2019

Benchmarking Data: The Truth About Your Business

Comprehensive records exist to allow us to benchmark reproductive and growing pig performance. A “benchmark” is, in the truest definition, a point of comparison. We often think this is a modern concept, but in fact the practice dates back to ancient times when Japan sent teams to China in 1607 A.D. to learn the best practices for business. This dataset includes multiple genetics and is not stratified by herd size or health. These quotes put the value of benchmarking in perspective:

“All Good to Great companies began the process of finding a path to greatness by confronting the brutal facts about the reality of their business. When you start with an honest and diligent effort to determine the truth of your situation, the right decisions often become self-evident.”
— Jim Collins
“The opportunity is in the variance…You can have your head in the freezer and your rear in the oven and still be at an average temperature. Averages don’t really tell you what’s going on. If you understand the distributions, you can make better decisions.”
— Dennis DiPietre, Vita Plus Swine Summit 2012

The PigCHAMP® 2018 database includes 375 farms. Benchmarking compares the key performance indicators to the mean, standard deviation, and 90th and 10th percentiles. This allows you to identify your individual herd productivity strengths and weaknesses and the industry trends. Anticipating that genetic improvement occurs at a given rate per year, one anticipates that outcomes such as total born and pigs weaned per sow farrowed will increase year over year. Comparing your own individual herds to those trend lines gives you areas of focus. Inherent within evaluating outcomes is the understanding that a records system such as PigCHAMP® utilizes meaningful reference standards, which is important when comparing different databases. Herds with standardized inputs and processes have inherent advantage over systems without standardization.

As you analyze benchmarking data, remember that each variable is independent of another variable. An example is percent repeat services with the mean at 6.6%, upper 10th percentile at 11.80%, and the lower 10th percentile at 1.54%. All herds would prefer to be in the lower 10th percentile in this category. The upper and lower percentiles will also shift by variable. Within the dataset, variation within each variable is illustrated by the standard deviation. This variation provides significant motivation for improvement.

Optimizing sow reproduction is challenging because of the variability of the numerous drivers. When evaluating performance, it is advantageous for veterinarians and producers to compare outcomes over time and by key drivers such as genetics, lactation length, age of gilt at first service, production system, climate, nutritional program, housing, etc. Many of these drivers are interrelated compounding positive or negative effects. Benchmarking quantifies these drivers and their effect on the system as a whole and on each individual farm. Linear time charts allow us to compare periods to minimize some of the negative drivers. With the ongoing changes in several of these parameters, the outcome will change, making the use of benchmarking comparisons more valuable. Each producer and system needs to establish their individual goals.

Optimizing sow reproduction is challenging because of the variability of the numerous drivers. When evaluating performance, it is advantageous for veterinarians and producers to compare outcomes over time.

In the 2018 PigCHAMP® data, producers in the upper 10th percentile had their herds delivering total born of 15.66 per litter and live born of 13.94 while the mean total born was 14.43 per litter and live born of 12.90 per litter. This equates to total born per sow per year of 38.83 and live born per sow per year of 32.38 respectively at the upper 10th percentile, and the mean at 35.78 total born and 31.99 live born per female farrowed per year.

Stillbirths per litter in the upper 10th percentile averaged 0.72 compared to 1.13 for the mean and 1.56 for the lower 10th percentile. Stillbirth reduction remains an opportunity on most herds. Numerous interventions continue to be applied, but the largest impact often resides with the staff and hours of staffing. With very few herds able to effectively staff 24 hours per day, and with 60% of the farrowings occurring in the non-attended hours, this opportunity is difficult to capture. Induction of farrowing is an effective strategy if farrowings are closely attended. Mummies per litter for the upper 10th percentile were 0.10 compared to 0.40 for the mean and 0.71 in the lower 10th percentile. Mummy rates are highly dependent on the staff’s desire to identify early gestation losses. Mummies form at approximately 40-45 days of development when skeletal calcification occurs. These early mummies are frequently lost in the placentas or naturally discarded below the flooring.

Piglet survivability is another large economic opportunity for producers. Preweaning mortality in the database in the upper 10th percentile was 10.31% compared to mean percent at 14.85% and the lower 10th percentile at 19.81%. As total born increases, average birthweights generally decline and individual birthweight variability increases contributing to lower survivability. In good health herds, 60-70% of the mortality occurs within the first 72 hours post birth. Strategies improving survivability include proper piglet environment, colostrum management via split-suckling and balancing pigs per sow based on functional teats. These strategies are well understood, but rely on consistent staffing and are hindered by staffing hours. Herds that staff 24/7 have opportunity to improve this parameter.

Year over Year Comparison

Pigs weaned per litter for the upper 10th percentile was 12.15 with the mean at 11.23 and the lower 10th percentile at 10.45. Pigs weaned per mated female per year for the upper 10th percentile was 29.22 with the mean at 25.28 and the lower 10th percentile at 21.20. Average age in days for weaning for the lower 10th percentile was 18.80, mean at 20.74, and the upper 10th percentile at 23.00 days. It is interesting that given the correlation of weaning weight to weaning age and post weaning performance, this parameter remains relatively static. Farrowing rate for upper 10th percentile was 90.30% with the mean at 83.90% and the lower 10th percentile at 76.98%. The number of producers recording litter weights was 20% of the total farms within this database. These herds had a mean litter wean weight of 132.95 pounds with the upper 10th percentile at 167.15 pounds and the lower 10th percentile at 98.74 pounds.

