​

Abstract: The current information flow from big data analytics to marketing decision making is a one-way channel in that there is actually no feedback mechanism from the implemented decisions back to the big data analytics based customer behavior assessment. To address this inadequacy, a two-way channel is proposed in this study. Deep-Q-Network (DQN) reinforced learning would be used together with big data analytics to provide not only accurate customer classification, but also the generation of precise differentiated marketing strategies for each customer segment, thus providing a natural solution for binding big data analytics even closer with AI techniques because of its designed feedback architecture. This approach would be further evaluated based on risk, cost, Dupont and SWOT analysis. As a result, the novel precise marketing approach proposed here, integrating vividly big data analytics models and artificial intelligence algorithms with strictly designed business model, could not only create value for the firms having implemented the business model, but also promote efficiency for customers making their purchases as well.
 

Keywords: Intelligent decision making; Purchasing pattern; Precise marketing strategies; Architecture; Evaluation