For a long time, the record systems which we keep have given us only the broadest and most limited look at system performance. The biggest confounders have been the emergence of multi-site production and the odd primitiveness that dogs farm or biological production.
Ever since we moved production to separate sites, the records systems have focused collection and evaluation of data via metrics like FE and ADG on the separate sites. There are few if any full system metrics which connect the performance of the sow farm to finishing or even a nursery to finishing performance. Perhaps most critical is the lack of tie-in of the final outcome of performance, the kill data to any production data. Second, because our systems do not allow for any individual animal performance data (except at kill), we work constantly with averages which hide the variation which spreads itself throughout the production process.
We normally cannot know ADG or FE for instance until a building empties and then we only have a grand average of all the animals in the building. As building sizes have increased, this average is now hiding ever larger deviations from the central tendency.
We are living in the age of variation and its control. This is the next major focus of cost reduction and quality improvement but we are in the infancy of creating the kind of systems and records which both mitigate variation or have any hope of properly measuring it.
For instance, everyone has seen kill sheets with identical average weights which yield far different monetary outcomes. This of course is due to the sort loss differences caused by widely differing distributions of weights on the two loads. If you would go back into the production records for those two loads you could just as easily find nearly equal ADG and FE though very different costs of production. This happens because the cost effects of outcomes above and below the mean generated FE and ADG create very different costs and these costs are not symmetric on each side of the mean. That is, the better FE's on the left side of the mean do not off set the costs of the high FE's on the right side of the mean.
This leads us to focus on the profit maximizing average load weight first as a means to generate the best outcome possible given the production process that is currently in place. Over time we can improve the production process and the target weight will then change. If that average weight is properly estimated and adhered to, the FE and ADG which occur on that load are the profit maximizing FE and ADG. Rather than compare them to an outside database average or benchmark, it is best to focus on the profit maximizing average weight of the load and let the FE and ADG just happen. This doesn't mean there is not waste in the FE or that the ADG is the best the farm can do over time but these metrics do not drive the profit maximizing decision, they result when the profit maximizing weight is selected.
As an example, if you select the profit maximizing weight, you can reduce FE by lowering that weight. You certainly will not make more money but you may have an FE that is more worthy of bragging about among those who avidly compare those things. You would be much better off my reducing the variation of weights around the profit maximizing average or focusing on better feeder management. By producing a truck load with the proper average weight with fewer heavy and light weight animals in the mix, FE will improve even at the same target weight. Because the changes are not symmetric, the reductions in FE which come from eliminating heavy weight animals while keeping average marketed weights the same, in almost all cases will outweigh the increases in FE generated from increasing the weights of smaller animals in the mix. If it doesn't, that only means the optimal FE on that day is very likely higher than the one you are targeting.
