Milking as a Service

One of the companies that I’d love to join is DeLaval in Tumba. Not just because of the short commute (I still miss my daily bike commute from when I was living in Copenhagen), but also because of the challenge they are addressing: Optimizing dairy farming using IT.

Humans have kept cows for 5,000 years and dairy farming has been optimized quite a lot already.

DeLaval is one of a handful companies who produce and sell milking robots — or voluntary milking systems. I’ve seen these in action at different sites in Denmark — the milking process is fully automated, the robot places the suction cups and milks the cow while the cow chews away on some power grain. Key figures like milk temperature, time since last milking, and quantum milked per udder are shown on a nearby display.

Kosläpp in Skåne 2016. Swedish national television broadcasts live from the yearly kosläpp events where cows are let loose on the spring green pastures after a winter in the stable. Copyright Frederik Jensen.

Milking robots have been around long enough for the technology to mature. When you have 500 cows not being milked while your system is down, reliability and availability are not just about money but also animal welfare. Once the system is in production, it needs to be in operation 24/7. Cows don’t go on weekend or take time off for Christmas.

Dairy stables are hostile environments for computers, they need to be protected from moisture, heat, and being stepped on by animals weighing nearly a ton.

Robots have moving parts that needs maintenance. Hence, one challenge is to service the robots. So far this has involved a service technician visiting the farm and measuring the machines with handheld devices to help decide which parts to replace. As farms often are located in remote areas, having the right spare part at the right place and time is a problem to be optimized. Replacing a component too soon because you are not sure it will last until the next scheduled maintenance visit costs money. Having a wide selection of spare parts in the service car just in case they may be needed costs money. Scheduling an extra service trip to replace a broken part costs money – and a lot of money if it has to be done express due to a production stop.

Enter the cloud: What if diagnostic data from the milking robots is collected automatically 24/7 and uploaded into a database in the cloud? Service technicians and the farmer can then monitor and analyse the data remotely to plan service visits and be sure to have exactly the spare parts needed.

This is what DeLaval is working on right now. In itself a desirable goal. But the implications down the line can be huge as it allows changing the business model:

What if a farmer does not invest in a milking robot but buys milking as a service? The operational risk will be on the milking service provider, not the farmer. Financing dairy production will move from CAPEX (farmer investing in machines financed by a bank loan) to OPEX (farmer paying for a service). The farmer needs to know less about operating milking robots and can speciliase in other areas like breeding cows, optimizing fodder, stable facilities, product development or doing marketing events like kosläpp.

Just like the Cloud has lowered the threshold for new companies to put a new product or service on the market, the Cloud can disrupt the business models around dairy farming. The bank may become less of a gatekeeper for young farmers to enter the business. Independent service technicians will have fewer opportunities to improvise fixes to mechanical problems and upsell while they are visiting.

Times are changing and it would be cool to help create optimal outcomes for the involved stakeholders.

The cows will probably not notice though.


I have actually worked with dairy farming previously: Back in my university days, I had a part time job to optimize scientific calculations to run on super computers. One of the programs did breeding planning for dairy farms.

Advertisements