Death rate percentage for upper 10th percentile was 7.20% with the mean at 11.68% and the lower 10th percentile at 16.20%. Culling rate for the upper 10th percentile was 28.30% with the mean at 45.06% and the lower 10th percentile at 61.60%. Culling rate variable is highly herd specific and driven by genetics, health, and economics. Genetic herds target annual replacement rates of 65%, while herds with closure for PRRS or Mycoplasma hyopneumoniae will have low replacement rates for a period.

Trends over time are an important aspect of benchmarking and when linked to the yearly database provide additional value of your herd’s performance to other producers. Comparing the upper 10th percentile to the mean over time also identifies if biological and management limitations are being reached. The following table illustrates key production variables of the mean and the upper 10th percentile from the historical PigCHAMP® database over the period of 2014–2018. Comparing the upper 10th percentile to the mean over time also identifies if biological and management limitations are being reached. If the upper 10th percentile improvement delta is higher than the mean, the best herds are improving at a faster rate.

Total born increases year on year are approximately 0.235 piglets per litter while born alive is 0.16 piglets per litter. Improvement rate is slightly higher over the 5-year period between the mean and the upper 10th percentile. Liveborn per female per year increased by 0.8 pigs for the mean and 1.04 for the upper 10th percentile herds over the five-year period. Stillborns increased 0.23 for mean and similarly for the upper 10th percentile during these five years. Increasing total born with an increasing stillbirth rate resulted in the increase in live births per litter. Mummies per litter increased 0.10 for mean, 0.03 for upper 10th percentile.

Pre-weaning mortality decreased by 0.77% for mean and increased by 1.08% for upper 10th percentile. Average age at weaning increased 0.57 for mean, and 1.20 days respectively for upper 10th percentile. Average litter weaning weight increased 1.38 pounds for mean and 7.58 pounds for upper 10th percentile. Pigs weaned per litter weaned increased 0.42 for mean and 0.53 for upper 10th percentile. Pigs weaned per mated female per year increased 0.90 for mean, 1.21 for upper 10th percentile.

Farrowing rate decreased 0.37% for the mean and 0.49% for upper 10th percentile. There has been a general industry concern that farrowing rates have declined. This dataset would support this decline. The goal of a 93% farrowing rate seemingly is difficult to achieve. Death rate increased 2.87% for mean and 1.63% for upper 10th percentile. This trend continues year over year and needs to be an area of focus. The reasons for this trend are related to earlier euthanasia of rectal prolapses, the increase in uterine prolapses, and the increase in open pen gestation. In our comparison, open pen gestation averages 2% higher mortality than stall housing. Culling rate increased 1.60% for mean and decreased for the upper 10th percentile.

Trend variables worth comparing between the 10th and 90th percentile are stillborn, mummies, death rate, and culling rate as these do not necessarily correlate within the same herds. For stillborn, the lower 10th percentile herds have increased 0.1 per litter while the 90th percentile herds increased 0.23 per litter. This difference is logical if associated with increasing total born. Mummies per litter increased 0.03 per litter for the upper 10th percentile while increasing 0.21 per litter for the lower 10th percentile.

Death rate of sows interestingly increased during the 5-year period in both 10th and 90th percentile categories indicating a generalized shift. Culling rate varied year by year with some decline or increase trend within the 10th and 90th percentile suggesting other factors such as markets’ influence. If variables such as stillborn, mummies, and death rate remain at the 90th percentile (low) year after year within the same herd, these are the ones to deep dive into their processes, procedures, and drivers. Lastly, mean and 90th percentile sow inventory trended up during the 5-year analysis period.

In summary, producers have to be excited to have productivity continuing to increase. These increases are critical to maintaining or lowering costs in our globally competitive industry. This variation in productivity pales in comparison to variation in financial performance. Linking production data with financial data is key. Opportunities exist for continual increases with staffing, staff training, and health being critical components to capture the genetic potential in productivity.


Dr. Joe Connor
Dr. Connor obtained his D.V.M. from the University of Illinois in 1976, obtained a Master of Science in Veterinary Medicine from the University of Minnesota in 2006, and completed the Executive Veterinary Program in 2009 with the University of Illinois. Dr. Connor is the founder and past president of Carthage Veterinary Service, Ltd. (CVS). Carthage Veterinary Service, Ltd. provides consulting to producers throughout the United States, the Americas, Europe, and Asia. Dr. Connor focuses on disease management and elimination and production economics. Dr. Connor was a leader in segregated production and wean–to-finish technology.

Dr. Connor was recognized as a member of the PIC Hall of Fame in 2016, and received the Master’s Award from The National Hog Farmer in 2009, the Leman Science in Practice Award in 2004 from the University of Minnesota, and American Association of Swine Veterinarians (AASV) Swine Practitioner of the Year Award in 1995 from the American Association of Swine Veterinarians.