Payment for Ecosystem Service schemes have become widely discussed in the academic literature and in policy circles over the past 10 years, and indeed an increasing number of schemes have been put in place across the world. This paper has four objectives. First, to explain the idea of a Payment for Ecosystem Service scheme to those not working in this area. Second, to set out a formal model which shows the effects of incomplete information in terms of deviations from a first-best outcome, and which allows the relative merits of outcome- and actionsbased schemes to be compared. Third, to explain why measuring the actual performance of any Payment for Ecosystem Service scheme is difficult. Finally, to set out what we see as the main research needs to inform policy development in the near future.