This episode was recorded in Fort Wayne, Indiana, during the 2025 Tri-State Dairy Conference. Dr. Fessenden gives an overview of his presentation which covered both research and field information on automated milking systems. He recommends going back to basics and formulating a rumen-friendly PMR with a complementary palatable feed that encourages the cows into the robot system. (5:57)
This episode was recorded in Fort Wayne, Indiana, during the 2025 Tri-State Dairy Conference.
Dr. Fessenden gives an overview of his presentation which covered both research and field information on automated milking systems. He recommends going back to basics and formulating a rumen-friendly PMR with a complementary palatable feed that encourages the cows into the robot system. (5:57)
The panel discusses ideas for driving cows to the robot on different types of PMRs, management of transition and late lactation cows in automated milking systems, and the use of custom pellets versus other supplemental feeds in the robot. (9:04)
Dr. Fessenden talks about some of his experiences visiting automated milking systems in Europe and some of the differences between European and North American approaches to diet formulations in automated systems. The group goes on to talk about different options for supplemental feed formulations in the robot. (18:38)
Dr. Fessenden and Dr. Tylutki share ideas for how feeding technology could help both traditional and automated milking farms in the future. They discuss more precise grouping of cows and targeted feeding of those groups to better match requirements, as well as how movement to different pens and diet changes can impact milk production. (23:45)
Dr. Tylutki updates the group on advancements in the AMTS balancing tools for multiple robot feeds. He and Dr. Fessenden describe the challenges of developing new tools that are streamlined and user-friendly for nutritionists. They also delve into the role artificial intelligence and machine learning might play in ration balancing in the future. (29:32)
Dr. Fessenden encourages nutritionists to think outside the box when working with automated milking herds to make decisions that are right for that particular farm. Dr. Tylutki chimes in with advice to spend time actually watching the cows on the farm, and Dr. Van Soest echoes this sentiment when it comes to troubleshooting issues with the robots themselves. Don’t just assume it’s an issue with the ration - ask or observe for yourself what may have changed on the farm that could contribute to the issue the farm is facing. (40:02)
Panelists share their take-home thoughts. (45:20)
Please subscribe and share with your industry friends to invite more people to join us at the Real Science Exchange virtual pub table.
If you want one of our Real Science Exchange t-shirts, screenshot your rating, review, or subscription, and email a picture to anh.marketing@balchem.com. Include your size and mailing address, and we’ll mail you a shirt.
Scott (00:07):
Good evening everyone, and welcome to the Real Science Exchange, the podcast where leading scientists and industry professionals meet over a few drinks to discuss the latest ideas and trends in animal nutrition. Hi, I am Scott Sorrell, gonna be your host, uh, this evening for the podcast. Uh, we've been at the Tri-State Nutrition Conference this week, and this is our 10th podcast and final podcast. So we, we, we saved the, the best for last. This one's titled What We Have Learned With Feeding, uh, with Feeding and Automatic Milking Systems and Opportunities for Advancements in the Future. And that's with Dr. Samuel Fein from a MTS. Samuel. Welcome. Uh, looking forward to the, the conversation here. Um, before we get started into the conversation, would you mind just kind of giving us an overview of your background and, and what you do for a MTS?
Dr. Samual Fessenden (00:58):
Yeah, sure. Um, so yeah, my background is I, I grew up on a dairy in Central New York, um, and then did my education in a variety of places, uh, at, at Cornell University, and then also came out to Minnesota, uh, university of Minnesota to work with, uh, Dr. Marshall Stern out there. And then after graduating from my PhD, um, which was primarily in the C-N-C-P-S development, uh, I then joined a MTS since that was in 2016, and spent a few years, uh, bouncing around North America and Europe and a few other parts of the world, kind of helping people understand how to use the model and, and really kind of refine their, their approach to, um, yeah, nutrition management on farms. So, uh, did that up until 2020. And then, um, yeah, I decided to kind of take a little bit of a, a change of course. And, uh, my wife and I started up a robot dairy in Central Minnesota or, uh, uh, Southeast Minnesota. And so we farmed with her parents there. Uh, milking a hundred and, well, right now about 135 cows on, on two le a five robots. Um, and I still do some occasional work for a MTS, um, on the side as I can fit it in, so, okay,
Scott (02:03):
Good, good. Yeah.
Dr. Tom Tylutki (02:04):
Overcrowding the robot.
Dr. Samual Fessenden (02:07):
Yeah, dad, man, I should have dried off a few cows before I left. That's what I got going. So,
Scott (02:11):
And we're listening to my co-host for this session, Dr. Tom Lucci. Tom, welcome this. You're, you're, uh, uh, uh, an expert at the, the pub. This is at least your second time
Dr. Tom Tylutki (02:22):
Expert at the pub. I like that.
Scott (02:24):
Dr. Tom Tylutki (02:27):
Sort of like Sam grew up on a 60 cow farm in Eastern New York. Ended up going to Cornell, uh, bachelor's, masters, then went into extension for a few years. Uh, went back for my PhD, stayed there a few years working on the model, and 2005, three of us spun out with a license to commercialize it. And since then, you know, we went from three people with no money, no product, just the big dream. And now we're in like, I don't know, 63, 64 countries. I always have to ask Lynn. Um, and yeah, it's been a blast touching a lot of cows and sheep and goats around the world.
Scott (03:04):
Yeah. Awesome. Awesome. Welcome. Glad to have you back. Thank you. Got two color, uh, commentators with me today. The first one's gonna be Nate Elzinga. Nate, welcome. Welcome. Yep. Uh, give us a little bit background about yourself.
Nathan Elzinga (03:17):
Uh, so I grew up in, uh, dairy in West Michigan. Um, started young out in the barn, uh, went kicking and streaming to Michigan State for dairy management. Um, I actually really enjoyed it, but, uh, went back to the farm for 10 years and then, uh, my seven year anniversary coming to Caledonia and, uh, doing nutrition work for them. So yeah.
