Editor’s note: This article is the first of a two-part series entitled “Marginal Thinking.”

In the end, food producers are the primary influence on food prices. Quotas and trade regulation may shape the context within which supply and demand operates, but global markets and local conditions still inevitably respond to supply and demand. Those who can produce food at the highest profit will continue to expand, and those who do not maintain adequate profitability will decline in their contribution to total production. As production by the profitable enterprises expands and more food is available, the price will eventually fall. The most profitable producers will be willing to gain market share for a slightly lower price.

As one sector of food production, the milk production business is not substantially different. Across the United States, successful dairies have expanded, and other dairies are disappearing. Since 1965, the number of dairy farms in the United States has fallen by nearly 90 percent, from almost 1.2 million to 120,000.

In other industries, successful businesses have been classified into one of three categories:

•product supremacy
•customer intimacy
•operational excellence


A dairy sells a largely homogeneous product to distant (anonymous) customers. With some exceptions (e.g., organic dairies, on-farm cheese production to local customers), most dairies will not compete via the first two strategies. A successful dairy most likely will be categorized as operationally excellent. Operational-excellence businesses tend to do most of the following activities well:

•early, wise adoption of technology
•economies of scale
•cost control
•efficient use of resources
•good decisions and problem solving

Technology adoption has been categorized into phases:

•early adopters
•mass adopters
•late or nonadopters

The most money is made by the early adopters of an effective, profitable, new technology; they gain advantage before the market adjusts. Eventually, a sufficient number of businesses adopt the technology, and the milk price decreases as discussed above. After that point, only the technology supplier makes a profit.

The impact of economies of scale is easily observed in the U.S. dairy industry. In 1965, the average dairy herd size was approximately 15 cows. In 2000, it was approximately 70. Very few dairies are built today in the United States with fewer than 600 cows; most are more than 1,000. Economies of scale and specialization of function is driving this transformation and the change has been ongoing over the past five decades in a nearly linear fashion and has been accompanied by an equally dramatic and linear increase in milk production per cow.

As dairies have become larger and more specialized, they have also afforded the manager more time to shift his time and efforts from daily labor to business management. Coupled with increased leverage with suppliers due to increased scale, dairy managers have been able to capture significant cost savings and improve profitability.

Increased scale has also made it possible to spread overhead costs (facility investment, especially parlors, tractors and other large equipment, consultants, manure management, etc.) over more cows and increased the efficiency of use of these fixed assets and costs of operation.

Good business and problem-solving decisions fall into two broad categories:

•Decisions made to increase profitability directly
•Decisions made to reduce risk (and in the longer term indirectly improve profit)

The remainder of this [article] will focus on how a dairy becomes more profitable.

Making money on a dairy farm
How does a dairy make more money? Simplistically, this is achieved by either lowering costs or increasing income, or both. In spite of nearly every trade magazine headline, keeping costs low is not the goal – higher profit is. In the dairy industry, it is almost always much more profitable to focus on the income side rather than to try to decrease expenses. Of course, cost control is important. However, dairies can dilute their fixed costs by more milk per cow or more cows. Not planting crops will lower costs, as will not feeding cows. Fishing without bait is lower cost fishing.

Since individual dairy farms have little control over milk prices (some opportunity exists for improving milk components and milk quality), increasing income almost always means increased milk sales. In nearly every dairy meeting, someone shows a scatter graph of profit per cow versus herd average. It appears like a shotgun blast. There is no obvious correlation between herd average and profit.

Unfortunately, few people remember that correlation does not mean cause and effect; even fewer realize that lack of correlation does not guarantee lack of cause and effect. While it is clear high milk production does not guarantee profitability, to state that milk production does not matter is foolish. On an individual dairy, increased milk production almost always results in increased profits.

To evaluate the financial impact of changes, one must be able to estimate the increased costs and the increased returns. This requires differentiating marginal costs from average costs, which implies a quick review of fixed versus variable costs.

Fixed costs do not change if production changes. Examples of fixed costs include insurance, land taxes, labor, trucks, etc. Variable costs typically include feed and manure management. (In a quota system, once a dairy is producing milk above their quota, quota purchase becomes a marginal cost or the reduced price of milk for over-quota production must be considered.)

Marginal feed costs
It is currently in vogue to insist a dairy know its average daily feed cost per cow. Typically, this is done by dividing the feed bill by the number of cows. This can also be calculated per hundred pounds of milk (cwt): divide the feed bill by the number of cwt. Unfortunately, calculating the average feed cost per cwt is almost useless. The major determinant of average feed cost per cwt is milk production per cow, and that number is more readily available by looking at how much milk the dairy ships each day.

In order to make decisions, what is actually needed is the marginal feed cost – the cost for the feed it takes for an existing cow to make more milk. This number is necessary to evaluate the financial impact of potential changes, such as semen purchases, ration changes, etc. Average feed cost includes maintenance, which approaches 50 percent of the feed on many dairies.

