Under the Data Money Extension program of the Dairy Brain project at the University of Wisconsin-Madison, in collaboration with a participating farmer and following farm needs, we developed a Cow Selection Index Tool for automatic milking systems (AMS). The Dairy Brain Data Money is a farm-directed, team-based program (farmer, management team, vendors and consultants, extension professional, and Dairy Team members) for farmers to improve management by adopting integrated data, models, and tools specifically designed for them based on their needs.

Looking for a better outcome

A farmer participating in the Data Money Extension program who milks using AMS had the need to select cows according to their performance so that production can be more efficient. A rational goal on AMS farms is to maximize the total production per milking system in a day, which depends on the speed cows are being milked. Milk harvested per unit of time from each cow in a day is critical. Some cows could be good producers but slow in releasing milk; other cows, even being fast milking, could take a long time to settle in the milking machine. Thus, the farmer's question was: “Is there a way to select cows according to their production performance in the AMS?”

After several discussions, data exploration, and conversations with experts, a principal component analysis identified four seven-day daily average variables to define a dynamic cow-specific “index”:

• Milk harvesting flow
• Milking gate passages
• Milk yield
• Number of milkings

The index varies from 0 to 1, with a higher number being better, relating to cows’ ability to efficiently produce milk within the AMS. Thus, the tool can rank cows from top to bottom and, weekly, display low, top, and importantly, flagged cows that require attention.

Figure: Screenshot of Selection Index Tool showing the herd average index and a specific cow index throughout 68 weeks of time. Tabs are available on the left to explore lists of top, low, and flagged cows.

The tool has been available and updated weekly since late spring 2021, a time in which the farmer has been able to use it and apply it effectively in decision-making. Here are some of the decisions the farmer makes using the tool:

• Make sure that no cow being designated as do not breed belongs to the list of top cows because those top cows are quality AMS performing animals and should have extra chances to become pregnant and have a next lactation.
• Impose a stricter breeding cutoff for those cows that appear as a “flag cow” for not breeding them anymore.
• Further investigate and find the reasons why cows appear flagged in the last two extractions “flag cow +1” as these cows may have additional challenges that could be related to hoof issues, udder issues, injuries, or sores that are revealing further production degradation.
• Decide on culling cows out of the herd by looking at the “low cows” when the pens are full, and the AMS utilization drops to make room for the better cows to reach their full production potential capacity.

In closing, it is important to highlight the power of applied research through the Data Money Extension program that allows the participative development of needed tools that are immediately being adopted and having an impact.