Scott (03:37):
Good, good. Glad to have you here. Looking forward to the conversation. And then we've got, uh, we've got Brandon Van Seuss. Now, that's a famous name, but I I'm imagining that you're not related to the, the fabled, Peter.
Dr. Brandon Van Soest (03:51):
Well, it's not, uh, been confirmed or denied yet, so,
Scott (03:54):
Okay. I see
Dr. Brandon Van Soest (04:01):
Yeah, so I traditionally didn't come from a dairy background per se, um, family owned landscape company, actually. And I decided it wasn't for me, and I wanted to go to vet school. Originally went to Purdue for undergrad, uh, animal science, and then decided he wanted to go to grad school, so went to Michigan State, um, under Dr. Van Deha, looked at some calf stuff there, and then ended up going to University of Guelph for a PhD, uh, working with Dr. Trevor Reese on robotic, uh, milking system, so mm-hmm
Scott (04:29):
Yeah, we had an interesting podcast earlier with, uh, a young lady, Anna Schwake, that, uh, that was very interesting. Yeah. Sam, uh, you gave a presentation earlier today. Would you kind of give us just kind of a background of some of the key elements that you covered and kind of give us a foundation for our discussion today?
Speaker 6 (04:53):
New research is changing everything we thought we knew about Coline's impact on the cow and her calf and top scientists have a lot to say about it. They're presenting new research that supports choline as a required nutrient to optimize milk production choline as a required nutrient to support a healthy transition choline as a required nutrient to improve calf health and growth, and choline as a required nutrient to increase colostrum quantity. This new research is solidifying cho's role as a required nutrient for essentially every cow, regardless of health status, milk production level, or body condition score. Learn more about the science that is changing the game and the choline source that is making it happen. Reassure precision release choline from Bal Chem, visit bal chem.com/scientists say to learn more.
Dr. Brandon Van Soest (05:56):
Yeah. Yeah.
Dr. Samual Fessenden (05:57):
So when we were, when I was pro or approached for, for doing this talk, um, you know, I, I didn't really know exactly how to title it, and it was something that, uh, yeah, I, I, I was asked to look at, you know, what, what have we learned and what's in the future? And I was like, well, geez, I don't know. I, I've learned a few things just myself, but there's a lot of people that feed a lot more robot cows than I do. But, uh, yeah, I figured I'd take a stab at it and kind of start going through the research side of things and, and working and talking to different nutritionists to kind of try to get a little bit bigger picture. Um, so yeah, that's what I kind of did with the talk is, is talk through, you know, this is what I generally see out there, and then maybe some of the gaps that we have in, in knowledge and in just, you know, do we need to sort of change our overall approach to feeding cows in a an a MS system.
Dr. Samual Fessenden (06:39):
Um, and so yeah, that's kinda what I did focus a little bit on the, the robot feed itself and then a little bit on the PMR formulation. And, um, then kind of a little bit of a look to the future of what we might be able to anticipate coming from formulation programs or from just kind of general thought process of how do we feed cows in a, in a, you know, maybe slightly different environment than we're used to with the TMR fed herd. What do you see as the biggest challenge? Uh, well, there's a lot of data and a lot of, uh, people, a lot of ways to look at it, and we don't really have a clear idea of what stands out as the most important data. So I think people get lost in the weeds when they're trying to formulate, and they get, I think, a little bit too in their heads with, uh, well, maybe I need to have exactly this pounds or this, you know, whatever metric I want to use when I think in reality, a lot of it's just letting the cow kind of be the cow and let her, you know, do her thing.
Dr. Samual Fessenden (07:26):
So, so I think I try to take a little more simplistic approach as I'm, I'm approaching formulating those herds, so, yeah. Cool. Yeah. Yeah.
Scott (07:34):
So you mentioned gaps. What are some of the biggest gaps you've seen?
Dr. Samual Fessenden (07:37):
Well, so like, what I see from, especially on, on translating, I'll say robot research into some of the, uh, um, actual on farm use or on farm, uh, formulation is, is I think some people try to get a little bit tuned into understanding, you know, if I feed X amount more pellet or robot feed, what happens to PMR intake? And, you know, there's a lot of studies that focus on that substitution ratio. Uh, and, and if you look at some of the more recent reviews, it's all over the board, right? We have substitution ratios that go from let's say 0.5 to 1.5. So when you are trying to approach it using those sort of ratios, or even just the ratio of energy in the, in the PMR versus the, uh, robot feed, you're kind of looking at, I'd say maybe the wrong numbers, you know, I think we're focusing too much on these, these, uh, metrics that are so highly integrated, you know, cow behavior and, and, and, and intake behavior is so, such a highly integrated outcome.
Dr. Samual Fessenden (08:32):
And trying to model or predict that from a formulation standpoint is, is almost an impossible task. And so, so that gap there, I think is one of the largest ones that I see where, where people kind of get wrapped up in some of these ratios, but really we just need to kind of go back to basics and say, Hey, I, I just gotta formulate a good PMRA good room, friendly PMR, and then work on a good palatable feed that gets the cows into the robot. Okay. Yeah. So, alright. That's, that's where I see some of the gaps.
Scott (08:55):
Yeah. So you gentlemen work with, uh, robotic dairies. Yeah. So what's some of the things, uh, some gaps that you guys have seen?
Nathan Elzinga (09:04):
So I'd be curious if in a scenario where you have, uh, basically corn silage to work with and it's a high starch corn silage, how do you formulate a PMR to drive visits to drive cows to the, to the robot?
