The two closely related mammalian Notch receptors Notch1 and Notch2 have been shown to play different, and sometimes opposing, roles in development and disease. But what is the mechanistic basis of these differences? Here, Raphael Kopan and colleagues address this question using mice in which the intracellular domains (ICDs) of these two Notch receptors have been swapped (p. 2452). They first show that ICD swapping has little effect on the development of organs in which either Notch1 (T cells, skin, the inner ear and endocardium) or Notch2 (the liver, eye, cardiac neural crest and lung) is known to act alone or is dominant over it paralogue, suggesting that the ICDs are interchangeable. In the case of Notch dosage-sensitive tissues, the researchers further show that the phenotypes observed are due to haploinsufficiency and not due to ICD composition. Together, these and other findings lead the authors to conclude that both the strength of Notch signalling (defined by the number of ICD molecules that get cleaved from the receptor and reach the nucleus) and the duration of signalling (the half-life of active ICD complexes) contribute to the differences between Notch1 and Notch2 functions in many developmental contexts.