Customer ordering behaviorin the form of size and timing of orders are critical for manufacturing systems such as make-to-order or assemble-to-order systems. To be competitive a Make-to-order manufacture must be flexible, both with regards to volume and product mix. A particularly critical parameter in determining the flexibility needs in the case of responsive manufacturing environments is the distribution of order sizes. In this paper several methods for describing the volatility of order sizes are presented. A discrete event simulation of a single server batch manufacturing system is subsequently conducted using various distributions of order sizes. The aim is to investigate 1) which measures for order size volatility best relates to the volatility of the output rate from the manufacturing system 2) how does the output stability from the system deteriorate as order sizes become more volatile.