Market-based compute grids encompass service providers offering limited resources to potential users with varying demands and willingness to pay. Providers face difficult decisions about which jobs to admit and when to schedule admitted jobs. For this reason, researchers investigate various heuristics for admission control and scheduling that aim to yield high revenue for providers. Such research has no framework within which to understand the revenue bounds associated with various workloads. This paper proposes a tractable analytical model for joint optimization of job admission and scheduling strategies aimed at provider revenue maximization. We show how solving this model yields maximum provider revenue given a linear user utility function. Our model can be used to understand the operating limits of heuristics for admission control and scheduling, and can also be used to investigate the implication of varying job.