This paper revisits the information sharing strategies of retail platforms through the lens of Bayesian persuasion. Traditional studies have mainly focused on scenarios where the low bound of market size (LBMS) is lower. However, this research relaxes this limiting assumption. A game-theoretic model is constructed consisting of a supply chain with a supplier and a retail platform. The analysis reveals that when LBMS is high, the platform's optimal strategy is to withhold the demand information if the supplier does not engage in encroachment. Conversely, when LBMS is relatively low, a Bayesian persuasion mode of information sharing emerges as the optimal approach. The study further examines the strategic dynamics when the supplier undertakes agency encroachment. If the commission rate is sufficiently high, the platform's best response is to adopt a full information sharing policy. Otherwise, the platform's optimal strategy depends on the market size lower bound, withholding information for high LBMS, and resorting to Bayesian persuasion for low LBMS. This work contributes to a deeper understanding of information management strategies in supply chain especially when the market size is low.