In a job shop, the intervals between job arrivals and the job processing times fluctuate and cannot be forecast. When many jobs (customers) arrive at a job shop at the same time, some customers may not place an order because they do not want to wait. To prevent such a situation, the job shop may temporarily increase the process capacity, thereby reducing the waiting time. If its timing is not correct, however, such an increase can also unexpectedly increase the idle time in the process. In this paper, the effects of fluctuations in the quantity of work arriving on waiting time, idle time and the rate of losing customers is studied using numerical calculations which assume extreme fluctuations. These characteristics show different delays in their response to fluctuations in job arrivals. The paper deals with the index and timing for temporarily increasing and decreasing the process capacity to absorb the fluctuations in the quantity of work arriving, that is under what circumstances to do so and when.