Milking Parlor

If you have individual milk meters on your parlor and they are importing into BoviSync, there are a couple of good initial reports to look at. 

Average Milk and Deviation by Milking Unit 

Sample Report Output: 

 I review this report on a weekly basis.  Sorting by each column and reviewing the extreme ends to identify possible unit issues.  And if there are issues that are recurring, I shorten the days range to make it more sensitive (but noisier) and check daily.  I have also changed the format to a graph over time, then looked at each unit as a series.  This allows you to see rapid changes in a particular unit(s). 

When reviewing this data you need to keep in mind that there is no such thing as 100% unit calibration.  Keeping units within 5% of the average is good and generally achievable.  The milk that is measured per cow individually will never sum to the milk that leaves the farm.  You can compare these regularly and most parlors will allow you to make a general adjustment of all unit calibrations to get it to be in-line with the milk that leaves the farm. 

The way I interpret this report is first looking at the yields.  The milk yield per shift is what is the overall average milk yield measured by each unit.  In parallel parlors, where the distribution of cows to stalls isn’t random due to animal behavior, also pay close attention to the deviation from expected.  This unit uses the cow’s expected milk yield and the deviation from expectation.  When the behavior of cows stratifies them to particular units the overall unit averages can be expected to be a little different, the deviation can help with reducing that noise in the data. 

The peaks can help in two ways.  First, if I see a unit that is 20% below average milk, the peak is 20% below average, the deviation is 20% below average, and the duration is the same as the others.  That points first toward a calibration issue.  That unit is likely performing fine, but the milk readings are off by 20%.  On the other hand when I see lower peak, similar milk yield compared to the average, and longer duration; then I can see there is something going on with this unit that is causing poor milking resulting in long milking times and lower yield. 

In parlors that milk in batches (e.g. parallel, herringbone, not box-stall robots or rotaries), one other interesting report is to ungroup this report and look at long durations.  How many cows are being milked for long durations and all the other cows are waiting on that side for one more cow to finish up.  Eliminating these very long durations, can be beneficial to that individual cow, but more importantly it will increase cows per hour through the parlor.  This reduces the standing time milking and in the holding area, in addition to increasing parlor capacity and driving down milking labor hours. 

Milk Let-down

The report also shows the average milk by shift, first two minute yield, and percentage of milk in the first two minutes.  As well as the peak milk and duration.  You want to see the first two minute milk to be greater than 50%. Maggie Gilles has an article in Hoards reviewing this rule of thumb: Aim for 50% milk letdown in the first two minutes (hoards.com)  This first report helps you to understand issues with the parlor equipment, and the logic to use here is similar to the logic described above.  This also adds in another dimension to that data. 

Milking Shift Performance - BoviSync

The first two-minute yield is going to be primarily impacted by milking prep procedures.  This report uses the same data, but instead of grouping by milking unit, we are going to group by milking shift number.  Adding the manual and manual detach counts that some parlor interfaces send, can also be useful. 

Here I want to look at individual shifts and using my knowledge of the dairy apply my expectations of differences in shifts.  In this particular sample report above, you can see that the third (night) shift has about 4% less milk in the first two minutes, and has 5 seconds longer duration with 1.2 pounds greater yield (3.7% greater). 

When there are differences between shifts that aren’t explainable, look to milking procedure protocol drift between shifts. And when overall you aren’t achieving your goals, look at the protocol.  What changes need to be made in your milking procedure to get the milk yield you want. 

Mastitis Detection 

Parlors that measure conductivity can leverage this data to identify animals that may have mastitis.  A difference in conductivity of the milk for an animal is an indicator of mastitis, but not a detection method.  You can use this to identify animals as suspect cows to evaluate with a report like this: Potential Mastitis Conductivity Dev Report or a more restrictive list like this: Cows to Strip Report

Additionally, some parlors allow you to input reports like those into their software to turn on lights on the milking points so that a milker can be notified of suspect animals that should be stripped. 

Review Milking Shifts 

Reviewing how milking shifts perform compared to each other can help you use data to demonstrate efficacy of milking preparation.  Differences in two-minute milk letdown, durations, and detaches can all be indicators of sub-optimal milking preparations. 

Two minute Milk by Day

 Teat Scoring 

This is a great article written in Dairy Herd, in which Dr. Paul Virkler and Dr. Bill Koffman summarizing the need for Teat End scoring and how to use it: Teat End Talk. 

BoviSync offers many types of pen scoring these are good resources to learn more about pen scoring: