Join the Newsletter

Stay up-to date with food+ag+climate tech and investment trends, and industry-leading news and analysis, globally.

Subscribe to receive the AFN & AgFunder
newsletter each week.

Image credit: iStock

What agtech can learn from the Indy 500: a conversation about data and uncertainty with Purdue’s Chad Fiechter

June 24, 2024

What’s easier: landing on the moon or raising crops?

So asked Chad Fiechter, assistant professor of agricultural economics at Purdue University, last week during Bushel‘s Buddy Seat conference in Fargo, North Dakota.

Chad

The answer to the question, as he noted, is pretty darn obvious. But the point of the comparison was more about getting the audience to consider “the similarities between the two and how both of them are dealing with really complex uncertainty.” And that complexity is an opportunity to improve the data going to farmers to provide actual value for their most pressing problems.

“The farmer is interacting with this uncertain system,” said Fiechter, who grew up in agriculture and spent a decade farming before moving to university.

By way of example of this uncertain system, he explained that uncertain weather can impact the kinds of pests present on a farm, which in turn impacts decisions growers make. Those decisions impact yield. There is also distribution of price throughout the growing season, and growers must make decisions about that.

“The whole point of this is to show you this is a messy, messy process,” he said.

Data is the backbone of many agtech startups trying to tidy up some of this mess. But right now, it’s siloed amongst various apps and platforms (e.g., Bushel, John Deere, etc.), and companies have yet to provide growers with a truly holistic view of the farm, says Fiechter.

“Lots and lots of people have talked about data and collected data and tried to bring value to producers,” he said onstage. “To me, it seems like we need to be working together to get all of the data in one spot so that we can actually understand what is the uncertainty and how farmers [can] manage the uncertainty.”

Fiechter currently teaches an advanced farm management course to undergrads at Purdue as well as a data analytics course for the MBA program. In a follow-up conversation with AgFunderNews, he discusses how that might become a reality, the difference between risk and uncertainty, and how to improve the way the agriculture industry deals with data.

AgFunderNews (AFN): Where did you get the idea to put crop production and the moon landing in the same conversation?

Chad Fiechter (CF): I was in an academic seminar once and a very humble ag economist was making a joke that his work was not rocket science. In the middle of it, someone leaned over to me and said it actually was: “It’s a dynamic, optimal control problem.” And I laughed out loud because it’s a joke that could only be made in a very weird setting.

But in that moment, it landed with me and I thought, “It’s really interesting to think about the complexity of both. It’s about the stakes. If you mess up landing on the moon, everybody dies. If you mess up food production at scale . . .

AFN: You discussed risk vs. uncertainty in your keynote at the Bushel event. What’s the difference?

CF: I would answer it from experience. It was frustrating me as a producer to understand that I couldn’t have a good grasp of the actual risk. I had to create some sort of an approximation for the risk that I was facing.

With risk, the probabilities are known. You have uncertain outcomes, but you know all the probabilities. In uncertainty, you have uncertain outcomes and you have to approximate what those will be.

AFN: Is the agriculture industry effectively using data to tackle this?

CF: In some ways, we’re using data inefficiently. From my perspective, there’s a need for deeper engagement with the data and, simultaneously, a backing away from real-time insights on every single aspect of farm management or risk management.

We just need general guidelines and parameters and rules of thumb. Of course, those rules of thumb should be informed by data.

Fiechter giving a keynote at the recent Bushel Buddy Seat conference in Fargo, North Dakota

Historically, ag economists have had access to unparalleled farm management datasets that the government put together. Now, because of the explosion in consumer data, we lag behind consumer economics on access to data to inform modern farm management in the way that I traditionally think about it. That’s because a lot of this data is held by apps like Bushel (I’m not faulting them) or John Deere or another service provider.

I think there’s still a place to do some public good through research, because I think that those rules of thumb are probably more impactful than the real-time insights that sometimes come out of these technology-driven companies.

As a producer, you don’t really have time to get an individual daily insight on where you’re at. You just need a metric to show you what you should do next.

I think the government still has great datasets. However, as a researcher, what I’m interested in is the digital data from representative growers.

Now we also have production data that’s literally collected at a minute scale. That’s all held by the data service providers, and these companies seem to perceive that the long-term value of their company is in that data.

But we still have to unlock some of the insights in those systems, because farms are managed holistically. You’re managing your finances while you’re planting in the field, and making decisions about your chemicals. If [a company] is interested in input decisions, but your input decisions have lasting effects on your marketing decisions and the way you finance, you need the whole package.

I still think land grant institutions have a role to play in aggregating that data.

AFN: Is there too much data in agriculture?

CF: I think farmers are begging for something that helps them quantify costs and benefits of adopting a technology. It seems to me that everybody tries to address that, but it’s pretty obvious when someone’s trying to make a profit, whenever they’re presenting you with the cost or the benefit that their service provides.

In some sense, I feel like we do have so much data — probably too much given the level of research that’s happened to try to distill out how we can actually build some insights.

AFN: Many folks at the show suggested agtech (including data) can’t replace human intuition on the farm. How do you feel?

CF: Data can be supportive of a producer.

As a researcher, it’s valuable, but it’s really only valuable if I have 100,000 other farmers that I can also observe what they’re doing, and we can say, “Hey, farmer x, if you do this, this is what we expect to happen on average.”

Right now, we’ve told everybody their data is valuable, there are a lot of tech startups built on the data, but we haven’t really used that data to provide the value. The value that’s created by aggregating the data has not been fully fleshed out.

Plus, everybody has put up walls in between all of these data slices. So now we just have an entire culture of producers who have been told that their data is really valuable, but no one’s ever figured out how to like functionally put it together to build insights.

AFN: What might that look like if they did?

CF: I recently went to the Indy 500 practice day. It was early enough that they let people down into the pits, so I just went up to one of the crew and asked, “so how many of the adjustments that you’re making on the car when they come back in are related to sensors and data that’s flowing back? How much of it is from the driver saying something like ‘I don’t like the way the back end of the car feels?'”

The crew person said it’s 100% what the driver says.

That blew my mind that you could be going 240 miles an hour, and control is still literally just the person sitting in the seat saying [something like] we need to change something in the rear-end because they don’t like the way it feels.

The race to digitize agriculture has forgotten that: the producer is still the person sitting in the seat, maneuvering it. There’s still a lot of intuition. My hope is to inform the intuition with data. My dream is that there’s more collaboration and scientific insights that are coming out and informing decisions.

Most people look at me and say, “You’re nuts.” They don’t trust that we can potentially maintain those boundaries around the value that they create because they own all this data. Everybody’s fearful that if we pull down the wall, their company is going to lose 30% – 40% of the value.

AFN: Thoughts on the future of agtech?

CF: Agtech feels a little like the Wild West, and it will settle out. My personal hope is that there’s still a place for academic research within the space, and that we could work with the agtech startups.

Join the Newsletter

Get the latest news & research from AFN and AgFunder in your inbox.

Join the Newsletter
Get the latest news and research from AFN & AgFunder in your inbox.

Follow us:

AgFunder Research
Advertisement
Advertisement
Advertisement
Join Newsletter