The determination of photospheric abundances in late-type stars from spectroscopic observations is a well-established field, built on solid theoretical foundations. Improving those foundations to refine the accuracy of the inferred abundances has proven challenging, but progress has been made. In parallel, developments on instrumentation, chiefly regarding multi-object spectroscopy, have been spectacular, and a number of projects are collecting large numbers of observations for stars across the Milky Way and nearby galaxies, promising important advances in our understanding of galaxy formation and evolution. After providing a brief description of the basic physics and input data involved in the analysis of stellar spectra, a review is made of the analysis steps, and the available tools to cope with large observational efforts. The paper closes with a quick overview of relevant ongoing and planned spectroscopic surveys, and highlights of recent research on photospheric abundances.