Current Progressive Dairy digital edition

3 open minutes with Matt Utt

Progressive Dairyman Editor Walt Cooley Published on 18 April 2018

Matt Utt has been leading research projects at Select Sires for the past several years. He was recently promoted to a newly created position to analyze data and find new, beneficial uses for it.

Progressive Dairyman editor Walt Cooley interviewed him about what it’s like to analyze data every day and how his work will benefit producers in the future.



You have an interesting title. What does a director of data innovation do on a daily basis?

UTT: I spend a lot of time in front of a computer and a lot of time on the phone. I work with people from our genetics, production and processing, and marketing departments. I might interact with someone who is in charge of the department or a field staff member.

I also get to collaborate with universities, focusing on management and reproductive management. I work with many datasets, including genetic evaluations; herd management data, which is a big one; and fertility data. I’m starting to play around with a little bit of production data.

Why is there a need for data innovation?

UTT: It wasn’t like overnight we decided we needed a director of data innovation. It kind of progressed over time. I had started doing some work in datasets while I was working as director of research. It just got to the point where I kept getting dragged more and more into data analysis. It was then I realized we had a need for someone to have a handle on the various kinds of data we work with and look for opportunities to use that data to enhance what we do but, even more importantly to help dairymen, who are our customers and owners.


When you say you worknwith data analytics, what does that mean?

UTT: I will work in various software platforms, including SQL or structured query language. When working with large datasets that have millions and millions of records, you can’t analyze and innovate effectively in a spreadsheet. I still do some work with some dairy management systems, like PC Dart and DairyComp305, when we dig into data on the farm. I also work with advanced analytical programs for analysis, visualization and reporting of data.

What are the opportunities you see hidden in today’s dairy data?

UTT: I think dairymen need better tools to make decisions. There are two pieces to that. First is gathering data together. We are in a world where we get data from many different sources. We are in data overload. There hasn’t been the ability to pull a lot of the data sources together to help make decisions. One piece of the data innovation puzzle is getting the data sources together to help a dairyman make better decisions. The second piece of that is combining different pieces of information that can tell us different things about an animal.

This is where we could get into talking about making better decisions about how to breed an animal, when to breed an animal, when to cull an animal. Different pieces of information play into that decision. There is genetic information that alludes to the potential value of an animal or its potential offspring. That is one piece of information that comes from one dataset. But another piece of information for that decision could be production level. Data innovation is getting data together and then improving how dairies make profitable decisions.

What kind of data is available to improve dairy cow reproduction?


UTT: Data available for improved reproduction today comes from two main sources. The first is going to be data collected in herd management software. And the second big one is activity systems. From the herd management data, we can look at past reproduction performance. We can look at performance by lactation groups and all the classic things we have done in the past. Activity systems give us the ability to monitor the animal more closely.

What is the future of reproductive research?

UTT: I think one of the questions to be answered is: What is happening to that animal before it receives its first service, and what effects does that have on the outcome of that service? Activity monitoring will help us answer those questions.

Another interesting one I see the industry moving toward in the future is in-line milk system data. Milk has got a lot of information in it. As we gain the ability to analyze those samples, like for example progesterone, we will know if animals are cycling coming into a reproductive management program. We will also have that information based on progesterone concentrations. Collecting data at the individual animal with milk is kind of pie-in-the-sky, but it may be a pretty big advancement.

What are the challenges you face to mine, refine and turn all this individual cow data into a helpful tool?

UTT: The biggest issue as we collect more and more data from herds is how different herds define different events that can happen to an animal. Dairies can have different definitions of how they describe the same event. For example, let’s say you call a case of milk fever “MILKFEVR” and I call it “MLKFVR.” I can go through the data and standardize it.

That is one piece of mining, but the other piece of it is the accuracy of how many times a milk fever event gets recorded. We’ll need some standardization if we want to do analytics across different herds.

Based on your experience looking at different datasets, give on-farm datasets a letter grade as to their quality for use in data analysis.

UTT: Let’s assume those are a C, or average, right now. I think if people start to see value in data, you will have people who will change, who will potentially adopt more standardized management practices and recording practices. How can that C become a B grade?

That has to do with how the user inputs the data. There are two pieces to how data users could improve: the standardization of the data labels and the standardization of the event definitions. As an industry, we are working toward a B by establishing some standardization. The ultimate goal, which is going to be hard to get to, is an A grade. That would require some sort of standardization for practices and protocols on the farm.

What kinds of data innovation will generate value for producers?

UTT: Producers want multiple sources pulled together so they don’t have to go to several different software platforms to look at their herd or to make decisions. They want the data in one place, like looking at weather data.

You’re obviously excited about data innovation. What will it take to get dairy farmers excited about it?

UTT: I think the first thing in all data innovation projects is to find something useful to dairymen, something they can really gravitate toward. Then the data output has to be simple, something they can visualize. If it can give them a piece of information they don’t readily get out of a single system, then they usually get excited about it.

What ongoing reproductive research data are you most interested in?

UTT: If we have the ability to gather more information from a single animal, as I mentioned before with milk progesterone concentrations and activity monitoring systems as examples, then we will be able to do a better job defining reproductive inefficiency or wastage. With that more finite information, we can determine variables related to reproduction: for example, why an embryo is lost at a certain point during gestation. end mark

Walt Cooley
  • Walt Cooley

  • Editor-in-chief
  • Progressive Dairyman
  • Email Walt Cooley