We consider performance characteristics of stochastic cyclic flow lines where identical sets of jobs are repeatedly produced in the same loading and processing sequence. Each machine has an input buffer with finite capacity. We characterize the performance with regard to mean processing time variation between jobs and random variation of processing times. We show that job variation as well as random variation significantly increases the cycle time. We also identify that random variation could absorb job variation to some degree and hence an appropriate degree of random variation could reduce the cycle time for a line with a given job variation. We compare the performance of cyclic order production with that of random order production. We discuss effect of job processing sequence briefly. Finally, we explain the implications of the performance characteristics on design of cyclic flow lines.