Making Informed Decisions When Replacing Cows

I frequently get asked the following questions;

When should I stop trying to breed a cow?
Below what level of production should I replace a cow?
And should I sell this cow even though she is in-calf?

In essence the question that is being asked is how valuable a cow or lactation is compared to the value of a replacement. Obviously there are a plethora of factors that interact and should be taken into account when trying to determine the correct answer. However, it must also be recognised that there is no one rule that will apply to every cow as each ones age, level of production and fertility status is different. To date the most useful tool I have come across for estimating the value of both lactation and pregnancy is the CowVal and PregVal modules in DairyComp305 software which were developed by Connor Jameson, John Fetrow and Steve Eicker.

The “Cow Value” module in Dairy Comp 305 estimates the value of each cow in a dairy herd relative to an average fresh heifer. A positive Cow Value means the animal is worth more than an average heifer in that herd. A negative Cow Value suggests the animal is worth less than an average heifer. In addition to estimating the value of the cow, the module also calculates the current value of a pregnancy for each adult, to help managers decide if an open animal should be inseminated. There are two important rules the program uses to estimate cow values:

  1. A dairy running at capacity is the most profitable.
  2. Less profitable animals should be replaced with more profitable animals.

Every decision made on an animal in a commercial dairy is based on improving the herd’s profitability. You keep an animal because she is more profitable to keep than to replace. You breed animals because they will be more profitable if they become pregnant.

Cow Value estimates can help the herd manager assign a value to animals more consistently. When calculating the value of a cow DC305 realizes that a commercial dairy cow’s current value is the sum of her value for beef and the value of the milk she is likely to produce in the future. The value of an animal’s future production is based on;

  • The amount of milk she is likely to produce in the future
  • The price of milk in the future
  • The value of the money you must invest in the animal

The amount of milk an animal is likely to produce in the future is based on:

  • Age – The younger the animal, the longer she is likely to stay in the herd and produce milk.
  • Stage of lactation – an animal that is 30 days in milk is likely to produce more milk in the future than an animal that is 200 days in milk
  • Reproductive status – an animal that is pregnant is likely to produce more future milk, than an animal that is open. An animal that was just inseminated is more likely to be pregnant than an animal that has not been inseminated.
  • Production level – a high producing animal is likely to produce more milk than a low producing animal. 

The following tables are examples taking from the DC305 software which is why they are in dollars:
The future value of a cow is of course, unknown.

Table 1

The module is based on models that estimate the future profitability of a cow based on the parameters listed above. But certainly, pregnant cows are much more likely to remain in the herd, and higher producing cows are more likely to be more profitable next lactation too. Thus, these predictions should be used as guidelines. They are not meant to replace sound judgment, but to augment it.

Table 2

The value of a cow is always relative to that of a replacement heifer. Thus, a cow with a negative value is a potential cull. A cow with a value of £150 that is diagnosed with a displaced abomasum may be more profitable culled than treated. 
The estimated value of a pregnancy can assist a dairy in deciding whether it is worth the effort to breed a cow. Likewise, for a pregnant cow, the pregnancy value can help estimate the cost of an abortion. Open cows with negative pregnancy values should not be bred, as spending money on
pregnancy will lower their value!
Perhaps the most thought-provoking concept arises when an open cow has a negative cow value, and also has a negative pregnancy value. This means she is worth less pregnant than if she remains open. However, the software algorithm assumes that cows that are not coded DNB (do not breed cows) are still trying to get pregnant, and that a percentage of the time they will. Thus, this cow will have a lower cow value while she is still eligible to be breed. Her value should increase once she is flagged as a DNB. This makes sense – it is sometimes a profitable decision to flag a cow as a DNB cow. Note that a cow flagged as DNB may still have a positive cow value, until her milk production decreases below that cull/cut off value.
The flip side is also of interest. Any DNB cow that has a positive pregnancy value has hopefully been flagged DNB because of some reason other than current milk production. This pregnancy value may be a crude estimate of the cost of culling her.
By necessity, we are modelling the future to predict the future production of each cow. We make lots of assumptions, such as eventually, all cows leave the dairy, and when they do, a replacement enters the herd. Crucial to the model is that a dairy farm will operate to maximize profitability. Again, DNB cows demonstrate some of the fundamental concepts.
Cows flagged as do not breed cows (DNB) or cows which fail to conceive are probably a bit more straight forward and are the cows which dairyman usually want to know at what daily milk yield should I sell this cow.
A cow should be sold once her milk production generates less profit than the daily replacement cost. If we take a rough estimate that our replacement cost for every cow in the herd is around 50p a day and our profit per litre is 5p she should be sold once her yield drops below 25 litres. This is because after all costs a heifer will generate around 80-90p profit per day based on a milk price of 28p and a gross profit of 4p. 
In order to calculate the cut off yield per day all that is required is the average cull value, replacement rate, cost of a replacement, milk price, milk yield and profit per litre being generated, and is something that can easily be done for cows in your herd by one of the Wynnstay Technical Team.
The power of these prediction models are only as good as the data they use to generate the values and must be interpreted with caution. Take the example below from one of my clients herd.
Table 3
Take table 3 for example, cows 356, 162, and 157, all three are heifers and all have negative CWVAL values. This is mainly due to the fact that they have not been bred (NO BRED) are empty (0 DCC) and are well into their lactation (DIM). If they were in-calf however they would have positive PGVAL and would probably not be sold. With these three cows it would be important to ensure that they are actually empty and have been scanned empty or are indeed truly DNB cows. Missing information or inaccurate recording has a profound effect on the information generated by the model. Wynnstay can offer a CowVal assessment of a customer’s herd that milk record and record reproductive events on their monthly milk records.
Written by Huw McConochie – Head of Dairy Technical Services
Follow on @HuwMcConochie
For more information contact

Leave a Reply