The standard in the field says that any variation found within the data from a gage R & R study that falls below 10% is acceptable. Once above 10%, but still below 30% variation, your system may still be used, but only under specific circumstances where addressing the issue is not possible at that time. With a variation above 30%, you should no longer be using your measurement system, as some part of it needs immediate attention in order to assure accuracy. The variation found by a gage R & R study is calculated using a ratio of the precision of the measurement system to the tolerance of the manufacturing process.