How to make the most of technical data from robots

robotic milking machine

© Tim Scrivener

As well as easing labour pressures and increasing milk yields, robotic milking systems have the potential to generate a huge amount of data that can be used to inform key management decisions.

Below, Nigel Hardie – consultant at The Dairy Group – shares three management areas that can be improved by making more of technical data collected by robots.

See also: WIYLS visits the only off-grid robots in the UK 

Health and welfare

When it comes to indicators of health and welfare issues, there are hundreds of things farmers can look at in terms of robot data.

While this can be quite overwhelming, it is important to remember that your robot – or robots – are effectively your eyes for the herd, so you should monitor key areas in the same way you would if you were physically handling cows twice a day.

Timing: One of the key things to look at is any anomalies in the times individual cows visit the robot. Animals tend to stick to the same sort of routine every day, so if she is late, for example, this could be a sign of issues with feeding, lameness or even mastitis.

Yield: Once she is at the robot, you can look for any changes in her milk yield as this could also be a sign of illness.

Weight: Some robotic systems have weigh platforms, and these can be incredibly useful tools to see how well a cow is performing health-wise.

Changes to weight – both gain and loss – can be indicative of health issues, but could also be a sign of diet or feeding imbalances.


Milk yields are the obvious thing to monitor when it comes to productivity, but other information can be beneficial to base decisions upon.

To get the best out of robots, it is important that milking times are as quick and efficient as possible. The most important thing is to establish the physical milking time of each cow – basically, how much milk is produced and over what timescale.

For example, if cow A takes four minutes a milking, giving 10-11kg each time, and milks three times a day, she is on the robot for a total of 12 minutes a day.

However, if cow B averages 18 minutes a milking, but produces the same amount of milk, then she is spending about an hour a day at the robot, without any commercial benefit.

From a production point of view, this data would suggest that it could be more efficient to take this cow out of the herd and replace her with better performing cows, so farmers can use this kind of data as the basis of their herd management decisions.

For productivity, keep targets and decisions in line with your contract requirements and remove cows from the herd that continually fall outside these parameters.


With farmers missing the opportunity to manually handle cows on a routine basis, activity levels are the best way of monitoring herd fertility.

Both high activity levels and low rumination can be a very strong indicator of heat if you are using combined activity and rumination collars.

Look for a diamond-shape on data graphs – a peak of activity paired with a V-shaped drop in rumination often means a cow is on heat.

As the data generates trends over time, this can also be used to make predictions about cow behaviour, and from a fertility perspective, means you can set an alert two days prior to an expected heat cycle to remind you to look out for signs that the cow needs serving.

Having access to this information also means it is really easy to identify cows who are not cycling and may need to be checked by the vet.

What’s more, the fact that the data is available on an app means precise detail can be accessed instantly, meaning far more accurate decisions can be made.

As well as this, farmers can use their own farm information to benchmark their herd and the data sets make it very easy to see exactly where you are, as well as helping to monitor progress to reach targets – whether those are for fertility, productivity or animal health.

Case study: Michael and Tony Ball, Coton Wood farm

Tony and Michael Ball

Tony, left, and Michael Ball © AHDB

At Coton Wood farm near Ashbourne, Derbyshire, a desire to expand their herd of Holstein Friesian cows led brothers Michael and Tony Ball to invest in eight Lely robots in 2015.

“Originally we looked at putting a rotary parlour in, but having looked at some robots at a dairy event, we decided this was a better option for us,” says Michael Ball.

One of the main drivers behind the investment was to save on labour, he explains. “With stepping up cow numbers there was obviously going to be an increasing demand on labour, which can be difficult to find, and also to get extra milkings out of the cows without extra cost and make the labour more flexible.

Farm facts

Coton Wood farm, Ashbourne, Derbyshire

  • 500 Holstein Friesian cows
  • Averaging 9,500 litres a cow a year, 4% fat, 3.35% protein
  • Milk sold to Freshways
  • All-year-round, keeping own replacements
  • Milked on eight Lely robotic milkers
  • Low-yielders grazed April-October; high-yielders fed a mix of cut and carry, and total mixed ration

As well as easing the man-hour pressures, the automated system has also helped Coton Wood manage the herd better through the data reports collated by the robots and being part of AHDB’s Strategic Farm network has helped to utilise this information better, says Michael.

“On a day-to-day basis we use the reports to monitor cow activity – so we know how many times each cow has visited the robot – as well as fertility lists, rumination and sick cow reports. These are standard lists and are easy to use and tweak to your individual farm needs.

Michael Ball with robotic milking machine


“From a longer-term management perspective, we have also been making use of weighing floors in the robots. We use this measure of cow weight as a guide to how she is progressing through her lactation – looking particularly at weight gain or weight loss to keep an eye on her body condition score prior to calving.”

This was something the farm was not doing on its previous system, but has found it to be incredibly beneficial since switching to robots, adds Michael. “In my view, robots are the way forward. They are an excellent choice from both a cow and a management perspective.”