Dr. Samual Fessenden (09:17):
Yeah, so like, like in my situation, what I do on my dairy myself is I, I feed a fairly high corn silage diet. I think we're turning around 24 pounds of, uh, corn silage dry matter, um, per cow. And so, yeah, and, and not a lot of haage just given agronomics and how, what makes sense to grow on my land. Um, so yeah, I, I tend to have a fairly heavy corn silage, PMR, uh, my robot feed is ground corn, just just ground corn being fed through the robots. And then I do have a gluten pellet coming in to kind of help, basically provide some feed to those late, later lactation cows and, and be able to get a little bit more feed into 'em that way. Um, so in my case, if I look at my ground, my, my total robot feed, especially for some of those high producing cows, it's a fairly high energy, you know, set up because it is just straight ground corn. Um, and so I think if, if we start to look at the difference between the PMR energy content and the pellet energy content, you know, maybe I'm kind of ending up in that situation where I don't have to worry too much about a high starch load in my PMR. I've, I've still got that energy drive for the, the pellets and the, and the robot or the, the robot feed. Yeah.
Dr. Tom Tylutki (10:20):
So, you know, it, it's kind of interesting if I jump in there talking to some, uh, nutritionists in Canada that feed a lot of robot cows, um, they've kind of done this break point of if the herd's doing less than 35, 37 liters, so yeah, less than 80 pounds, then they'll, they'll start with that PMR being 80% of the energy level of, of, but as they get herds going 40 up, so, you know, 90 pounds up, they're actually bringing that closer to where in some of the cases they're over 90% of the energy coming from the PMR and they aren't seeing that this drop in visits or anything. It's, she's just got the demand to go get milked.
Nathan Elzinga (11:03):
It's like the chicken or the egg thing. Like it's, I've noticed that too. I've got one herd that's a real high producing herd and I can get away with murder on the PMR. I can have, it's basically like a TMR for most other herds, and we're feeding a ground meal through the robot and limited amounts of it, but I'm getting the visits and, and the production's there and it's not a factor. But I think you're right. I think you've, as that production level goes down, it's like you need to drive the production, you need the visits to do it, but what do you do first?
Dr. Samual Fessenden (11:29):
Yeah. I think what I see, you know, and I don't necessarily have this problem in my herd or, or in some of these, let's say higher producing herds, is you get those, those late stale cows or the lazy cows, you know, that sort of thing. And it's, I, I don't really see that. I, I, at the end of the day, I don't really, a late lactation cow, I don't really care if she's only getting milked, you know, maybe even less than two times a day
Dr. Tom Tylutki (12:12):
Think the repro issues is a big one because what's your calving interval,
Dr. Samual Fessenden (12:14):
Right? So for us, I mean, we're right around that 13 months calving interval. Yeah.
Dr. Tom Tylutki (12:19):
And that's probably, I think, starting to push it in terms of maybe running into some issues with some, some of those lazy cows.
Dr. Samual Fessenden (12:26):
Yeah. Yeah. And like for us, we, we don't really, I don't really seem to have those lazy cows. Yeah. Um, but I think in, in situations where maybe it's just not a, a higher level production in the herd that yeah, you end up having cows that are just kind of camping out at 70 to 80 pounds for a long part of their lactation and you know, they're not, not really coming into the robot. They don't have that that drive to, to be coming into the robot to eat and to get that energy that,
Nathan Elzinga (12:49):
So you have two fees going to your system. How do you manage that throughout lactation, right?
Dr. Samual Fessenden (12:53):
Yeah. Yeah. So right now,
Dr. Tom Tylutki (13:43):
Hmm. How'd you end up just doing the gluten pellets?
Dr. Samual Fessenden (13:46):
Um, meaning when we made the switch over or? Yeah. Why, why? Uh, well, so the reason I have gluten pellets in there still, I guess is because of just bulk, bulk density. I mean, if I want to be pushing, because I, I typically like to be somewhere between 10 to 12 pounds of, of robot feed per a hundred pounds of milk. And so if I am, you know, trying to get that much feed into a high producing cow and I'm feeding it with ground corn, you know, especially that high producing cow is getting quite a bit more ground corn than than gluten pellets. Um, I still need something to help bring in a little bit of, of, you know, more feed and a, and a denser form so that
Dr. Tom Tylutki (14:20):
Yeah. Yeah. But I guess I was going in slightly different way. Know you go to a lot of places they'll have these fancy,
Dr. Samual Fessenden (14:27):
All custom pellets. Custom
Dr. Tom Tylutki (14:28):
Pellets and like that.
Dr. Samual Fessenden (14:29):
Well, for me it was cost. I didn't really wanna pay, pay for custom pellets.
Dr. Tom Tylutki (14:32):
And did you see, did you see any difference when you went from
Dr. Samual Fessenden (14:34):
The custom pellets? No, not really. I mean, we went from, uh, yeah, 'cause I, I, we started the dairy in, in fall of 2020 and we fed a custom pellet to start off just 'cause I, I didn't really wanna rock the ship too much. I was still in learning process. I'm, I'm always still in the learning process
Dr. Samual Fessenden (15:09):
And so yeah, that's where I just said if I can use my own corn on farm, then that's going to be a, a good cost savings. At the time it was probably about 85 cents ahead per day. So I, I, I could, I could spend a lot to, to figure out how to make this work. And, um, yeah, so, so we did that and, and honestly what I first saw, you know, without changing much, and we were at a pretty stable point when I changed over. Um, so I, I could kind of watch my numbers. Uh, I didn't really change much on milk access or anything. So when we switched over the, and maybe it was just the feed switch itself, but I, right. I had kind of an increase in refusals, so I had cows that were more interested in coming into the robot anyway.
