Answers are grounded in this paper and cite their sources.
Revenue optimization represents a major strategic challenge for large-scale companies and e commerce platforms.In environments characterized by massive and heterogeneous user populations, the real - time adaptation of marketing strategies becomes a critical driver of performance. This paper proposes an adaptive intelligent algorithm based on online learning to maximize revenue generated through premium subscription plans within a platform serving millions of users. The problem is formulated as a sequential decision-making process under uncertainty. Nine distinct marketing strategies, differentiated by promotional offers and commercial incentives, are dynamically presented to users during their navigation sessions. No prior knowledge of conversion rates is assumed, and each user interaction produces binary feedback corresponding to success or failure. The objective is to rapidly identify the strategy yielding the highest conversion rate while minimizing exploration costs. Experimental results demonstrate that the Thompson Sampling based algorithm outperforms both random selection and Upper Confidence Bound (UCB) approaches in identifying the optimal strategy. After only a limited number of interactions, the proposed method efficiently detects the most profitable strategy, thereby reducing advertising and operational costs.
No takes yet. Share an insight, caveat, or question.
Bienvenue et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: