There's no average cow, says Jay Johnston of Fermentrics Technologies, so it's time for us to marry today's science to old-fashioned cow sense and observation.
Luther: Jay Johnston is the CEO of Fermentrics Technologies and the chairman of the board of Ritchie Feed and Seed, a regional feed manufacturer located in Eastern Ontario, Canada. The company has developed a unique analytical system as a means of better defining the characteristics of forages and feed ingredients. Its innovative use of gas production technology allows for the design of more cost-effective diets and is presently used in 25 countries worldwide. Johnston operates a cash crop farm with his family. Thank you for joining us.
Jay: Nice to be here.
Luther: What is the emerging disruptor in the dairy industry?
Jay: Wow. That’s a large question. Well, from our point of view, it’s one of being able to figure out why cows actually do what they do and to marry that to the technology that we’ve developed for fermentation to measure ingredients. What typically happens is, people make the assumption that cows are going to do what they think they should do and nobody has bothered to go ask the cows. So, we went and asked the cows and married it to biology.
Luther: Can you tell us a little bit more about the gas production technology?
Jay: We certainly didn’t invent it, but we’ve refined it.
In a typical analysis, you measure ingredients at a set point of, say, 30 minutes, or 30 hours, or 48 hours, or something like that, and assume from that you can perfectly describe how diets can work.
The reality is, all ingredients interact and interact at different time points, and you’re just guessing if you pick a set time point. What we did is we took that and we built in a method of measuring CO2 with methane so you can actually come up with a very, very accurate prediction of how it’s going to ferment and do it very quickly. So, we’ve got it down to a matter of minutes in many cases.
It’s unashamedly a diagnostic tool, and it’s used by — oh, golly — we started out for our own self-serving purposes and now the largest herd is milking 120,000 in China. The smallest herd is 13 cows on an Amish farm in upstate New York. In 26 countries and, I don’t know, there’s three-quarters of a million cows that are using it.
Luther: Wow. So, I take it cattle are kind of like humans. Some of them are picky and some of them don’t care for the feeding program. How does this work for them?
Jay: Cows are going to do what they want. And, you know, there was a wonderful scientific paper put out by Mike Allen, who teaches at Michigan State, and (it) basically said, shut off your computers and go look at the cows, because there’s no such thing. All these computer models are average cows, average this, average that. Well, there is no such thing. It’s just like there’s no average humans. So, the better thing to do is to go and ask the cows.
With this facial recognition system that my son and his company developed, you’re taking 28 frames per second and you actually — you’re doing it live. So, you can actually see what they’re doing, when they’re doing it, how much they’re eating. And then if you marry that to how it’s fermenting, it’s really fascinating. And you can go and, you know, measure how things are going to actually be — how productive they’re going to be.
Quite by accident, we changed how feed was distributed in one of the research herds we work with. And it was pure accident. The guy that was feeding that Sunday hated backing up a feed wagon. So, instead of backing up, he threw a whole barley just back to the pathway. So, he changed the distribution pattern and he totally changed how the cows ate. And it was pretty cool.
From there, it’s like, “Oh, golly, you know, all the stuff we thought we knew we don’t know very much.” It’s time to go rethink everything. If that’s not disruptive, I don’t know what is.
Luther: So, what were the effects of that change?
Jay: The cows changed the distribution of where they ate in the barn, and it’s pretty neat because, you know, barns are not cheap at home. They’re like $7,000 per stall.
A lot of mixers can mix, but they can’t distribute. So you end up having a good portion of your feed bunk that’s not actually being used, which is a total waste of resources.
We actually managed to change who went where, and we’re using more bunk space. So, in theory, you’ve got to change your stocking density, and the fascinating thing is, we’re trying to figure out why they moved. And we measured at Karl Dawson’s lab. In Alltech, we tried looking at organic compounds, like volatile compounds. It’s not that, but they definitely move. And it’s the heifers that move. Some of it is competition. They don’t like getting beat up by the big girls, but they move down. They’re smart enough to know that’s where the good groceries are. Whether it’s a tactile thing, I don’t know. We’re trying to figure that out.
Luther: So, this technology is even able to make judgments or to observe and see a change in pattern even to the point where you’re not even sure exactly why it’s working, but you know it is.
Jay: Oh, yeah. The fun part in statistics is you run an experiment. You say, “Okay. Now, fine with this, this, and this is my control… You know, 5% or whatever it is. 95% assurance this is what’s going to happen.”
Well, with facial recognition and AI, it’s live. It is what it is.