Dr. Samual Fessenden (15:46):
Um, they just didn't have milking permissions to allow it at that time. Um, and yeah, milk production stayed pretty steady. Um, and, and components actually slowly started to come up as I was able to tailor that, that energy levels for those high producing cows. Um, so yeah, it was, um, definitely a good switch for us to, to go to that ground feed, um, side. I've been slowly reducing overall gluten pellets since then just to Oh, really? Yeah. Just to see, just to see what I can do and, you know, it's, it's working so far
Dr. Brandon Van Soest (16:13):
When you swapped from the robot pellet, did you go down and feed or did you stay same on how much you fed through the robot? Uh,
Dr. Samual Fessenden (16:20):
When I swapped to pretty much about the same, I was trying to target that same kind of level just on a, on a pounds per a hundred pounds of milk. Um, but yeah, beyond that I, I've, I've been slowly reducing it just to kind of see what I can do, but I don't really have an incentive to reduce down to like eight or or six, you know, what you kind of see maybe even towards more of the, the guided flow tech traffic systems. Um, just because it's not like I have the extra cost of a custom pellet that I'm trying to offset. If I'm gonna take corn outta the robot, it's just gonna go into the PMR. So
Dr. Tom Tylutki (16:56):
You, you're not gonna try like some of the people you're doing with the guided flows of not feeding any pellets in the robot.
Dr. Samual Fessenden (17:01):
Uh, not in my farm mill.
Dr. Brandon Van Soest (17:03):
Dr. Brandon Van Soest (17:45):
Um, but still target those, uh, fresh cows, high cows, but also not overfeed the lake cows with a high starch like you're doing with the gluten. Yep. And it saves costs too, right? I mean, at the end of the day, you're gonna have a more expensive pellet or feed that you're gonna feed the fresh cows or feed the high cows, but then you're lower lactating animals that don't need a much, but you're still gonna be able to give them enough to incentivize them to come in instead of just trickling a little bit in there or having to affect your PMR so much that you're now balancing it too low.
Dr. Tom Tylutki (18:14):
It's almost like we're going back to good old component feeding.
Scott (18:17):
Yeah. Yeah. So you mentioned you traveled a bit in your early career. Did that also include, uh, Europe and if so, right? They, they, they do a lot of robot feeding over there, uh, a lot of pellets, but seems like I could be wrong, but they're putting a lot more nutrition through those, uh, feeders than just carbohydrate. Is that Yeah. That be a fair assessment?
Dr. Samual Fessenden (18:38):
Um, I, I'd say probably generally, um, there's, there's a lot of differences in the, in the programs over there in terms of, um, yeah, what sort of feeds are on farm, what sort of feeds they have access to, uh, you know, a lot less soil and a lot more rape seed, um, just given, yeah. Different regulations over there. Um, and yeah, so yeah, when I was over there I was traveling quite a bit in, in continental, continental Europe and then even, you know, a little bit in Ireland and, and England a little bit as well. Um, but yeah, going around Germany and the Netherlands, lots of robot herds, um, and a lot of 'em feeding a custom pellet. But what's interesting is a lot of 'em are, are, it's not like a tailored pellet. I say custom pellet, the mill's gonna carry one or two or maybe three robot pellets and all the robot farms are gonna get the same formulation. Um, and so, so you see a little bit less, uh, uh, kind of customization in that way. What's interesting is some places also feeling out like palm kernel, which is not a very palatable feed.
Dr. Tom Tylutki (19:30):
Oh, palm ker. Well, it's cheap
Dr. Samual Fessenden (19:32):
And it's cheap. That's, yeah. So basically just trying
Dr. Tom Tylutki (19:34):
To completely indigestible Yeah,
Dr. Samual Fessenden (19:36):
Beding, yeah. So it just was, was just interesting seeing that that difference there. But they are pushing more total pounds through the robot, um, in, in, in some instances. Some places are not. But, um, yeah, it was kind of good for me to see that because I kind of understand, okay, you know, this is what, why they're all mostly on pellets over there and very few non pelleted feeds. Um, and that's right how the systems were built, right. So, so deal island,
Dr. Tom Tylutki (20:01):
I think just how the systems were built, it's really just a difference in, in the, the industry. Yeah. In the overall industry. They are still much more reliant upon the more traditional feed mill relationship versus US and Canada where it's so much driven by consultants or, you know, what farms and consultants had done to the feed industry of decoupling everything. Yeah. Uh, that, that it, it's just widely, wildly different than than yeah, north America.
Dr. Samual Fessenden (20:32):
Yeah. And like in, in North America, you'd be rare to see a a, I'll say custom formulated pellet in the, in the PMR or in the TMR, but in, in Europe that's, you'll see that, you'll see a protein, you know? Yeah. Everything's protein pelleted.
Scott (20:43):
It's actually a frustration for us. Right? Yeah. So balms in the, uh, ru protected nutrient business and those are lipid encapsulates. Yeah. You're screwed with they don't last in pellets and they are bound and determined to put it in the pellet. Yep, yep. And I can't get 'em to put it in the PMR and so, I don't know, it's just kind of cultural. It
Dr. Tom Tylutki (21:01):
Is. Yeah, it is. Yeah.
Scott (21:02):
And the feed companies, right, they wanna put it in the pellet 'cause they wanna sell more pellets. Yeah.
Dr. Samual Fessenden (21:06):
Yep. Sell more pellets, the easier to move pellets, you know,
Dr. Tom Tylutki (21:10):
I'll make no additional comments about app pellets.
Dr. Brandon Van Soest (21:13):
Well, and there's a chances gonna be great. Less shrink, right? If it's in app pellet form, it's not blown away. Yeah. Truth true. Potentially could be more consistent if the pelleting and mill does Well, good job. But if's a good pellet
Dr. Tom Tylutki (21:36):
Yeah.
Dr. Brandon Van Soest (21:36):
Put it in a bin. And
Dr. Tom Tylutki (21:37):
The pelleted mineral products are flipping awesome. One of our, our customers actually, they're a distributor in Argentina, they do a lot of pelleted minerals and it is just truly, I wish, if anything that would be really cool to have in the US because that, that handling and shrink is just way much easier.
