Five steps are required to design organizational performance measures for human decisionmaking: construction of the objective matrix, assessment of information preference functions, specification of the parameters of the information volume penalty function, and solving the instrument design problem. The implementation scenario is based on the quality control function of an industrial firm. An activity value function is generated for each individual to describe the quality control manager's information preferences. Next, penalty function parameters are specified. Once preference functions measuring the value of different information sources are assessed, and the parameters of the information volume penalty functions are estimated, the design problem can be solved. The implementation exercise shows that the approach for modeling performance interdependencies must be based on conditions surrounding specific problems. This rule applies to the formation of independent activity and component subgroups, and the form of models for aggregating performance information preferences over these subgroups. Tables and references are included. (Author abstract modified)