There’s a big barn that’s using (it) in California. We changed how the cows were eating just by putting citrus pulp out. We changed the pattern. There was one odd pattern that was coming up every time a blue truck went through the barn. We thought, “That’s a bit odd. What’s up with blue trucks?” Suddenly, (it) dawned on us that the blue truck was driven by the guy doing the artificial insemination, and they’re not stupid. They headed for the hills. We went and got another blue truck just to see, and they were smart enough to know it wasn’t the offending blue truck.
There’s a whole world of ethology, of how cattle do things and why they do it.
If you take it to a feedlot, the biggest problem you’ve got is acidosis. The problem with acidosis is, you’ve got 30,000 steers. Which ones are acidotic? So, you get a bunk rider that has to pick it out. Well, every time you get it wrong, you take these animals out. You’ve got to reintroduce them and that never goes that well. If you can come up with an idea of how to measure the diet and make it most productive and then to measure which animals are having the symptoms of acidosis, you could save yourself a lot of time and a lot of grief. That’s why there’s a lot of interest in it.
Luther: Does this change as time goes along, too: as the herd changes, what used to work or used to be the best option may not be the best option in the next six months, year, 18 months?
Jay: Well, you know, it’s funny. If you go and talk to the old timers, they’ll tell you, “Well, you know, I didn’t have all the university degrees, but I’ve been watching cows all my life, and this is how they work.” It’s turned out they’re pretty darn accurate. We should probably shut up and listen to them rather than, you know, look at our university degrees on the wall. Some of the old timers are very intuitive and they may not know why, but cows will do certain things.
It’s a perfect example in one of the farms that we worked with. They thought a cow got banged up because she was in heat and somebody had mounted her. It wasn’t that. She was just eating quietly and some other (cow) came along and knocked her down. Well, the floor was the problem because it wasn’t grooved properly. So, here, they would have made the decision that she got hurt because she was in heat when in fact she wasn’t and it was like, “Oh, golly, that’s a really good cow. We’ve got to fix this problem.”
The neat part is, you get to see what’s going on when you’re not there because it’s taking so many pictures so accurately. We had one example of two heifers quite happily eating away at the TMR and two mature cows — I don’t know, maybe 15 feet away — got in a real fistfight, and the heifers just said, “I’m out of here,” and they went and laid down. They didn’t come back and eat. Or, if you don’t have enough feed in the bunk. The timid cows come up at midnight; there’s no feed there. They’re just going to go and lay down. Well, there goes your dry matter intake. And you know, you make an assumption that, okay, this is a brilliantly designed ration with a certain dry matter intake. Well, guess what? It’s not equal, and therein lies the problem.
Luther: Given the variability that you’re discussing, how can diet formulation and distribution be tailored to situations like that?
Jay: Well, there’s some really cool work that just came out of Penn State and it’s called “temporal diets.” It’s a fascinating idea because, obviously, there’s a diurnal pattern in how hormones work, and how cows eat, and so forth, and so forth. So they’re trying to match up having periods of high-starch/low-fiber diets for one part of the day and then low-starch/high-fiber diets for another part of the day. And the basis for it was not just physiology, but actual intake data, but the intake data was garnered from some of this research equipment where the cow has a collar on. She sticks her head in the feed bunk and it gets measured. Well, the problem with that is, if she doesn’t like the cow that’s next to her, she’s not going to show her true side.
With this system that we’ve got — it’s whatever the cows do, they do, and you design the ration accordingly. So, you should be able to cut a fair bit of money out of how it’s done. I mean, the best we’ve done in Dubai, we had a herd that was down to just 14% protein and still banging along at 38.5 liters. It just takes a little bit of thought, but you can actually do it and save a lot of money.
Luther: So, how do you merge nutrition, technological innovation, digital management all together in this new future?
Jay: We’re making it up as we go. Every day, you go, “Gee, I didn’t know that.”
I suspect if the question is what’s it going to look like, it’s going to be a live system where what the cows are actually doing — if we get it right — they’re going to be their own digestibility metric. In other words, this is what the cows are doing. If you see this action in the cows, this is the type of diet you should have. So, there won’t be any more highfalutin research on how individual ingredients actually ferment. It’s going to be the cow who is going to be the teller of the tale. And it’s going to be fun. It’s going to annoy a lot of people.
Luther: So essentially, what you’re saying is, we’ll be able to test and actually see from the cow itself exactly the result and then adjust accordingly.
Jay: Let’s say you got a huge pile of corn silage. I mean, it’s like being a drunk. One sample is too many in a million. It’s not enough. You could take samples all day long. And so, the best you’ve got to do is guess. So, why not go ask the person that’s actually eating this stuff? And they’ll tell you pretty quickly. And then you get to adjust how things are distributed. That’s turned out to be the real shock. You know, everyone designs these rations and then they go and look at how it actually gets distributed. That usually gets messed up pretty quickly.
