Chewing the data cud with SmartHolstein Lab’s Dr. Jeffrey Bewley

Share This Post


  • WKU SmartHolstein Lab was established by the Holstein Association USA and Western Kentucky University to explore new technologies and genetics. 
  • This operational dairy provides an easy-to-access demonstration and development farm to evaluate new technologies designed to collect novel phenotypic traits, including wearable or indwelling sensors and milk-based biomarkers. 
  • A major emphasis will be placed on practical use of data for on-farm decision making. Inclusion of data into AgriTech Analytics (a DHI dairy records processing center owned by Holstein Association) will also be explored.
  • In addition to providing wearable sensor technologies, HerdDogg’s Animal Record Collection (ARC) is serving as the central platform for collation of individual animal data.
  • We asked Dr. Bewley to sketch out the road ahead for this unique dairy industry showcase.
SmartHolstein setting

“We expect things to go wrong. That’s how we improve them.”

Positioned at the confluence of dairy technology, research, and education, the WKU SmartHolstein Lab is focused on herd health, barn management, and milk quality — with a heavy emphasis on data-driven science. Led by Holstein Association USA Analytics and Innovation Scientist Dr. Jeffrey Bewley, the project has assembled the dairy industry’s most advanced technologies, providing WKU researchers with a first-of-its-kind opportunity to advance techniques in precision dairy farming, cow comfort and health, and calf welfare.

Working alongside Dr. Bewley are WKU Dairy Manager Adam Blessinger, WKU SmartHolstein Lab Fellow Gretchen Colón Suau, and Dr. Fred J. DeGraves, D.V.M., who is the WKU Department of Agriculture and Food Science Chair. The project is also supported by the Kentucky Agricultural Development fund, which has provided a matching grant for the WKU SmartHolstein Lab to foster development of dairy and agricultural technology in Kentucky.

Rumor has it that you’re a dairy nerd. How did that happen?

I’ve been a dairy data nerd for at least 35 years. And I’m proud of it! I grew up on a dairy farm in Kentucky, and I was interested in data from the time I was really too young to be interested in data. When I was about 10 or 12 years old, my dad, who was an accountant, introduced me to spreadsheets and I immediately thought, I could do some really neat things with data from my grandfather’s dairy.. So while my friends were playing video games, I was building spreadsheets for our dairy farm.

As for my formal training, I was an undergraduate at University of Kentucky in animal sciences, followed by a degree in dairy management from the University of Wisconsin at Madison. I spent some time managing a dairy in South Africa, then worked for a dairy nutrition company, PerforMix Nutrition, in Idaho. 

I got my PhD at Purdue, where I focused on technology and decision economics on dairy farms. That’s where I got started in the technology arena, which led me joining the faculty at the University of Kentucky, where I was focused on issues such as cow comfort and mastitis. And then I spent some time with Alltech, a global animal nutrition company.

What makes the SmartHolstein Lab unique?

It’s the only place in the country where the dairy industry’s top talent and technology has been convened to explore the latest thinking in how animal genetics and animal welfare can produce better milk. Our mission is to lead Holstein and dairy advancements through research, development, and outreach in technologies, analytics, and genetics. Our scientists are studying novel phenotypic traits using the latest technologies such as wearable technologies, computer vision powered by artificial intelligence, and milk-based biomarkers. We’re looking at how data gathered during the studies can improve on-farm decision making.

Our goal is to optimize production for the best that the Holstein breed has to offer. So therefore we’re looking at how we can employ genetic selection strategies. Data that we collect, along with expert observations by the science and veterinary team are at the heart of this; we are collecting data that is used to determine how the animal performs or how the animal looks or how the animal acts. 

Historically milk production metrics have been the focus of dairy science. It’s relatively easy to measure how much milk each cow gives per day, and so for well over a century the efforts of the industry have increased milk production per cow. But there’s more to breeding a great cow than just how much milk she gives. 

Sure, we want an animal that provides a lot of milk. But more important is that she does it in a healthy manner. And that she is efficient and fertile. In order to make genetic selections to pursue these goals, we need data on a wide array of attributes and observations. Our goal is to assess how we can use data from various new technologies to make better genetic selections.

Are there specific genetic traits that you’re looking to improve?  

One area of focus is the TPI index, which is a combination of various different traits that are designed to really produce a profitable animal. The Holstein Association owns the TPI formula and publishes that information. TPI is designed to select for a balanced cow: one that is healthy, presents well, and has a long lifespan — all in one package. The TPI formula is one way of trying to select for that, and then to breed animals in the top 20% of that index.

