Seeding an Uncertain Technology

2025-06-12

AI Summary

This paper addresses the optimal selection of initial seed users by startups under uncertainty regarding product quality and unknown underlying social network structure, aiming to maximize diffusion of high-quality products while containing risks from low-quality ones. Method: We propose the first sequential information design framework integrating Bayesian learning with stochastic network diffusion, formally modeling how beliefs about product quality evolve through observed adoption cascades. Contribution/Results: We prove that the optimal seed set size grows logarithmically with network size. In heterogeneous user networks, the asymptotically optimal strategy selects seeds exclusively from the user type with the lowest per-unit “viral probability” cost. This yields the first verifiable, quantitative criterion for seed selection—rigorously justifying empirical practices such as targeting key opinion leaders—while ensuring theoretical soundness and practical applicability.

Technology Category

• Reasoning under Uncertainty: Stochastic Optimization
• Search and Optimization: Learning to Search
• Game Theory and Economic Paradigms: Imperfect Information

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Web Mining and Content Analysis: Content-based information diffusion
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

I study how a startup with uncertainty over product quality and no knowledge of the underlying diffusion network optimally chooses initial seeds. To ensure widespread adoption when the product is good while minimizing negative perceptions when it is bad, the optimal number of initial seeds should grow logarithmically with network size. When there are agents of different types that govern their connectivity, it is asymptotically optimal to seed agents of a single type: the type that minimizes the marginal cost per probability of making the product go viral. These results rationalize startup behavior in practice.

Problem

• Optimal seed selection for uncertain product quality
• Logarithmic seed growth with network size
• Single-type seeding for viral probability maximization

Innovation

• Optimal seed number grows logarithmically with network
• Seed single agent type for viral probability
• Minimize cost per viral probability margin