Sorry, you need to enable JavaScript to visit this website.

Data Dork: How Intelligence and Data Work On the Farm

08:23AM Dec 06, 2014

By Rich Kottmeyer

When I ask farmers to describe on-farm intelligence systems and how they work, there’s often more confusion than understanding. In reality, intelligence is a simple concept gone amok by people trying to sell shiny new things (in almost every industry). 

The easiest way to break through the clutter is to picture yourself as a party of one in a restaurant. There’s no way you can listen to every conversation going on around you. To do so, you need a machine to capture, store and analyze all of the conversations so you can focus on one conversation at a time. That’s big data. 

As a farmer, you can review the data generated from your operation on your own or with your advisers (which you have done for years). By definition, you don’t have big data nor need a big data solution. 

Back at the restaurant, you overhear the couple next to you fighting, but since you lack the ability to actively probe deeper for context, you use signals intelligence (also known as spying or insight). Signals work better when they are clear (e.g. girl throws water in face of guy and storms out) or when you can put lots of signals 
together. We like to call signals, which come in all sorts of forms—from weather, hydrology and nitrogen 
models to marketing, price, news and the science of agronomy—data. 

After dinner, you follow the couple to the movie theater. The next day, you check their receipt from the store for “makeup” food and plant a bug in their kitchen. That is layering. But the more data you layer, the more difficult it is to define the right picture, much like the difference in doing a 50-piece puzzle versus 500 pieces. You get a better picture with more pieces—if you piece them together correctly. 

At some point, data analytics is required, which is a machine that handles the volume of information you can’t. Remember, an accurate but somewhat blurry picture is better than a high definition picture of a potato when you are planting corn. 

Signals intelligence is very hard and expensive to put together correctly. The difference between a well-funded and thought-out source and everything else is enormous. Focus on credibility and the resources of the source. Likewise, it is far better to have a good consolidated source of signals than try to incorporate lots of little data models for specific parameters. 

Today that is what is being sold to farmers—signals intelligence and data analytics. But that is only one side of the equation. It isn’t something new (though the tools are), and it doesn’t replace what you already know. 

To get more context about the fight between the couple in the restaurant, you could go buy the poor sap a drink and become his friend. Now you have human intelligence, which often takes the form of personal relationships, interrogating and learning. Of course, there might not be time for friendship (a big problem in human intelligence is the time it takes to set it up). 

So, you follow him to the restroom, throw him up against a wall and interrogate him. At that point, he is likely to tell you whatever you want to hear versus the truth. Human intelligence is full of bias, exaggeration and “group think,” but it can be extremely insightful compared to signals intelligence. 

Finally, you decide to recruit help, so you use the wait staff and bartender who see the couple every week to 
determine if their fighting is common or a rare issue. Now, you are tapping into institutional knowledge.

The fact is, as a farmer, you have enormous human intelligence assets. You have worked the land for years. Your agronomist and cooperative relationships are years old. They’ve worked in the area and know the land and its performance. There’s also the Extension service and more collective knowledge on a localized basis. 

You need both sources of intelligence to get an accurate read. Countless stories of war and national security point to deep problems that occur when a nation or army overemphasizes one side of intelligence versus creating a mosaic. 

Now, as you listen to the couple argue, what good is the intelligence? Outside from keeping you from being bored, not a thing. Farmers need actionable intelligence. 

That means before you buy any new service, make sure your advisers and equipment can take action based on it. Remember when we all bought detailed maps of our farms only to wonder what we should do with them or what they told us that we didn’t already know? Today, your first stop should be to your equipment dealer for two reasons: Most of the signals coming off the farm are captured by the equipment, and second, most data services on the market emphasize agronomy support and “prescriptions.” You need to know you can execute the prescription, especially if it calls for variable-rate anything.

Finally, you want to predict, based on your new intelligence, the likelihood of your actions resulting in a positive outcome. After harvest, you want to use the same tool to determine if it was your new on-farm intelligence network that added value or, say, Mother Nature being kind. That is benchmarking. Without it, intelligence tools can be like black box answers. It is fundamental to every intelligence gathering exercise. 

The concepts of building your on-farm intelligence network are simple and full of common sense, which should be remembered when confronting shiny bells and whistles. Different sources balance out each other. 
Human intelligence has blinders and can be less innovative (fewer “aha” moments), but it gathers deep insight. Signals require sorting out garbage and challenging yourself. 

The worst intelligence networks are those with too much information and not enough insight. In short, each tool should add to what you already know (which is considerable). When in doubt, wait, or you are likely to over-rely on the newest tools at the expense of what you already know. 

However, don’t wait to evaluate your weaknesses and find the right tools for your farming operation. Just be measured. After all, can you even imagine Wal-Mart without its analytic abilities to know what you want or what to place in each store and where  to place it for each customer segment? Or the pharmaceutical industry’s ability to pick which molecules to spend hundreds of millions of dollars of research on to create one new Food and Drug Administration-approved drug? Or counter-terrorism without a network of devices to capture and sort through chatter?

The key is to have a trusted adviser that can help you create this new intelligence capability for your farm (at least initially until you become extremely familiar with how the pieces work together). 

Next up, we will evaluate a typical U.S. grain farmer and what intelligence they generally have and where they are often weak. Following that, we will: 1) look at the type of services being offered and what potential they have for you; 2) provide key insights of what to look for and avoid; and 3) create a general order of operations detailing what you need to focus on first, second, etc. to have an effective intelligence network versus a mess of random shiny objects.