We say that a social choice rule is implementable with (small) transfers if one can design a mechanism whose set of equilibrium outcomes coincides with that specified by the rule but the mechanism allows for (small) ex post transfers among the players. We then show in private-value environments that any incentive compatible rule is implementable with small transfers. Furthermore, in order to showcase the applicability of our results, we relate them to the recent developments in implementation theory, such as (i) continuous implementation in Oury and Tercieux (2012), (ii) robust undominated Nash implementation in Chung and Ely (2003), (iii) the full surplus extraction property in Cr ́emer and McLean (1988), and (iv) the role of honesty in Dutta and Sen (2012) and Matsushima (2008). Next we argue by example that our mechanism fails to work in interdependent-value environments. We therefore identify a condition under which our results can be extended to interdependent-value environments and connect this identified condition to the notion of strategic distinguishability due to Bergemann and Morris (2009b). This shows that the extension of Abreu and Matsushima (1994) to incomplete information environments generates a substantial constraint, which can be completely dispensed with for the case of virtual implementation, as Abreu and Matsushima (1992b) show.