How´s your SHRINK? Are you throwing money away?
The Dairy market continues to be very dynamic with a focus on profitability. One way to improve profitability is to reduce waste by not placing more of the valuable components in your products than necessary. At Eurofins DQCI, we have analyzed several finished products from all over the United States for protein and butterfat content, and we have found that there is up to 11% variation among plants.
Whole milk, skim milk, and heavy cream products were purchased at supermarkets across the country, originating from a wide variety of processing and bottling plants, and subsequently analyzed for their butterfat and protein content by reference methods (Kjeldahl and Mojonnier, respectively). Although only butterfat content in those products is regulated by the Pasteurized Milk Ordinance (PMO), we have chosen to focus on protein as well, since this component often is more valuable ($ per pound) than butterfat.
Once the data was collected, we evaluated the difference between the minimum required levels of butterfat as dictated by the PMO as well as the difference between the product label claim and the true protein content of the products. This is usually referred to as a residual value (i.e., the residue between the optimal level and the observed value). The residual values were then plotted into a graph that easily shows the difference between the various products.
The PMO regulated minimum butterfat value for whole milk is 3.25%, heavy cream is 36%, and for skim (or fat free) milk the maximum value is 0.5%. It should be noted that these exact values cannot always be practically achieved in day-to-day production, and some variation should be considered a tolerance level. Another goal of this study was to establish an industry acceptance level of these variances; in other words, if we cannot achieve exactly 36% in heavy cream every day, what should we accept as adequate?
Although protein levels are not regulated by the PMO, we based the study on the product specifications; whole and skim milk being 3.33% and heavy cream at 0% (although the true protein values for heavy cream will always be higher than zero).
We have observed that most of the samples (8 through 25) are within 1% of the declared value for butterfat, whereas the first seven samples are much higher than expected levels, indicating that these productions might need to review their procedures.
Figure 1: A residual plot of heavy cream samples collected and analyzed for butterfat (blue) and protein (orange). The differences between the declared and true values are presented here.
Heavy cream is declared to have 0% protein per serving, although that is not practically achievable, there seems to be a consensus of 2% total protein in heavy cream, which does not seem to be an issue in this product type.
Figure 2: A residual plot of whole milk samples collected and analyzed for butterfat (blue) and protein (orange). The differences between the declared and true values are presented here.
Whole milk closely follows the same pattern as heavy cream; however, here we do not see two distinct groups, like we observed in the samples above. We do see that we once again have many processors who are very close (0.1%) to the desired levels, whereas others are far away from these levels (we have observed samples of more than 0.3% above desired levels).
Although there is no correlation between protein and butterfat levels, it is interesting that the worst performers in whole milk samples regarding butterfat are also the worst performers on the protein levels. This might indicate that the neglect on monitoring the levels is not only limited to one parameter, but both.
Figure 3: Total fat levels of skim milk samples (blue line). The maximum level of fat in skim milk (or fat free milk), illustrated as a dotted line, is 0.5% as regulated by the PMO.
Because the PMO does not dictate minimum levels, but rather maximum levels of butterfat for Skim milk, the goal for the production is to remove or skim as much of the butterfat as possible to be used for other products. Therefore, the residual plot in Figure 3 shows that some processors are more successful in skimming the milk than others. We have observed that most samples contain 0.1% butterfat or less but there are some samples who have much more butterfat than the best performers. Although none of the samples had more than the maximum levels of 0.5% (dotted line in Figure 3), this demonstrates that some processing facilities still have improvements to make.
The anonymity of the processing facilities does not allow us to establish an actual monetary loss. However, if we consider these a mid-sized production facility with an average daily intake of 1.8M lbs of milk, we can estimate the daily loss of $78,000 for protein and butterfat. This loss is only for a single production day; if these losses continue, the annual loss would be substantially higher. Furthermore, this does not account for the loss that might occur on the intake side, where the estimate might be equally off, either overpaying for the raw material or in the case of a cooperative, underpaying their members.
As a producer of certified reference material (calibration standards) for the dairy industry, Eurofins DQCI has observed very divergent behavior on instrument calibration procedures. Most production facilities rely on instruments (generally FTIR technology) to control and verify their levels in various products before bottling or otherwise making the product ready for the market. Although most processing facilities are maintaining the calibration of their instruments daily with proper maintenance and pilot sample monitoring, we have observed some who are only calibrating their instruments once every three or six months. The result of too infrequent or improper instrument calibration is that their instruments and levels might have been misaligned for all this time. In the example above a maladjustment for three months will have resulted in a staggering loss of $7,000,000!
Our recommendation, which most of the industry is following, is to calibrate the instruments once every 14 days and analyze pilot samples once every production day to monitor drift. This procedure will not totally prevent loss but will prevent fluctuation and reveal loss much sooner. The few thousand dollars spent on calibration standards and pilot samples on an annual basis might be the best investment in the whole dairy production.
The Author:
Dino Demirovic Holmquist
Vice President Sales, Eurofins DQCI
