In this talk I will briefly review the approach to renormalization called causal perturbation theory that avoids dealing with ill-defined objects. This approach started with the seminal paper of Epstein and Glaser (1973) and recently got new interest because of applications to QFT on curved spacetimes. It also allows to uncover some interesting algebraic structures present in perturbative renormalization, like for example Hopf algebras and BV structures.