Milking cows is a complex and time-consuming process that requires significant effort, attention to detail, and plenty of consistency. As technology advances, more and more farmers are turning to robotic milking systems to streamline the milking process and improve efficiency. However, not all robotic milking systems are created equal with some systems being more efficient than others.

One factor that can impact efficiency is the genetic traits of the cows being milked, but how important are Robotic traits in reality?

In the UK currently, 10% of all cows are milked robotically! Lely alone will have an installation base of just over 3,000 robots by the end of 2023, and our growth model based on orderbook and historic numbers, predicts our installation number to have doubled by the end of 2028.

The demand for automation within the dairy sector is clearly growing at a rapid rate, and the thing to bear in mind is that a lot of the traits to keep an eye on for future robots, are also of benefit in your current conventional milking system.

Traits influencing Efficiency

While many farmers are focused on breeding cows for traits like milk production, health improvements and disease resistance, there are other genetic traits that can have a significant impact on robotic milking efficiency. These traits are often overlooked or forgotten, but they can make a big difference in the performance of robotic milking systems.

Whilst the first two that probably always spring to mind are teat length and teat placement, there are several others that play their part, such as milking speed, udder depth, stature, mastitis resistance, lameness and temperament.

Currently, of all of those, the only two traits directly included in the make-up of PLI in a positive way are mastitis and lameness. Our suggestion is to carry on with your usual breeding criteria, but when you have a shortlist of sires that fit, then perhaps use some of the robot traits to differentiate between them.

Teat Length

The physical length of the teat wants to be ideally over 3 cm in length, but isn’t the whole story, as you can have much longer teats that have no width, meaning cluster slip can be just as bad. The robot fairs better with shorter teats than a lot of conventional milking systems as there is less weight involved, due to having no claw to support.

The heritability of teat length is 29% (0.29), and so is worth putting some positive emphasis on where possible.

Teat placement

Teat placement is an important factor to consider when breeding cows for robotic milking systems. Cows with teats that are evenly spaced and hang straight down are easier for robots to milk. This is because the robot can more easily identify the teat and attach the milking cups without the need for human intervention.

The trait rear teat placement is nearly always the initial consideration, but the outcome is also hugely influenced by udder support and the angle of the teat. Other factors to look at are how the rear and fore teats line up, eg, if you have very wide rear teat placement and very close fore teat placement then it can become difficult for the robot to differentiate.

Again, the lack of claw and therefore pipe length restrictions, means that the robot can often handle uneven udders a little easier with cleaner milk out, due to the fact that the teat cups are always hanging vertically.

Rear teat placement and side teat placement both have a heritability of 29% (0.29).

Udder depth

The traditional view has always been to breed heifers with as shallow udders as possible. Shallow udders are not an issue for robots, but extremely shallow udders in combination with other traits at extreme levels (such as body depth and stature) can make attachment more difficult.

Other traits that play their part to influence overall udder depth are fore udder attachment and udder support.

Udder depth is a very heritable trait for the type traits at 35% (0.35) and so if you breed for extremes, you will see the results very rapidly in genetic terms.

Milk Speed

For many years the thought process has been that quick milking cows are very often the high cell count members of the herd. Very little data exists to prove or disprove the theory in conventionally milked herds, as very often the way milk speed is recorded leaves a lot of inaccuracy. The scale of 1 – 9 identifies the fastest and slowest milkers, but 2 – 8 often just get recorded as average, because we only notice the cows that come off early or are the last in the line to finish. The main reason for the link to high cells was that the cow had open teat ends and would leak milk in the bed and collecting areas.

In robotic systems we do not see a link between milk speed and cell counts, due to the cow’s ability to go and milk when she wishes, meaning less is leaked in the bed, and there is no collecting area to wait in when in a free access system.

Milk speed is measured in kg’s per minute on a robot system, with average being around 2.6, but several herds are achieving nearly 4.0 when the management and breeding in combination together have been focused on.

The conversation often goes along the lines of, lets make sure we don’t use any slow milkers, the issue being that we are not putting enough positive emphasis on faster milking genetics. Milk speed is a huge driver of efficiency in robot milked systems.

Milk speed has a very high heritability in comparison to other management traits at 21% (0.21), whereas something like fertility index which is focused on regularly would be a much lower heritability in comparison at 4% (0.04).

Other influential traits

Stature

Extreme stature in combination with extreme shallow udder depth can mean that robots could have difficulty seeing where the udder actually is. There is no suggestion that poor udder support sires should be used, just perhaps steer away from the extremes in combination.

Mastitis resistance

It doesn’t matter whether milking on a conventional system or robotic system, cases of mastitis cause a great deal of financial loss and wasted labour hours dealing with the cases.

Lameness

If genomic testing, then emphasis should be put on the lameness trait, as cows with any disease in the foot will ultimately visit the robot less because its simply too painful for them to want to go. These same cows go unnoticed in conventional systems as the whole group is taken to the collecting area for every milking.

Temperament

Temperament is another important genetic trait to consider when breeding cows for robotic milking systems. As a breed, the Holstein has become quieter and quieter over the years. Combine that with the fact that robot milked cows become so easy to handle anyway, and those outliers really do stand out. Don’t think of temperament as an issue for the robot to milk the cow, as it is a robust machine that can handle it. The area that often people don’t think about, is when you are in the pens working with the cows, that’s where those flighty temperament cows really show up, and can risk upsetting the whole group.

The take home messages

Use the data…

Every minute of every day there will be live data streaming back from your robots, giving you a vital insight into your herd and its performance, on an individual and whole group basis. We certainly don’t recommend death by data, but if you don’t measure something, then how are you going to manage it?

Make use of all the health traits if you are genomic testing or using genomic bulls, lameness and mastitis will have an even greater impact on your business than they did before.

No drastic measures…

Do not throw the baby out with the bath water! Your genetic decisions have got you to a good place so far, and so keep with your existing balanced approach, but incorporate these traits where possible. Breeding has always been a steady process, but one that reaps great rewards in the long term, breeding for robot efficiency will be no different. Keep looking at the profit drivers, then if you need to decide between two very similar sires, see which have the better robot traits.

Remember – less than 1% of cows in new robot start ups have to be changed out of the herd. You haven’t been breeding for robot traits for the last 30 years, and so there is no need to have a huge panic about them now. However, a little direction and focus on them going forward will make a huge difference to output in the years to follow. All of those small increments every generation will add up in the long run.

Photography Credit @Jenny Wood Photography