Nathan Elzinga (21:56):
So you mentioned feeding palm kernel over in Europe. One of the things that I plan on toying with here in the future, high oleic soybeans through the robot in combination with something like corn. Uh, you guys have thoughts on that?
Dr. Samual Fessenden (22:08):
Yeah, we've, we've got some farms around us in, in Southeast Minnesota doing that. Um, yeah. Using some of those high oleic beans. Um, and they say cows love it, it works really well. Mix it in with corn, you know, and grind it or, or even more, you know, a larger particle size. Yeah,
Dr. Tom Tylutki (22:24):
Yeah. I'd still grind it because you want to
Dr. Brandon Van Soest (22:25):
For
Dr. Samual Fessenden (22:26):
The Yeah. For how it flows outta the feeders, you know, that's, I think trying to find a, a happy medium there and make sure you're not getting
Dr. Tom Tylutki (22:31):
Touch quarters at least. Yeah. That'll still flow.
Dr. Samual Fessenden (22:33):
Yeah. Yeah. But if you have, you know, a lot of fines in there too, if it, if it bridges up and that's, that's the biggest struggle with ground feed in a, a lot of the robot feeders is not
Dr. Tom Tylutki (22:40):
My problem
Dr. Samual Fessenden (22:41):
But it is my problem.
Scott (22:42):
It's
Nathan Elzinga (22:46):
Well then you could control dose over lactation, um, and, and cut the, the starch, you know, hopefully is palatable enough that that's to help, uh, draw cow the robot too at the same time. Oh,
Dr. Tom Tylutki (22:57):
Well, roasted beans. Yeah. Especially roasted. I love roasted beans. Yeah. Yeah. And I've done some stuff with, with high ole beans with the bypass plant in New York. Those high ole beans are flipping cool. Yeah.
Dr. Brandon Van Soest (23:10):
Yeah. We've got some, several, several farms putting it in the robot. Um, you know, it's one of those things that allows the farmer if they're growing it on farm, um, to make it on farm grain mix. And that's something, you know, we're totally fine with, you know, you know, taking it off the plate of taking pellet out. But yeah, we've, we've done that where we're mixing it with on farm grain, um, and they just kind of make that and put it up in their bin. Yeah. So it's worked, worked well.
Scott (23:33):
Yeah. Good, good. So you said there's some things that, uh, you're gonna talk about what we need to change, and maybe we've touched on some of those, but kind of gimme an idea of some of the things that you talked about
Dr. Tom Tylutki (23:44):
In, in your presentation.
Dr. Samual Fessenden (23:45):
Yeah, I think, you know, kind of looking towards the future, if you talk with most nutritionists, you know, they're almost always gonna be balancing on a group of cows sometimes, you know, even just taking 'em out the robot side, you're talking about A PMR that they're, sorry, A TMR that's fed to, sometimes it's a single group, you know, for all milking cows get the same TMR or if we're talking high and low groups or, um, whatever types of splits you might have. Uh, you know, we are always using our ration balancing programs on averages of cows or, or groups or herd averages or one standard deviation above if you're kind of doing the lead feeding style, um, all those methods or, or kinda the predominant method that's out there as we move more and more towards the, the use of this technology and, and kind of get towards what people would call precision feeding.
Dr. Samual Fessenden (24:30):
Um, can we be targeting at least maybe smaller groups of cows and maybe getting down to individual cow ration balancing. And I think a quite a bit of that's on kind of putting on my MTS hat, us
Dr. Tom Tylutki (25:06):
That sort of thing. Well, just, here's, here's a great example. Looking at some of this stuff over the years, uh, when we balance for, for the average of a group, okay. Or, okay, we'll throw a lead, the lead factor on there so that we're 10% above average, or whatever number you want, when you actually look at the individual population within that group, and I don't care if it, if if it's the high group or if you group by age, whatever the assumption is that that group is following a normal distribution. And when you dig into it, I think in all the valuations I've done, I may have seen one pen that followed a normal distribution. Most of the time we're dealing with a, with a bimodal or even a trimodal population. So we've got these two or three subgroup subpopulations and we're trying to shoot for one of them.
Dr. Tom Tylutki (25:58):
And it could actually be where we've got this population that's way to the extremes. And what we're balancing for doesn't really exist. So the kind of the approach that, that we're working on is, is to be able to do, look at this with the individual animals and, you know, the simplest form, even if we didn't get to the individual animal basis, we could identify those subpopulations. And actually we may be formulating, okay, we're gonna formulate for this, this group, but behind the scenes we're actually evaluating it and formulating it for two or three subpopulations so we can get an overall, a better overall picture of, of what those animals are gonna do. It's pretty wild as, as I've dug into this, what we can do it, it's pretty neat.
Dr. Samual Fessenden (26:43):
Yeah. And I think as you, as you look at even individual cows, you know, not that we're necessarily ready to balance on an individual basis, but if you start looking at like, lactation curves of individual cows, I mean, we all can picture a standard lactation curve, but you start looking at individuals and you're like, all right, well this cow, she came up, you know, and she peaked let's say 140, 150 pounds, and she stayed there for 200 days. She got pregnant in that time. You know, there are cows out there that do that. And pretty much in every herd there's those cows that just have this, this atypical lactation curve. And so if we go ahead and adopt an overall feeding, you know, strategy of saying, okay, yeah, I've got a high cow and a low cow group, let's say just for a TMR type herd, you know what you gonna do? You're gonna, you're gonna have her in a high cow group for a while, and then once she's, you know, checked pregnant or confirmed pregnant, you move her to a low cow group. Well, is she actually a low cow? Maybe not. You know, she might still be producing really quite well. And then you move her on to this low energy ration and guess what? You make her a low cow. And
Dr. Tom Tylutki (27:37):
So that's a great point, Sam. I actually did, we were doing that at a farm I worked with closely for a number of years. We actually grouped more by age. Yeah. And, and they had daily milk weights when it was kind of wild. I was watching cows, we would move cows. Okay. Just normal cow behavior, that whole social reset, they've gotta find their place in the pecking order. If they're on the same diet, just that move, they'll come down and milk for three or four days and then come back up. Depending on the social structure, they may end up higher on their lactation curve. If you change diets though, and this is where I've, I've found that, oh man, I'm not changing diets anymore than three points of forage between groups. If you change that diet, five, six points of forage, like a lot of people do, going from high to a low or more, they come down with that social crash. And there's so much more NDF in the diet that they're full. They never come back to that. So we, we've just clipped those cows and wow, now we got fat, fat like lactation cows. There you go, folks.
