The paper presents a methodology for the analysis of distributed business information systems, which is intended to support the design of availability improvement solutions; e.g., the redundancy of equipment, the replication of databases, or the installation of diagnosis aid tools. The methodology presented combines an availability model, representing the behavior of the individual components and the consequences of their failures upon the information services, with a cost model, representing the economic damages that the unavailability of those services will cause. The paper includes a case study concerning a large retail company, in which two alternative solutions for the architecture of the information system are assessed and compared.