Marginal feed cost in Canada is typically near $7.50 per additional 100 kilograms (220.5 pounds) of milk sold. Average feed costs per 100 kilograms (220.5 pounds) range from $12 to $15, and they are greatly affected by milk production, which dilutes maintenance costs.

Why is this important? Because at $7 marginal feed and $28 milk, there is a four-to-one return on increased milk production. Using the assumptions in the example ($0.66 per pound milk price; $70 per ton for the total mixed ration (TMR), etc.), a herd with 100 cows milking can increase their profit by more than $10,000 per year if dry matter intake (DMI) increases by only 0.5 kilograms (1.1 pound) per cow per day! The marginal profit per 100 kilograms (220.5 pounds) is more than $20.

Note this profit per 100 kilograms for marginal milk is the same regardless of whether the dairy is currently already profitable or is losing money. The marginal economic impact of improved production does not depend on the farm’s current average status.

Dairy cows live and die by the following rule: They will not make more milk unless they are fed more. There is a direct relationship between DMI and milk production. Examples are everywhere:

•Heat stress decreases DMI and milk production drops
•High-producing cows eat more than low-producing cows

Decisions on a dairy farm involve making a change. A change incurs added expenses and, hopefully, added income. Because added income usually results from increased milk production, it is important to account for the increased feed costs associated with the increased milk production per cow.

Additional milk production from an existing cow costs only extra feed for milk production – there is no change in maintenance feed costs. Thus, using average feed cost per cwt will overestimate these additional costs because average feed costs include the cost of maintenance. The marginal feed cost must be used. In certain cases, using the wrong estimate will prevent dairies from making profitable changes.

Much of the data or information available to the practicing veterinarian is in the form of averages. Dairies are replete with averages: rolling herd average milk, average days-open, average milk per cow per day, average feed cost per hundredweight of milk, average age at first calving, etc. Averages are valuable and useful “first-cut” parameters of a dairy’s performance. Dairy veterinarians and their clients use them daily as a shorthand summary of the dairy’s biological (milk per cow per day) and economic (net farm income from operations per cow) performance. Averages can be a useful measure of a farm’s status, but it is only one measure. When making specific management decisions, averages can be misleading sources of information.

Lag, momentum, bias and variation
First, the averages themselves may not accurately reflect the farm’s real status. Averages are vulnerable to several types of error. Averages may suffer from significant lag. The measured effect of a change in management status reflected in a calculated average may lag far behind the actual physical change on the dairy. For example, a failure of the breeding program in replacement heifers may not be reflected for more than a year in the average age at first calving.

Many of the averages used on dairies contain a great deal of momentum built into the manner in which they are calculated. Rolling herd average milk production, for example, includes the production of cows over the past 12 months in its calculation. Extreme changes in current production are required to change the course of rolling herd average in a month’s time.

Many of the averages used by dairies are also subject to bias. Services per conception is biased toward a favorable view of the reproductive status of the dairy, since only those cows that conceive enter into the calculation, ignoring the undesirable outcomes in open, repeat breeding cows. Culling and the deliberate exclusion of “do not breed” cows from many reproductive calculations can play havoc with the accurate expression of a dairy’s status.

Beyond these potential sources of error, averages also are deficient in characterizing a dairy because they, of necessity, express only the central point on a distribution. Averages do not describe or display the degree of spread, or variation, in the population. The degree of variation is a key element in evaluating a dairy farm’s performance, and identifying cows at the extremes makes it possible to affect their future.

For example, average days open of 120 days may not be serious if all cows are open between 80 and 150 days. If the spread is from 45 to 250 days, the problem is likely more severe and costly. Even if a herd has a days open of 110 days, the individual cow left unbred for 180 days deserves identification and intervention.

Averages can be particularly dangerous when used in making economic decisions. Economic decisions based on averages can be seductively appealing; at first glance, they can seem like common sense. In fact, dairies often make decisions based on such common sense. Often, these common sense decisions are wrong and very costly. The remainder of this discussion will consider some examples of this faulty thinking.

Before leaving the issue of averages, however, it seems prudent to step back and acknowledge that all averages are not useless. They can and should be used in evaluating a dairy, provided the person doing so understands their shortcomings. In economic terms, the average performance of the dairy is critical. The bank will get paid from the average net income per cow. The dairy family will buy food and clothing from the average profits. Equity in the farm will be built on the average retained earnings.

The point being made is not that averages can be ignored, but rather that one should think of averages as the goal to be attained, not the basis for managerial decision making. Management must act based on the specific case; averages happen as the sum of the experience of a series of specific cases.

Keep in mind that on all dairies the average cow is healthy, pregnant, milking an average amount of milk and happily bored. Today’s management cannot be based on her status. As a crude example, if on average cows do not get milk fever on a farm, it does not mean the farm should not keep calcium on hand. PD

References are available upon request.

—Excerpts from 2006 Western Canadian Dairy Seminar Proceedings