The Late Coniacian, shallow-marine Bad Heart Formation of the Western Canada foreland basin is very unusual in that it contains economically significant ooidal ironstone. Deposition of shallow-water and iron-rich facies appears to have been localized over the crest and flanks of a subtle intrabasinal arch, in part interpreted as a forebulge and partly attributed to reactivation of the long-lived Peace River Arch. The formation comprises two upward-shoaling allomembers, typically 5–10 m thick, that are bounded by regionally mappable ravinement surfaces. The lower unit, allomember 1, grades up from laminated mudstone to bioturbated silty sandstone, which is abruptly overlain by bioturbated ooidal silty sandstone grading into an almost clastic-free ooidal ironstone up to 7 m thick. Ooidal ironstone was concentrated into NW- to SE-trending ridges, kilometres wide and tens of kilometres long. Ironstone formation appears to have been promoted by: (a) drowning of the arch, which progressively curtailed sediment supply; and (b) enhanced reworking over the shallowly submerged arch and over a fault-bounded block that underwent episodic vertical movement of 10–20 m during Bad Heart deposition. Allomember 2 also shoals upwards from mudstone to bioturbated and laminated silty sandstone but lacks ooids, apparently reflecting a rejuvenated supply of detrital sediment from the arch. The marine ravinement surface above allomember 2 is a Skolithos firmground, above which is developed a regional blanket of ooidal sediment. In the east, ooids are dispersed in a bioturbated silty sandstone with abundant evidence of repeated reworking and early siderite and phosphate cements. Westwards, this facies grades, over about 40 km, into almost clastic-free ooidal ironstone about 5 m thick; the lateral facies change may reflect progressive clastic starvation distal to a low-relief source area. The two allomembers are interpreted to reflect eustatic oscillations of about 10 m, superimposed on episodic tectonic warping and block-faulting events. The development of ooidal ironstone immediately above initial marine flooding surfaces indicates a close relationship to marine transgression, reflecting sediment-starved conditions. Ironstone does not appear to be related to either sequence boundaries or maximum flooding surfaces. The Bad Heart Formation is blanketed by marine mudstone deposited in response to major flexural subsidence and rejuvenation of clastic sources in the Cordillera to the SW.