This paper is a reflection on the conditions required to use Optimal Matching Analysis (OMA) in sociology. The success of OMA in biology is not related to any supposed similarity of the method with biological processes but comes from setting costs in OMA in accordance with biological theory. As sequences in sociology are made of events and time, the determination of costs should be guided by sociological theories of time. After a discussion of the sociological meaning and consequences of costs, this paper comes back on the Dynamic Hamming Distance and the body of social theories of time (Durkheim, Elias, Bourdieu) from which it is derived as an example of how sociological theory can inform cost setting in using OMA in sociology.