Decentralized decision making is a fact in the modern business world accompanied by extensive research that looks into its consequences for overall firm profits. We study the interactions of decentralized marketing and operations divisions in a corporation and explore their impact on overall firm profits in the case with and without coordination of the two decentralized units. We assume that the marketing department is responsible for the price that influences the demand (sales), and the operations department is responsible for the production rate. We allow for backlogging over time. We model the interdependence involving marketing and operations decisions as a non-cooperative differential game, with the two divisions as strategically interacting players. We find that, without coordination, strategic interactions of marketing and production result in inefficiencies that can quantitatively be substantial. Next, we introduce a dynamic transfer pricing scheme as a coordination device and evaluate if it establishes efficient (first best and fully coordinated) outcomes. We show that if production and marketing play a game with pre-commitment strategies, there exists a dynamic transfer price that efficiently (fully) coordinates decentralized decision making and hence results in Pareto-efficient company profits. If the two decentralized divisions play a game without pre-commitment, dynamic transfer prices can partially coordinate decentralized decision making but fail to fully eliminate overall inefficiencies arising from strategic interactions among decentralized divisions.