Miocene carbonate reservoirs in Central Luconia, offshore Sarawak, Malaysia, have been delivering gas for over 30 yr. In this paper, learnings from that period of production are used to understand the key drivers affecting flow during production and recovery optimization in existing fields as well as development decisions for new discoveries. The large data set, generated over more than 40 yr, was analyzed in a consistent manner through a holistic database, constrained by a stratigraphic framework, to allow reservoir units to be compared like-for-like (“integrated knowledge base” [IKB] concept). Carbonate reservoir heterogeneities impacting flow are grouped into “horizontal–heterogeneities”—argillaceous flooding layers and exposure-related karst—and “vertical–heterogeneities”—large-scale architectural elements, found especially along platform margins. Both types of heterogeneities control water ingress during production and influence the recovery mechanism. Argillaceous flooding layers can act as baffles, holding back water rise during production, or can form pressure compartments. Long-lived, fault-bounded reef margins, carbonate shoals, islands, and karsts can be vertical conduits for aquifer inflow. Platform shape and architecture impact column height and hence recovery efficiency. Additional drivers impacting recovery were found to be gas-column height, aquifer size and permeability, pressure connection to neighboring fields, and field development concepts. All drivers identified impact decisions throughout the field life, e.g., well count and design, intervention capabilities, evaluation and mitigation of early-water breakthrough, reservoir management, selecting enhanced recovery methods, and abandonment pressure. The IKB allowed to derive “big rules” on what matters for flow, which were used to decide on development strategies for greenfields in Central Luconia. The presented outcomes can be extrapolated to comparable carbonate systems, whereas the IKB approach can be adapted and applied to other mature basins and reservoir types where equally vast and historic data sets are awaiting to be used in the current era of digitalization.