Targeting the grid with computer vision
Many pigs consistently fall outside the “sweet spot” at the slaughter plant
By Charles Grant
In 1996, I worked with Bill Maclean, the plant manager at Schneider’s on Marion Street in St. Boniface, Manitoba, not far from the old Manitoba Pork office and delivery docks. We were figuring out the best way to ship fresh-chilled pork to Japan. It also happened to be "Take Your Kid to Work Day," so I brought my daughter along for a tour of the plant and an introduction to Mr. Maclean.
After the tour, we sat down in Bill’s office. Having just completed filling an order for pork shoulders, Bill turned to my daughter and asked if she had any questions. After a moment of thought, she asked, “What’s the most frustrating part of your job, day in and day out?” Bill paused before responding, “I can only make money on 60% of the pigs I buy.” Intrigued, my daughter asked, “Why is that the case, Mr. Maclean?” He explained, “Because there’s too much variability in size and quality from one carcass to the next.”
Story from today:Fast forward 28 years to 2024, when I found myself at Johnny Prejet’s feeder barn in Notre Dame du Lourdes, Manitoba. We were working on a project to improve the speed and efficiency of pig weighing with less labor involved. Gustavo, the barn manager, had immigrated from Mexico and was exceptionally skilled in barn operations.
Recalling my daughter’s question for Bill Maclean, I asked Gustavo, “What’s the most frustrating part of your job, day in and day out?” Without hesitation, he replied, “The biggest challenge is that, despite my best efforts, I can’t seem to get all the pigs to hit the sweet spot on the slaughter plant’s grid. As a result, Johnny gets less money, and I get a smaller bonus.” I asked, “Why is it so difficult to hit that sweet spot with all the pigs?” Gustavo explained, “We don’t have an easy way to weigh them quickly and accurately.”
It’s clear that both Bill Maclean’s and Gustavo’s frustrations stem from the same issue: many pigs consistently fall outside the “sweet spot” on the slaughter plant’s grid—a problem that has persisted for some time. Typically, 10 percent of pigs shipped are too heavy and 20 percent of pigs shipped are too light, with both the heavies and lights facing stern discounts of up to 30 dollars per pig when it comes to settlements.
The grid and its impactThe “grid” system, introduced in the early 1990s, was designed to bring more consistency to the weights and characteristics of pig carcasses. It aimed to send a financial signal to producers, rewarding those whose pigs met the ideal weight and characteristics and penalizing those who didn’t. At the time, pigs sold through Manitoba Pork, the single desk seller, had their market prices adjusted by a factor ranging from 0.87 to 1.14, depending on the carcass weight and characteristics.
In February 1997, the Chicago Mercantile Exchange transitioned from live hog futures to lean hog futures to better align with the changing needs of the swine industry.
While these changes improved the consistency of carcasses arriving at slaughter plants, the percentage of pigs hitting the “sweet spot” has remained stubbornly low—typically below 80% and often below 70%. Despite efforts to shift toward flat-rate premiums and discounts per hog, the issue persists. Slower line speeds at slaughter plants, caused by carcasses falling outside the desired weight range, result in costly inefficiencies. Pork retailers are also frustrated by inconsistent pork cut sizes, which complicate the supply chain and make it difficult to develop branded products. This carcass inconsistency remains a significant, ongoing challenge and is holding back the industry.
A third storyIn the early 2020s, Manitoba faced areas of lower-than-normal protein levels in its soybean harvest. Despite a surge in soybean acreage in the province, surpassing 1 million acres, buyers were frustrated by the low protein content of some of the beans. These soybeans were being transported to Quebec for processing into meal and oil. The processors, who were committed to meeting the required protein specifications in their soybean meal products, faced a significant challenge. The low-protein soybeans made it difficult to meet those specifications, risking costly claims and reputational damage.
In response, soybean buyers introduced discounts for low-protein soybeans, which some producers felt were excessive. To address these concerns, the Manitoba Pulse and Soybean Producers commissioned me to investigate the issue. I traveled to Quebec to visit the Vitara plant, which was receiving the low-protein soybeans. There, I learned about the complexities of blending low-protein soybeans with higher-protein soybeans from Quebec and Ontario to create a soybean meal that met the required specifications. Could they source enough high-protein soybeans to make the blend work? After blending, would the finished meal meet or exceed the protein specifications? The blending process was costly, uncertain, and slowed line speeds.
My investigation concluded that the discounts for low-protein soybeans were a fair sharing of the burden between processors and producers. Similarly, in the swine industry, some pig producers may feel that grid discounts are excessive. However, an investigation would likely reveal that the costs associated with oversized and undersized carcasses are fairly distributed between producers and processors, especially with the competition among slaughter plants to procure the pigs. In some cases, processors may even displace their own deliveries to allow independent producers to haul more pigs than initially booked, because their pigs are getting too heavy and any delay in shipping will cause discounts. What this says is that the 30-dollar discount per pig to the producer probably represents roughly half the total of 60 dollars in inefficiencies if the processor’s burden is included in the calculation.
This highlights the broader issue of consistency and how pricing mechanisms are designed to fairly share the costs of that inconsistency across the supply chain. While the challenge is not new, the solutions to address it have remained elusive until now in the swine industry.
SolutionA solution to the challenge of pigs missing the sweet spot on the grid lies in a new, low-labor, high-speed method for weighing pigs that is coming to market under the product line called Porcus Optimus™ from a start-up called Moccus Maximus Inc. The patented livestock assessment system (WIPO: WO 2024/174045 A1) involves walking pigs in groups of 20 to 30 through an assessment system where they are weighed using computer vision technology. Pigs ready for market are automatically marked with livestock spray, so they can be loaded onto trucks the next day. The percentage of carcasses hitting the sweet spot of the grids rises. Also, slaughter plants can be notified of the pigs’ weights in advance, enabling them to organize arrivals more efficiently and maintain top line speeds.
The dynamics of the movement through the livestock assessment system of groups of pigs aligns with their herding instinct and desire for mental stimulation. They quickly learn to look forward to the walk through the system and embrace the activity as part of their day. We have found that with a three-wide system that we can weigh 30 pigs in about 20 seconds. The practice of walking together as a group through the system also makes the pigs easy to guide onto a truck for transport when they are ready to go to market.
This system also simplifies settlements by counting and weighing pigs as they move into and out of feeder barns, providing clearer inventory tracking and enabling more frequent weighing to monitor growth rates. This innovation offers a promising solution to a long-standing issue in the industry and can help bring the grid closer to its ideal, improving efficiency and profitability for both producers and processors.
Change in thinkingIn the pig industry, new methods sometimes emerge that require a shift in perspective. This is one such case. The previous method of weighing pigs on a load scale involved confining them individually, allowing time for each pig to settle into a fixed stance to obtain an accurate weight. This process was challenging for both the pigs and the herdsmen, which is why it was only used when necessary.
In contrast, the assessment system operates differently. Pigs, in groups of 20 to 30, are walked briskly through the system, ensuring both accurate weight and count. The pigs benefit from the exercise and mental stimulation of the free walk. This new approach makes weighing pigs easier for both the animals and the herdsmen, enabling regular use of the method for more precise herd management with data. Once the industry adopts this new mindset and embraces the technology, it will lead to greater consistency in carcass quality and a significant boost in efficiency across the swine supply chain.
Grant is a senior instructor and chair of the agribusiness program in the Department of Agribusiness and Agricultural Economic at the University of Manitoba.
