Producing milk to fit the demands of today’s consumers should be a central focus of the nutritionist, producer and milk processor alike. Advances in milk processing, such as ultrafiltration, have allowed for unique product offerings that target the high-protein, reduced-sugar, lactose-free market.

Vanamburgh mike
Professor / Department of Animal Science / Cornell University

A well-known example of this type of product is produced by fairlife LLC, a joint venture of Coca-Cola and Select Milk Producers.

fairlife milkUltrafiltration is a process by which milk is passed through a series of filters to separate major components and then recombined in different ratios to produce a product that fits perceived consumer demand.

Table 1 shows the composition of fairlife ultrafiltered milk compared to conventional cows’ milk.

Milk fat protein and lactose composition of varous fluid milk types

So what would it take for a cow to produce a similar product without the filtration process? First, lactose is the major osmotic regulator in the cow’s mammary gland and as such, cows cannot produce lactose-free milk.


While different breeds of cattle produce varying percentages of fat and protein, the ratio of fat-to-protein is generally greater than 1. In fact, when cows produce milk with a fat-to-protein ratio less than 1, it is commonly a sign of a nutritional disorder and is not biologically optimal.

Genetic selection can change milk components over time; however, sole focus on a single trait can be detrimental to health and well-being.

So while it is not practically feasible to produce raw milk with the same composition of fairlife, innovative milk processors will allow the industry to remain flexible to consumer demand.

Producers should therefore focus on achieving high-component yield. This article will focus on formulating for milk components using field-usable nutrition models such as the Cornell Net Carbohydrate and Protein System (CNCPS). Particular emphasis will be placed on formulating for milk protein yield.

Formulating rations for milk components

The process of formulating for milk components begins long before the nutrition model or balancing software is opened. First and foremost, producers and nutritionists need to have effective protocols to collect relevant farm, herd, group and individual cow information in a consistent manner over time.

A key example can be seen with bodyweight (BW) and body condition score (BCS). While many nutritionists may score a selection of cattle while walking a pen, it is more effective to train key farm personnel to score cattle at times when they are handling individual cows.

Calving, pen movement, first breeding, foot trims and dry-off are all excellent opportunities to record BCS. Bodyweight of the mature herd should be determined on third and greater lactation cows between 60 and 180 days since calving.

Improperly characterized changes in BW and BCS are the most common reasons for disparity between model predicted and actual production levels.

Forages and feeds need to be analyzed to determine nutrient composition and availability of carbohydrate and protein. Forages should be analyzed for fiber digestibility, preferably using in vitro aNDFom disappearance at 30, 120 and 240 hours.

Feedstuffs contributing high amounts of rumen-undegradable protein may need to be evaluated for intestinal digestibility. Poor process control in the manufacturing of animal byproduct meals can turn a normally good source of metabolizable amino acids (AA) into a very expensive, indigestible product.

Once the characterization of the cattle and feeds is done, actual formulation can begin. When using models such as the CNCPS, users should generally follow a stepwise process:

  1. Gather as much information about dry matter intake as possible. Many intakes will be determined on a pen basis, so it is important to know the range of days in milk and energy-corrected milk yield of the pen when considering intake potential of the animals you are targeting.

  2. Satisfy the rumen microbial energy and nitrogen needs first. Seek to reduce excess rumen-available protein by providing fermentable carbohydrate while reducing inclusion of highly degradable protein sources, especially purchased feedstuffs.

    Another strategy is to evaluate urinary nitrogen (N) excretion relative to milk N excretion. The objective is to reduce urinary N excretion to the point that milk N: urinary N is greater than 1:1. This minimizes the feed N that is being lost to the environment and makes room in the ration for more fermentable carbohydrates and protein sources that supply a higher level of metabolizable protein (MP).

  3. Evaluate metabolizable energy (ME) adequacy. Remember, milk protein synthesis is an energy- driven event; supplying additional escape protein or rumen-protected AA when ME is limiting can lead to varied production responses and expensive feed bills.

  4. The choice to use fat supplementation should center on the energy needs of specific groups of cows according to their lactation cycle. Intakes and condition score should be monitored to avoid overconditioning or intake depression.

  5. Evaluate MP and individual AA adequacy. Current recommendations for individual AA as a percent of MP are model-specific. Dr. Thomas Overton provided optimal values for users of Nutrient Requirements of Dairy Cattle (2001) and CNCPS (v.6.5) in the February 7, 2016, issue of Progressive Dairyman.

    In the future, CNCPS recommendations will be given relative to energy. Current evaluations are indicating that diets should provide 1.1 to 1.2 grams of metabolizable methionine (Met) per Mcal of ME. Optimal lysine (Lys) supply can then be targeted using a Lys:Met ratio of 2.7 to 2.8:1.

  6. If supplementation of individual AA is necessary, as is often the case with Met or Lys, look for independent and unbiased sources of information regarding rumen protection and intestinal digestibility of rumen-protected AA products.

Choices on individual feedstuffs should center on cost, quality, availability and variability. Some byproduct feedstuffs may be inexpensive and readily available; however, their variation day to day may result in inconsistent rations.

Researchers at Ohio State have demonstrated the major areas of feed variation, and they have developed useful calculators to determine optimal sampling schedules based on many important factors.

In practice, the most effective way to control a variable feedstuff is to blend individual batches (if possible) or keep inclusion rates of any single highly variable feedstuff to less than 10 percent.

Finally, successful deployment of any ration formulated for milk components will be determined at the feedbunk. Researchers at Miner Institute have clearly demonstrated the effect of overcrowding on production and health.

Even the most well-balanced ration, if sorted or slug fed, will fail to promote high milk component yield. Milk fat depression and subacute rumen acidosis are two nutritional disorders that have a strong effect on milk components. These conditions are most often the result of a combination of poor ration formulation and inconsistent feeding management.

In summary, the innovative advancements in milk processing will continue to meet the growing demands for quality milk products. Optimizing the supply chain for milk components starts with biologically sound management of the cow and her feedstuffs, all while reducing resource use.

The continued development of field-usable nutrition models has allowed nutritionists to balance specifically for high milk component yield. However, the successful deployment of any ration will depend on consistent and attentive management at the farm level.  end mark

Samuel Fessenden is a Ph.D. candidate at Cornell University.

Michael Van Amburgh is a professor at Cornell University. Email Michael Van Amburgh.