Widely implemented at national metrology institutes (NMIs), the coincidence method is a technique to assay a wide variety of radionuclides which decay through two or more types of radiation. Through a survey of the literature, this paper seeks to describe the main aspects of one of the most powerful direct methods available in radionuclide metrology. The basics of coincidence counting and the efficiency extrapolation method are covered. The problem of non-linearities in the extrapolation curve is also considered. The main characteristics of variants to the conventional coincidence instrumentation are presented.