Accretion of minor satellites has been postulated as the most likely mechanism to explain the significant size evolution of massive galaxies over cosmic time. A direct way of probing this scenario is to measure the frequency of satellites around massive galaxies at different redshifts. Here we present our study of satellites around massive galaxies (M star ~ 1011 M⊙) up to z ~ 2. We find (Fig. 1) that the fraction of massive galaxies with satellites down to 1:10 mass ratio is ~15% (~30% down to 1:100), not varying with redshift (Mármol-Queraltó et al., (2012)). We also find that our satellites are younger than their central galaxies at low z (Mármol-Queraltó et al., 2013). Then, if minor merging is acting to form massive galaxies, their ourtskirts should be younger than their cores. The challenge to find this age gradient in nearby massive galaxies is opened.