The increasing amount of information flowing through commercial social networks offers clear advantages for companies who can take a valuable feedback from community actions. In particular, the identification of influential users in on-line social network can support companies in designing and targeting marketing campaigns, as influential gate-keepers and diffusers of information can ignite epidemics through word-of-mouth. In this paper, we model a time-dependent commercial social network as a time-varying weighted directed graph. Moreover, we propose an approach to determine opinion leaders and their contributions to a temporal business value, by taking into account behavioural and structural aspects of the commercial social network.