The endosomal sorting complex required for transport (ESCRT) machinery mediates cargo sorting, membrane deformation and membrane scission on the surface of endosomes, generating intraluminal vesicles (ILVs) to degrade signaling receptors. By live-cell imaging of individual endosomes in human cells, we find that ESCRT proteins are recruited in a repetitive pattern: ESCRT-0 and -I show a gradual and linear recruitment and dissociation, whereas ESCRT-III and its regulatory ATPase VPS4 display fast and transient dynamics. Electron microscopy shows that ILVs are formed consecutively, starting immediately after endocytic uptake of cargo proteins and correlating with the repeated ESCRT recruitment waves, unraveling the timing of ILV formation. Clathrin, recruited by ESCRT-0, is required for timely ESCRT-0 dissociation, efficient ILV formation, correct ILV size and cargo degradation. Thus, cargo sorting and ILV formation occur by concerted, coordinated and repetitive recruitment waves of individual ESCRT subcomplexes and are controlled by clathrin.