From the NMC Newsletter "Udder Topics", October 2005

Dairy operations can be viewed as a system of processes directly or indirectly impacting one another and ultimately creating the quality of the end product, raw milk. This makes the quality of the milk produced on a farm a reflection of the performance of all the processes that constitute a dairy operation. Therefore monitoring that quality is critical for the herd manager and all employees to know how the herd is doing.

Dairy managers receive data on a daily basis regarding dairy productivity and quality. However it is often difficult to take full advantage of this information. Test results in the form of rows and columns of numbers are hard to read or interpret. Summaries in the form of monthly or yearly averages enable us to compare one month or year to another. They do not however help us distinguish between natural month to month variation and/or notice a trend that we should be concerned about. In addition, these summaries usually come too late and are therefore taken out of context making it hard for a herd manager to tie any shift in fat, protein or SCC to a specific day, event or change on the farm.

Daily monitoring of milk quality represented in a more "comprehensible" form than tables of numbers would help provide information about herd's past, present and future performance in time for managers and other dairy employees to make appropriate decision.

Statistical process control (SPC) is a tool that has been used for a long time in manufacturing industries to manage quality. SPC measures process behavior with control charts. It looks at the operation as a system of processes and allows us to know when real change has occurred. By examining a control chart we are able to answer one very important question: Are we improving, staying the same, or getting worse? The sole purpose of SPC techniques is to allow production managers to distinguish with statistical certainty the difference between "normal" (common cause) and "abnormal" (special cause) variation.

Recently these techniques have been adapted and applied to the dairy industry. Daily and weekly herd management data can be graphed into easily understood control charts that provide powerful day-to-day decisions aids. The steps to develop a control chart are as follows:
1) Start with a times series chart.
2) Add a centerline for central reference
3) Add control limits computed from the data
4) Apply tests to distinguish real change from random variation

If we look at a dairy operation as a system of processes we will find that there are numerous applications for SPC in dairy herd management; from monitoring feed intake, health or reproductive performance to milk quality monitoring. On most dairies, daily milk quality data are readily available, but the rows and columns of numbers can be overwhelming and make it difficult to draw any conclusions. Looking at a chart, however, makes it easier to identify where the process is heading. Examples might include: Did we have a significant increase in BTSCC in the last month? Did the new teat dip that we purchased improve milk quality on the dairy? Has a change in our bedding routine affected BTSCC? Control charts have been used for years in other industries and offer great potential for improved decision making in dairy management.

Source: “Are We Improving, Staying the Same, or Getting Worse” (www.dairyperformance.com) and the 2000 NMC Annual Meeting Proceedings, pg. 140 (Reneau).