Averages are Not Sufficient Benchmarks for The Future
Producers must learn to master high average performance within a narrow variance of outcomes.
By Dennis DiPietre
We know that almost all benchmarking is carried out utilizing the average value of the various performance metrics of the farm over a period of time. The average is a very important and revealing number, and while it is still necessary, it is no longer sufficient to tune production systems for the challenge laid down by the Food and Agriculture Organization (FAO). We know that averages by themselves hide a lot of critical information in the production process.
Because we did not see the future clearly, the modern production system is laid out and built to attain high average performance, but it is incapable of measuring variance around the average. You would be hard-pressed to find a single producer of any size in the United States that actually weighs every single animal in a building at any time in its production lifespan. Without individual measurements, we cannot measure variation.
The exception is the sow herd, where many individual animal measurements are already made and recorded but the benefit is rarely mined. What insight could be obtained by benchmarking not only the average number of pigs weaned per female, average live born and average pre-weaning mortality, but also their associated variances or standard deviations? What does it likely imply about a system if the average number of pigs weaned is 28 per year, but this is achieved with a five-pig standard deviation versus a two-pig standard deviation? The former implies the system is alternatively surged and starved compared to a steadier flow in the second case.
Beyond the sow farm, we know almost nothing about individual animal feed or water consumption, daily growth, comfort level or weight gain, as our technology does not allow these things to be measured in a cost-effective way.
What we do know is that for most farms in the United States, by the time a contemporary group is fully marketed, you need a 100-lb. weight range or more to capture 95 percent of the market weights coming out of the system. A good rule of thumb seems to be that the day before the first marketing; a 200-lb. range is needed within a finishing building to capture the largest and smallest animals. That's evened out to the 100-lb. range by marketing cuts and extra time for stragglers. When someone says they market their animals at 280 lbs., what they really mean is that about 20 percent of them were within a few pounds of 280 and the rest were scattered around the mean at varying, larger distances.
The standard deviation is a now a critical statistic needed to accompany the mean since it seems clear that some of the biggest gains in efficiency are likely to come through reduction in variance, even when maintaining the same mean performance (though let's not rule out improving mean performance too!). For many years we relied on the back-of-the-envelope calculation of the single-pig optimum weight as a function of the price of pigs, the price of feed and the efficiency of the animal. What we now know is that when you market groups larger than one pig at a time (like a truckload, building, site, flow etc.), the profit-optimal average weight also critically depends on the variance of weights in the marketing group.
Understanding this opens up $1to $5 profit improvements per head when only the low hanging fruit available through this insight is captured. That may not sound earth-shaking, but consider that the average profitper- head-sold during the period from 2000 to 2009 was just over $4, as estimated by the Iowa State University profit series. The standard deviation of profit per head was over $20.
So-called low-cost firms that believe they have extracted as much value as possible out of the resources they use will find that the revenue left on the table from the variation they have created - but currently don't measure - is perhaps the most significant opportunity to capture since pigs came out of mud puddles and moved into buildings. Scarce global resources will only be within the reach of those who master high average performance within a narrow variance of outcomes.