of competing directly with algorithm, can foster algorithmic learning. We test our propositions in the context of Instagram platform, where both algorithm and human influencers act as intermediaries between advertising brands (complementors) and users. Findings suggest that algorithm and human intermediaries have distinctive advantages in intermediating between complementors and users, and algorithm learns vicariously from observing human intermediaries.
