Understanding variation in a pig production process is a data thing. In order to manage something or improve it, it must be measured. Measuring variation has not been a high priority in the first phase (if you will) of the modern pork industry, at least at the production level. Most record systems and the available technology to-date have focused on measuring and recording group averages. Kill sheets traditionally have provided more information about variation but until relatively recently, most producers have not had access to individual pig outcomes even from the packer (which is the only place in the vast majority of operations, where individual pigs are weighed and evaluated for quality). Typical kill sheets group pigs in ranges without giving individual animal weights and lean percents.
To play in the modern swine industry and to have a chance at managing variation with some degree of effectiveness, measurements of individual pigs is required. While averages can be calculated by weighing or measuring the entire group and dividing by the number of individuals, the classic measures of variation (variance, standard deviation and coefficient of variation—a measure of relative variation) require individual pig measurements in order to calculate. This is performed by the packer in the ordinary course of business as they kill pigs and determine what to pay for them, but there just really has not been much possibility, in a cost effective manner, to measure a sufficient number of individual pigs in the production process, given current building and equipment technology to measure variance. That is changing. Sort scales and individual feeding stations with electronic id of the animals are beginning to provide the first convenient and cost effective way to gain access to some of the first data needed for analysis of the variation present.
Random sampling of pigs coupled with measuring the sampled individual’s weight or some other characteristic is another method which can give valuable insight into variation for the entire population but it is difficult and costly to do properly in the field. How are we going to use the information that is available while transitioning to more effective measuring technologies? We’ll take a stab at that shortly.
