John Deere: From Equipment Manufacturer to AgTech Firm
John Deere invests heavily as machine learning technology offers the next revolution in agricultural production.
By 2050, the UN projects that there will be over 9.7 billion people living on Earth. Several studies estimate that in order to feed our growing population and increasing biofuels consumption, global crop production will have to increase by at least 50%. [1] Meanwhile, an aging population and urbanization are decreasing the supply of human farm labor. Firms are investing in improvements to technology and artificial intelligence to close some of these labor gaps, while increasing yields for farmers. One firm positioning itself as a leader in the agricultural technology space is Deere & Company, the 180-year old manufacturer of agricultural equipment and other types of machinery.
Data Collection and iOT
John Deere has long been focused on enabling precision farming, the practice of measuring and responding to field variability to optimize resource allocation and crop yield. To enable data collection, every one of their large agricultural machines has a 4G LTE modem built in, as well as sensors capturing data be uploaded to the cloud. [2] Deere also built its own digital platform where farmers can access data about their crops and equipment or export data for further analysis to over 90 software partners. Based on their insights, farmers can customize plans and coordinate operations across machines to execute on them. [3] The ability to closely control farm operations is supported by Deere’s semi-autonomous tractors. While the tractors still require a human in the cab, they are able to steer along precisely defined lines, preventing overlap and increasing efficiency. [4] As their VP of technology, Stone, has said: “We’ve got computer-vision systems now, internally developed, on basically all of our large ag equipment…It’s on tractors, on our sprayers, on our harvesters. These vision systems have deep neural nets underneath them. That is definitely the future of our equipment. I think machine learning is going to be as core to John Deere as the diesel engine.” [2]
Acquiring Talent from the Outside
John Deere’s latest investment in ML came in the form of a $305 million acquisition of Blue River Technology, a Sunnyvale startup which specializes in AI for Agriculture. Their “See and Spray” technology recognizes weeds and selectively spr (Food and Agricultural Organization of the United Nations)ays them, reducing herbicide inputs by up to 90%. They’re also developing a “LettuceBot” for “precision lettuce thinning”, though there are many other potential agricultural applications of the technology. “Blue River is advancing precision agriculture by moving farm management decisions from the field level to the plant level,” said Jorge Heraud, co-founder and CEO of Blue River Technology. “We are using computer vision, robotics, and machine learning to help smart machines detect, identify, and make management decisions about every single plant in the field.” [5]
Deere is also investing in a 33,000-square-foot design and test lab in a partnership with Iowa State University. Much of the research coming out of the university will undoubtedly make its way into Deere’s future products. One such example is a machine-learning program that can find patterns in soybean leaf images and match them with eight common sources of stress (disease, nutrient deficiency, herbicide injury, etc.). [6]
Additional Opportunities
One area that John Deere lacks investment is in the drone hardware space. Agricultural drones are becoming increasingly popular as quick surveillance and data collection agents. As they become cheaper and increasingly autonomous, more and more farmers will have access to drones. Deere has partnered with Krespy, an aerial intelligence drone company, to integrate their drones into Deere’s operations optimization offerings. [7] However, relying on a partnership with an outside hardware provider can be risky. As a company that specializes in hardware solutions themselves, building drone technology in-house feels like a natural addition to their product line.
Another area that may warrant additionally investment is in the analysis of farmers’ data. Currently, when it comes to intelligent recommendations, companies such as AgDNA are much more advanced in their machine learning capabilities. There is a risk in the future that smaller, more specialized machines will replace Deere’s large multi-functional tractors. At that point, Deere will be more dependent on revenue from their cloud service than large equipment purchases. However, when data is more commoditized and analysis is the only point of differentiation, will farmers still have an incentive to keep their data with John Deere, or will they exit the ecosystem? Even if Deere is looking to invest further, will they be able to recruit top talent to a non-traditionally “tech” firm?
Words: 726
[1] Food and Agricultural Organization of the United Nations, How to Feed the World in 2050, http://www.fao.org/fileadmin/templates/wsfs/docs/expert_paper/How_to_Feed_the_World_in_2050.pdf, p. 12, accessed November 2018.
[2] Brody, L. and Steinberg, D. (2018) ‘DIGITAL REINVENTION. (cover story)’, Forbes, 201(7), pp. 90–91. Available at: http://ezproxy-prod.hbs.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=131523230&site=ehost-live&scope=site (Accessed: 13 November 2018).
[3] Deere. “Operations Center.” https://www.deere.com/en/technology-products/precision-ag-technology/data-management/operations-center/, accessed November 2018.
[4] Deere. “John Deere Rolls Out Smarter S700 Combines & Front-End Equipment.” https://www.deere.com/en_US/corporate/our_company/news_and_media/press_releases/2017/agriculture/2017jun1_s700_combine.page, accessed November 2018.
[5] Biofuels Digest: Deere and Cargill say “Domo Arigato Mr. Roboto” – Artificial intelligence infiltrating agriculture and farming 2017, , Newstex, Chatham.
[6] John deere invests in new testing lab at ISU. (2018). Farm Industry News, Retrieved from http://search.proquest.com.ezp-prod1.hul.harvard.edu/docview/2076865307?accountid=11311
Very cool topic, pretty unique blend of old-school and new-tech. But with just one harvest per year, farmers can be risk-averse and thus reluctant to try new technology, especially when they have generations of experience & proven growing techniques… should Deere even try to convince the small family farms (90% of US farms are < $350k annual net income) to buy in or is the ROI for this tech infeasible for anyone but the largest players? Longer-term, will high-density, controlled-environment vertical greenhouses replace traditional farms?
Thanks for this Jane Deere! Its really exciting to see a 180-year old company reinventing itself to match the technology of the current day and age. I think the M&A approach John Deere leveraged to acquire Blue Mountain’s technology was a smart way to get around the difficulty they have attracting tech talent as a non-traditional technology firm. I’m curious on the quality of the data that exists in the agriculture industry (which is critical for machine learning) and whether John Deere, as a large established firm, could play a role in improving data collection.
Great article, Jane. With regards to talent acquisition, I do not think that this should be an issue as long as they keep innovating new equipment that can keep up or be ahead of other technology enabled agriculture equipment in the market. The risk of farmers churning, however, is more concerning in my opinion. John Deere must find a way to maintain the learning curve from the data they collect, translated into higher yield rate YoY for their farmers. I would be interested also to see if they can apply these learnings in non-farming intensive geographies, and if it is possible to use the data to grow certain crops in inadequate soil/climates.