Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How a Vineyard Cut Water Use by 60 Percent with Sensor Operations
Transcript
- Lucas: Luna, I spent last week down in the Central Valley looking at a vineyard that cut its water use by sixty percent over two years — not by ripping out vines, but by changing how and when they irrigate. Luna: Sixty percent with no yield hit? That's a big number. Most growers I've talked to are happy with ten, maybe fifteen percent. Lucas: Right, and that's exactly what makes this case interesting. This isn't a tech startup with a pilot project. It's a family-run operation called Trebbiano Ridge — about two hundred acres of Chardonnay and Pinot Noir, been around since the seventies. The owner, a guy named Frank DiMarco, told me he used to irrigate on a timer. Every Monday, Wednesday, Friday, same amount, year after year. Luna: So the classic 'set it and forget it' approach. I bet that led to a lot of overwatering on cooler weeks and underwatering during heat spikes. Lucas: Exactly. And the thing is, Frank knew his soil held water differently in different blocks — sandy loam on the east side, heavier clay on the west — but he had no way to track it block by block. So he watered everything the same, which meant the clay blocks were getting drenched while the sandy blocks were still dry. That alone was wasting millions of gallons a season. Luna: So what was the operational shift? They put in sensors? Lucas: They did. But it wasn't just buying a box of sensors and sticking them in the ground. Frank worked with an ag-tech consultant named Maria Santos, and they designed a three-layer sensor network. At each monitoring station, they placed a sensor at six inches, eighteen inches, and thirty-six inches deep — that's the root zone for mature grapevines. They installed one station per five-acre block — forty stations total across the vineyard. Luna: And that gave them a real-time picture of moisture at different depths, not just surface wetness. So they could see when the deep roots were still getting water and hold off on irrigating. Lucas: Exactly. But the real operational innovation was how they turned that data into a decision rule. Maria built a simple algorithm that takes three inputs: the soil moisture readings, the local evapotranspiration rate from a nearby CIMIS weather station, and the forecast for the next three days. The algorithm then calculates exactly how much water to apply — not in inches per hour, but in gallons per vine. Luna: So they moved from a time-based schedule to a demand-based schedule. That's a fundamental process change. How did the crew adapt to that? Lucas: That was the hardest part. Frank told me the first season, his irrigation foreman kept overriding the algorithm. He'd look at the numbers and think, 'No way the vines need that little water. It's August. I'm turning it on.' They had to run side-by-side comparisons — two rows, same varietal, same soil type, one watered by the algorithm, one by the foreman's gut. After three months, the algorithm block had the same yield, better sugar content, and used forty percent less water. Luna: That visual proof must have been powerful. Nothing convinces like a side-by-side test. Lucas: It did. But Frank also changed the incentive structure. He started rewarding the foreman for water savings, not just for keeping vines green. He said, 'If you hit the seasonal water budget, there's a bonus.' That flipped the mindset from 'more water is safer' to 'less water is smarter.' Luna: That's a classic operations move — align incentives with the outcome you actually want. So after two years, they went from using about two million gallons per season down to eight hundred thousand. That's the sixty percent figure. Lucas: Right. And Frank calculated the payback period on the sensor network and software — the whole thing cost about four hundred thousand dollars to install, including the consultant. With water prices in California, he's saving roughly a hundred and twenty thousand dollars a year in water costs alone. That's a three-and-a-half-year payback, not counting the energy savings from running pumps less. Luna: So the economics work even before you factor in any yield quality improvements. I'm curious — did they see any other operational benefits beyond water? Lucas: Yeah, a couple. One is that the crew spends much less time moving irrigation lines. Because the system is automated drip — they have valves that open and close based on the algorithm — the foreman can manage the whole vineyard from a tablet. That freed up about three hundred hours of labor per season. The other benefit is that they now have three years of soil moisture data, so they can predict which blocks will need more or less water in a given year based on early-season readings. That's a planning tool they never had before. Luna: If these conversations have moved your work forward in some small way, consider supporting the show at buy me a coffee dot com slash fexingo. It's a simple way to keep us ad-free and focused on real operations cases. We're listener-supported, and every contribution helps. Lucas: Yeah, it makes a difference. And we don't run ads, so that's really the only way the show stays on the air. Thanks to everyone who's chipped in. Luna: Alright, back to the vineyard. I want to talk about the human side — the pickers. How did they get buy-in from the seasonal crew? Lucas: Frank and Maria trained a group of ten lead pickers to recognize visual stress signs — leaf curling, color changes — that indicate the vine is actually underwatered. The idea was that the algorithm might miss a localized issue, like a clogged dripper or a broken valve, but a trained eye could catch it. So the pickers became a human sensor network. If they saw a stressed vine, they'd flag it, and the foreman would check the data. Luna: So you get the best of both worlds — data-driven decisions with human override for edge cases. That's a good process design principle. Lucas: Totally. And it also gave the crew a sense of ownership. They weren't just following orders; they were part of the system. Frank said that after the first year, the lead pickers started competing to see who could spot stress patterns first. It changed the culture. Luna: Alright, let's zoom in on that specific heat spike day you mentioned — August 15th, 2025. What happened? Lucas: That day, the forecast called for a high of one hundred and eight degrees. The algorithm had been cutting back irrigation for the previous three days because the soil had good moisture. But on the morning of the 15th, Maria noticed that the six-inch sensors on the sandy loam blocks were dropping fast. The algorithm automatically triggered a mid-day drip cycle for those blocks only — a short pulse, about fifteen minutes. On the clay blocks, the deep sensors still showed adequate moisture, so the system held off. Luna: So instead of watering the entire vineyard — which they would have done on the old schedule — they gave a targeted pulse to the blocks that actually needed it. Lucas: Exactly. That single day saved an estimated one point two million gallons compared to what the old schedule would have applied. And the vines handled the heat fine. Frank checked them the next morning — no sunburn, no leaf drop. The old approach would have blasted water everywhere, much of it running off or evaporating. Luna: That's a powerful example of what precision operations can do. But I'm thinking about the broader adoption question. Most farms — especially small and medium ones — don't have four hundred thousand dollars to drop on a sensor network. Lucas: That's a fair point. Frank actually financed part of it through a USDA conservation grant that covered about forty percent of the cost. And Maria told me they're working on a lower-cost version using off-the-shelf IoT sensors that would bring the price down to around a hundred and fifty thousand for a similar-sized vineyard. Still not cheap, but the payback would be under two years at current water prices. Luna: So the economics are improving. And with water prices only going up in the West, this kind of operational shift is going to become more common. Lucas: I think so. But what struck me is how much of the success came down to process design, not just technology. The sensors are just data. The algorithm is just math. The real operational win was changing the decision-making rhythm — moving from a fixed calendar to a responsive system, and training the humans to trust the data while also using their eyes. Luna: It's a classic operations story: the technology enables the change, but the change itself is about workflow, incentives, and culture. Lucas: Exactly. And Frank put it well. He said, 'I spent forty years thinking I was a good farmer because I watered every week. Turns out I was just lucky the vines are tough.' Now he's got a system that lets him farm block by block, vine by vine. That's the operational edge. Luna: Great case. Next time, I want to look at how a different type of operation — maybe a warehouse or a hospital — could learn from this. The principle of moving from fixed to demand-based scheduling is pretty universal. Lucas: That's a good idea. Let's set that up for next week.