Nathan Elzinga (28:36):
Dr. Samual Fessenden (28:36):
Yeah. And that's, I I think when you kind of translate that into the robot herds, if we look at robot feed tables and you kind of end up, and sometimes you look at what people put together on their, on their feed tables and you say, well, there's, there's some traps here. And, and by traps I mean like where a cow's gonna fall off a tea feed table, or she's not gonna be able to really, all of a sudden she's gonna cross over this magic number of 200 or 210 days in milk, gets onto a different part of the feed table, and all of a sudden gets, you know, a bunch of feed pulled away from her when she really actually should be entitled to more. 'cause she's still cranking away on milk.
Dr. Tom Tylutki (29:04):
I think you hit a good point there. I think people overcomplicate their feed tables. Yes, absolutely. And they, you know, and this is a big thing in Europe and, and Japan where, where they'll have 'em on a flat flat feeding basis for 50, 60 days in milk, and then they go on the feed table and it's like, well hell, we're already post peak. What, what did we do? Yeah. Or, or they, they just do some really crazy things that, that just make no sense. They're overcomplicating it.
Dr. Samual Fessenden (29:30):
Yeah. Yep.
Nathan Elzinga (29:32):
So the, the, the tool in a MTS for balancing for robot herds or where are we at with that? And like, do you have plans to add more than one robot
Dr. Tom Tylutki (29:41):
Feed?
Dr. Samual Fessenden (29:41):
Oh yeah. Tom, when's my, when's my second feed coming in?
Dr. Tom Tylutki (29:44):
Dr. Samual Fessenden (30:44):
You know, I think our challenge with, as we're developing these tools and maybe as you've played with the, the robot feeding tool in the old program, um, it's, it's a little bit tricky because we we're trying to sort of be able to give people this, this whole different way of thinking about ration balancing, but we also have a lot, right? A lot of users are very, very used to doing something one way. So when we build out this new part of the program, it's, it, it looks, it's really messy and complicated in the development side of things. Yeah. And Tom and I know how to run it, maybe a few other people, very few. But then we gotta figure out how to make it actually usable for nutritionist. And so that you guys aren't just banging your heads against table saying, what, what, what is going on here? What are they trying to do?
Dr. Tom Tylutki (31:23):
Nathan Elzinga (32:47):
So is that why in nutrition, like you've scaled all the tabs back to like one or two and now all of a sudden there's a bunch more like you got up molded or
Dr. Tom Tylutki (32:56):
Yes, absolutely. That, that was part of the simple, simple test. It was my decision. I'm like, okay, we don't know which ones people use, so let's get rid of 'em all
Dr. Tom Tylutki (33:53):
No. Okay. Um, seven's kind of a different story than how the model development has been in the past. Uh, even like with now, with, with where we are with 6, 5, 6, the introduction of the branch chains, they would give us, you know, like we will do the branch chains is a great example since we just did that. Here's the, here's the equations in a spreadsheet form. Okay. So that we could then figure out how to build it in. Um, and as long as the answers we were getting are compliant with, with what they had in a spreadsheet, all's good to go. Seven, since it goes dynamic and it's a lot more complicated that they actually have had it coded themselves by a third party, and then we basically are getting a pack, a runtime engine of it that we then have to figure out how to communicate with it and build an interface around it.
Dr. Tom Tylutki (34:54):
We got the first version of that package middle of last year, and it was like, woo. Yeah. Um, you coders, um, only thinking the world of web apps. This, this is unworkable for us. So they had to go back and have a bunch of stuff redone to get it into a form that would actually be workable for several of us license holders. We just got word two or three weeks ago that we now have access to that package, um, when we're right in the middle of trying to finish up, you know, as we're rolling out nutrition and, and finishing up getting the new advanced optimi optimization routines in, uh, that we haven't even started that. And so, I don't even know at this point how workable it's gonna be. I don't know how long it's gonna be. It's, it's gonna be a pretty big project because we've gotta do all this interface, um, as well as we don't even know how fast it's gonna run.
Dr. Tom Tylutki (35:58):
You know, if it's going five, five minute increments, over 300 hours, you know, it's gonna be a challenge. 'cause like now in, in nutrition, be on a diet, you know, change of feed amount, you get the updated results. Well now if it's, if you hit a number and if, if it takes any significant amount of time to run through all 400 hours worth of simulations that, so we don't even know how, we don't even know any of that yet. Um, I'm hoping probably in the next six, eight weeks, uh, our, our coding team will start looking at it and trying to figure it out. My guess is if, if it really is as where it's supposed to be, where we can actually use it, we're, it'll be later this year.
Dr. Samual Fessenden (36:43):
And that'll be probably a little bit more limited to in the release to start, because we need to kind of understand what sort of anticipated changes we'd see on predictions. Well,
Dr. Tom Tylutki (36:49):
So kind of the plan too is we're not, it's not gonna wholesale replace 6, 5, 6. You're right. Yeah. Okay. We're actually gonna do it as, as, as an add-on like the advanced optimizer and it'll be, you know, an evaluation screen to start with because people aren't gonna have the, the data to do all of the calculations, all the slow pull fast below all the, you know, all that off of non forges, you know, there, there's, there's gonna be some challenges with rolling it out. So Yeah. I see it as a multi-year rollout, actually. Well, yeah, yeah. Yeah.
