The important thing to understand about carbonate diagenesis is not how porosity is created, but how it is destroyed. Detailed core observations from two deeply-buried carbonate platform successions (the Finnmark platform, offshore north Norway, and the Khuff Formation, offshore Iran) show that in both cases most vertical porosity variation can be accounted for by only two or three factors, namely: (1) stylolite frequency, (2) proportion of argillaceous beds, and P) anhydrite cement. The spatial distribution of these factors is determined by the depositional distribution of clay minerals (important for localizing chemical compaction) and the occurrence of hypersaline depositional conditions and associated brine reflux (important for localizing anhydrite precipitation and dolomitisation). However, the intensity of chemical compaction and consequent porosity loss in adjacent beds by carbonate cementation also depend upon thermal exposure (temperature as a function of time). Evidence from the Finnmark platform and other examples indicate that the stratigraphic distribution of early-formed dolomite is also important for porosity preservation during burial, but this factor is not apparent in the Khuff dataset. Insofar as the Finnmark and Khuff platforms can be regarded as representative of carbonate reservoirs in general, recognition of the above porosity-controlling factors may provide the basis for general models predicting carbonate reservoir potential both locally (reservoir-model scale) and regionally (exploration-scale). Distributions of clay, anhydrite, and dolomitization should be predictable from stratigraphic architecture, whereas variations in thermal exposure can be mapped from basin analysis. In the present examples at least, factors that do not need to be considered include eogenetic carbonate cementation and dissolution, depositional facies (other than aspects related to clay and anhydrite content), and mesogenetic leaching to create late secondary porosity.