On November 21, 2000 CE, the c. 48 × 106 m3 Paatuut landslide in West Greenland triggered a tsunami with a maximum runup height of c. 45 m. Although a field team examined the landslide in the immediate aftermath, prior events and processes, in addition to the cause of the landslide, were never studied. We combined field data, satellite images, and historical photos to bridge this knowledge gap. Our investigation unveiled that a hitherto unknown c. 55 × 106 m3 landslide occurred at the same slope in May or June of 1996. This landslide was a frozen debris avalanche, and we suggest a result of permafrost degradation since c. 1949. The subsequent 2000 landslide and tsunami removed and obscured the traces of the 1996 landslide. Interestingly, the 1996 landslide caused a tsunami with a runup height of only 15 m near the landslide impact area, one-third of the 2000 tsunami. We applied tsunami modelling and interpretation of morphological field evidence to explore why these volumetrically similar landslides on the same slope could produce markedly different tsunami runup heights. The deposit of the 1996 landslide on the coastal slope produced a large, unconsolidated, wet sediment volume that could be entrained in the 2000 landslide, and in addition, reducing the basal friction of this later event. Furthermore, differences in drop height and rheology between the two landslides may explain the different tsunamigenic potential. We see evidence of much older post-glacial landslide activity on the slope, constituting a static preconditioning factor for the landslides. The 1996 and 2000 landslides demonstrate the incomplete record of large landslides in the Arctic and the importance of considering the runout path, substrate, and entrainment in determining the tsunamigenic potential of landslides. Above all, they also demonstrate the sensitivity of these Arctic slopes to global warming and associated permafrost degradation.