As we think about the product—milk—there are specific qualities for which we want to select. For example, A2, which is a protein in milk. Some animals express A1 while others express A2. And while the science is still not fully vetted out at this point, there is some indication that there may be some benefits to human health and human digestion of dairy products when it comes from animals with the A2 genes as opposed to those with the A1 gene. This, in turn, may increase the marketability of the product and potentially could improve the acceptance of dairy products by consumers.

How does animal well-being factor into the research?

Well, first of all, I think animal well-being is extremely important. We have a responsibility to take care of the animals that we’re working with in the dairy industry or any livestock industry. And so ultimately the main driver for dairy producers focusing on animal well-being is just that it’s the right thing to do. Certainly it’s true that consumers are interested in animal welfare, but the bottom line is basically, we just want to take good care of their animals. The love that dairy producers have for the animals that they work with is very, very high. And everybody likes to walk out to the barn and see happy, healthy cows.

Animal well-being is an area where the HerdDogg system can really help us. Like all livestock farmers, dairy farmers are continually monitoring their animals, and there’s no doubt that technology cannot replace a trained eye. But nobody has the ability to watch every animal 24 hours a day. 

How else can HerdDogg’s technology help with this study?

With HerdDogg, we have a device that is watching that animal behavior all the time. One of the challenges that we’ve always had with animals is that, unlike people, when something’s wrong, they don’t walk up to you and say, “Hey, my stomach hurts” or “Look, my foot hurts.” We believe HerdDogg’s wearable technology can provide insights into whether or not something’s wrong with that particular animal. So, yes, there are things we can surmise just by looking at an animal, but with HerdDogg’s activity and wellness tags we now have the ability to measure across time and identify changes in behavior. 

So, for example, an animal that’s starting to have a foot problem or stomach discomfort or an infection will likely exhibit subtle changes in behavior. Perhaps their activity level may decrease slightly. Or their eating time or rumination time may decrease. The HerdDogg system can help us identify these trends faster to indicate that something’s wrong with an animal. 

It may not tell us necessarily what’s going on with the animal, but it tells us we need to go look at her in more depth or take some course of action to help her deal with whatever challenge she’s been confronted with. Signals and alerts such as these are of huge value to the farm and the cow, because if we can intervene sooner, we’re more likely to have a positive outcome than if we wait until it’s too late. A busy ranch owner needs all the help they can get to ensure the well-being of their herd.

This graph shows a clear decline in activity measured by our accelerometer. This is a very good indication of the onset of an adverse health event that begins about 6 am on January 12 and sustains.

Because we have a variety of technologies available here, we can determine how well each of these technologies helps us figure out when something’s going on with an animal. It may well be that a combination of technologies will provide a better indicator of animal health issues.

How does the HerdDogg Animal Record Collection (ARC) factor into this research?

Most technologies have moved towards storing data on the cloud, which provides a lot of advantages. Historically a lot of that data was sitting on a desktop somewhere and if you weren’t there at the farm, you weren’t able to access that data. But now we can access the data from basically anywhere and provide multiple users access — it’s ideal for research purposes.

We’re also integrating data from multiple sources. One of the challenges that the dairy producer has today is that they have multiple systems monitoring different things. And oftentimes those systems are standalone, so the information doesn’t communicate among the technologies. We are working with some groups that are trying to integrate that data so that they don’t have to go to so many sources to be able to understand what’s going on for that animal. We want to go to one location and figure out this particular animal’s situation based on all the data that we have from multiple sources.

How do you see dairies evolving through technology? 

There’s been a pretty rapid acceleration of technology in the dairy, mirroring a lot of what’s happening in the rest of the world. Take your average grocery store or shopping in general and you’ll see how data has completely changed those industries. It’s also very apparent in the automobile industry, where computers are now everywhere in our automobiles. A lot of what we do is take some of those basic concepts from industries like those and bring them into the dairy industry.

Another major change that’s underway, not data per se, but there’s a lot of automation taking root in the dairy to handle tasks that were formerly done by people. I think ultimately the overall effect is that it increases the level of sophistication of the people that are on the farm. They need to understand a lot of different aspects of the operation and it requires a more knowledgeable person to be a successful manager or assistant manager or worker on the dairy farm to be able to work with these kinds of technologies and data.

But by the same token, probably that’s a good thing in many ways, because if you take somebody in their early 20s who is just entering the job market, they’d much prefer to work with data, smartphone apps, and that sort of thing than to perform monotonous manual tasks. Certainly, data, analytics and insights are more engaging than herding animals. But maybe that’s just the data nerd in me talking.

Subscribe To Our Newsletter

Get updates and learn from the best

More To Explore

By using this website, you agree to our use of cookies. We use cookies to provide you with a great experience and to help our website run effectively.