The two key metrics are speed and accuracy. The accuracy can be measured in terms of the amount of time that the couriers, when they arrive at their trucks, need to spend to re-arrange their trucks into the correct order. A reduction in this time can be used as a proxy for the accuracy of the loaders. The speed of the loaders can also be measured. We will likely speed the belt up to accommodate for the five minutes of stretching and warm-up exercises. Thus, we will measure the number of packages down the belt per minute to assess speed improvements. Our scanning devices along the belt give us an accurate count of package throughput that we can use for this measure. We will also add another measure to account for the extra staff member, by measuring throughput per minute per employee as well, in order that we may understand the degree to which any improvements in throughput are attributed to the extra staff member and which are attributed to other factors.

We will thus measure turnover rates in order to evaluate whether or not our strategies to reduce turnover have been effective. We will also measure turnover rates vs. performance improvements to ascertain the degree to which the former affects the latter. At present we only suspect a relationship. This is a null hypothesis that we will need to either confirm or reject by measuring the turnover rates against the performance improvements.
Works Cited:

Murphy, Lawrence R. & Cooper, Cary L. (2000). Healthy and Productive Work: An International Perspective. Retrieved May 15, 2009 from http://books.google.com/books?id=gVIVAAAAIAAJ

No author. (2009). Ramp -- Manual Loading and Unloading. Occupational Safety and Health Administration. Retrieved May 15, 2009 from http://www.osha.gov/SLTC/etools/baggagehandling/ramp_manual.html