Why does a giant agricultural co-op need a chief technology officer? Farming has always relied on data (just think of the Farmer’s Almanac), but since assuming the CTO role at Land O’Lakes in 2018, Teddy Bekele has been busy modernizing the co-op’s data practice.
Bekele has overseen reorganizations of Land O’ Lakes’s IT departments, helped spearhead new ways of using the co-op’s treasure troves of agronomic data, and is working on an AI platform to help farmers plant and harvest more effectively. Big data is increasingly central to the crop inputs business, in which Land O’ Lakes buys farming supplies and sells them to retailers, he said—and it’s invaluable when providing insights to farmers.
At CES 2025, Bekele sat down with IT Brew to chat about what digital transformation at Land O’ Lakes looks like.
This interview has been lightly edited for content and clarity.
Can you walk me through how you’ve developed the data and analytics since you got there?
Obviously, insights were always a big thing for us. But first of all, in order to be able to generate the insight, you have to have a model—and then in order to create the model, you have to have the data in place to be able to put that in place. And so we’ve really put a lot of rigor over the last decade of really putting our data in a place where we can make it easily accessible and also do something with it.
So, whether it’s our operational, transactional data from our ERP [enterprise resource planning]—which is kind of traditional IT in general, and that’s where we do the normal reporting—[or] things like our agronomic data. What I mean is in our crop inputs business, we actually have about 150 research plots where they plant all the seed varieties. They plant in all the different conditions, different soil types, different practices. And now that data is captured. It’s always been available, but not been available from an analytical standpoint. So, we’ve cleansed that and made that available.
Was it a technical struggle modernizing all that old agronomic data? How was it stored before?
Let’s call it semi-spreadsheets, because it is captured by a lot of it. Now, we’re capturing things for drones and weather sensors, weather stations that are out in the field, etc. But in the past, people actually walked the field and captured the data on a piece of paper, and they’ll now digitize it and put it in an Excel spreadsheet so there’s something to work on. Now, we have it in more of a traditional data warehouse. There’s some kind of cleansing tools up front, and then that can then export it again to our data platform, where we then run our analytics and everything else on top of that.
Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
Was it at least relatively structured? I talk sometimes with people at enterprises where they basically inherited the equivalent of the Indiana Jones warehouse.
It wasn’t as bad as that, right? As you create the model, you start to figure out what else you’re missing in the data set itself. So, we’re able to kind of go back and say, “Hey, you weren’t capturing this. Can you capture this along the way as well?” And so we’re able to do that now. Where it was messy for us was not our data within the four walls of Land O’ Lakes. It’s the data coming from the outside. Think about retailers that we work with, and some of them are very large—but there was no master data management practice put in place. So, it would be the Wild Wild West. The same customer would exist six different times in their one system, and then they had multiple systems.
Have you had any challenges getting buy-in for any of these initiatives?
There’s always challenges, because farmers are, in general, rightly so, are very skeptical about any new thing that comes their way. And the reason being, it’s not because they’re behind the times or obtuse, it’s every year [they] are literally betting the farm. The amount of inputs you buy and the amount of money you spend on new equipment and new capabilities is so vast, that it could be one bad year or two bad years—and then you completely wipe generational farming altogether.
So, in general, they sort of have this attitude of like, okay, what snake oil are you bringing? Trust is important at that point in time. What you bring is new. You also have to let them know the level of risk they’re signing up for and let them sort of make the decision about how much they want to put their foot on the pedal versus not, and then start small. A lot of times we say, “Let’s try it on one or two of your 30 fields, see how that goes.” And then if you want to see if you want to expand from there.