Scott (37:25):
What role do you see AI plan and rice and balancing software going forward? Or what role should it play
Dr. Tom Tylutki (37:32):
Today?
Scott (37:33):
Nothing. Nothing.
Dr. Tom Tylutki (37:34):
Yeah. Uh, because all these AI models are,
Dr. Samual Fessenden (37:38):
Well, they're trained differently. They're right. They're language models.
Dr. Tom Tylutki (37:40):
They're language models. Yeah. They're an infant. Yeah. Um, ask me that question. In five years,
Scott (37:46):
It's gonna take that long.
Dr. Tom Tylutki (37:48):
Uh, probably not, but I'm hedging
Dr. Samual Fessenden (38:46):
You know? But I think much further down the road as we refine our relationship with ai, it, it definitely will become a part of what we do.
Dr. Tom Tylutki (38:52):
Oh,
Dr. Samual Fessenden (38:53):
Absolutely. From data management, I, it's gonna have to, not even just ration balancing, but just how do we, how do we understand and analyze the data coming outta some of air sensor systems on farms?
Scott (39:01):
That, that's what I was gonna
Dr. Tom Tylutki (39:02):
Ask as well. I think the first one is, how could you use AI in evaluating all the raw data coming outta the robot,
Nathan Elzinga (39:06):
Right? If you dump all the, all the outputs for A MTS and all the cow data and make this massive data set, then you could do something, right. Really
Scott (39:15):
Dr. Samual Fessenden (39:18):
And
Dr. Tom Tylutki (39:18):
Like I said, asian okay. Three years
Nathan Elzinga (39:48):
All my farms are perfect
Dr. Tom Tylutki (39:50):
Yep. Yep. That's what I thought too.
Scott (39:56):
Sam, is there anything that we've left uncovered that, uh, you presented in your presentation today?
Dr. Samual Fessenden (40:02):
Um, well, I guess I'd just say, you know, as, as nutritionists are, are looking at trying to understand Yeah. Feeding robots. I, I, I think especially with young nutritionists, there's always a, a challenge of, of feeling confident enough to try something. And so you always tend to default back to what's safe. And, and, and nutritionists in general are, you don't wanna lose a client, you don't wanna make a mess, you know, make a change that, that messes up the farm. And a TMR Fed heard, okay, you know, yes, you can pretty easily make some big mistakes, but when you, when you make some mistakes in a robot fed herd or, or a robot herd, right? All of a sudden the cows aren't coming into milk, at least with right. Classical farms, they're getting milked three times a day or two times a day. So, so I think there's sometimes a little bit of stress put on, especially those younger nutritionists to feel confident in what they're doing.
Dr. Samual Fessenden (40:43):
Right. My advice would be just go ahead and try it, you know, sort of fake it till you make it type thing and
Dr. Tom Tylutki (41:24):
Well, I want to add something to that, and it's something that you just inherently do. Okay. Okay. Um, and young nutritionists would typically don't know this, just go watch the cows. Well, yeah, watch their behavior. You can tell so much about even within robots with, with what flow is, what's, what's all this, you know, what kind of attracts cows are like that. Watch the cows. We don't train people in grad school or undergrad to be nutritionist and, and we never even talk about that type of stuff of watch the cows.
Dr. Brandon Van Soest (41:59):
That's something you run into a lot is like, I'll get called in tech support for, from some farms and some of our, my colleagues and they'll be like, I looked over the rations, they look good, you know, what's going on? And my first thing is, is well, what's changed on the farm? You know, come winter time some guys like to put stuff around the robots to try to keep it a little warm. Well, is this the first year they've done it? You know what, you know, there's definitely things that we do that really, you don't think anything of it, but it dramatically affects, you know, it's a new situ, it's a new situation, new environment. You're putting something up to, you know, block stuff while you're blocking their vision. You know, like you said, if you, if you sit back and kind of watch cows how they're interacting within the pen, you can pretty quickly see like, oh yeah, we did, do you changed that? We moved to water, we moved, uh, whatever. You know, just kind of see, oh, well we're all clunked up here. Well that fan's broken. Well that's why they're all, you know, whatever, you know, I think sometimes we get too hyperfocused on the rations. It's gotta be perfect. Well, I think it goes back to feed management. Well, the juno's been down, or you know, the feed pushups haven't been done. If it's a manual or Oh, the mixer wagon is not mixing like it should. It's like, well the pay
Dr. Tom Tylutki (43:08):
Loader's broke, so we're using a skidsteer to load the mixer and it takes three times as long, so we're not paying attention or the
Dr. Brandon Van Soest (43:13):
Scale's off, you know, the weights way bars, but scales. Yeah. Right. The way bar on the mix wagon's off or something. You know, there's a lot of things like you said that just get on the farm and just, you know, take a few minutes to, you know, instead of doing a quick walkthrough, hey, let's take a few seconds to kind of visualize things. I think some nutritionists, and I mean, I can get caught up in it too, going back to the ration and saying, oh, milk's down, milk's up. Visits are down. Visits are up and try to tweak something and it's like, look at the bigger picture. You know, there's, it's a, it's a full system. It's not just what you're feeding. So Yeah,
Dr. Samual Fessenden (43:44):
I mean we're, we're trying to balance in these programs to grams of amino acids. Right, right, right. The weight of a piece of paper and the cows eating, you know, 60 pounds of dry matter a day.
Dr. Tom Tylutki (43:54):
Oh, even worse. When was the last time when you did dry matters, right?
