What the Plant Floor Taught Me That No Model Ever Could

There is a version of my career that never included Tulare, California.

In that version, I went straight from a supply chain internship into a corporate analytics role, built models from a desk in Minneapolis, and became very good at telling stories about milk I had never actually seen move. I would have been fine. Probably pretty good, even.

But I would have been missing something I didn't know I needed.

When I joined Land O'Lakes through their Supply Chain Talent Acceleration Program, my first rotation put me on the floor of a butter manufacturing plant in Tulare. I was twenty-two, I had a finance degree and a lot of Excel confidence, and I was about to learn that the dairy industry has a way of humbling both of those things very quickly.

The gap between the model and the milk

Here is something nobody tells you when you are studying agribusiness economics in a classroom: the numbers on a spreadsheet have a smell.

Milk comes in warm. Separation takes time. Equipment breaks down at 2am on a Tuesday for no reason the maintenance log can fully explain. A yield assumption that looks perfectly reasonable in a model can fall apart completely when the vat temperatures aren't cooperating, or when a supplier's component levels shift mid-month, or when three people call in sick on the same day and the plant has to make decisions in real time that no optimization algorithm anticipated.

I spent a year learning that. Not reading about it — learning it, on the floor, in the middle of it.

The project I worked on during that rotation — improving material yield in UF milk processing — required me to understand not just the math of what should happen, but the operational reality of what actually happens and why. The gap between those two things is where the real work lives. And you cannot see that gap from a spreadsheet. You have to go stand next to it.

What I carried forward

I have been building models and analytics for a decade now. I have forecasted commodity markets, optimized mass balance across multi-plant networks, built risk exposure tools and supply planning frameworks. The technical work is what I am known for professionally.

But the thing that makes that technical work actually useful — the thing I credit for whatever effectiveness I have as an analyst and as a leader — is the year I spent learning that there are people behind every number I put on a page.

When I build a production plan today, I know what it feels like to be the plant receiving it. When I model a balancing decision, I know what it means operationally for the team that has to execute it. When I recommend a risk management strategy, I can hold the market signal and the supply reality and the human constraint all in the same frame — because I have lived in each of those frames at different points in my career.

That is not something a model can teach you. It is not something a certification can give you. It is something you have to earn by showing up somewhere uncomfortable and paying attention.

The thing I tell every analyst on my team

I lead a planning and analytics organization now. When I bring new analysts onto the team, one of the first things I encourage them to do — as early as possible in their career — is get out from behind the screen.

Visit a plant. Ride along with a field rep to visit a member farm. Sit in on a scheduling call with the production team. Talk to the milk planner who is routing trucks at 5am when the numbers don't add up. Ask questions and then stop talking and actually listen to the answers.

Not because the technical skills don't matter — they do, enormously. But because analytics without operational empathy is just math. And math, on its own, does not move milk.

The best analysts I have ever worked with are the ones who understand that the data is a representation of something real. Real farms. Real plants. Real people making real decisions under real pressure. When you keep that truth inside the work — not as a footnote, but as the actual point — the work gets better. The models get more honest. The recommendations get more useful.

And the stories you tell with the numbers start to actually land.

A note on humility

I want to be careful not to romanticize the plant floor experience into something it isn't. It was hard. It was sometimes frustrating. There were days I felt very far from the career I thought I wanted.

But looking back, that discomfort was the education. The moments where the model didn't work and I had to figure out why. The conversations where an operator with thirty years of experience taught me something in five minutes that no textbook had managed in four years. The slow realization that expertise is not something you download — it is something you accumulate, layer by layer, by being willing to go where the work actually happens.

I got into dairy analytics because I love the complexity of it. The markets, the pricing mechanics, the optimization problems, the strategy.

But I stayed because of what I learned in Tulare: that behind every billion pounds of milk is a web of people doing hard, important work — and that the best thing an analyst can do is understand that work well enough to make it easier.