In order to pin down the fundamental nature of dark energy, and thus to understand what most of the Universe is actually made of, new and more precise observations are required, along with more efficient and reliable statistical techniques to interpret those observations correctly and to understand the implications they have for our theoretical models of the Universe. The outstanding challenge posed by the nature and properties of dark energy is giving rise to a flourishing of proposals for new observational campaigns. Type Ia supernovae, gravitational lensing, cluster counts and baryonic acoustic oscillations are some of the techniques available to study dark energy, through its influence on the past expansion history of the Universe and on the growth of structures. I will review the current observational status and outline the promise of various proposals which in the next decade should considerably improve our understanding of dark energy. I will comment on the prospects of each technique and on the possibility of an optimal strategy for dark energy investigation.