Luther: We’ll talk about that just for a moment since it is so important, the distribution side. You know, when you say it gets messed up, how is a system like this able to improve that? You’ve touched upon it I know, but just from a—
Jay: Well, you get—
Luther: Concrete examples.
Jay: You get to calibrate how you distribute things. Mixers are designed to mix. You know, they’ve left out the bit about how they distribute. And it’s almost illogical to think — say you’ve got 2 to 3 tons in a mixer — that they’re going to be distributed equally all the way down a bunker. Probably isn’t. If you see how the cows are reacting to it, you can change just how you distribute, where you start and stop. And you can actually manipulate it and move the cows around. That’s the absolute fascinating part.
Luther: What does the future hold for diet formulation, distribution, for your system? Where do you see it going?
Jay: Well, starting with the diet formulation, I think what’s going to happen is, instead of living in a world where you are predicting what should or shouldn’t happen with the formulation, you’re actually going to be able to measure it. And why would you want to predict something you can actually measure? Instead of measuring things at 48 hours, or 30 hours, or something like that, we’re going to be able to measure it literally live from a digestibility point of view.
If we get this all right, let’s say, hypothetically, you’ve got a dairy herd that’s having problems, you could start the assay at 8 o’clock in the morning. And by 9 o’clock, you pretty well know what the problem is and how to fix it. You could fix it by the next feeding.
That, in the perfect world, is where we’re going, and it will be a function of not just Fermentrics stuff and the gas fermentation, but the cattle will actually be telling you. “Okay, they’ve changed their feeding pattern. There’s something screwy. Can you see if it’s this or this?” It will be basically a live diagnostic system.
Luther: Are there other applications for this technology beyond maybe determining a behavior pattern that a cow maybe needs to be looked at to see if there’s something wrong? As you said, there’s bullying going on that’s causing disruption in the herd.
Jay: Oh, I think there’s a myriad of things. I mean, it’s going to be things like barn design, ventilation, where the waterers are. It’s a multiplicity of things that interact and how cows work and don’t work.
There’s some really cool work out of University of British Columbia where you can predict subclinical metritis. We’ve already done it — predict subclinical lameness. There’s a bunch of preventive measures that have nothing to do with nutrition, but have everything to do with ethology and cow management. And that’s the really cool part. They are now working on estrus prediction, and there are some markers that happen well in advance of normal estrus prediction, either by a human or by a pedometer, that they change their patterns of action and eating and so forth.
There will be a lot of management things that are affected. A lot of humps and hollows are going to be taken out of the system. And the really funny part is, we’re going back to (what) the really good all-time managers say: “Well, I wouldn’t have done that anyway, because cows don’t like that.” We’ve just spent a myriad of time and money to come back to the beginning.
Luther: Do you think there’s application outside of just cows, or maybe obviously pigs and poultry?
Jay: Oh yeah. Anything that moves that you can measure. It doesn’t have anything to do with agriculture, but the real big interest is in athletics. You’ve got a million-dollar basketball player pounding up and down the floor. And if you can tell that he’s about to blow a hamstring or something just by the way he’s moving, there’s an awful lot of interest. Anything that could be measured and have a metric put against it will work.
Luther: Well, let’s bring this to a little bit higher level and literally bring it home. How does this technology affect the average consumer’s table at the end of the day?
Jay: Oh, golly. Well, starting with a friend who wants to do it in Europe, with the way food is distributed in grocery stores there, if you can ever come up with a system, which has identifiable metrics, which measures animal welfare, the whole 9 yards, they’re all over it.
If you’re a consumer and you’d say, “Well, golly, if I had two products, one I know the animal was absolutely treated humanely and the very best possible with the very best nutrients versus, well, I don’t know, it’s just somebody else’s,” guess which one you’re going to pick.
This is a way of accommodating the need for huge amounts of data unobtrusively — like, it’s not invasive in any way, shape or form. You’re not sticking something in the cow or tying something on the cow. You’re just sitting back, digitally watching the cows. You’re just letting them be cows, or pigs, or chickens, or whatever. So, if that doesn’t elicit a degree of enthusiasm from consumers, I’m not sure what will.
Luther: Jay Johnston is the CEO of Fermentrics Technologies and the chairman of the board of Ritchie Feed and Seed, a regional feed manufacturer located in Eastern Ontario, Canada. Thank you very much for joining me.
Jay Johnston spoke at ONE: The Alltech Ideas Conference (ONE17). To hear more talks from the conference, sign up for the Alltech Idea Lab. For access, click on the button below.