In the pork industry, no one wants to leave money on the table, especially as tight margins and rising input costs squeeze every head. For Keystone Cooperative, the search for unrealized opportunity led to a high-tech solution for a familiar challenge: marketing timing.
By leveraging artificial intelligence to create a custom top-weight prediction model, Keystone discovered that while weight prediction isn’t new, the way you analyze it can change the bottom line.
“Our operations were already performing well, so this wasn’t about fixing a problem,” says Nathan Hedden, vice president of swine and animal nutrition at Keystone Cooperative. “It was about asking, ‘What are we leaving on the table?’ When you market thousands of head each week, even small improvements in timing can add up quickly.”
Optimization Over Repair
Keystone’s move to better harness the power of AI was driven by a desire to discover incremental gains in an already high-performing system, rather than a need to fix existing failures. Tapping into the “cooperative advantage” allows internal efficiencies to directly benefit member-owners.
The project aims to use “top weight” prediction to capture unrealized revenue through better marketing timing.
“While it doesn’t change daily work inside the barn, better marketing decisions improve Keystone’s financial performance,” Hedden says. “That value flows back to member owners through patronage. It’s strengthening the entire system so growers share in the success.”
Data Integration and Collaboration
The project didn’t require a total overhaul of their records. Instead, it required connecting the dots between silos of information that had been growing since 2017.
“We didn’t start from scratch,” Hedden adds. “Keystone already had years of trusted production data going back to 2017, we just hadn’t fully connected it. This included weights, feed consumption, group lifecycle records, and more.”
The project was built through close collaboration between Keystone and a third-party vendor.
“Keystone operations brought the production knowledge, IT brought the AI expertise, and together we focused on one clear use case: predicting the optimal marketing window for top weight,” he says.
The non-negotiable? Clean data. Hedden says there was no need for guessing when they had historical records at their fingertips.
“The model was also built to support, not replace, human judgment,” he says. “It provides earlier, better signals so our team can make confident marketing decisions.”
Uncovering Invisible Patterns
Early results indicate that while the necessary data was always present, AI provides the “connective tissue” required to see patterns that escape manual analysis.
“We’re still early and haven’t fully implemented the model yet, but one thing is clear: the opportunity is real,” Hedden says. “The study showed that while the data existed, it hadn’t been connected in a way that allowed us to pinpoint ideal market timing.”
By combining feed consumption, environmental data and performance flow, Lindsay Sankey, director of public relations at Keystone Cooperative, says the model provides a more granular look at the ideal marketing window. It reveals patterns people can’t see manually. It’s also designed to learn over time, becoming more tailored to Keystone’s system with each new data set.
“The value is in better decisions made earlier,” she says. “Our team still makes the call, but with greater confidence and clarity.”
Tracy Soper, senior director of data excellence at Keystone Cooperative, says that being laser focused on one thing helped.
“We had that one problem we wanted to solve – help me get the right pig on the truck. I want them at this weight. Help me make that better,” Soper says. “Because every percent that we reduce where we’re outside of that boundary is basically margin for us as an organization.”
Soper believes anyone who thinks AI will solve all their problems will likely end up frustrated. But focusing on small projects along the way will add up in the end.
Enhancing Grower Profitability
The ultimate goal of the technology is to drive financial performance that flows back to the growers through patronage, reinforcing the strength of the cooperative, Hedden points out.
“This tool helps our production management team better optimize when pigs are marketed,” he explains. “More precise timing improves the profitability of Keystone’s pig division, and that value flows directly back to growers.”
This custom approach ensures the results are relevant to its specific member base.
“While others may use similar technology, this is built the Keystone way: using our data, our operational knowledge, and strong collaboration between operations and IT,” Sankey says. “Even small improvements create meaningful revenue at scale.”
When Keystone performs better financially, member owners benefit through stronger patronage. Sankey says that’s the cooperative advantage: using technology to strengthen the whole system, not just one part.
