The time dependent response of a polar solvent to a changing charge distribution is studied in solvation dynamics. The change in the energy of the solute is measured by a time domain Stokes shift in the fluorescence spectrum of the solute. Alternatively, one can use sophisticated non-linear optical spectroscopic techniques to measure the energy fluctuation of the solute at equilibrium. In both methods, the measured dynamic response is expressed by the normalized solvation time correlation function, S(t). The latter is found to exhibit unique features reflecting both the static and dynamic characteristics of each solvent. For water, S(t) consists of a dominant sub-50 fs ultrafast component, followed by a multi-exponential decay. Acetonitrile exhibits a sub-100 fs ultrafast component, followed by an exponential decay. Alcohols and amides show features unique to each solvent and solvent series. However, understanding and interpretation of these results have proven to be difficult, and often controversial. Theoretical studies and computer simulations have greatly facilitated the understanding of S(t) in simple systems. Recently solvation dynamics has been used extensively to explore dynamics of complex systems, like micelles and reverse micelles, protein and DNA hydration layers, sol–gel mixtures and polymers. In each case one observes rich dynamical features, characterized again by multi-exponential decays but the initial and final time constants are now widely separated. In this tutorial review, we discuss the difficulties in interpreting the origin of the observed behaviour in complex systems.