Automatic Milking Systems: Key Performance Indicators

Important user information

This is a dynamic and evolving document that will be updated regularly as we increase our understanding of the incorporation of Automatic Milking Systems (AMS) in Australian dairy farms. Please check online for regular updates.

The whole farming system should be taken into consideration recognising that farm management practices have an impact on several key performance indicators (KPI). It is also true that circumstances of every aspect of the operation influence the management decisions made on farm which then influence the key performance indicators of the operations.

The purpose of this document is to highlight some indicators that can be used by farmers who have AMS to look at the performance of different aspects of their system as a 'snap-shot' at a given point in time.

The context of the whole operation must be considered, as the data can only show the 'electronic events' and does not give any indication of what the people power are doing behind the scenes.

What information do AMS need to monitor?

A farmer reviews the AMS software

Farmers that invest in AMS are largely freed from the task of routine milking. However, there are still a lot of jobs that need to be done on the farm.

In addition to spending time on other on-farm tasks such as feeding cows and taking care of breeding and animal health of their herd, they still need to devote some time each day to spend in front of the computer to monitor reports and alerts that are generated by the software.

This is an integral aspect of the routines that cannot be dismissed as it will ensure that appropriate management decisions are made in a timely manner through monitoring the performance of the individual cows and the broader system.

AMS farmers still monitor the traditional KPI such as days in milk, average daily milk production or pregnancy rate. However, they also have some KPI that are specific to this type of operation.

The purpose of this document is to introduce some of those KPI specific to AMS. They should not be analysed in isolation as they are generally interrelated.

KPI for AMS Farms

Table 1. Some common KPI that are used by AMS farmers.



Common target

Impacted by

Milking frequency

Number of milking events/cow in any 24 h period

>2 milkings/cow/d

(higher in early lactation, lower in late lactation)

Cow factors: parity, days in milk, milk production level, health status, animal behaviour, milking speed, previous experience and training, cup attachment success, udder conformation

System factors: herd size, cow traffic, fetching times, waiting times, number and size of feed allocations, feeding level, gate times, walking distance, condition of laneways and weather conditions

Equipment factors: throughput rate of equipment, accuracy of cup attachment, downtime for alarms, servicing and maintenance and breakdowns

Milking interval

Interval between consecutive milking events; measured in hours since the previous milking

<12 hours (>60%)

<16 hours (>85%)

<20 hours (>95%)

Cow factors: parity, days in milk, milk production level, cow health, reproductive status – on heat, previous experience and training

System factors: herd size, cow traffic, fetching times, waiting times, number of feed allocations, proportion of feed in each allocation, gate times, walking distance, condition of laneways and weather conditions

Box duration time (min) (cow in – cow out)

Time spent per a cow (in a single or multi-box robot) from the time she walks in till she walks out; measured in minutes

6 – 8 minutes/cow/visit

Cow factors: behaviour – temperament, training, production level, milking frequency, milk flow, udder shape and teat positions, teat visibility related to hair, mud or tail

Equipment factors: camera and lens cleanliness, vacuum and pulsation levels and feed left in bin after milking

Waiting time in the pre milking holding yard

Time elapsed from entry to the premilking waiting yard to milking start; measured in hours

<2 hours/milking

Herd average <1 hour

Cow factors: previous experience, training and motivation level, fetched or voluntary, forced waiting times including down time for system washes

System factors: cow traffic, large groups of cows coming together and weather conditions, timing of arrival in relation to system washes, feed availability during milking and/or in post milking area

Equipment factors: throughput rate of equipment, equipment downtime

Milkings per hour and distribution of milkings

Number of milking that each

AMS unit performed per hour; measured in milkings/h across the day

6 – 8 milkings/h (box robot) and 70 - 80 milkings/h (robotic rotary )

Flat distribution of milkings

System factors: cow traffic, fetching times, waiting times, feed allocation management, type of feeding system and weather conditions

Equipment factors: cleaning cycles, number of washes and rinses per day and times the washes take place

Daily milkings

Number of milkings that eachAMS unit performed per day

150 – 170 milkings/d (Box robot) and 1200 – 1600 milkings/d (Rotary robot)

Cow factors: cow traffic, milk yields/cow/day and milk yields/milking, proportion of incomplete milkings

System factors: herd size and milking frequency, management strategies around incomplete milkings

Equipment factors: robot performance, cleaning cycles, number of washes/rinses per day and times the washes take place

Incomplete milkings

Number of milking events that are flagged as incomplete due to missed attachments or premature milk cupremoval

<10% daily milkings

Cow factors: milking interval, behaviour, temperament, training, udder shape, teat positions and teat visibility related to hair, mud or tail, previous experiences

Equipment factors: camera and lens cleanliness, vacuum and pulsation levels and take off levels

Robot utilisation

Proportion of a 24 h period that the robot is performing milking related tasks

> 85% daily time conducting milking/cleaning and/or <10% free time

Systems factors: herd size, cow traffic and milking frequency

Equipment factors: activities performed on cows when in the robot such as cleaning or treatments, cleaning cycles and number of washes per day

Stop alarms – Call outs

Number of times the systems stops completely and calls the operator to attend the dairy to conduct a physical task or visual inspection

< 2 stop alarms per week

Cow factors: previous experience and training

Equipment factors: functioning, cleanliness, service and maintenance

Data summary

Due to variability in AMS, the average values of several days are commonly used for variables such as milking frequency.

Whilst the above mentioned is a list of KPI that are commonly monitored on a regular basis on most robotic farms, it is not an exhaustive list and many farmers will have additional parameters that are of interest to them and which might be monitored on a regular basis.

Many exception reports indicate cows or equipment that breach thresholds set within the equipment support software. It is these reports that allow the herdsperson to identify cows that require assistance or inspection, for example; cows that require an udder health check, cows that are not consuming an acceptable level of their feed allocation, any cows that have dropped milk production dramatically.

It is a common observation that some of the most successful AMS operators use the time freed up from milking to monitor cow and system performance indicators and shift their focus to improving an aspect of their operation that has previously been of lower priority due to time constraints.


Dr. Nicolas Lyons
Development Officer Robotic Milking Systems
Mobile: 0401 650 073


Prepared by Dr. Nicolas Lyons together with Assoc. Prof. Kendra Kerrisk (Project Leader FutureDairy, The University of Sydney). The authors would also like to recognise the valuable input from commercial companies and farmers.