Thermoelectric materials are opening a promising pathway to address energy conversion issues governed by a competition between thermal and electronic transport. Improving the efficiency is a difficult task, a challenge that requires new strategies to unearth optimized compounds. We present a theory of thermoelectric transport in electron doped SrTiO3, based on a realistic tight binding model that includes relevant scattering processes. We compare our calculations against a wide panel of experimental data, both bulk and thin films. We find a qualitative and quantitative agreement over both a wide range of temperatures and carrier concentrations, from light to heavily doped. Moreover, the results appear insensitive to the nature of the dopant La, B, Gd and Nb. Thus, the quantitative success found in the case of SrTiO3, reveals an efficient procedure to explore new routes to improve the thermoelectric properties in oxides.