Consumer Acquisition for Recommender Systems : A Theoretical Framework and Empirical Evaluations

We study the problem of consumer acquisition in recommender systems, by considering a marketplace where a recommender system provider offers personalized incentives to encourage consumer participation. We investigate a model of consumer acquisition that accounts for several key market dynamics, including the value that a consumer contributes to the recommender system, the cost of participation (e.g., privacy loss), as well as the value that a consumer can derive from the system due to network externality created by existing consumers. Importantly, our model enables the recommender system provider to maximize its utility by optimizing the sequence of consumer acquisitions. While solving the consumer acquisition model is computationally prohibitive when there is a large number of consumers, we propose a heuristic algorithm that can obtain near-optimal solutions efficiently. We evaluate the performance of our model on two real-world datasets. We find that our model can properly identify and incentivize "high value" consumers, whose early participation can accelerate the performance improvement of the recommender system and bring significant utility to the system provider both directly and via accumulated network externality. Our work contributes to the design of market mechanisms that facilitate effective consumer acquisition in recommender systems