Dr. Samual Fessenden (43:57):
Oh yeah. Yeah. That's still, still a challenge on farms. Yeah. Yeah. Uh,
Dr. Tom Tylutki (44:01):
I, one of, of the favorite things I love to do on farms is, you know, I should carry a lawn chair with me. Just go sit in the, in the feed alley and watch cows eat. Mm-hmm
Nathan Elzinga (44:13):
Starting to bring a camper, like total camper with me and just spend a week
Dr. Samual Fessenden (44:16):
Dr. Tom Tylutki (44:17):
That's not a bad idea.
Dr. Brandon Van Soest (44:18):
Hunting. Hunting. Cameras work,
Dr. Samual Fessenden (44:19):
Hunting cameras work for those that have kids that need to get home. Yeah.
Dr. Tom Tylutki (44:23):
But no, the campers even better because they're, there's no cell signal. No, no internet, no one can bother you. Tiny house. Tiny house. Yep. Yeah. One farm a week.
Scott (44:34):
Gentlemen, this has been a great, uh, great discussion. I've enjoyed it, but they flickered the lights and that means it is last call. And so what we do here at last call, what I'd like to do is have each of you to come up with a couple takeaways or couple key learnings that you had, just something you'd like to share with the audience. And, uh, and then we'll wrap it up. And I'll start with my cohost down here. Tom, if you wanna lead us off,
Speaker 6 (44:58):
Tonight's last call question is brought to you by Nitro Precision Release nitrogen. Nitro delivers a complete TMR for the room microbiome, helping you feed the microbes that feed your cows. To learn more about maximizing microbial protein output while reducing your carbon footprint, visit balcom.com/nitro. Sure,
Dr. Tom Tylutki (45:20):
Sure. I think it, it's really important, again, we, we gotta focus on the cows no matter what system we got. And, and I think it's a little bit easier to do with, with MS systems because they tend to be more cow friendly. Um, but we still gotta watch the cows. It's still all about the people and, and not be afraid, like the way Sam put it, don't be afraid to try new things and, and really look at the cost side of, of how we're doing these formulations and they're really challenging, but the technology will improve to where we can make these things easier to, to formulate and think about.
Scott (45:55):
Yeah. Good. Nate, what would you like to share with us?
Nathan Elzinga (46:00):
Well, I think this discussion's been helpful for me. I mean it's, I have all these things that I wanna try and, like you say, the confidence to do it and without screwing things up, but it's, I think that's where the industry has to go. Uh, margins are too tight. We gotta be able to feed individual cows. It takes more than one feeding a robot to do that if you wanna do it right. So, um, yeah, for me, this discussion's been a lot of fun and yeah, focus on the cows. The cows will tell you everything.
Scott (46:25):
Yep. Absolutely. Good comments, Brandon.
Dr. Brandon Van Soest (46:28):
Yeah, obviously echo the cows. I mean, you gotta be able to watch them. Um, but another thing, we touched on it a bit, but just, you know, not over, um, you know, focusing on feed tables all too much. You wanna make them make 'em user friendly. You don't want to hyper-focus on, you know, oh, they need to go up, they need to go down. Like you said, don't, don't create traps for 'em. Um, but I think another thing that we don't look at, and we didn't really talk today about, 'cause we're nutrition focused, but milk access. I mean, I think that's another one that I know internally we talk a lot about. And as nutritionists don't usually get an opportunity to make too many changes on those, but I think sometimes we set up cows up for failure there too. Like, you kind of touched on, you know, well we, we got more refusals. Well it's like, you know, is that a good thing or are we teaching these cows they can't go? And now you end up fetched cows down the road. So, um, but yeah, I think just not overcomplicating some of those settings, but still understanding what's happening so that you can use those as problem solving situations rather than, oh, I needed to go change the ration. I think it's looking, I mean, big thing is, is look at other things other than run back to the ration every time something's going wrong.
Scott (47:40):
Yeah. Thanks Brendan. Wise advice, Sam? Yeah.
Dr. Samual Fessenden (47:45):
Um, I think, yeah, for my couple key take takeaways and the first one I'll, I'll borrow from all three other panelists. Um, and I have no problem doing it because it's, I think that it's, that should be your first fo focus is on those cows. Um, just identifying those individual, you know, barriers to those, those cows being the high producers that they can be. Um, and really just taking time to listen to 'em and, and, and look and watch the cows. Um, and then, yeah, my second one would be, I'll say maybe a little more futuristic looking. Um, I think there's a lot of really cool things coming on the horizon for how nutritionists can help solve some of those, those persistent questions that we have on farms. Um, and so yeah, I think if it's, if it's from a ration balancing or even just large debt analytics, um, I think, yeah, there's some really cool things coming down the road, so, yeah.
Scott (48:26):
Yeah. Cool. I said it at the top that, uh, we saved, uh, the best for last. I still think that's true. So you guys did a great job. I want to thank you for joining us and, uh, sharing your time and expertise. It's been good. We've really enjoyed it. To our loyal listeners out there, as always, we want to thank you for joining us for, uh, yet another episode. We hope you learned something. We hope you had some fun, and I hope to see you next time here at Real Science Exchange, where it's always happy hour and you're always among friends.
Speaker 6 (48:53):
We'd love to hear your comments or ideas for topics and guests. So please reach out via email to a h.marketing at val chem.com with any suggestions, and we'll work hard to add them to the schedule. Don't forget to leave a five star rating on your way out. You can request your Real Science Exchange t-shirt in just a few easy steps, just like or subscribe to the Real Science Exchange. And send us a screenshot along with your address and t-shirt size to a h.marketing at bal chem.com. Bal Chems real science lecture series of webinars continues with ruminant focused topics on the first Tuesday of every month. Monogastric focused topics on the second Tuesday of each month, and quarterly topics for the companion animal segment. Visit b chem.com/real science to see the latest schedule and to register for upcoming webinars.