This paper examines capacity-driven Web services from a transactional perspective. Such services are empowered with different sets of operations that are selectively triggered in response to requirements posed on them and hence, need to be satisfied. Each set of operations forms a capacity that either fails or succeeds at run-time. In case of failure and for the sake of business continuity, a capacity driven Web service binds to some transactional properties such as pivot, retriable, and compensatable, and behaves in accordance with the guidelines that each transactional property sets. This paper, also, shows that the particular nature of capacity-driven Web services calls for reviewing these transactional properties' guidelines. © 2011 IEEE.