Big Dairy Enters the Era of Big Data
After Steve and Lori Dockendorf’s two oldest children left their dairy farm to go to college, the husband-and-wife owners of a 100-cow farm in Watkins, Minn., had to figure out how to replace the labor they’d lost. The traditional solution would have been to hire a couple of extra hands. Instead, the Dockendorfs went with robots: robots to help feed the cows, robots to help clean the barn, even robots that can milk the cows. The farmers used to spend three hours milking, twice a day. “Now,” says Lori, “we wake up in the morning, and the robots have already milked all the cows.”
There’s a notion that life on the farm is simple, but running a dairy farm is wickedly complicated. Unlike with other businesses that make things, the machines that produce milk are, well, cows. Cows need to be fed. They need to sleep. They gain and lose weight. The number of variables a dairy farmer has to stay on top of is staggering.
Fortunately, technology has made deep inroads into the world of dairy farming, helping manage day-to-day chores and collecting and analyzing the large batches of data a herd of cattle generates daily. In early October at the World Dairy Expo in Madison, Wis., the U.S. industry’s largest trade show, some of the latest technologies were on display.
A Canadian company, Dairy Quality, unveiled a new product called Milk Guardian, a small black box that slides onto the back of an iPhone. A farmer inserts a plastic slide containing a milk sample from one of his cows, and the device counts the number of somatic cells (a high somatic cell count can be an indicator of mastitis, an infection of the udder tissue). Counting somatic cells used to require sending milk to an offsite lab and waiting a week or more for results; using a microscope and an app, Milk Guardian can analyze a sample on location in six seconds or less. The accessory and app together cost $1,800.
Steve Mangan, technical director for Dairy Quality, had originally thought of a larger, more expensive solution. He figured farmers could circumvent shipping a sample to a lab by outfitting barns with a larger, stationary device and a network connection to a centralized processing center. After connecting with Gary Jonas, co-founder and president of app developer Mpengo, Mangan realized that most of what he needed could be found in a smartphone. “We needed a camera; a smartphone has that,” Mangan says. “We needed software; we could do that through an app. We needed to export the data; a smartphone’s designed to do just that.”
One of the best-attended booths at the Expo was that of Dutch dairy-equipment manufacturer Lely, where the highlight was the company’s Astronaut A4 robotic milker, an updated version of the system the Dockendorfs use on their farm. A key virtue of having a robotic milker in your barn—provided you have about $200,000 to buy one—is that you don’t have to be there for cows to get milked. The A4 has a small gated area with a feed trough at one end that the machine refills between milking sessions. A cow enters the area on its own, knowing that it’ll find grain to snack on. While it’s munching, a stainless-steel-and-carbon-fiber arm extends underneath it, automatically attaching milking machinery to its four teats (the A4 uses lasers to scan the underside of the cow and identify where the teats are). When the milking is done, the trough swings away, the gate opens, and the cow rejoins the herd in the barn. The A4 then cleans and resets itself for the next cow.
This saves a dairy farmer that fun daily ritual of waking up at the crack of dawn to milk the cows, but what’s just as important is the data collected from each cow. The A4 scans a cow’s collar, using either radio waves or infrared light to tell one animal from another. Next, the system tracks several parameters while the cow is milked: its weight, its milk production, the time required to milk it, the amount of feed the cow eats—even how long the cow chews on its cud (determined through audio sensors on certain collars).
The machine collects data on the cow’s milk as well. Milk coming out of each teat (or quarter, in dairy parlance) is checked for color, fat and protein content, temperature, conductivity (an indicator of possible infection), and somatic cell count. This information is pulled together into reports for each cow; if the A4 detects a problem, the machine can alert the farmer on his phone.
Even better, by early next year the cow may text the farmer. More than 4,400 miles away from Madison, near the Swiss city of Bern, researchers are testing sensors that can be implanted in cows to determine when they are in heat. Using algorithms and a cellular chip, the cow can text its owner to let him know it’s time to set up a quiet corner in the barn with low lights and some Teddy Pendergrass records.